Introduction

Micronutrient malnutrition, caused by inadequate consumption of iron, zinc, and vitamin A, compromises the health of many populations (Kennedy et al. 2003). Worldwide, two billion people suffer from micronutrient deficiencies and about 815 million people are under-nourished (Global Nutrition Report 2017: https://www.globalnutritionreport.org). Vitamin A deficiency (VAD) accounts for about 70% of the childhood deaths globally by affecting over 250 million pre-school children (Muthusamy et al. 2015). Also, protein energy malnutrition (PEM) due to inadequate proteins in diet affects billions of people worldwide (Bain et al. 2013). Deficiency of the essential amino acids, lysine and tryptophan, causes fatigue, delayed growth, loss of appetite, depression, and anxiety in children (Nuss and Tanumihardjo 2010; Jompuk et al. 2011). Quality protein maize (QPM) is a significant promise for solving problem of PEM across the world (Nyakurwa et al. 2017), as it contains higher tryptophan and lysine in endosperm. Although cereals are rich source of energy, they are ill balanced or possesses inadequate amount of micronutrients (Nuss and Tanumihardjo 2010). To combat dietary deficiencies of protein, vitamins, and essential micronutrients, different strategies, such as industrial food-fortification, supplementation, and dietary diversification, have been tried worldwide; but none of these approaches has been found viable in the long run (Tanumihardjo et al. 2007). The development of micronutrient-enriched staple plant foods through breeding approaches, i.e., “biofortification,” is a food-based solutions to combat micronutrient deficiencies (Gupta et al. 2015; Neeraja et al. 2017).

Maize is one of the most important food crops in the world, and together with rice and wheat it provides 30% of the food calories to more than 4.5 billion people in 94 developing countries (Babu and Prasanna 2014). Maize has attained importance in world economy. Maize kernels lack the essential amino acids lysine and tryptophan and are also deficient in vitamin A, C, B, iron, zinc, and iodine. Among the cereals, only the yellow kernel maize exhibits tremendous natural variation for provitamin A (proA) carotenoids to exploit through plant breeding, and being a carotenogenic plant (Rodriguez-Amaya 2001), it is considered as one of the potential model cereal crop, showing promise for proA biofortification (Zhang et al. 2012; Sagare et al. 2018). Naturally existing mutant alleles of key genes, lycE (lycopene-ε-cyclase) and crtRB1 (β-carotene hydroxylase) of carotenoid biosynthesis pathway, significantly enhances accumulation of proA compounds in the maize endosperm (Harjes et al. 2008; Yan et al. 2010; Babu et al. 2013; Azmach et al. 2013), therefore called as favorable alleles. Also, the recessive allele of opaque2 gene (o2) enhances endosperm lysine and tryptophan by two folds (Mertz et al. 1964). PCR-based co-dominant markers for lcyE, crtRB1, and o2 have been developed for their utilization in molecular breeding.

The use of molecular markers in marker-assisted backcross breeding (MABB) helps for introgression of target genes into a recipient genetic background (Babu et al. 2005) and significantly reduces the breeding cycles required for recovery of recurrent parent genome (RPG) (Gupta et al. 2013). Also, molecular markers serves as an alternative for highly expensive HPLC (high-performance liquid chromatography) analyses required for estimation of micronutrient contents among the individuals of segregating populations in MABB (Zunjare et al. 2018). A number of inbred lines/hybrids possessing favorable allele constitution at either or both the loci of lcyE/crtRB1 have been developed employing MABB (Vignesh et al. 2012; Babu et al. 2013; Azmach et al. 2013; Muthusamy et al. 2014; Liu et al. 2015; Muthusamy et al. 2015; Yang et al. 2018; Zunjare et al. 2018).

The present study was undertaken to identify promising β-carotene donor using gene-based markers, introgression of favorable allele of crtRB1 into QPM inbreds through MABB, and evaluation of MABB-derived inbreds for nutritional and other agronomic traits.

Material and methods

Plant material

Maize germplasm lines were obtained from the Indian Institute of Maize Research (IIMR), New Delhi; Vivekananda Parvatiya Krishi Anusandhan Sansthan (VPKAS), Almora; Maize Research Centre (MRC); and Institute of Biotechnology (IBT), Rajendranagar, Hyderabad. QPM inbreds, five of CBML6 family, and eight of CBML7 family (CBML6 and CBML7 are QPM versions of BML6 and BML7, respectively—parental lines of maize hybrid DHM117; developed at IBT using CML181 as QPM donor) were used for identification of suitable recurrent parent (online resource 1) and seventy-two (72) maize germplasm lines (online resource 2) were used for identification of promising donor for β-carotene. The study was carried out during kharif 2013 to rabi 2015–16 at IBT and MRC, Rajendranagar, Hyderabad.

Recurrent parents and donor parent

Light box screening was carried out to check the kernel opacity. The o2 gene–based SSR umc1066 was used to screen QPM germplasm lines and further tryptophan/lysine content was determined. Final selection of recurrent parents (RPs) was carried out on the basis of combining ability and heterosis of the QPM inbreds of CBML6 and CBML7 families (Surender et al. 2017).

Seventy-two maize inbreds were screened with functional markers of crtRB1-5′TE, crtRB1-3′TE, and lcyE-5′TE to study allelic variations at β-carotene hydroxylase (crtRB1)-5′TE, lycopene-ε-cyclase (lcyE)-5′TE and β-carotene hydroxylase (crtRB1)-3′TE. Several studies reported that allele1 of crtRB1-3′TE alone can double the β-carotene concentration in maize kernel irrespective of genetic constitution of crtRB1-5′TE and lcyE; therefore, biochemical analysis for β-carotene estimation of the maize inbreds possessing favorable allele of crtRB1-3′TE was carried out to identify promising donor parent (DP).

Molecular analysis

DNA was extracted from young leaves of 20-day-old plants using a modified CTAB method. PCR reaction was carried out in 10 μl reactions containing, 2 μl (25 ng/μl) maize genomic DNA, 0.3 μl (2.5pmole) of forward and reverse primers each, 1 μl (2.5 mM) dNTPs, 1 μl Taq buffer containing 15 mM MgCl2 (10×), 1 μl (1 U/μl) of Taq polymerase, and 4.6 μl of sterile distilled water using the Eppendorf Thermocycler. After initial denaturation for 5 min at 94 °C, each cycle comprised 1-min denaturation at 94 °C, 45-s annealing at 55°–57 °C, and 1-min extension at 72 °C with a final extension for 10 min at 72 °C at the end of 35 cycles. The PCR products were mixed with bromophenol blue gel loading dye and were resolved by electrophoresis on 3% Metaphor™ Agarose gel (containing 0.5 mg/ml ethidium bromide) and visualized on GelDoc™ XR (Bio-Rad Laboratories Inc., USA).

Breeding scheme

For MABB scheme, both RPs were crossed with DP to obtain F1 seeds (Fig. 1). The F1s were backcrossed with respective RP to obtain a large number of BC1F1 seeds. In the BC1F1 generation, individual plants that were heterozygous at the target loci were identified (foreground selection) reducing the population size for further screening. From the individual plants that were heterozygous for target loci, those that were homozygous for the recipient allele at one marker locus flanking the target loci, i.e. recombinants, were identified (recombinant selection). From the recombinant plants, individuals with recurrent-type allele at amino acid and endosperm modifier loci were selected (modifier selection). After modifier selection, individuals with the less number of donor-type genic SSRs were selected (background selection). In the second backcross (BC) generation, the same strategy was followed for selection of individual plants with the desired allele combination at the target loci, flanking marker, modifier loci, and RP-type genomic composition at the non-target loci; and crosses were made with the respective RPs to obtain next generation. Selected BC2 plants were self-pollinated in two successive generations for further analysis.

Fig. 1
figure 1

MABB scheme for enrichment of β-carotene into QPM inbreds

Full size image

Marker assay

Functional marker of crtRB1-3′TE (Yan et al. 2010) and o2 gene–specific umc1066 marker (Babu et al. 2005) were used for F1 confirmation and foreground (FG) selection in BC1F1, BC2F1, BC2F2, and BC2F3 generations. Flanking markers (FMs) of crtRB1, viz., umc1506, and bnlg1028 (Chander et al. 2008), were used for recombinant (RC) selection in BC1F1 and BC2F1 generations. Six SSR markers, viz., mmc0241, umc1216, phi072, bnlg1633, bmc1382, and phi075 (Danson et al. 2006) for endosperm and amino acid modifiers, were used. The sequence information of these markers (online resource 3) was taken from maize genome database (http://www.maizegdb.org). Polymorphism between DP and two RPs was surveyed using these markers. A total of 500 genomic SSRs unlinked to crtRB1 and o2 covering all ten chromosomes, including the o2 and crtRB1 carrier chromosome 7 and 10, respectively, were screened and the markers showing polymorphism between the DP and RPs were used for background (BG) selection to recover the recipient genome (online resource 6). The sequence information of these markers was taken from maize genome database. The SSRs that revealed fixed (homozygous) alleles at non-target loci at one generation were not screened at the next BC generation. Only those markers that were not fixed for the RP allele were analyzed in the following generations.

Phenotyping and biochemical analysis

Biochemical analysis for estimation of β-carotene was carried out using water alliance HPLC System. Tryptophan content of maize kernels was estimated using the method described by Hernandez and Bates (1969) and the value of lysine was calculated by multiplying tryptophan value by four as tryptophan and lysine possess strong correlation (r = 0.99) in o2-based germplasm, with value of lysine being four times that of tryptophan (Vivek et al. 2008).

The phenotypic data was recorded in BC2F3-generated plants during rabi 2015–16. Observations on different agronomic characteristics, viz., days to 50% tasseling, days to 50% silking, days to maturity, plant height, ear height, ear girth, ear length, number of kernel rows per ear, number of kernels per row, 100 seed weight, and grain yield per plant, were recorded.

Data analysis

Scoring of marker alleles was carried out for BG selection. Heterozygous alleles were represented by H and scored as 0.5. Homozygous recipient allele was represented by A and scored as 1 while homozygous donor allele was represented by letter B and scored as 0. Based on the scoring data, RPG was calculated using the formula given by Kumar et al. (2012).

$$ \mathrm{RPG}=\frac{\left[A+1/2H\right]\ }{N}x100 $$

(N, total number of parental polymorphic markers screened)

For comparing agronomic traits, tryptophan/lysine content and β-carotene content between individuals of the BC2F3 families and the corresponding RP line statistical analysis were performed using Windostat 8.0.

Results

Allelic variation at crtRB1-3′TE, lcyE-5′TE, and crtRB1-5′TE and identification of promising DP

Among all the 72 inbreds screened for allelic variability study; four inbreds, viz., MGU23379, MGU23207, CM150, and BAJIM12-11, were found to contain favorable allele, i.e., allele1 (543 bp) of crtRB1-3′TE gene, one inbred BML13 was found to possess favorable allele, i.e., allele4 (993 bp) of lcyE-5′TE, and no other inbreds were found with favorable allele of crtRB1-5′TE (online resource 4).

According to Azmach et al. (2013), inbred lines harboring the favorable alleles of the crtRB1-5′TE and 3′TE produce higher levels of proA, whereas favorable allele of lcyE gene has weak association with levels of proA carotenoids. Babu et al. (2013) reported that allele1 of crtRB1-3′TE alone is responsible to double the β-carotene concentration in maize kernel irrespective of genetic constitution of crtRB1-5′TE and lcyE gene. Marker-assisted selection (MAS) for favorable allele1 of crtRB1 alone leads to rapid doubling of total proA concentration (Babu et al. 2013). Indian maize germplasm for this favorable allele has been screened by many researchers (Selvi et al. 2014; Zunjare et al. 2017) to identify proA donor to be used in breeding program.

Biochemical analysis for β-carotene estimation of the four inbreds containing allele1 of crtRB1-3′TE was carried out using HPLC method. Among the 4 inbreds, MGU23379 was found to contain high amount of β-carotene (6.31 μg/g), followed by MGU23207 (5.25 μg/g), CM150 (3.01 μg/g), and BAJIM12-11 (3.17 μg/g). Total carotenoid content was recorded highest in CM150 (7.23 μg/g), followed by MGU23379 (9.18 μg/g), MGU23207 (4.27 μg/g), and BAJIM12-11 (3.34 μg/g), respectively. As the β-carotene content was found to be high in MGU23379, it was selected as a promising β-carotene donor in MABB program to enrich β-carotene of the QPM inbreds. The β-carotene content in maize inbred BML13 possessing allele4 (993 bp) of lcyE-5′TE (1.31 μg/g) was lower than four inbreds carrying allele1 of crtRB1-3′TE.

Identification of promising QPM inbreds for conversion into β-carotene rich lines

Seeds of five QPM inbreds of CBML6 family and eight of CBML7 family were found to possess < 25% kernel opacity in light box screening. These QPM inbreds were further subjected to genotyping with o2-specific SSR umc1066 and showed amplicon size of 155 bp. Tryptophan/lysine estimation of QPM inbreds of both the families was carried out.

In CBML6 family, tryptophan and lysine content (online resource 1) was high in CB6-36 (0.89%; 3.56%) and in CBML7 family it was high in CB7-28 (0.84%; 3.36%). Also, general combining ability (GCA) and heterosis studies proved that cross combination CB6-36 × CB7-28 exhibits on par grain yield per plant when compared to check DHM117 (Surender 2015). Biochemical analysis for kernel iron (Fe) and zinc (Zn) of QPM showed that CB6-36 and CB7-28 are rich sources of Fe (42 mg/kg) and Zn (48 mg/kg), respectively (Surender 2015). The β-carotene content of CB6-36 and CB7-28 is 0.71 and 1.29 μg/g, respectively. Therefore, these lines were selected for enhancing β-carotene and named them as CBML6 and CBML7, respectively.

Parental polymorphism survey

MGU23379 (β-carotene donor–DP), CBML6, and CBML7 (RPs) were screened with crtRB1-3′TE (Forward, Reverse1, and Reverse2), o2-specific umc1066 (Forward and Reverse), five amino acid–linked SSRs (mmc0241, umc1216, phi072, bnlg1633, bmc1382), one endosperm modifier SSR phi075, two FMs (umc1506 and bnlg1028), and 500 genomic SSRs to check parental polymorphism. DP showed distinct polymorphism with both RPs at crtRB1-3′TE, umc1066, mmc0241, phi075, umc1506, and bnlg1028 loci (online resource 5). The rest of the modifier loci were monomorphic between DP and both RPs, and out of 500 genomic SSRs screened, DP showed polymorphism at 161 and 82 loci with CBML7 and CBML6, respectively (online resource 6).

Generation of F1, BC populations, genotyping, and phenotyping

During kharif 2013, F1 seed materials were developed by making crosses between MGU23379 and both RPs (CBML6 and CBML7). The F1 seeds obtained from both crosses were sown as per ear to row method during rabi 2013–14 to generate F1 population. Each individual of F1 population was subjected to FG selection with umc1066 and crtRB1-3′TE. Both the F1 populations exhibited heterozygosity for crtRB1-3′TE and o2 allele. FG-selected true F1s of both the crosses were backcrossed with respective RPs to generate BC1F1 seed material. BC1F1 population was generated during kharif 2014. In BC1F1 population, 500 plants of each cross were tagged and FG selection was performed for o2 and crtRB1-3′TE loci (Fig. 2). Individuals with recessive-type allele at o2 were subjected to screen with crtRB1-3′TE loci and individuals possessing heterozygosity at crtRB1-3′TE loci were further screened with FM umc1506 to select single recombinants (SRCs). SRC individuals were subjected to screen with polymorphic amino acid (mmc0241) and endosperm (phi075) modifiers. As the RPs possess favorable alleles for modifiers, individuals showing RP-type alleles at both of these loci were selected (Table 1). A total of 21 CB6-BC1F1 and 14 CB7-BC1F1 individuals were subjected to BG selection with 82 and 161 polymorphic SSRs to recover maximum RPG of CB6 and CB7, respectively. In the CB6-derived BC1 progenies, the average RPG content was 76.73%, which ranged from 60.33 to 83.00%: approximately 1.73% higher than the theoretical value (75%). Whereas in CB7-BC1F1 individuals, the average recovery of RPG was 74.87% which ranged from 64.25 to 83.22%: approximately 0.13% lower than the theoretical value (75%). In CB6-BC1F1 and CB7-BC1F1 population, all the plants with the highest RPG (P6-193: 83.00, P6-329: 82.66, P6-407: 82.25 and P7-19: 83.22, P7-487: 82.75) were backcrossed with respective RPs for generating BC2F1 seed material (Table 1). Chromosome-wise recovery of RPG in the best lines (CBML6: P6-193; CBML7: P7-19) is shown in online resource 7.

Fig. 2
figure 2

Foreground selection of CBML6-BC1F1 with crtRB1-3′TE. M: 100 bp; D: MGU23379; R: CBML6; Lanes 1–74,: CBML6-BC1F1 population

Full size image
Table 1 Molecular progress in MABB
Full size table

During rabi 2014–15, backcrossed seeds of P6-193 and P719 individuals of CB6-BC1F1 and CB7-BC1F1, respectively, were sown to generate BC2F1 populations. In BC2F1 population (250 plants), individuals possessing recessive allele at o2 and heterozygosity at crtRB1-3′TE loci were screened with FM bnlg1028 to select double recombinants (DRCs) (online resource 8). DRCs were subjected to modifier selection and all DRCs were found to possess recurrent-type allele at modifier loci. Further, BG selection with 14 and 27 polymorphic SSRs which were not recovered in BC1F1 populations of CB6 and CB7, respectively, was carried out (Table 1). For CB6- and CB7-derived BC2F1 population, the RPG varied from 88.25 to 93.25% and 87.98 to 91.06%, with average of 90.35% and 88.09%, respectively (Table 1). The average RPG value was 2.85% and 0.59% higher than the theoretical value (87.5%) in CB6 and CB7-BC2F1, respectively. All the plants of CB6 and CB7-BC2F1 with the highest RPG (P6-13: 91.75, P6-178: 93.25, P6-229: 91.34 and P7-5: 90.25, P7-37: 90.75, P7-93: 89.67, P7-112: 91.00, P7-154: 89.20, P7-193: 88.33, P7-216: 91.06, P7-238: 90.15) were selfed for generating BC2F2 seed material (Table 1). Chromosome-wise recovery of RPG in the best lines (CBML6: P6-178; CBML7: P7-216) is shown in online resource 9. The Chi square (χ2) test for the segregation ratio of donor type and heterozygote alleles at crtRB1-3′TE locus was performed, which confirmed the goodness of fit of our data to the theoretical ratio of 1:1 in BC1F1 and BC2F1 generation (Table 1). BC2F2 seeds derived from BC2F1-CB6-P6-178 and BC2F1-CB7-P7-216 were used to generate BC2F2 population during kharif 2015.

During kharif 2015, 300 plants of each population were tagged and FG selection was performed for crtRB1-3′TE (Table 1). The χ2 test for the segregation ratio of donor type, heterozygotes, and recurrent-type alleles at crtRB1-3′TE locus was performed, which confirmed the goodness of fit of our data to the theoretical ratio of 1:2:1 in BC2F2 generation (Table 1). Individuals with recurrent-type allele at o2 loci and donor-type alleles at crtRB1-3′TE loci (online resource 10) were selfed to generate BC2F3 seed material. Harvesting was done and cobs showing similar morphology with RPs in terms of texture, size of the grains, and row arrangements were selected for recording cob phenotypic data. Further, seeds of these cobs (19 and 21 selfed cobs of CB6 and CB7-BC2F2, respectively) were used for estimation of β-carotene, tryptophan/lysine content. Tryptophan and lysine content varied from 0.88 to 0.90%; and 0.83 to 0.85% in CBML6- and CBML7-derived BC2F3 seeds, respectively (online resource 11). Tryptophan/lysine content was found to be on par with respective RPs. Content of β-carotene varied from 5.00 to 6.25 μg/g and 5.85 to 6.80 μg/g in CB6- and CB7-derived BC2F3 seeds, respectively (online resource 11), which is significantly high compared to RPs (Table 2). BC2F3-CB6 seeds derived from individuals P6-3 and P6-69 were found with higher β-carotene content (6.25 μg/g), followed by P6-178 (6.24μg/g) and P6-23 (6.23μg/g); whereas BC2F3-CB7-derived seeds of P7-29 were found with higher β-carotene content (6.80 μg/g) followed by P7-158 (6.79 μg/g).

Table 2 Statistical analysis
Full size table

Seeds with higher content of tryptophan and β-carotene (CB6: P6-3 and P6-69; CB7: P7-29) were named as PVCBML6 and PVCBML7 and sown during rabi 2015–16 to generate BC2F3 population. A total of 60 plants for each BC2F3 population (PVCBML6 and PVCBML7) were tagged and FG selection was performed for o2 and crtRB1-3′TE loci (Table 1). All plants were found to possess recurrent-type allele at o2 loci. At crtRB1-3′TE loci, one plant was found with RP-type allele, the possible reason for this could be a contamination by outcrossing with pollens of RP. Plants possessing recurrent-type allele at o2 and donor-type allele crtRB1-3′TE loci were subjected for plant phenotyping. The DUS test guidelines for maize suggested by protection of plant varieties and farmer’s right authority, India (http://www.plantauthority.gov.in), were used as a reference for phenotyping. In PVCBML6, stem anthocyanin pigmentation was found to be reduced than the original QPM line. Also, the silk color was dark pink in original QPM line which was changed to light pink in PVCBML6 (Fig. 3, online resource 12). In PVCBML7, tassel density was found to be more sparse and tassel was more erect than original QPM line. Silk color was changed to light pink in PVCBML7, which was pure white in original QPM line (Fig. 3, online resource 13). Plants possessing favorable phenotypic characters were selfed to generate BC2F4 seeds.

Fig. 3
figure 3

Plant characters of original and converted version of CBML6 and CBML7

Full size image

The phenotypic data was recorded on different phenotypic characters and differences in the phenotypic traits between the BC2F3 progenies and their RP lines were analyzed using Windostat 8.0 (Table 2). There were non-significant differences in the phenotypic traits between the BC2F3 progenies. The tryptophan and lysine content in converted inbreds (PVCBML6 and PVCBML7) remained on par with respective RPs. The β-carotene content in converted lines was significantly higher than that of RPs (CB6 and CB7) but not as high as donor (MGU23379).

Discussion

QPM possesses almost the double amount of tryptophan and lysine in kernels than the normal maize (Vasal 2000); however, QPM genotypes also possess very low proA carotenoids (Gupta et al. 2015). Mutant allele of crtRB1 and lcyE significantly enhances proA carotenoids in maize kernels (Harjes et al. 2008; Yan et al. 2010; Babu et al. 2013). The present study used MABB for combining the favorable allele of crtRB1-3′TE into QPM inbreds (CBML6 and CBML7—parental lines of popular maize hybrid DHM117). The functional marker for crtRB1-3′TE (Yan et al. 2010; Vignesh 2012; Babu et al. 2013; Selvi et al. 2014; Liu et al. 2015; Muthusamy et al. 2015; Zunjare et al. 2017) and o2-based umc1066 marker (Gupta et al. 2013; Hossain et al. 2018; Surender et al. 2017) showed codominance in an earlier investigations. In this study, the functional marker for crtRB1-3′TE and o2-based umc1066 marker showed co-dominant polymorphisms between the DP and the RP lines, allowing us to trace the favorable crtRB1-3′TE and umc1066 alleles in each segregating backcross generation and validate homozygous and heterozygous genotypes.

Previous studies reported severe segregation distortion (SD) for crtRB1-3′TE across the generations and crosses (Babu et al. 2013; Muthusamy et al. 2014; Liu et al. 2015; Zunjare et al. 2017). In current study, crtRB1-3′TE segregated as per Mendelian inheritance because large enough population was screened for achieving sufficient foreground positive genotypes. In majority of populations, o2 also segregated as per Mendelian inheritance (Gupta et al. 2013; Hossain et al. 2018) with an exception in one population (Jompuk et al. 2011).

Realizing the importance of the FMs for RC selection to minimize linkage drag, FM umc1506 and bnlg1028 were used. With these FMs, linkage drag around crtRB1 locus was minimized, and crtRB1 transfer was precisely monitored and it made possible to recover DRCs with crtRB1-3′TE allele1. FMs for RC selection to minimize linkage drag in MABB program of o2 introgression have been reported (Babu et al. 2005; Gupta et al. 2009; Gupta et al. 2013; Singh and Ram 2014; Surender et al. 2017), but none of the earlier investigations for enhancing β-carotene in QPM background have used FMs. For the first time, we have reported the use of FMs for crtRB1 locus to select recombinants.

In present study, SSR markers used in MABB program resulted in high recovery of RPG in two backcross generations and a continuous increase in genomic recovery ratio in contrast to the theoretical ratio was observed in each backcross generation (Gupta et al. 2013; Muthusamy et al. 2014; Feng et al. 2015; Liu et al. 2015; Hossain et al. 2018; Yang et al. 2018). The introgressed progenies with high RPG attributed to a high degree of resemblance with their corresponding RPs (Muthusamy et al. 2014; Hossain et al. 2018; Yang et al. 2018).

The variation for lysine and tryptophan among o2-introgressed progenies was observed in various studies (Gupta et al. 2013; Hossain et al. 2018; Zunjare et al. 2018) in their MABB programs, due to various modifier loci that affect regulation of amino acid biosynthesis. But in current study, tryptophan/lysine content in introgressed inbreds was on par with respective RPs. And the nutritional benefit of QPM with enhanced lysine and tryptophan was conserved in the MABB-derived lines. The reason for this is selection for recurrent-type alleles at modifier loci in BC1 and BC2 progenies along with o2 allele.

Introgressed inbreds possessed significantly higher β-carotene than their respective RPs, but was not as high as DP. This may be due to the contribution of other genetic loci or minor QTLs apart from allele1 of the crtRB1 gene to the accumulation of β-carotene (Wong et al. 2004; Chander et al. 2008; Zhou et al. 2012; Kandianis et al. 2013). The crtRB1 and lcyE genes show endosperm-specific expression (Babu et al. 2013); and reduced transcript of mutant crtRB1 leads to lesser amount of β-hydroxylase, and thus lesser conversion of β-carotene to further pathway components (Vallabhaneni et al. 2009; Yan et al. 2010). The advantage of crtRB1 and lcyE together for proA enhancement over individual effects has been reported (Azmach et al. 2013; Gebremeskel et al. 2018; Zunjare et al. 2018). But the favorable alleles of the both genes (crtRB1 and lcyE) occur in low frequency in the maize germplasm (Azmach et al. 2013; Babu et al. 2013; Muthusamy et al. 2015; Gebremeske et al. 2017). Babu et al. (2013) reported a very strong association between allele 1 of crtRB1 and the endosperm content of β-carotene and proA concentrations irrespective of the genotypic constitution for lcyE. The crtRB1 has much larger effect than lcyE on β-carotene and proA concentrations; thus, MAS for the crtRB1 locus alone is a reliable strategy for rapidly achieving genetic gains for β-carotene and proA carotenoids in tropical maize breeding programs (Babu et al. 2013). The results of current study showed that the favorable crtRB1-3′TE allele significantly improved β-carotene. These findings are in accordance with Liu et al. (2015), Muthusamy et al. (2015), and Zunjare et al. (2018) who attempted introgression of crtRB1 gene into quality protein maize inbred lines using molecular markers and observed increase in β-carotene content of converted lines than the original QPM lines.

Although a significant increase in the β-carotene content was observed in these converted lines, the β-carotene content is less than the target level 15 μg/g proA carotenoids, which emphasize the need of additional efforts to increase these levels. Therefore, further introgression of genetic loci like crtRB3, CCD1, and ZEP1 (Zhou et al. 2012; Suwarno et al. 2015) can be attempted to achieve the target levels of proA carotenoids in maize kernel.

The normal and converted lines of CB6 and CB7 were found to be significant for all the phenotypic characters. Agronomic traits showed the stable expression in BC2F3 families, and thus represented the final converted versions of the original recipient lines (Babu et al. 2005). The grain yield of converted inbreds was also on par with the original inbreds. The similarity was due to indirect selection of loci for yield potential and various agronomic traits through BG selection (Gupta et al. 2013; Muthusamy et al. 2014; Hossain et al. 2018; Zunjare et al. 2018).

Conclusion

The study implemented a successful demonstration of MABB coupled with rigorous phenotypic selection for agro-morphological characters. These converted inbreds can be used for reconstitution of DHM117 maize hybrid rich in tryptophan, lysine, and β-carotene, which will alleviate VAD and PEM. Converted lines also provides an ideal platform for stacking number of nutritionally important traits such as enhanced Fe, Zn, vitamin E, folate, ascorbic acids, and low phytate. The QPM lines with improved β-carotene concentration serve as an important breeding material for developing QPM maize hybrids with high β-carotene for people dependent on maize as a major component of their diet. Therefore, the development and production of QPM hybrids employing MABB, with high β-carotene concentrations, is an efficient way to deliver biofortified maize for improved health and development benefits to the people of developing countries.