Assessment of the phenolic compounds of pearl and finger millets obtained from South Africa and Zimbabwe

Abstract Millet grains are rich in phenolic compounds which have health benefits. This study aims to elucidate the phenolic properties of pearl and finger millet grown in South Africa and Zimbabwe. The milled samples were analyzed by Waters Synapt G2 Quadrupole time‐of‐flight (QTOF) mass spectrometer (MS). A total of eight phenolic compounds were detected and quantified in the millet varieties, which included derivatives of benzoic acid such as protocatechuic and p‐hydroxybenzoic acids. Flavonoids such as catechin, epicatechin, procyanidin B1, procyanidin B2, and kaempferol glycoside were also detected. Generally, catechin was the dominant phenolic compound, followed by epicatechin. The mean values for catechin ranged from 2.50 to 12.6 mg/kg for the pearl millet and 610.4 to 675.1 mg/kg for the finger millet. While the epicatechin mean values ranged between 1.2 to 1.8 for pearl millet and 99.1 to 139.5 for finger millet. Procyanidins B1 and B2 were only detected and quantified in the finger millet types, While Kaempferol glycoside was only recorded in the pearl millets with mean values of 196.0 mg/kg for pearl millet South Africa and 213.6 mg/kg for pearl millet Zimbabwe. There was a difference among the varieties for the content of kaempferol glycoside. Protocatechuic and p‐hydroxybenzoic acids were only present in the finger millet types, their mean values were (20.9, 23.7 mg/kg) and (16.8,13.5 mg/kg) respectively. It can be substantiated from the outcome of this study that millet can be used as a source of valuable phenolic compounds and that the variety of millet is the determining factor of the phenolic compound content.

. Among the physiognomies of millet which need to be explored are the phenolic compounds.
Phenolic compounds are a class of secondary metabolites found in plants and further divided into phenolic acids and polyphenols (Giada, 2013). Furthermore, phenolic acids are divided into two classes: hydroxybenzoic and hydroxycinnamic acids (Dykes & Rooney, 2006). The main dietary phenolic compounds include the phenolic acids, flavonoids, and tannins (King & Young, 1999). The phenolic acids and flavonoids are also considered vital in promoting health by reducing the risk of metabolic syndrome and the related complications of type 2 diabetes (Lin et al., 2010). To date, there is no uniform set of identified and quantified phenolic compounds in different millet varieties. Chethan, Sreerama, and Malleshi (2008) identified nine phenolic acids which include gallic acid, protocatechuic acid, p-hydroxybenzoic acid, vanillic acid, ferulic acid, syringic acid, trans-cinnamic acid, and p-coumaric acid, in millets. On the other hand, Chandrasekara and Shahidi, (2011) found that hydroxycinnamic acids and their derivatives were the main contributors to the total phenolic compounds of insoluble-bound phenolic fraction of millet varieties. However, in another study by Xiang, Apea-Bah, Ndolo, Katundu, and Beta (2019), flavonoids were found to be the predominant phenolic compound in different millet varieties, whereas Sharma, Sharma, Handa, and Pathania (2017) reported higher amounts of phenolic content and antioxidant activity in methanolic extracts of kodo millet grains. In like manner, Pradeep and Sreerama (2015) found that kaempferol was the most abundant flavonoid in raw millet varieties. In comparison, the most dominant phenolic compounds in maize varieties are phenolic acids followed by flavonoids (García-Salinas, García-Salinas, Alemán-de la Torre, & Ramírez-Díaz, 2017). It is, however, important to realize that different analysis methods also affect the total phenolic compound contents of plants. Chethan and Malleshi (2007) argue that although different solvents are being used to extract phenolic compounds from plant foods, acidified methanol is the best organic solvent for extraction of phenolic compounds from millets. Other factors such as environmental conditions, cultivar of a plant, processing conditions, and storage are also found to affect the quantity of phenolic compounds in plants. Environmental factors such as sun exposure, soil type, and rainfall influence the phenolic content of plants (Manach, Augustin, Morand, Remesy, & Jimenez, 2004). Similarly, Shahidi and Naczk (2004) also reported that the type and the content of the phenolic compounds of grains depend on the type of millet, variety, part of the grain, climatic conditions, and cultivation practices.
More studies to expansively profile the phenolic compounds of millet types in different locations are therefore warranted. With all these interesting incites and the importance of millet grains mentioned above, studies on phenolic compounds in millet are limited, particularly in the Southern African region. This study aims to investigate the phenolic compounds of pearl and finger millet grains obtained from South Africa and Zimbabwe by LC-MS method using the Synapt G2 qTOF from Waters (Milford, USA).  The Zimbabwean millet samples were obtained from local market in Harare, Zimbabwe. The samples were milled using a grinder and sifted to produce a fine texture. The samples were prepared in duplicate and ready for further analysis.

| Extraction of phenolic compounds
The extracts were prepared by using 2 g dry millet material + 15 ml 50% methanol/1% formic acid in water with ultrasonication for 1 hr and standing overnight, followed by centrifugation and transfer of the supernatant to a glass vial ready for the LC-MS analysis.

| LC-MS analysis
The samples were analyzed by LC/MS quadrupole time-of-flight (QTOF) mass spectrometer (MS) connected to a Waters Acquity ultra-performance liquid chromatograph (UPLC) (Waters, Milford, MA, USA) was used for high-resolution UPLC-MS analysis. Electrospray ionization was used in negative mode with a cone voltage of 15 V, desolvation temperature of 275°C, desolvation gas at 650 L/h, and the rest of the MS settings optimized for best resolution and sensitivity. Data were acquired by scanning from 150 to 1,500 m/z in resolution mode as well as in MSE mode. In MSE mode, two channels of MS data were acquired, one at a low collision energy (4 V) and the second using a collision energy ramp (20−60 V) to obtain fragmentation data as well. Leucine enkephalin was used as lock mass (reference mass) for accurate mass determination, and the instrument was calibrated with sodium formate. Separation was achieved on a Waters HSS T3, 2.1 × 100 mm, 1.7 μm column. An injection volume of 2 μL was used, and the mobile phase consisted of 0.1% formic acid (solvent A) and acetonitrile containing 0.1% formic acid as solvent B. The gradient started at 100% solvent A for 1 min and changed to 28% B over 22 min in a linear way. It then went to 40% B over 50 s and a wash step of 1.5 min at 100% B, followed by re-equilibration to initial conditions for 4 min. The flow rate was 0.3 ml/min, and the column temperature was maintained at 55°C.

| Statistical analysis
Data obtained were subjected to a one-way analysis of variance of the statistical analysis software (SAS, 2010). Where mean values were statistically significant at p < .05, Duncan's multiple range test was used. The Principal component analysis (PCA) analysis was conducted using PAST version 4.02, a software for scientific data analysis, with functions for data manipulation, plotting, univariate and multivariate statistics analysis. Table 1 shows the retention time, fragments, elemental composition, and tentative identification of phenolic compounds detected in the four millet varieties. Total of eight (8) phenolic compounds were detected and quantified in the millet types, which included derivatives of benzoic acid such as protocatechuic acid and p-hydroxybenzoic acid. Flavonoids such as catechin, epicatechin, procyanidin B1, procyanidin B2, and kaempferol glycoside were also detected. Tryptophan was the only amino acid detected in the millet varieties. Citric acid was also detected but not quantified in any of the varieties. Numerous unknown compounds were also detected during the analysis which were later assigned to suggested compounds using the elemental composition, through SCI-FINDER application. Generally, finger millet varieties had more phenolic compounds than the pearl millet types, irrespective of where they were grown.

| Total flavonoids
The contents of flavonoids in the millet varieties determined by the LC-MS method are shown in (Table 2). Finger millets were recorded to have higher flavonoids than those found in the pearl millet types, indicating varietal differences. Catechin was the dominant flavonoid ( Figure 1a

| Phenolic acids
The first compounds to be identified were citric acid at a retention time of 1.69 min; however, its fragment was not detected, and it was not quantified in any of the millet types, as shown in Table 1

| Principal component analysis (PCA)
Principal component analysis was carried out using the methods of Hammer, Harper, and Ryan (2001) to further investigate possible variations among the millet varieties. Figure 3 shows a loading of procyanidin B1 and B2, and p-hydroxybenzoic acid in the same area on the right middle side of the score plot, indicating possible link between these phenolic compounds. In the same way, catechin, tryptophan, and kaempferol glycoside clustered at the right topside of the plot. In the model plotted, the principal components (PCs) explained variance of 99.96% of the data. The two principal components PC1 and PC2 had variabilities of 99.8990 and 0.05817, respectively, as listed in Table 3. The loading plots shown in Figure  Finger millet varieties were seen as the main contributor to the clustering on the right side of the cluster confirming that they contain most of the phenolic compounds.
As shown in Table 3

| Pearson correlation (r) between the phenolic compounds
Pearson correlation was performed to investigate the correlations among the individual phenolic compounds ( Figure 5). The results indicate that strong positive correlation exists between protocatechuic acid and catechin, epicatechin, p-hydroxybenzoic acid, procyanidin B1, and B2. Exception was observed between protocatechuic acid and both tryptophan and kaempferol where a strong negative correlation existed. The same results were observed for catechin and epicatechin. There was also strong correlation between p-hydroxybenzoic acid strongly correlated positively with procyanidin B1 and B2 and strong negative correlation with kaempferol. Pearson correlation plot in Figure 5 showed a significant negative correlation between p-hydroxybenzoic acid and both epicatechin and tryptophan as indicated by the blank space in the figure. A general positive correlation among the phenolic compounds is evident as indicated by the blue circles.

| Unknown compounds
A total of 4 unknown compounds with identified elemental formulas were also detected during the analysis. These compounds as presented in Figure 6 were assigned to possible compounds using SCI-FINDER®. They include acetic acid, 2,2',2''-(5.28-dioxopentacyclo), glutaric acid, benzopyran, and b-d-glucopyranoside.

| CON CLUS ION
This study reports the profile of the phenolic compounds in the millet grains and confirms that the LC-MS-based profiling is a good technique for the phenolic characterization. It also substantiates the fact the millet grains are good source of nutrition and also beneficial for the health status of those who rely on it as staple food, due to the presence of nutraceutical components such as antioxidants and polyphenols. Because millet flavonoids play important roles in the prevention and management of type 2 diabetes, the finger millet varieties fit perfectly in terms of health benefits, as it contains the flavonoids more than the pearl millet types. The results of this study confirm that the type of millet was the determining factor of the phenolic contents. These results have provided useful information for effective utilization of millets as functional food ingredients for promoting health. Broader profiling of different millet varieties to include most of the varieties present in the Southern African region is encouraged.
The results of the current study demonstrated difference among the varieties of millet rather than the environmental effects, since finger millet varieties had more phenolic compounds than the pearl millet varieties regardless of the area of cultivation.