Introduction

Protein-energy malnutrition (PEM) has emerged as a major health problem, especially in the developing world (Temba et al. 2016). Maize provides a significant amount of total calorie to the human populations worldwide. Energy requirement to the tune of 62% in Mesoamerica, 43% in eastern and southern Africa, 22% in west and central Africa, and 28% in the Andean region, comes from maize (Shiferaw et al. 2011). It is also a preferred choice as a food in many of the tribal belts, especially in the north eastern states. A major portion (60–70%) of maize grains produced worldwide is used for animal consumption. Maize thus serves as an important source of plant protein and total calorie both directly and indirectly. Maize endosperm protein is, however, known to be poor in nutritional value due to low amount of essential amino acids, lysine (2.0% in protein) and tryptophan (0.4% in protein), and their concentration is nearly half of the level recommended for human nutrition (Mertz et al. 1964; Prasanna et al. 2001). Since, humans and monogastric animals like poultry cannot synthesize these amino acids in their body, healthy diets therefore must include the alternate sources of lysine and tryptophan (Bjarnason and Vasal 1992; Gupta et al. 2013). Among various strategies, biofortification turns out to be the most sustainable and cost-effective solution to provide micronutrients in natural form (Bouis et al. 2011; Gupta et al. 2015).

The recessive opaque2 (o2) mutant has been successfully utilized in the breeding programme for enhancement of protein quality (Vasal et al. 1980). Initially, maize cultivars with o2 mutation was not preferred by farmers and consumers, due to soft and opaque endosperm, increased susceptibility to insect-pest and diseases, and breakage of grains during mechanical processing (Bjarnason and Vasal 1992). Later, endosperm modifier genes that confer hard endosperm in the o2 background were introgressed at CIMMYT, Mexico (Villegas et al. 1992) and University of Natal, South Africa (Geevers and Lake 1992). This eventually led to the development of nutritionally enriched hard endosperm maize, popularly phrased as ‘quality protein maize (QPM)’ (Vasal et al. 1980).

Globally, a large number of normal maize hybrids have been released and commercialized. In contrast, the germplasm base of QPM is quite narrow, and substantially lesser number of genetically diverse QPM hybrids are available. In India, nearly a dozen QPM hybrids has been released, compared to more than hundred non-QPM / normal maize hybrids (Yadav et al. 2015). It is therefore necessary to develop diverse QPM varieties across different maturity groups and agroecology. The conventional approach takes 10–15 years or more to implement this programme. Conversion of elite normal maize hybrids into QPM requires significantly lesser time, primarily due to tested combining ability, heterosis and adaptability of the already released hybrid (Prasanna et al. 2010). Introgression of a recessive allele through conventional backcross breeding involves 6–7 generations of backcrossing. This time can be significantly reduced to two backcrosses by molecular marker-assisted backcross breeding (MABB) (Babu et al. 2005; Muthusamy et al. 2014). Here, we report rapid conversion of three popular medium-maturity single-cross hybrids released in India, HM4, HM8 and HM9, to QPM using MABB.

Materials and methods

Plant materials

The genetic materials comprised of three elite normal/nonQPM maize inbreds, HKI323, HKI1105 and HKI1128 having low level of lysine (1.76–2.08% in protein) and tryptophan (0.35–0.52% in protein) (table 1). These are the parents of three commercial medium maturing single cross maize hybrids in India (HM4 (\(\textit{HKI1105}{\times }{} \textit{HKI323}\)), HM8 (\(\textit{HKI1105}{\times }{} \textit{HKI161}\)) and HM9 (\(\textit{HKI1105}{\times }{} \textit{HKI1128}\))). HKI161 possesses o2 allele and is a QPM inbred. HM4 is adapted to north western plain zone (NWPZ), while HM8 and HM9 are for peninsular zone (PZ) and north eastern plain zone (NEPZ) of India (Kaul et al. 2009). CML161 (CIMMYT QPM inbred), HKI161 and HKI193-1 (QPM inbreds selected from CML161 and CML193, respectively at Uchani Centre, Chaudhary Charan Singh Haryana Agricultural University (CCSHAU), India) with high levels of endosperm lysine (3.32–3.80% in protein) and tryptophan (0.74–0.85% in protein), and desirable endosperm modification served as the donors for introgression of o2 allele.

Table 1 Details of the recurrent and donor parents used in the study.
Full size table

Target allele for introgression

Two SSRs, umc1066 and phi057 are present in first exon and sixth exon, respectively, of O2 gene present on chromosome 7 (Yang et al. 2004). The primer sequences are: umc1066 = forward (F): \(5^\prime \)-ATGGAGCACGTCATCTCAATGG-\(3^\prime \); reverse (R): \(5^\prime \)-AGCAGCAGCAACGTCTATGACACT-\(3^\prime \) and phi057 = F: \(5^\prime \)-CTCATCAGTGCCGTCGTCCAT-\(3^\prime \) and R: \(5^\prime \)-CAGTCGCAAGAAACCGTTGCC-\(3^\prime \). Polymorphic SSRs between respective recurrent and donor parents were used for marker-assisted selection (MAS) of the o2 allele.

Fig. 1
figure 1

Marker-assisted backcross breeding scheme adopted for conversion of normal maize hybrids into QPM versions.

Full size image

DNA isolation and polymerase chain reaction (PCR)

Genomic DNA was isolated from young seedlings using the standard CTAB procedure (Murray and Thompson 1980). PCR reaction (Bio-Rad, California, USA) was carried out in 10 \(\mu \hbox {L}\) of a reaction mixture containing 2 \(\mu \)L of 20 ng/\(\mu \hbox {L}\) genomic DNA as the template, 2 mM \(\hbox {MgCl}_{2}\), 1 mM dNTPs, 2 \(\mu \hbox {M}\) of the primer pair (F and R), and 1.5 U Taq polymerase (GeNei, Mumbai, India). Touch down procedure standardized at Maize Genetics Unit, ICAR-Indian Agricultural Research Institute (IARI), New Delhi, was used for PCR amplification (Pandey et al. 2015). The resulting PCR amplicons were resolved in 4% agarose gel for 4 h. The resolved amplified products were visualized using a gel documentation system (AlphaInnotech, California, USA).

MABB strategy

For MABB (figure 1), recurrent parents (as females) and donors (as males) were crossed in 2009 rainy season (July to October) at IARI experimental farm, New Delhi. The \(\hbox {F}_{1}\)s were raised at the Maize Winter Nursery Centre (off season nursery), Hyderabad, India, during 2009–2010 winter season (December to April). Heterozygosity of the \(\hbox {F}_{1}\)s was tested using the o2-specific marker, and the true \(\hbox {F}_{1}\)s were backcrossed as male parents to their respective recurrent parents. \(\hbox {BC}_{1}\hbox {F}_{1}\) progenies were grown at Delhi during the rainy season in 2010, and foreground selection was carried out using the o2-gene-specific marker(s). Desirable plants with high recovery of the recurrent parent genome (RPG) and morphological similarity to recurrent parents were further backcrossed to raise the \(\hbox {BC}_{2}\hbox {F}_{1}\) population at Hyderabad during 2010–2011 winter season. The selected plants with high RPG and phenotypic similarity to their recurrent parents were selfed. \(\hbox {BC}_{2}\hbox {F}_{2}\) populations were raised at Delhi during 2011 rainy season, and selected plants were selfed to advance progenies. Backcross progenies of HKI323\(\times \)HKI161 and HKI1105\(\times \)CML161 were generated as per the above procedure. However, for HKI1128\(\times \)HKI193-1, \(\hbox {BC}_{1}\hbox {F}_{1}\) seeds could not be generated during winter of 2009–2010 owing to the nonsynchrony of flowering; hence, fresh crosses were generated and \(\hbox {F}_{1}\)s were planted at Delhi during 2010 rainy season, and all backcross progenies were eventually raised one generation later compared to other two inbreds.

Marker-assisted foreground selection

Foreground selection was employed in \(\hbox {BC}_{1}\hbox {F}_{1}\), \(\hbox {BC}_{2}\hbox {F}_{1}\) and \(\hbox {BC}_{2}\hbox {F}_{2}\) generations using the marker specific to o2 allele. SSRs, umc1066 and phi057 were used for selection of the foreground positive plants. Heterozygous plants (O2 / o2) were selected in the \(\hbox {BC}_{1}\hbox {F}_{1}\) and \(\hbox {BC}_{2}\hbox {F}_{1}\), and homozygotes (o2 / o2) were selected in \(\hbox {BC}_{2}\hbox {F}_{2}\). Chi-square (\(\chi ^2\)) test was performed to test the goodness of fit of the observed segregation pattern at the o2 locus in each of the generations.

Marker-assisted background selection

A set of \(>350\) genomewide SSRs distributed throughout the maize genome was used for identifying polymorphic markers between the respective recurrent and donor parents. The sequences of the SSR primers were adapted from the maize genome database (www.maizegdb.org) and custom synthesized (Sigma Tech., USA). These polymorphic SSR markers were employed in each of the \(\hbox {BC}_{1}\hbox {F}_{1}\) and \(\hbox {BC}_{2}\hbox {F}_{1}\) generations of the three crosses to recover the RPG. The final recovery of RPG, across genetic backgrounds, was determined in the \(\hbox {BC}_{2}\hbox {F}_{4}\) generation.

Endosperm modification

Selfed seeds (\(\hbox {BC}_{2}\hbox {F}_{3}\) onwards) from homozygous plants (o2 / o2) were selected and analysed for the degree of opaqueness using a standard light box (Vasal et al. 1980). For analysis of endosperm modification, the back-lit kernels were rated on a scale of 0 to 100, with each number indicating per cent opaqueness. For instance, ‘100’ stands for 100% opaque while ‘0’ for 100% vitreous/hard (Hossain et al. 2008). Grains with minimal degree of opaqueness were selected and used for advancement of homozygous progenies (o2 / o2).

Phenotypic characterization of MAS-derived inbreds

The MAS-derived inbreds along with the original inbreds were evaluated at Delhi, during 2013 and 2014 rainy season in two replications each having two-rows/entry. Inbreds were characterized for 31 morphological characters (PPVFRA 2007) and 12 grain yield-related traits. Plants were raised in 3 m length row with a plant to plant distance of 20 cm, and row to row distance of 75 cm. Standard agronomic practices like application of 10–15 t of farmyard manure, 150: 80: 60 kg of N: P: K, and 20–25 kg \(\hbox {ZnSO}_{4}\) per ha in soil, and 6–8 irrigations depending upon the requirement were given, to raise the crop.

Evaluation of MAS-derived hybrids

Since the hybrids targeted for improvement are adaptable for cultivation in rainy season, crosses were generated in winter at off season nursery, and reconstituted hybrids were evaluated in the following rainy season. In rainy season, seeds of selected MAS-derived QPM inbreds were increased, and progenies (\(\hbox {BC}_{2}\hbox {F}_{5}\) for HKI323 and HKI1105, and \(\hbox {BC}_{2}\hbox {F}_{4}\) for HKI1128) were crossed to reconstitute the QPM version of original hybrid at Hyderabad, during 2012–2013 winter season. HKI161 being a QPM inbred, was directly used as a parent for generating QPM version of HM8. The crosses along with original hybrids were evaluated in two replications, each having two-rows/entry at Delhi during 2013 rainy season. Standard agronomic practices used for raising inbreds were also followed to raise the hybrids. The hybrids were characterized for 31 morphological characters (PPVFRA 2007) and 12 grain yield-related traits.

Estimation of lysine and tryptophan in endosperm protein

Inbred trials conducted for assessing the grain yield and component traits at Delhi during 2013 and 2014 were used for estimation of lysine and tryptophan in endosperm protein. For hybrids, the yield trial conducted in Delhi during 2013 was used for quality analyses. In addition, a separate hybrid trial was constituted at Delhi in 2014 only for quality analyses. Two to three plants in each entry of the trials were self-pollinated, and selfed grains were used for estimation of grain quality. Total endosperm protein was estimated using the micro-Kjeldahl procedure (AOAC 1965). The concentration of lysine and tryptophan in endosperm flour was measured using UPLC (Dionex Ultimate 3000, Thermo Scientific) at Maize Genetics Unit, IARI, New Delhi (Sarika et al. 2016). Degermed endosperm flour per sample with three replications was used. Acid hydrolysis was done for lysine, while for tryptophan alkaline hydrolysis was performed. Samples were eluted through Acclaim 120 \(\hbox {C}_{18}\) column (5 \(\mu \)m, 120 Å, 4.6 \(\times \) 150 mm, Thermo Scientific) and detected with RS photodiode array detector (PDA) with absorbance at 265 and 280 nm wavelength respectively. Concentration of amino acids in each sample was estimated by standard regression using external standards (AAS-18, Sigma Aldrich, USA).

Evaluation of MAS-derived hybrids in multilocation-based national trials

The MAS-derived experimental crosses evaluated at Delhi during 2013 rainy season were further nominated as ‘essentially derived variety’ (EDV) and tested under the AICRP-Maize, coordinated by the ICAR-Indian Institute of Maize Research (IIMR), New Delhi. Under this system, each entry was coded and the trial for each of the hybrids was undertaken at 3–12 designated locations of their respective zone of adaptation. The entries were evaluated in complete randomized block design (three replications and having four rows/entry/replications) for the two consecutive years (2014 and 2015 rainy seasons) (Annual Progress Report, Kharif Maize 2015, 2016). Various agronomic traits including grain yield were recorded. Hybrids were also evaluated for their responses to diseases like, maydis leaf blight, turcicum leaf blight, banded leaf and sheath blight, polysora rust, common rust, charcoal rot, fusarium stalk rot, sorghum downy mildew, Rajasthan downy mildew and bacterial stalk rot.

Results

Marker polymorphism

The umc1066 was polymorphic between recurrent parents (HKI323 and HKI1128) and donor QPM inbreds (HKI161 and HKI193-1), respectively. For HKI1105 and CML161, phi057 was used as a polymorphic marker. Genome-based SSR markers, 351–463, distributed among 10 chromosomes were screened between the recipient and donor parents and 30–42% was polymorphic across three crosses (table 2).

Marker-assisted introgression of o2 allele

\(\textit{BC}_\textit{1} \textit{F}_\textit{1}\) generation

Foreground selection in \(\hbox {BC}_{1}\hbox {F}_{1}\) resulted in the identification of 21 heterozygous plants in HKI323\(\times \)HKI161, 32 in HKI1105\(\times \)CML161 and 50 in HKI1128\(\times \)HKI193-1 populations (table 3). Segregation of o2 allele in all the three populations deviated from the expected Mendelian ratio of 1:1 (table 3), while recovery of RPG varied from 63.26 to 91.9% (table 4). Two plants each in HKI323-based (73.6 and 74.3% RPG), HKI1105-based (76.3 and 77.0% RPG) and HKI1128-based (85.5 and 90.0% RPG) populations were selected for further advancement (table 4).

\({\varvec{BC}}_\textit{2} \textit{F}_\textit{1}\) generation

A total of 53 heterozygous plants (O2 / o2) in HKI323\(\times \)HKI161, while 57 in HKI1105\(\times \)CML161 and 68 in HKI1128\(\times \)HKI193-1 were identified (table 3). Significant segregation distortion of o2 allele was observed in the first two crosses, while in third (HKI1128\(\times \)HKI193-1) it was 1:1 (table 3). Background selection in the heterozygous plants using polymorphic SSRs led to the recovery of 84.9–93.9% RPG in HKI323\(\times \)HKI161, 82.1–93.8% in HKI1105\(\times \)CML161 and 89.6–94.0% in HKI1128\(\times \)HKI193-1. Two plants each in HKI323- (93.2 and 93.9% RPG) and HKI1128- (93.3 and 94.0% RPG), while three in HKI1105-derived progenies were advanced (82.1%, 91.0% and 91.1% RPG) (table 4).

\({\varvec{BC}}_\textit{2} \textit{F}_\textit{2}\) generation

Foreground selection identified 58 homozygous plants (o2 / o2) in HKI323\(\times \)HKI161, while the same was 20 and 60 in HKI1105\(\times \)CML161 and HKI1128\(\times \)HKI193-1, respectively (table 3). All the three crosses deviated from the expected segregation pattern of 1:2:1 (table 3). Homozygous plants (o2 / o2) with similarity to their respective recurrent parents were selected for advancement.

\({\varvec{BC}}_\textit{2}{} \textit{F}_\textit{4}\) generation

The highest recovery of RPG observed was 98.0% in HKI323\(\times \)HKI161, 96.6% in HKI1105\(\times \)CML161 and 98.3% in HKI1128\(\times \)HKI193-1 (table 4). Based on higher recovery of RPG, phenotypic similarity to the recurrent parent and desirable degree of grain modification, HKI323-44-68-16, HKI1105-22-99-3 and HKI1128-48-1-14 were finally selected and used for advancement.

Table 2 Per cent polymorphism and distribution of SSRs used for background selection in the study.
Full size table
Table 3 Segregation of o2 allele in each of the backcross and selfed progenies across the three crosses.
Full size table

Evaluation of introgressed inbreds for grain quality and yield attributes

Endosperm lysine and tryptophan in MAS-derived inbreds

Across years, 58.5–70.0% increase in lysine, and 46.0–96.0% enhancement in tryptophan was observed in the selected progenies over their respective recurrent parents (table 5). The concentration of protein in endosperm remained almost same in both original and introgressed inbreds. However, the protein quality was significantly improved in MAS-derived inbreds (table 5).

Endosperm modification in introgressed progenies

For HKI323-based progenies, degree of opaqueness varied from 25–75%, while for HKI1105 and HKI1128, it was 25–50% and 50–100%, respectively. However, the MAS-derived inbred that was selected for generating the hybrid combinations possessed 25% opaqueness for HKI323-44-68-16 and 50% for each of \(\textit{HKI1105-}22\text {-}99\text {-}3\) and HKI1128-HKI1128-48-1-14 (table 4).

Morphological characteristics of MAS-derived inbreds

The MAS-derived inbreds showed high degree of resemblance with their respective recurrent parents (see table 1 in electronic supplementary material at http://www.ias.ac.in/jgenet/). However, the introgressed inbreds differed from their original inbreds for very few characters. For example, anthocyanin colouration in brace root and base of glume is present in HKI1105, while absent in QPM of HKI1105 (HKI1105-22-99-3). The MAS-derived inbreds however possessed similar grain yield as achieved in original inbreds (table 6).

Table 4 Recurrent parent genome (RPG) recovery and degree of opaqueness in different backcross generations.
Full size table

Evaluation of reconstituted hybrids for grain quality attributes, yield attributes and responses to diseases

Endosperm lysine and tryptophan in MAS-derived hybrids

MAS-derived QPM version of HM4 (HKI1105-22-99-3\(\times \)HKI323-44-68-16), HM8 (HKI1105-22-99-3\(\times \)HKI161) and HM9 (HKI1105-22-99-3\(\times \)HKI1128-48-1-14), were designated as HM4-Q, HM8-Q and HM9-Q, respectively. The concentration of lysine and tryptophan in endosperm of reconstituted hybrids also recorded significant improvement over their respective original hybrids. Based on both the years, 50–78% enhancement in lysine was recorded, while tryptophan showed 51–100% increase across hybrids (table 7). Protein content remained almost same in both MAS-derived and original hybrids, but the protein quality showed significant enhancement. The increase in lysine in protein was in the range of 49.5 to 77.0%, while, the same for tryptophan in protein was 61.0 to 97.5% (table 7).

Table 5 Performance of original and improved inbreds for lysine and tryptophan in endosperm.
Full size table
Table 6 Agronomic performance of improved inbreds vis-à-vis original inbreds at Delhi during rainy season 2013 and 2014.
Full size table
Table 7 Performance of original and improved hybrids for lysine and tryptophan in endosperm.
Full size table
Table 8 Agronomic performance of improved hybrids vis-à-vis original hybrids at Delhi during kharif 2013.
Full size table

Evaluation of MAS-derived hybrids for morphological characteristics

The reconstituted hybrids resembled their respective original hybrids with high degree of similarity except a few (see table 2 in electronic supplementary material). The selfed \(\hbox {F}_{2}\) seeds of HM4-Q, HM8-Q and HM9-Q showed desirable degree of endosperm modifications. Grain yield and other contributing traits were also similar among the original and MAS-derived hybrids (table 8).

The data generated during 2014 and 2015 under the multilocation trials of AICRP-Maize, clearly suggested that QPM version of reconstituted hybrids showed high degree of resemblance to their original hybrids for grain yield and yield traits (see table 3 in electronic supplementary material). The flowering behaviour and maturity of the reconstituted hybrids were similar to the original hybrids as well (see table 3 in electronic supplementary material).

Further, the reconstituted and original hybrids also showed high degree of resemblance for resistance to important diseases of maize in the country (see table 4 in electronic supplementary material). For example, HM4 during 2014 had the disease score of 3.0 against maydis leaf blight, while the same was 2.4 for HM4-Q. HM8 and HM9 recorded score of 3.0 and 2.8, while their QPM versions had 2.7 and 2.3, respectively. Similarly, all the hybrids recorded moderate resistance against turcicum leaf blight during 2015 (see table 4 in electronic supplementary material).

Discussion

MABB has been employed to introgress low phytic acid (lpa) (Naidoo et al. 2012), o2 (Gupta et al. 2013) and crtRB1 (Muthusamy et al. 2014) alleles to improve nutritional quality traits in maize. In the present investigation, distinct marker polymorphism between the respective recurrent and donor parents facilitated the introgression of recessive o2 allele into the elite inbreds. Since we used o2-specific SSRs (located within the target gene), segregants could be selected with 100% efficacy (Babu et al. 2005). We also found severe segregation distortion (SD) for the o2 as reported by Jompuk et al. (2011). The reason for the SD could be the presence of many segregation distortion regions (SDRs) throughout the maize genome (Lu et al. 2002). Frequent occurrence of SD thus necessitates generation of large populations for obtaining sufficient numbers of foreground-positive plants (Muthusamy et al. 2014).

Introgression of o2 allele resulted in significant increase in protein quality by enhancing the concentration of lysine and tryptophan in the endosperm of both MAS-derived inbreds and hybrids. Lysine in protein recorded 48–95% enhancement, while tryptophan in protein showed 47–118% increase across inbreds/hybrids over their respective genotypes. The mechanism of enhancement of lysine and tryptophan in o2 genotypes are of diverse type. The enhancement of nutritional quality in o2 mutant is mainly due to (i) reduction of lysine deficient zein proteins followed by enhanced synthesis of lysine-rich non-zein proteins (Habben et al. 1993), (ii) reduced transcription of lysine catabolizing enzyme, lysine keto-reductase, (Kemper et al. 1999), and (iii) enhanced synthesis of various lysine-rich proteins and enzymes (Jia et al. 2013).

Though introgression of o2 allele alone has a major effect on accumulation of lysine and tryptophan in higher concentration, the levels of the same varied substantially across genetic background. For example, lysine ranged from 0.277 to 0.373%, while tryptophan varied from 0.067 to 0.082% across three inbreds. The variation is due to amino acid modifier loci that influence the accumulation lysine and tryptophan in o2 genetic background (Pandey et al. 2015). The levels of lysine and tryptophan in MAS-derived hybrids also showed large variation, and interactions of amino acid modifiers contributed by both the parents possibly determined the final level of the targeted amino acids. Further, the lysine and tryptophan levels were lesser in HKI323-44-68-16 and HKI1105-22-99-3, compared to their respective donor parents (HKI161 and HKI193-1, respectively). It is possibly due to the fact that favourable modifier loci present in the QPM donor was lost owing to repeated backcrossing to recurrent parent. QTLs for these modifier loci in QPM background have been identified recently, and can be used along with o2 allele in future breeding programme (Babu et al. 2015).

SSRs covering all 10 chromosomes were used for recovering the major proportion of the RPG within two backcross generations. The highest RPG recovery among the selected progenies was 96.58% in HKI1105\(\times \)CML161, while it was 97.97% and 98.35% in HKI323\(\times \)HKI161 and HKI1128\(\times \)HKI193-1, respectively. To achieve comparable results, conventional breeding would take five backcrosses since o2 is recessive in nature. In conventional method, \(\hbox {BC}_{5}\hbox {F}_{3}\) progeny would be crossed with its other parent to reconstitute the hybrid. Thus, it would require 14 seasons from crossing the recipient and donor to the evaluation of hybrids. In contrast, MABB approach used here took 8–9 seasons since \(\hbox {BC}_{2}\hbox {F}_{4}\) (HKI323- and HKI1105-based)/ \(\hbox {BC}_{2}\hbox {F}_{5}\) (HKI1128-based) progenies were crossed to reconstitute the hybrids. The MABB approach thus clearly saved time of raising 5–6 additional seasons, and therefore accelerated the pace of breeding (Gupta et al. 2013; Muthusamy et al. 2014).

Phenotypic features such as plant, ear and grain characteristics used in combination with MAS helped to recover the RPG even rapidly (Manna et al. 2005; Muthusamy et al. 2014). Although, introgressed inbreds and reconstituted hybrids resembled their respective recurrent parents or original hybrids for majority of characters, they also differed for a few characters. In fact, morphological characteristics that show sharp contrast are highly useful for registration of genotypes (Gunjaca et al. 2008). Besides they also act as morphological marker to unambiguously differentiate the QPM-versions from the original inbreds during seed production and certification. The contrasting features are possibly due to the effects of minor proportion of donor genome (2.03–4.89%) present in the introgressed progenies (Choudhary et al. 2014).

The selected MAS-derived o2 inbreds across three genetic backgrounds showed desirable degree (25–50% opaqueness) of endosperm modification. CML161, HKI161 and HKI193-1 are the popular QPM inbreds and possess hard endosperm (50% opaqueness) due to the presence of favourable endosperm modifier loci (Paez 1973; Pandey et al. 2015). Although no marker system was used for the selection of endosperm modifiers in the backcross generations, the screening of kernels from o2 / o2 plants on light box could successfully select the desirable progenies that possessed higher modification (Gupta et al. 2013).

The grain yield potential of the reconstituted hybrids was at par with the original hybrids across multiple locations. Selection of SSR alleles specific to recurrent parents indirectly led to the selection of unknown loci associated with yield, agronomic traits and heterosis (Gupta et al. 2013; Muthusamy et al. 2014). Further, a similar response recorded in reconstituted and original hybrids against various diseases is also due to similar genomic constitution achieved through high recovery of RPG. Here also, the loci involved in responses to various diseases were not selected in the backcross progenies, but the background selection helped in retaining those unknown loci. However, in some cases the responses in QPM versions were slightly different from the original hybrids. These minor changes in response to diseases are due to the presence of minor fragments of the donor parent genome. QPM versions of HM4, HM8 and HM9 have now been released and notified for commercial cultivation in their respective zones. The newly developed QPM hybrids with better protein quality, high grain yield and diverse adaptation offer promise in reducing nutritional insecurity.