Food spoilage is of major concern to the food industry, because it leads to considerable
economic losses, a deteriorated environmental food-print, and to possible public health
hazards. In order to limit food spoilage, research on the preservation of food products has
always received particular attention by the food industry. Traditionally, such efforts have
mainly relied on the application of chemical preservatives or drastic physical treatments.
However, chemical preservatives are becoming increasingly unpopular by the consumers,
and some have even proven to be toxic and linked to cancer and other health problems.
Physical treatments of the products, on the other hand, can deteriorate the sensory
properties of the products, and may even destroy some of the nutrients and vitamins. In this
context, biopreservation, which is defined as the use of safe antibacterial/antifungal
microorganism (so-called protective cultures) has unexploited potential to inhibit the
growth of pathogenic microorganisms and enhance the shelf life of the final food product.
In order to apply biopreservation in food products effectively, detailed knowledge on the
metabolism of protective cultures is required. The present PhD project is mainly focused on
the application of in vitro NMR spectroscopy for studying the metabolism of protective
cultures. As an important part of this work, an analytical protocol was developed for realtime
in vitro NMR measurements of bacterial fermentation, which includes guidelines from
the sample preparation to the data processing and the modelling of the metabolic profiles.
The protocol is applied in an experimental design with two strains of lactic acid bacteria.
The results highlight some of the metabolic differences between the strains, in terms of
nutrients consumption and metabolites kinetics. As a part of this work, an NMR data
preprocessing technique, called ‘Reference Deconvolution’, was employed for the first time
to improve the multivariate analysis of the in vitro real-time metabolomics data and proved
a necessary and elegant solution to the inherent inhomogeneity problem of the samples in
the in vitro NMR measurements of cells. A second objective of the project was to develop
an accurate approach for quantifying mold growth and inhibition. A new method was
presented for quantifying mold growth and measuring different segments of mold colonies,
based on multispectral images and k-means clustering. The method was developed into a
software package called ‘PCLUSTER’, and was demonstrated to be very helpful in two other
biopreservation related metabolomic studies. In one case, PCLUSTER was used to quantify
how the concentration of diacetyl affects inhibition of the indicator molds and in the second
case PCLUSTER served as an efficient tool for quantifying inhibition assays, and finding
antifungal metabolites and metabolites that correlated positively/negatively with the
inhibition. The developed analytical tools are expected to be very beneficial in the studies
related to the biopreservation, and will be used in the future investigations of the protective
cultures.