Influence of distinct process-induced (micro)structures on the in vitro starch digestibility of common beans: A kinetic approach

Transitioning towards food systems favourable for both human health and environmental sustainability has been recently identified as an immediate challenge. In this context, a significant increase in consumption of plant-based foods (e.g. cereals, fruits, legumes, nuts, vegetables) is being strongly promoted to improve health and environmental benefits. Consequently, the development of safe and nutritious food products from sustainable sources is an important action path to which food scientists and technologists could and should contribute. Pulses, dry seeds of the legume family, constitute an interesting food group for tackling the diet-health-environment challenge. Different macro- and micro‑nutrients are present in these seeds, making part of an intrinsically complex structural network. Changes to this network are induced by processing, which is required to induce palatable conditions and improve digestive functionality. Process-induced structural changes are likely to have a determinant effect on the digestive response of pulses. Nevertheless, fundamental studies aiming to unravel their process‑structure-digestive functionality relationships are scarce.

In this doctoral research, it was hypothesised that processing could be used as engineering strategy to generate common bean (micro)structures with tailored in vitro starch digestion kinetics. Common beans were chosen as representative case study due to their importance in terms of production and human consumption. Starch was selected as component of interest because it is the most bio-encapsulated nutrient in pulses. Initially, process-structure relationships were established by applying different (combinations of) processing variables to common beans, followed by characterisation of structural properties at the macro and micro levels as a function of processing time. Next, (process-induced) structure‑digestive function relationships were studied by in vitro simulated experiments where the kinetics of starch digestion during the small intestinal phase was qualitative and/or qualitatively evaluated in samples with specific (micro)structural properties.

Application of conventional (95°C, 0.1 MPa, f(t)) and alternative (25°C, 600 MPa, f(t) or 95°C, 600 MPa, f(t)) treatments showed that palatable common beans were only achieved when high temperature was included as process variable. Contrarily, common beans subjected to 600 MPa and 25°C failed to soften regardless of processing time. Similar softening profiles were obtained upon application of the two treatments at high temperature, either in combination with high hydrostatic pressure or not. These two treatments, followed by standardised mechanical disintegration, also resulted in similar microstructural properties. Specifically, cell clusters and open cells were characteristic microstructures as a result of short processing times, probably because pectin in the middle lamella and primary cell wall was not yet (extensively) solubilised. An evolution towards well separated, individual closed cells was observed as processing time increased, most likely driven by thermally-induced pectin solubilisation. Given that this microstructural characterisation was always performed after a standardised step of particle size reduction, an open question remained regarding the role of human mastication as mechanical disintegration technique. Therefore, a mastication study including 20 participants was performed. Samples subjected to the study were thermally processed common beans with different process-induced hardness levels (95°C, 0.1 MPa, f(t)). From this experiment, it was concluded that individual mastication behaviour did not significantly influence the particle size distribution of oral boluses, while thermal process intensity (i.e. residual hardness level) did. The most representative starch-rich fraction of oral boluses was constituted by individual closed cells, which relative contribution to the total bolus mass was higher at longer processing times. A simultaneous decrease in the mass percentage of larger starch‑containing fractions (e.g. cell clusters) occurred with the increase of individual cells.

Individual cotyledon closed cells were the most characteristic fraction of palatable common beans after being subjected to in vitro as well as to in vivo mechanical disintegration. This anticipated a key role of the remaining starch‑surrounding barriers (cell wall and protein matrix) and their process‑induced status during subsequent digestion. In vitro simulated digestion of both heterogeneous oral boluses and homogeneous isolated cotyledon cells from thermally processed common beans resulted in significant differences at the level of starch digestion kinetics. These differences were quantified by mathematical modelling of the kinetic data using a reparametrized logistic model, which was selected for the first time in the context of starch digestion in this work. The quantified differences were linked to process-induced modifications at the level of starch-surrounding barriers. More in detail, longer lag phases in samples obtained after shorter processing times were related to cells with lower cell wall permeability and/or a more compact protein network. Similarly, higher reaction rate constants in samples obtained after longer thermal treatments were linked to the amount of enzyme binding and breaking down starch chains per minute, following its movement across and around the above mentioned starch-surrounding barriers.

In the last part of this doctoral research, mechanistic insight into the in vitro starch digestion process of cotyledon cells isolated after thermal treatment of common beans (95°C, 0.1 MPa, f(t)) was obtained. To this end, digests were separated into supernatant and pellet fractions. Subsequently, the isolated fractions were subjected to different qualitative and quantitative characterisations. It was demonstrated for the first time that digestion of cellular structures from common bean cotyledons occurs inside the cell, regardless of thermal processing time. Additionally, a digestion mechanism of three steps was postulated. This mechanism comprises (i) enzyme diffusion through the cell wall followed by overcoming of any hindrance exerted by the cytoplasmatic contents (e.g. protein matrix), (ii) adsorption on and hydrolysis of starch granules from the periphery towards the core of the cell, combined with (iii) diffusion of digestion products towards the outside of the cellular space. Parallel protein disappearance, possibly following the same digestion mechanism, was also proven by microscopic evaluation. Finally, different rate‑limiting steps for starch digestion in common bean cotyledon cells were postulated depending on the intensity of the thermal process applied.

Overall, this PhD research evidences the potential of targeted processing as engineering strategy to modulate in situ the in vitro starch digestion kinetics of common beans through modification of their structural properties. The processing conditions applied in this work can be easily reproduced at both household and industrial levels, demonstrating its promising applicability potential.