Abstract
Mahewu is one of the most popular indigenous fermented meals consumed in Southern
Africa. It is usually produced through a natural fermentation process. Maize (Zea mays) is
one of the multi-purpose cereal crops grown in South Africa and is the key raw ingredient
used to make mahewu. The most important factors in the processing of mahewu are boiling
time and fermentation (time and temperature), as these are responsible for achieving a
consistent and quality product. This study thus optimised boiling time and fermentation (time
and temperature) of white and yellow maize mahewu, and investigated the physicochemical,
bacteria diversity and composition (nutritional constituents and metabolites) of the optimal
products. Optimization of the mahewu was done utilizing the Box Behnken design (BBD) of
the response surface methodology (RSM). The optimal processing conditions obtained based
on the physicochemical properties studied [pH, titratable acidity (TTA) and total soluble
solids (TSS)] were 25°C for 54 h and a boiling time of 19 min for white maize mahewu
(WM), and 29°C for 72 h and a boiling time of 13 min for yellow maize mahewu (YM).
Samples derived from these optimal conditions were made up for other objectives of this
study. The bacterial composition of the optimized mahewu products and raw materials used
were evaluated using amplicon sequencing. Total phenolic content (TPC), total flavonoid
content (TFC) and tannin content (TC), as well as antioxidant activities using (2,2-Azinobis
(3-Ethyl-Benzothiazone-6-Sulfonic acid)) (ABTS) and (2,2-diphenyl-1-picrylhydrazyl)
(DPPH) of the raw maize and derived mahewu samples were also measured. The
physicochemical characterization of mahewu samples and raw maize flour (RW and RY) was
analysed using scanning electron microscopy (SEM), Fourier transform infrared (FTIR)
spectroscopy and X-ray diffraction (XRD). In addition, proximate composition, minerals,
amino acids and untargeted phenolic composition were investigated. The last objective of this
study then adopted an untargeted metabolomics approach through the means of gas
chromatography high-resolution time of flight mass spectrometry (GC-HRTOF-MS) to
investigate significant (p ≤ 0.05) metabolites in the optimized mahewu samples and related
raw materials. The results showed significant (p ≤ 0.05) inverse relationships between the pH
and TTA among the samples. The optimal mahewu products’ pH varied between 3.41 and
4.51 for WM and 3.44 to 4.65 for YM, while their TTA ranged between 0.47 to 0.68 % for
WM and 0.38 to 0.62 % for YM. A reduction of the TSS from approximately 6.0 °Brix to
between 4.67 to 5.17 for WM and 4.80 to 5.53 for YM, was also observed. The study further
revealed the first report on the bacterial microbiota predominantly associated with optimized
ii
mahewu, which included Paenibacillus, Stenotrophomonas, Pseudomonas and Lactococcus.
The most abundant phylum in the flour samples was protobacteria, while Firmicutes was
most prevalent in the malt samples. The most abundant genus in the flour samples was
Pseudomonas, while Lactococcus dominated in the malt samples. The TPC and TFC were
higher than the TNC reported for the optimal mahewu samples. The increase in TPC and TFC
after fermentation led to an enhanced antioxidant capacity in the optimal mahewu samples.
The decrease in tannin content may have been due to acyl hydrolases activity produced by
fermenting microorganisms during fermentation. Profiling of the phenolic compounds
revealed the presence of coumarins and derivatives, quinic acid, terpene lactones,
macrolactams, tryptamines and derivatives, O-glycosyl compounds, O-glycosyl compounds,
hydroxyanthraquinones, taxanes and derivatives, long-chain fatty acids, resorcinols, 2,2-
dimethyl-1-benzopyrans, lineolic acids and derivatives. Variations in phenolic compounds
were also observed, possibly as a result of structural configuration, phenolic complexity and
enzyme specificity of these compounds. SEM, FTIR and XRD showed promising potential
evidence to understand the morphological changes, functional properties and peak diffraction
intensities in the mahewu samples as well as in the RW and RY flour. Fermentation
influenced the proximate composition of the optimized mahewu samples, with significant (p
≤ 0.05) improvement in the protein and carbohydrate content. Increases in mineral
composition of sodium, magnesium, phosphorous, potassium, calcium, manganese, iron,
copper and zinc were observed after fermentation of optimal mahewu samples in this study.
Potassium was the most abundant in the optimal mahewu samples, with values ranging from
3051.61 to 3283.38 (YM) and 2882.11 to 3129.97 (WM). Interestingly, heavy metal levels
were in compliance with the Joint FAO/WHO Expert Committee on Food Additives
recommended levels (JECFA). A trend of increase/decrease was observed in the essential and
non-essential amino acids. Arginine and leucine were the most abundant essential amino
acids, while for non-essential amino acids, glutamic acid, aspartic acid, alanine and proline
were found to be abundant. The GC-HRTOF-MS metabolomics technique further facilitated
an understanding of significant metabolites in the raw materials and optimal mahewu
samples. The unsupervised analysis technique utilized in this study, known as principal
component analysis (PCA), efficiently divided analyzed samples into several clusters based
on inocula, raw white and yellow maize type, boiling time and fermentation (time and
temperature). Thereafter, the dataset was subjected to a supervised analysis technique,
‘orthogonal partial least square discriminant analysis’ (OPLS-DA), which identified relevant
significant (p ≤ 0.05) metabolites, contributing to differences in the various samples clusters
iii
as revealed by the PCA. Notably, phenol, ketone, fatty acids, benzene, ester, hydrocarbon,
terpenoid, vitamin and phytosterol differentiated RW from RY flour. Classes of aromatic
compounds, phenols, ketones, esters, fatty acids, benzenes, hydrocarbons, vitamins,
terpenoids and amides were common to all inocula. Furthermore, following the preparation of
RW and RY flour into optimal mahewu samples, an increase of 2,4-Di-tert-butylphenol and
phenol, 2,2'-methylenebis[6-(1,1-dimethylethyl)-4-methyl- was observed in white maize
mahewu with either millet malt or malted white maize. Hexadecanoic acid, 2-hydroxy-1-
(hydroxymethyl)ethyl ester, hexadecanoic acid methyl ester, methyl stearate and n-
Hexadecanoic acid increased for yellow maize mahewu. In conclusion, optimal mahewu
products utilizing white and yellow maize with various inocula will help to increase the
supply of stable and nutritional fermented foods for reducing hunger and malnutrition. This
study also gives a thorough review of the significant metabolites present in optimal mahewu
products (WM and YM) and associated raw materials, which can serve as a foundation for
the quality control of these products for further studies.