The Six maize genotypes (planted in 3 replications and two seasons) were obtained from the research farms of the Maize Improvement Program at the International Institute of Tropical Agriculture (IITA), Ibadan, Nigeria. The genotypes were obtained through conventional breeding from different planting cycles and parent lineages and named as follows: PVA SYN HGBC₀ (provitamin A maize HGB cycle zero); PVA SYN HGBC₁ (provitamin A maize HGB cycle one); PVA SYN HGAC₀ (provitamin A maize HGA cycle zero), and DT STR SYN2-Y (yellow maize variety, used as control 1). The three PVA maize genotypes used for Kokoro production were selected from the group of six genotypes based on different levels of provitamin A content (high and low) in the maize flours. The three PVA maize genotypes (PVA SYN HGBC₀, PVA SYN HGBC₁, PVA SYN HGAC₀) were labeled Maize 1, Maize 2, Maize 3, respectively, and Maize 4 was DT STR SYN2-Y, a yellow landrace maize variety as control. Maize 1 to Maize 3 were used to produced Kokoro 1, Kokoro 2, and Kokoro 3, respectively, while Kokoro 4 is from the landrace yellow maize variety used as control sample 1. Commercial Kokoro (“Kokoro Egba”) produced from white maize was used as a control sample 2. The commercial Kokoro was purchased at Imashayi, Yewa North Local Government Area in Abeokuta, Ogun State, Nigeria, with coordinates 7°09′N 2°55′E bordering the Republic of Benin. It is a community known for its age-long tradition of processing Kokoro commercially.

Proximate composition analyses for moisture, ash, crude protein, crude fat, and crude fiber were carried out on the samples by the methods described by the Association of Official Analytical Chemists (AACC, 2005; AOAC, 2012). Carbohydrate content was determined by difference. All analyses were done in triplicate.

The method of Howe and Tanumihardjo (2006) was adapted for carotenoid analysis, and the quantity of β-cryptoxanthin, α-carotene and trans with cis-isomers of β-carotene were extracted as described by Alamu et al. (2014). Carotenoid extraction was done by weighing 0.6 g of the maize flour/Kokoro in a screw cap tube, 10 mL of ethanol, which contains 0.1% butylated hydroxyl toluene (BHT) was added and vortex thoroughly for 1 min. The tubes were placed in a water bath for 5 min at 85°C to achieve ethanol precipitation. After this, the tubes were taken out of the water bath, and 0.5 mL of 80% Potassium hydroxide (KOH) was added to the mixture for the saponification of any interfering oil. The samples were mixed using a vortex mixer for 30 s, after which the samples were placed in a water bath (85°C) for 5 min. The mixture was vortex again before returning to the water bath for another 5 min. The tubes were removed from the water bath and immersed immediately in an ice bath followed by the addition of 3 mL of cold deionized water. Separation of carotenoids was done three times with the addition of 3 ml of hexane using a vortex mixer for 10 s and then centrifuged for 10 s at 1,000 rpm using SORVALL GLC1 centrifuge (model 06470, USA). The mixture containing combined hexane fractions were washed with deionized water thrice, vortex, and centrifuged for 10 s at 1,000 rpm. The upper phase was pipette into a 15 mL concentrator tube and with the aid of the TurboVap LIV concentrator under nitrogen gas, the mixture containing hexane fractions was dried down. Reconstitution of the extract was done using methanol/dichloromethane (1 mL, 50:50 v/v), after which 100 μL aliquot was injected into the HPLC system for analyses of a-carotene, b-carotene (cis and trans isomers), and b-cryptoxanthin. Quantification of carotenoids was carried out by the use of Waters HPLC system (Water Corporation, Milford, MA), which consist of a guard column, C30 YMC Carotenoid column (4.6 9 250 mm, 3 lm), Waters 626 binary HPLC pump, 717 auto-sampler and a 2,996 photodiode array detector (PDA) with operating system Empower 1 software (Waters Corporation).

The constituent of solvent A includes; methanol: water (92:8 v/v) with 10 mmol/L ammonium acetate while solvent B includes; 100% methyl tertiary-butyl ether. The performance of gradient elution took place at 1 mL/min with other conditions such as; 29 min linear gradient from 83 to 59% A, 6 min linear gradient from 59 to 30% A, 1 min hold at 30% A, 4 min linear gradient from 30 to 83% A and a 4 min hold 83%. β-carotene eluted at 25 min. Generation of chromatograms was achieved at 450 nm, and identification of a-carotene, β-carotene (cis and trans isomers), and β-cryptoxanthin were determined using standard and external method based on the calibration curve from pure standards and verification of absorption spectrum and co-elution with available authentic standards. α-carotene, β-carotene, and β-cryptoxanthin standards were purchased from CaroteNature, GmbH (Lupsingen, Switzerland). The solvents used were HPLC grade.

Provitamin A (PVA) was obtained by the summation of the amount of β-carotene (13-cis-β-carotene, 9-cis-β-carotene, and trans isomers) to half of the amounts of a-carotene and β-cryptoxanthin (Alamu et al., 2014). α-carotene and β-cryptoxanthin have 50% of the provitamin A activity of β-carotene because when these compounds are centrally cleaved, there is a production of one molecule of retinal while β-carotene can produce two.

The percent apparent retention (AR) of carotenoids was determined by dividing the carotenoid content (μg g⁻¹) of the processed product (Kokoro) by the carotenoid content (μg/g) of unprocessed food (maize flour) (Bechoff et al., 2017) according to the formula:

The calculation for percentage contribution of vitamin A in Kokoro was done by using the provitamin A carotenoids values for each maize genotype through extrapolation. The Kokoro snack is usually packaged with an average weight of 100 or 150 g. In this study, we assumed these packaging sizes to be serving sizes. The serving size of Kokoro 100 g for boys and girls age 9–18 years and adult (female and male; >19) was used. Also, 150 g for adult (female and male; >19) per serving of Kokoro was estimated. The retinol equivalent was calculated using ratio 12:1, which is 12 μg of Provitamin A (PVA) is equal to 1 μg of Retinol Equivalent [Institute of Medicine (IOM), 2001].

The sensory evaluation was carried out on the Kokoro samples, and the commercially purchased Kokoro (control 2) for color, taste, crispness, and aroma sensory attributes. The overall acceptability of all samples was also rated in comparison with the control samples. The preference test based on the degree of likeness was used. In the test, 20 untrained panelists comprising male and female staff and students from Bowen University, Iwo, Osun State were engaged. These panelists were familiar with the quality attributes of Kokoro. The sensory evaluation questionnaires were based on a nine-point hedonic scale where 1 represents “dislike extremely,” and 9 represents “like extremely” for all the quality attributes employed (Iwe, 2002). The samples for evaluation were all placed at the same time in each of the well-illuminated booths, and the participants were seated into separate booth. The sitting of each panelist in a separate booth was to prevent bias decisions and produce an individual judgment. One ring each of three Kokoro samples (provitamin A maize), landrace yellow maize and commercially purchased Kokoro was presented on the white plastic plates and identified with a three-digit code. The samples were in triplicate in each panelist booth all at the same time making it a total of 15 samples per panelist with a serviette for wiping hands and a glass of clean water for rinsing the mouth in between the evaluations to prevent a carry-over effect.

All the data were statistically analyzed using SPSS 20. The analysis of variance (ANOVA) was used to establish a significant difference at p < 0.05. The means were separated using Duncan's multiple range test (DMRT). Tables 1, 3, 5 estimated four maize genotypes in two seasons with two replications in field and in three replications in the laboratory. Tables 2, 4 estimated four maize genotypes in two seasons with two replications in field, one commercial maize snack and three replications in the laboratory. Table 5 sensory evaluation had four maize genotypes, three replications in the sensory test with 20 panelist.

The carotenoid profile result before (maize flour) and after the production of Kokoro with the retinol equivalent (μg RE) from provitamin A maize genotypes, landrace yellow maize, and commercially purchased Kokoro are presented in Tables 3, 4. The carotenoids in PVA maize studied consist of the non-provitamin A (nPVA) and provitamin A (PVA) carotenoids. The nPVA carotenoids are lutein, zeaxanthin, and total xanthophylls, while the PVA carotenoids are made up of β-cryptoxanthin, α-carotene, and β-carotene (Alamu et al., 2014, 2015; Pillay et al., 2014; Awoyale et al., 2016). The result of variance (ANOVA) in Table 3 shows a significant difference in the carotenoid profile of maize flour.

Table 4 shows significant differences at p < 0.05 and can be attributed to carotenoid isomerization or degradation during Kokoro production caused by the various unit operations (boiling, steeping, wet milling, pre-frying, cooling, post-frying, and final cooling). The unit operations in the processing method led to the exposure of the food matrix to direct sunlight, heat, light, metals, and oxygen that resulted in the formation of cis-isomers, loss of provitamin A quantity (Aman et al., 2005; Ilg et al., 2009; Rodriguez-Amaya et al., 2011; De Moura et al., 2015). Oritz et al. (2018) stated that changes during dry milling of kernels might be related to thermal sensitive triggering isomerization/degradation reaction in carotenoid. Pillay et al. (2014) reported an increased concentration of available carotenoid retention in mealie meal, which is contrary to the result observed in this study. Awoyale et al. (2016) reported the production of ogi and its provitamin A retention. The retention results presented in the study are compared similarly before storage. In comparison with data on boiled maize presented by Alamu et al. (2015), reported much lower retention of β-carotene. The carotenoid retention in this study will be much lower because Kokoro undergoes series of processing stages.

Also, in Table 3, carotenoids were not detected in the commercially purchased Kokoro because it was produced from white maize, which lacks pigment (Ford, 2000). The retinol activity equivalent (RAE) in Table 3 shows that Maize 2 had the highest value while in Table 4, it was observed that Kokoro 1 maize snack's retinol equivalent RAE) at 100 g and 150 g/serving size was the highest while Kokoro 4 (yellow maize) at both serving sizes was the lowest.

The lutein content before processing ranged from 10.38 to 12.87 μg g⁻¹, while after processing, it ranged from 0.29 to 0.57 μg g⁻¹. It was observed that the Maize 1 genotype had the highest lutein content before and after processing, followed by Maize 2 with (cycle 1), which is from a different planting cycle. It implies that cycle 0 (Maize 1) can retain lutein compared to other maize samples. Maize 2 and Maize 4 flour samples showed high values of zeaxanthin of 10.97 and 11.05 μg g⁻¹, respectively, while Maize 1 had the lowest value of (6.03 μg g⁻¹). However, Kokoro from Maize 3 contained the highest zeaxanthin value (0.75 μg/g), but the original content of flour was (9.91 μg g⁻¹), while Maize 1 had the lowest zeaxanthin before and after processing, respectively (0.51; 6.03 μg g⁻¹). The result implies that Maize 3 genotype retained zeaxanthin content, although it had a low value initially before processing into Kokoro. Maize 1 was observed to be the lowest before and after processing. The retaining ability of Maize 3 might be attributed to the genetic variation and cycle of the seed. Zeaxanthin is reported to be isomeric with lutein, and it provides the primary yellow color in maize (Sajilata et al., 2008). Lutein and zeaxanthin are carotenoids with no vitamin A activity, but they possess anti-oxidative properties. Reports have been made as to its presence in the macular pigment of the human eye, reducing the risk of macular degeneration, cataracts, age-related macular degeneration, and other degenerative diseases (Pixley et al., 2013; Vishwanathan and Johnson, 2013; Bernstein et al., 2016).

The summation of lutein and zeaxanthin is xanthophylls, and after processing, severe degradation was observed in all the maize samples from the initial values. This could be due to the exposure to light, high temperature, and oxygen during Kokoro processing steps and degradation and isomerization have been reported to be caused by the factors mentioned above (Rodriguez-Amaya, 2001; Ambrosio et al., 2006; Boon et al., 2010). However, Kokoro 1 was observed to have had the highest value, which implies the highest degradation amongst the Kokoro from all other maize samples.

β-cryptoxanthin is a vitamin A precursor providing only one-half of the provitamin A activity of β-carotene. Maize flour β-cryptoxanthin ranged from 1.94 to 4.06 μg g⁻¹, and in Kokoro, it ranges from 1.25 to 1.50 μg g⁻¹. Maize 4 was observed to contain the highest β-cryptoxanthin amount (4.06 μg/g), and Maize 2 had the highest value (3.54 μg g⁻¹) among the PVA maize genotypes. Kokoro 2 and Kokoro 3 were not significantly different, while Kokoro 1 recorded the lowest value (1.25 μg g⁻¹). From the result, it was observed that β-cryptoxanthin was the most abundant provitamin A carotenoid among the Kokoro samples. Furthermore, it was observed that β-cryptoxanthin was generally higher in the Kokoro from PVA maize genotypes than the landrace yellow maize.

The α-carotene content before and after processing is presented in Tables 3, 4. It was observed that Maize 1 had the lowest content before and after processing, and Kokoro 3 had the highest content in Table 4. This implies that Kokoro 3 will contribute more to the α-carotene precursor during provitamin A activity. α-carotene is a precursor of vitamin A, providing provitamin A activity, although it is not as potent as β-carotene because it provides only one molecule of vitamin A [Institute of Medicine (IOM), 2001; Grune et al., 2010].

Trans-β-carotene composition in the maize flour samples ranged from 2.10 to 3.25 μg g⁻¹, while in the Kokoro samples, it was 0.74 to 1.54 μg g⁻¹. Maize 2 flour had the highest trans-β-carotene content (3.25 μg g⁻¹), and Maize 4 contains the lowest amount (2.10 μg g⁻¹). Kokoro 1 recorded the highest value of 1.54 μg g⁻¹ although, it had the lowest value (2.42 μg g⁻¹) among the PVA maize before processing, while Kokoro 4 (control 1) had the lowest value (0.74 μg g⁻¹). β-carotene, which is a precursor to vitamin A, and it is the most potent vitamin A active carotenoid because it can provide two molecules of retinol (Institute of Medicine (IOM), 2001; Grune et al., 2010). The result shows that Kokoro 1 has a high converting potential to vitamin A, and it would confer the highest provitamin A activity in comparison with other Kokoro samples from the maize genotypes.

For 13-cis-β-carotene, the maize flour samples ranged from 0.85 to 1.31 μg g⁻¹ while 9-cis-β-carotene ranges from 0.88 to 1.67 μg g⁻¹. The occurrence of cis-isomers (13-cis-β-carotene and 9-cis-β-carotene) can be attributed to the exposure and disintegration of the maize kernel matrix during pulverization to obtain the maize flour for the estimation of the carotenoid profile before processing. Maize 3 had the highest content (1.31 μg g⁻¹ of 13-cis-β-carotene, and Maize 1 had the lowest content (0.93 μg g⁻¹) among the PVA maize genotype while Maize 4 had 0.85 μg g⁻¹. Maize flour from Maize 2 recorded the highest 9-cis-β-carotene value (1.67 μg/g), and Maize 1 had the lowest content (1.34 μg g⁻¹) among the PVA maize genotypes. It can be implied from the result that Maize 3 and Maize 2 experienced isomerization faster to form a high content of 13-cis- and 9-cis, respectively, during the pulverization of maize grains to maize flour. In contrast, the isomerization formation rate during milling of Maize 1 grains was low. The 13-cis-β-carotene values for Kokoro ranged from 0.32 to 0.63 μg g⁻¹ while 9-cis-β-carotene was from 0.41 to 0.93 μg g⁻¹. Kokoro 1 was observed to have undergone the fastest isomer formation among the maize genotypes with the highest 13-cis value of (0.63 μg g⁻¹) and 9-cis (0.93 μg g⁻¹), respectively. In contrast, Kokoro 3 had the lowest isomer formation content for both 13-cis (0.57 μg g⁻¹) and 9-cis (0.79 μg g⁻¹). The exposure of the Maize 1 genotype to many unit operations during Kokoro processing brought about the high amount of 13-cis and 9-cis formation. In contrast, when exposed to short processing (pulverization to maize flour), isomer formation quantity was low. Aman et al. (2005) reported that the occurrence of carotenoid isomers like 13-cis and 9-cis and 15-cis β-carotene is attributed to varietal difference, sample storage, handling, and food processing. Also, the presence of elevated amounts of isomer formation in Kokoro 1 can cause loss of color, low/loss of vitamin A activity, and less absorption ability in humans compared with the trans component of the carotenoid.

Total β-carotene was found to be the highest in Maize 1 (3.10 μg g⁻¹) after processing to Kokoro 1 from the initial value of 4.69 μg g⁻¹ while Control 1 had the lowest total β-carotene in Kokoro of 1.47 μg g⁻¹ from 3.83 μg g⁻¹. Total β-carotene is the summation of trans, 13-cis, and 9-cis isomer. It can be implied from the result that Kokoro 1 carotenoids will be converted more to vitamin A.

Provitamin A (PVA) activity is derived from the summation of the amount of β-carotene (13-cis-β-carotene, 9-cis-β-carotene, and trans isomers) to half of the amounts of a-carotene and β-cryptoxanthin (Alamu et al., 2014). The provitamin A content of Kokoro samples ranged from 2.43 to 4.00 μg g⁻¹, and they are significantly different from one another. Kokoro 1 had the highest value (4.00 μg g⁻¹), while Kokoro 4 (control) had the lowest value (2.43 μg g⁻¹). Kokoro 1, with the highest PVA quantity among the Kokoro samples, will have more conversion ability to vitamin A in the body. However, the amount in μg g⁻¹ is lower in comparison with the HarvestPlus target value of 17 μg g⁻¹ (dry weight) (Mulualem, 2015). In this study, it was observed that Maize 1 flour had a low quantity (5.96 μg g⁻¹ before processing but retained more in Kokoro (4.00 μg g⁻¹) while the highest PVA content in maize flour was Maize 2 (8.43 μg g⁻¹) with a product of 3.87 μg g⁻¹. However, Kokoro 3 had the lowest PVA content (3.77 μg g⁻¹) amongst the PVA maize genotype with an initial value of 8.13 μg g⁻¹. Thus, it could be inferred that the retention of PVA is genotype dependent. Alamu et al. (2015) who reported that PVA retention for boiled maize was genotype dependent. This finding has corroborated the overall objective of this study that aimed to test the improved and promising pipelines genotypes for suitability for product development.

The carotenoid apparent retention result for Kokoro is presented in Table 5. The knowledge about the quantity of carotenoids present after processing is essential to determine the retention of the carotenoids and for increasing breeding targets during the creation of new breeds. It was observed in the profile that lutein and zeaxanthin had low percentage values compared to other carotenoids present. The degradation could be due to the polienic chain in the xanthophylls (lutein and zeaxanthin) causing an increased light absorption. However, this could result in high susceptibility to oxidative degradation and geometrical isomerization due to the exposure to light, high temperature, and oxygen in the many unit operations during processing (Britton et al., 1995). Kokoro 1 had an apparent PVA retention of 67.05%, Kokoro 2 had 45.88%, Kokoro 3 had 47.34%, and Kokoro 4 (control 1) 37.98%. The processing of these maize genotypes into Kokoro has many operations (boiling, steeping to ferment, wet milling, cutting, rolling, pre-frying, cooling, and post-frying) with a lengthy process duration (3 days). The increment in surface area of the maize resulted in the exposure of the carotenoid to the mechanism of isomerization/degradation through light, heat, atmospheric oxygen, etc. (Rodriguez-Amaya et al., 2011), which brought about a reduction in micronutrient level (carotenoid). Based on the result, Kokoro 1 had the highest retention value, which implies the presence of more micronutrient (carotenoid) compared to the others. The retention value observed could be a result of genes.

The percentage contribution of PVA-Kokoro per day for boys, girls, men, and women at 100 g and 150/serving sizes is presented in Table 6. Kokoro 1 (PVA SYN HGBC₀) for all the servings shows the highest percentage contribution of vitamin A while Kokoro 4 (control 1) was the lowest. The result in Table 6 illustrates that PVA-Kokoro has almost twice the percentage contribution ability than yellow maize. This is important because the PVA maize genotypes contain increased carotenoid quantity targeted at reducing vitamin A deficiency.

The sensory attributes of Kokoro are presented in Table 7. All attributes showed significant (p < 0.05) differences among all the samples except aroma and color characteristics. The color values ranged from 6.55 to 6.83 and were not significant from each other. The Kokoro samples from the PVA maize genotypes with Kokoro 4 (control 1) and commercial Kokoro (control 2) were liked slightly. This implies the panelists cared less about the color and had a slight likeness, whether it is white or yellow. The preference of the taste ranged from neither like nor dislike to liking moderately. It was observed that Kokoro 1 and 2 taste alike and were preferred among PVA maize genotypes, while Kokoro 4 (control 1) is similar in taste to commercial Kokoro (control 2). Based on the result, Kokoro 1, 2, and 4 (control 1) are familiar in taste to commercial Kokoro (control 2). The aroma parameter showed no significant differences when compared to the commercially purchased Kokoro due to the same range of mean scores obtained in Table 7. This can imply that the aroma was the same or in a close range as the commercially purchased Kokoro. The crispness of Kokoro 4 (control 1) and commercial Kokoro (control 2) were the most preferred. The least preferred was Kokoro 3, and it was neither liked nor disliked. Color and aroma did not influence the sensory perception of the panelists. However, taste and crispness might have had a substantial effect on the overall acceptability. Kokoro 4 (control 1) and commercial Kokoro (control 2) had a high mean sensory score in overall acceptability, and among the PVA maize genotypes, Kokoro 1 was the highest. Kokoro 3 sample was observed to be the least preferred in terms of taste, color, crispness, aroma, and overall acceptability. However, all the samples had general acceptability for all evaluated sensory attributes as none scored below the minimum acceptable rating of 5 on the 9-point hedonic scale reported by Adebowale and Komolafe (2018), Idowu (2015), and Ayinde et al. (2011).

Kokoro 4 (control 1) and commercially purchased Kokoro had values that were very close to each other for all parameters estimated. This can mean that Kokoro 4 was not significantly different from the purchased commercial Kokoro (control 2), Kokoro 4 is the most acceptable sample. However, among the PVA maize genotypes, Kokoro 2 had the highest overall acceptability mean score, implying it is acceptable. Various authors have reported the increase of nutrient quality acceptability of Kokoro with the aid of supplementation with legumes; Ayinde et al. (2011) reported that the Kokoro composition of 90:10 maize-beniseed blend was the most preferred while 60:40 was the least accepted and this could be due to the nutty beniseed. Abegunde et al. (2014) stated that the supplementation of Kokoro with cowpea caused a decreased acceptability with increase in cowpea flour.

Furthermore, Dauda et al. (2020) reported that maize snack (Kokoro) fortified with groundnut (10 and 20%) was the most preferred. Various authors have also reported the acceptability of other food product from biofortified crops. Ukpabi et al. (2012) stated the potential use of orange-fleshed sweet potato (OFSP) chips and reported low acceptability between varieties. Ukom et al. (2019) reported increased nutrient in maize ogi with supplementation of OFSP and African yam bean and Alamu et al. (2014) reported high acceptability of roasted orange maize hybrid without husk.