Search

Understanding taste is key for optimizing the palatability of seaweeds and other non-animal-based foods rich in protein. The lingual papillae in the mouth hold taste buds with taste receptors for the five gustatory taste qualities. Each taste bud contains three distinct cell types, of which Type II cells carry various G protein-coupled receptors that can detect sweet, bitter, or umami tastants, while type III cells detect sour, and likely salty stimuli. Upon ligand binding, receptor-linked intracellular heterotrimeric G proteins initiate a cascade of downstream events which activate the afferent nerve fibers for taste perception in the brain. The taste of amino acids depends on the hydrophobicity, size, charge, isoelectric point, chirality of the alpha carbon, and the functional groups on their side chains. The principal umami ingredient monosodium l-glutamate, broadly known as MSG, loses umami taste upon acetylation, esterification, or methylation, but is able to form flat configurations that bind well to the umami taste receptor. Ribonucleotides such as guanosine monophosphate and inosine monophosphate strongly enhance umami taste when l-glutamate is present. Ribonucleotides bind to the outer section of the venus flytrap domain of the receptor dimer and stabilize the closed conformation. Concentrations of glutamate, aspartate, arginate, and other compounds in food products may enhance saltiness and overall flavor. Umami ingredients may help to reduce the consumption of salts and fats in the general population and increase food consumption in the elderly.

Umami (“hartigheid”) is één van de vijf basissmaken en is in het begin van de vorige eeuw ontdekt in zeewier. Umami wordt veroorzaakt door bepaalde vormen van aminozuren, en daar wilden we het fijne van weten. In dit project hebben we eiwitten en de aminozuren waaruit die zijn opgebouwd chemisch gekarakteriseerd en sensorisch beoordeeld. Op termijn is het perspectief om uit zeewier, eendenkroos en reststromen van gewassen eiwitcomponenten te isoleren met toegevoegde waarde qua smaak.

Umami (“hartigheid”) is één van de vijf basissmaken en is in het begin van de vorige eeuw ontdekt in zeewier. Umami wordt veroorzaakt door bepaalde vormen van aminozuren, en daar wilden we het fijne van weten. In dit project hebben we eiwitten en de aminozuren waaruit die zijn opgebouwd chemisch gekarakteriseerd en sensorisch beoordeeld. Op termijn is het perspectief om uit zeewier, eendenkroos en reststromen van gewassen eiwitcomponenten te isoleren met toegevoegde waarde qua smaak.

In 2050 moeten grote zeewierplantages op zee voorzien in onze behoefte aan voedsel en grondstoffen. In de Oosterschelde doen Wageningse onderzoekers nu experimenten met duurzame zeewierteelt.

Hoe worden we gezond oud? Trainen en extra eiwitten werken, blijkt uit onderzoek van de afdeling Humane voeding. Over het effect van vitamines en omega-3 vetzuren is het laatste woord nog niet gezegd.

Protein undernutrition is a major health concern for older adults, especially for those who are ill. There is growing consensus for a protein intake target of 1.2 - 1.5 gram per kg bodyweight per day (g/kg/d) for these older adults. However, this target is not reached by the majority of older adults. Therefore, more effective and novel strategies to increase protein intake are warranted, including the use of protein-enriched foods and drinks. This thesis evaluated the impact of the developed protein-enriched foods and drinks on protein intake and physical performance among older adults. The studies in this thesis were done as part of the Cater with Care® project; a collaboration between the university, care organizations, and partners from the food industry. The industrial partners developed the products, focusing each on different product categories: Carezzo Nutrition developed bread, pastry, and fresh juices and soups; The Kraft Heinz Company focused on long shelf-life and convenience foods; and the Veal Promotion Foundation produced veal meat. To fit the products to the needs of the target group, interviews with undernourished older adults (at home or hospitalized) and with dietitians were conducted (chapter 2). These interviews showed that undernutrition awareness is low among older adults. To treat undernutrition by changing their eating habits, older adults need to be aware of their health problem, they need to be willing to change, and they need to be able to understand and implement the dietitian’s advices. This process takes time while undernutrition should be treated immediately. For immediate treatment, enriched products could be used, without first creating awareness. According to the interviewees, enriched products should fit within older adults’ eating habits, and have small portion sizes. To gain insights in food choices of hospitalized older adults (65 years and older) an observational study was conducted. In this study, energy and protein intakes of 80 hospitalized older patients at low and high risk of undernutrition were assessed (chapter 3). Patients who received an energy- and protein-rich menu, because of their risk of undernutrition, were better able to reach the protein and energy targets than patients with a low risk of undernutrition receiving a standard menu. Based on these results we proposed that all hospitalized older adults – both at low and high risk of undernutrition – should receive an energy- and protein-rich menu. Subsequently, a pilot study was done in a care home and a rehabilitation center with the aim to explore the potential of the developed protein-enriched products to increase protein intake (chapter 4). Participants did not compensate their consumption of regular protein-rich foods (e.g. dairy, cheese) upon the introduction of protein-enriched foods and drinks. The 22 institutionalized elderly (mean age 83 years) consumed 12 gram protein per day more than they did before the intervention. Consequently, more people met the protein target of 1.2 g/kg/d than before the intervention. We concluded that protein-enriched products enabled institutionalized elderly to reach protein intake targets. Furthermore, we gained valuable feedback to improve the assortment of protein-enriched products for the effectiveness study. In the final study, effects of the protein-enriched products on protein intake and physical performance were studied in a randomized controlled trial during hospitalization and subsequent recovery at home. During the hospital period in which 147 older patients participated, patients that received protein-enriched products increased their protein intake compared to the control group that already received a protein-rich hospital menu (chapter 5). As a result, 79% of the intervention group reached a protein intake of 1.2 g/kg/d, compared to 48% of the control group. Finally, effects of the protein-enriched products were tested at home, for a longer period (chapter 6). Half of the hospital phase participants (n = 75) continued the intervention at home for 12 weeks. The protein-enriched products were successfully implemented in the daily menu of the older adults: the intervention group had a higher average protein intake (1.5 ± 0.6 g/kg/d) than the control group (1.0 ± 0.4 g/kg/d) during the 12-week intervention period. Seventy-two percent of the intervention group reached a protein intake of 1.2 g/kg/d during the 12-week intervention, compared to 31% of the control group. Protein intake of the intervention group was mainly increased by the following protein-enriched products: bread, dairy drinks, dairy desserts, soups, and fruit juices. However, despite the successful improvement of protein intake, we found no added value on physical performance in the first 6 months after hospitalization. It was concluded that with the protein-enriched familiar foods and drinks, we have a feasible, acceptable, and appetizing long-term strategy to increase protein intake of older adults in various settings. We envisage a beneficial role of these protein-enriched products in combination with physical activity in older adults with lower protein intakes.

Artikel n.a.v. rapport ' Perspectieven van sojavervanging in voer : op zoek naar Europese alternatieven voor soja ' PPO -project 3250119600

Measurement errors in dietary assessment using duplicate portions as reference method Laura Trijsburg Background: As Food Frequency Questionnaires (FFQs) are subject to measurement error, associations between self-reported intake by FFQ and outcome measures should be corrected for measurement error with data from a reference method. Whether the correction is adequate depends on the characteristics of the reference method used in the validation study. The duplicate portion method (DP), compared to the often used 24h recall (24hR), seems a promising reference method as correlated errors between FFQ and DP, such as memory bias, errors in portion size estimations and food composition databases, are not expected. Aim: This thesis aimed to determine the validity of the DP compared to the 24hR as a reference method for FFQ validation. The second aim was to explore the validity of nutrient densities for DP, 24hR and FFQ. The third aim was to determine the factors associated with misreporting of energy, protein and potassium as estimated by DP, 24hR and FFQ. Methods: Within the DuPLO-study, a Dutch validation study which is part of the NQplus study, two DPs, two FFQs, two blood and urinary biomarkers and one to fifteen 24hRs (web-based and/or telephone-based) were collected in 198 subjects, within 1.5 years. Also, one or two doubly labelled water measurements were available for 69 participants. Multivariate measurement error models were used to assess proportional scaling bias, error correlations with the FFQ, validity coefficients and attenuation factors. Furthermore linear regression analysis was used to determine the association between misreporting and various factors. Results: The DP was less influenced by proportional scaling bias, had lower correlated errors with the FFQ and showed higher attenuation factors than the 24hR for potassium, sodium and protein. Also, the DP seemed a better reference method than the 24hR for the assessment of validity coefficients for the FFQ for various fatty acids. The attenuation factors for the FFQ, using either the DP or 24hR as reference method, agreed reasonably well. Furthermore, the DP showed, when using plasma fatty acids as reference, slightly better ranking of participants according to their intake of n-3 fatty acids (0.33) and the n‑3/LA ratio (0.34) than the 24hR (0.22 and 0.24, respectively). Less group level bias was observed for protein and sodium densities compared to their absolute intakes for FFQ, 24hR and DP, but not for potassium. Overall the validity coefficients and attenuation factors for DP, 24hR and FFQ did not improve for nutrient densities compared to absolute intakes, except for the attenuation factor for sodium density. Lastly, BMI proved to be the most consistent determinant associated with misreporting (group level bias) of energy, protein and potassium for DP, 24hR and FFQ. Men tended to underreport protein by the DP, FFQ and 24hR and persons of older age underreported potassium but only by the 24hR and FFQ. Other explorative determinants did not show a consistent association with misreporting of energy or nutrients by the different dietary assessment methods. Conclusion: With respect to error correlations and attenuation factors the DP performed slightly better than the 24hR as a reference method for validating FFQs in epidemiological research. Furthermore, the use of nutrient densities does not necessarily improve the validity of the dietary intake estimates from DP, 24hR and FFQ. Moreover, it was shown that BMI is an important determinant of misreporting of energy, protein and potassium for these three assessment methods.

Dit experiment was opgezet om het effect van eiwitniveau (normaal versus verlaagd) en aminozuursamenstelling in het rantsoen te evalueren op de technische prestaties van beren gehuisvest onder een tweetal sanitaire condities.

Measurement errors in dietary assessment using duplicate portions as reference method Laura Trijsburg Background: As Food Frequency Questionnaires (FFQs) are subject to measurement error, associations between self-reported intake by FFQ and outcome measures should be corrected for measurement error with data from a reference method. Whether the correction is adequate depends on the characteristics of the reference method used in the validation study. The duplicate portion method (DP), compared to the often used 24h recall (24hR), seems a promising reference method as correlated errors between FFQ and DP, such as memory bias, errors in portion size estimations and food composition databases, are not expected. Aim: This thesis aimed to determine the validity of the DP compared to the 24hR as a reference method for FFQ validation. The second aim was to explore the validity of nutrient densities for DP, 24hR and FFQ. The third aim was to determine the factors associated with misreporting of energy, protein and potassium as estimated by DP, 24hR and FFQ. Methods: Within the DuPLO-study, a Dutch validation study which is part of the NQplus study, two DPs, two FFQs, two blood and urinary biomarkers and one to fifteen 24hRs (web-based and/or telephone-based) were collected in 198 subjects, within 1.5 years. Also, one or two doubly labelled water measurements were available for 69 participants. Multivariate measurement error models were used to assess proportional scaling bias, error correlations with the FFQ, validity coefficients and attenuation factors. Furthermore linear regression analysis was used to determine the association between misreporting and various factors. Results: The DP was less influenced by proportional scaling bias, had lower correlated errors with the FFQ and showed higher attenuation factors than the 24hR for potassium, sodium and protein. Also, the DP seemed a better reference method than the 24hR for the assessment of validity coefficients for the FFQ for various fatty acids. The attenuation factors for the FFQ, using either the DP or 24hR as reference method, agreed reasonably well. Furthermore, the DP showed, when using plasma fatty acids as reference, slightly better ranking of participants according to their intake of n-3 fatty acids (0.33) and the n‑3/LA ratio (0.34) than the 24hR (0.22 and 0.24, respectively). Less group level bias was observed for protein and sodium densities compared to their absolute intakes for FFQ, 24hR and DP, but not for potassium. Overall the validity coefficients and attenuation factors for DP, 24hR and FFQ did not improve for nutrient densities compared to absolute intakes, except for the attenuation factor for sodium density. Lastly, BMI proved to be the most consistent determinant associated with misreporting (group level bias) of energy, protein and potassium for DP, 24hR and FFQ. Men tended to underreport protein by the DP, FFQ and 24hR and persons of older age underreported potassium but only by the 24hR and FFQ. Other explorative determinants did not show a consistent association with misreporting of energy or nutrients by the different dietary assessment methods. Conclusion: With respect to error correlations and attenuation factors the DP performed slightly better than the 24hR as a reference method for validating FFQs in epidemiological research. Furthermore, the use of nutrient densities does not necessarily improve the validity of the dietary intake estimates from DP, 24hR and FFQ. Moreover, it was shown that BMI is an important determinant of misreporting of energy, protein and potassium for these three assessment methods.

Ongeveer de helft van het eiwitrijke veevoer in Nederland wordt geïmporteerd van buiten Europa. Het overgrote deel van deze grondstoffen bestaat uit soja- en palmproducten. Op Europees niveau zijn we voor 96% van onze sojabehoefte en voor 70% van onze totale eiwitbehoefte afhankelijk van import van buiten Europa. In de teeltgebieden, met name in Zuid-Amerika, leidt de grootschalige teelt van deze gewassen tot grote ecologische en sociale schade; door ontbossing van natuurgebieden en daarmee gepaard gaande CO2- uitstoot en biodiversiteitsverlies, door uitputting van de bodem, vervuiling van drinkwater, bedreiging van de lokale voedselvoorziening en gedwongen landonteigening. Regionale eiwitteelt biedt een enorme kans voor de Nederlandse landbouw door het sluiten van kringlopen en het verminderen van milieuschade en sociaal onrechtvaardige omstandigheden. Om knelpunten en kennisvragen te identificeren die een (snelle) transitie naar het gebruik van meer regionaal eiwitrijk veevoer in de weg staan, richtte Milieudefensie in 2015 de Raad voor Regionaal Veevoer op. Dit rapport geeft de bevindingen weer van de Raad voor Regionaal Veevoer, inclusief haar aanbevelingen richting overheid, bedrijfsleven en maatschappelijke organisaties.

Dit experiment was opgezet om het effect van eiwitniveau (normaal versus verlaagd) en aminozuursamenstelling in het rantsoen te evalueren op de technische prestaties van beren gehuisvest onder een tweetal sanitaire condities.

Undernutrition is a major health problem in the growing elderly population. It is estimated that one in ten Dutch community-dwelling older adults is suffering from undernutrition, and one in three Dutch older adults who receive home care. Undernutrition may lead to many negative consequences, ranging from fatigue and falls to impaired immune function and death. This makes undernutrition an obvious target for preventive measures. Undernutrition can be defined as “a state of nutrition in which a deficiency or excess (or imbalance) of energy, protein, and other nutrients causes measurable adverse effects on tissue/body form (body shape, size and composition) and function, and clinical outcome”. In addition, it is often described as protein energy malnutrition. Adequate protein intake may to some extent prevent and reverse this process. However, throughout ageing, it becomes increasingly difficult to reach adequate protein intake due to higher protein needs and lower protein intakes. Finding solutions to assist older adults in reaching their optimal protein intake is necessary. In our overall research project, we considered 1.2g protein per kg weight per day (g/kg/d) as adequate protein intake. In Dutch community-dwelling older adults, protein intake is around 1.0 g/kg/d, implying room for improvement. However, it is possible that many of these older adults deal with physiological changes, medical conditions, and physical and mental limitations that impair their appetite and food provision. For these older adults with higher protein needs, merely recommending that they eat more would not be realistic. It would be more realistic to explore strategies that increase protein intake without having to increase food intake. This calls for the exploration of instruments that match the needs and preferences of older adults: protein-enriched regular products. One particular group that can be identified as a target group for such products, are older adults who receive home care. Undernutrition prevalence is high in this group, which may be explained by their health problems that led to this dependence on home care. Likewise, many of these older adults also depend on meals-on-wheels. These meals-on-wheels recipients, regardless of whether they receive home care or not, often risk undernutrition too. In both these (overlapping) care-dependent groups, difficulties in adhering to energy and protein recommendations can be discerned. For this reason, enriching the readymade meals that these older adults receive may contribute to the prevention of protein undernutrition by increasing protein intake while keeping food intake the same. Here, protein enrichment instruments can be used to prevent undernutrition, but only when implemented in a timely manner. Adequate undernutrition management systems are therefore necessary to facilitate timely intervention, ensuring that the developed protein-enriched meals are actually offered and effective. For this reason, the overall aim of our research project was to gain insight into the current state of undernutrition management in community-dwelling older adults in the Netherlands and explore the role of protein-enriched regular products as a supportive instrument in protein undernutrition management. In Study 1 (chapter 2) we explored the experiences of 22 Dutch nutrition and care professionals and researchers with undernutrition awareness, monitoring, and treatment among community-dwelling older adults. This qualitative study among, for example, dietitians, general practitioners, nurse practitioners, and home care nurses provided insight into the current bottlenecks within the existing undernutrition management guidelines. In these telephone interviews, these experts also discussed the current dietary behaviour problems of older adults and their impact on undernutrition risk. The experts’ experiences implied that undernutrition awareness is limited, among both older adults and care professionals. In addition, the interviewees were unclear about which professionals are responsible for monitoring and which monitoring procedures are preferred. The dietitians feel that they become involved too late, leading to decreased treatment effectiveness. In general, the interviewees desired more collaboration and a coherent and feasible allocation of responsibilities regarding undernutrition monitoring and treatment. This implied that the available guidelines on undernutrition management require more attention and facilitation. In the following mixed-methods study (chapter 3), with interviews, we qualitatively explored the dietary behaviour and undernutrition risk of 12 Dutch elderly meals-on-wheels clients, one of the largest at-risk groups. We followed up on this information by quantifying the topics that emerged from the qualitative exploration of experienced bottlenecks in performing adequate dietary behaviour. For this, we used a survey among 333 meals-on-wheels clients. The interviews with elderly meals-on-wheels clients made clear that they have fixed and habitual eating patterns, while at the same time their appetite had decreased throughout the years. This was confirmed by the survey finding that regular portion size meals were perceived as too large by the oldest group aged over 75y. In addition, as the professionals suggested earlier, the interviewed elderly clients indeed showed limited awareness of undernutrition risk. Simultaneously, the survey showed that almost one in four elderly meals-on-wheels clients was undernourished. These findings led to the conclusion that staying close to the identified dietary habits may facilitate small yet effective modifications within these habits to prevent inadequate nutritional intake. Still, the limited awareness of undernutrition risk was expected to play a limiting role in whether clients believe they need dietary modifications. Consequently, informing them about this need could facilitate their motivation to implement modifications. After learning about the general dietary behaviour of these older adults, we used this information for Study 3 (chapter 4). We developed two kinds of protein-enriched readymade meals that are in line with the needs and preferences of older adults: one of regular size (450g) and one of reduced size (400g). We tested these meals in a lab setting in 120 community-dwelling older adults in a single-blind randomised crossover trial. One day a week at lunchtime, for four weeks, participants had to consume and evaluate a readymade meal. Overall, regardless of portion size, the protein-enriched meals led to higher protein intakes in vital older adults in a lab setting during lunch. In this crossover study, the participants liked the protein-enriched meals and the regular meals equally. However, we did not find the expected lower ratings of satiety after the reduced-size meals, while one reduced-size enriched meal and another regular-size enriched meal led to higher ratings of subsequent satiety. This higher satiety in the enriched meals could lead to compensational behaviour on the remainder of the day. After establishing that the protein-enriched meals were effective and acceptable in the lab setting, we moved to the homes of older adults to test the meals in a longer-term study in Study 4 (chapter 5). In this double-blind randomised controlled trial of two weeks, we also included protein-enriched bread to assess whether both this bread and the meals could increase daily protein intake to 1.2g/kg/d in 42 community-dwelling older adults to reach optimal protein intake. We found that the enriched products again led to higher protein intakes and a high liking. The mean protein intake per day was 14.6g higher in the intervention group, which amounted to a protein intake of 1.25g/kg/d, compared with 0.99g/kg/d in the control group. In addition, the meals scored 7.7 out of 10, while the bread scored 7.8 out of 10, which both were comparable with their regular counterparts. Lastly, we found no negative effect of compensational behaviour throughout the day. These promising findings indicated that we achieved a good match between older adults’ needs and preferences regarding protein intake. In the general discussion of this thesis (chapter 6), we combined our learnings from the four studies to reflect on protein undernutrition management in community-dwelling older adults and the possible role of protein-enriched regular products. We have discussed a conceptual framework consisting of three wheels of protein undernutrition management. In the first wheel regarding awareness, we proposed that limited awareness of adequate nutrition and body composition forms the largest bottleneck in undernutrition management. When this awareness is generated among both older adults and professionals, it will benefit the second wheel of monitoring. Here, we argued that a policy and the actual facilitation of that policy are required for this monitoring to succeed. When the monitoring is performed adequately, in the third wheel, the appropriate treatment can be carried out. We discussed that personalisation and evaluation of this treatment are important conditions. All in all, the public health implications that we have discussed on the basis of our findings can be summarised by the three key messages that could help us ace in adequate protein undernutrition management: address awareness in both older adults and professionals, facilitate continuous collaboration between professionals, and offer protein-enriched products expediently.

Ongeveer de helft van het eiwitrijke veevoer in Nederland wordt geïmporteerd van buiten Europa. Het overgrote deel van deze grondstoffen bestaat uit soja- en palmproducten. Op Europees niveau zijn we voor 96% van onze sojabehoefte en voor 70% van onze totale eiwitbehoefte afhankelijk van import van buiten Europa. In de teeltgebieden, met name in Zuid-Amerika, leidt de grootschalige teelt van deze gewassen tot grote ecologische en sociale schade; door ontbossing van natuurgebieden en daarmee gepaard gaande CO2- uitstoot en biodiversiteitsverlies, door uitputting van de bodem, vervuiling van drinkwater, bedreiging van de lokale voedselvoorziening en gedwongen landonteigening. Regionale eiwitteelt biedt een enorme kans voor de Nederlandse landbouw door het sluiten van kringlopen en het verminderen van milieuschade en sociaal onrechtvaardige omstandigheden. Om knelpunten en kennisvragen te identificeren die een (snelle) transitie naar het gebruik van meer regionaal eiwitrijk veevoer in de weg staan, richtte Milieudefensie in 2015 de Raad voor Regionaal Veevoer op. Dit rapport geeft de bevindingen weer van de Raad voor Regionaal Veevoer, inclusief haar aanbevelingen richting overheid, bedrijfsleven en maatschappelijke organisaties.

Background and aim Dietary proteins are of special interest for the heterogeneous group of older adults, since these people do not always have an adequate protein intake. When protein-rich products are better aligned with the requirements of older persons, an adequate nutrient intake is more likely. In this thesis we therefore explored two approaches for tailoring protein-enriched meals to older consumer subgroups; emotion-based and functionality-based. We expected a better ‘product-cluster fit’ (i.e. a more positive meal experience) when the clusters’ meal associations are congruent to their mealtime expectations. Methods We conducted an online survey in which vital community-dwelling older adults (n=392) reported their mealtime-related emotions and mealtime functionality. Using a hierarchical clustering analysis we described clusters within our population. Subsequently, we explored the extent to which the expectations of these clusters can be applied for the development of tailored protein-enriched meals. For the emotion-based approach, we conducted two central location tests (CLTs, n=461) to explore older adults’ food-evoked emotions. For the functionality-based approach we conducted in-depth interviews in order to get further insights regarding functional mealtime expectations and attitudes towards proteins and protein-enrichment. Based on the latter insights we tailored PE meal concepts to two functionality-based segments. In a final home-use test, the members of the functionality-based segments (n=91) prepared and evaluated the tailored PE meal concepts. Results The emotion-based approach resulted in four clusters; pleasurable averages, adventurous arousals, convivial indulgers, and indifferent restrictives. These emotions that these segments associated with their mealtimes varied along the two dimensions valence and arousal. However, from both CLTs we learned that the variation in valence-arousal as observed for mealtime-related emotions was not observed for emotions related to actual foods. The latter makes it challenging to identify products that evoke emotions congruent to the mealtime expectations of the emotion-based clusters. With regard to the functionality-based approach, we encountered three clusters; physical nutritioners, cosy socialisers, and thoughtless averages. The cosy socialisers value the social interactions and cosiness during their mealtimes, whereas the physical nutritioners focus more on the health and nutrient aspects of meals. Thoughtless averages have the least distinctive mealtime expectations. We translated these functional mealtime expectations into two PE meal concepts; one tailored to cosy socialisers and one tailored to physical nutritioners. These meal concepts were well-accepted by the participants. However, congruency between mealtime expectations and functional meal associations did not result in a better ‘product-cluster fit’. Conclusions Given the challenge to identify congruency between the meal associations and the mealtime expectations of the emotion-based clusters, we consider the emotion-based approach to be not yet actionable enough as a basis for tailoring PE products to older consumers. In contrast, the functionality-based approach appeared to be more promising, since the functional meal expectations could be translated to well-accepted tailored PE meal concepts. However, the effectivity of our functionality-based approach was not yet confirmed in this thesis, since congruency between functional meal associations and functional meal expectations did not necessarily result in a better ‘product-cluster fit’. Future studies, focussing on e.g. other meal types, are recommended to further explore mealtime functionality as a basis for tailoring PE meals to older consumer subgroups.

Resultaten van een onderzoek naar het voorkomen van virussen (EVEX, Eel Virus European X; HVA, Herpes Virus Anguillea; EVE, Eel Virus European) bij palingen afkomstig uit diverse landen (zowel wilde paling als paling van kwekerijen) en van een zwemexperiment met geïnfecteerde en niet-geïnfecteerde palingen. Na voltooiing van de zwemreis werd in beide groepen het bloedbeeld onderzocht op diverse parameters (o.a. op hematocriet, het aantal rode bloedcellen). De resultaten zijn zeer verontrustend, omdat ze een aanwijzing vormen dat virus-infectie ertoe leidt dat palingen hun migratie naar de paaigronden niet kunnen voltooien door bloedarmoede, een beschadigde lever en een tekort aan eiwitten. Dit kan mede een oorzaak zijn voor het wereldwijd teruglopen van de palingstand

In 2050 moeten grote zeewierplantages op zee voorzien in onze behoefte aan voedsel en grondstoffen. In de Oosterschelde doen Wageningse onderzoekers nu experimenten met duurzame zeewierteelt.

Aeration of food is considered to be a good method to create a texture and mouthfeel of food products that is liked by the consumer. However, traditional foams are not stable for a prolonged time. Microbubbles are air bubbles covered with a shell that slows down disproportionation significantly and arrests coalescence. Protein stabilized microbubbles are seen as a promising new food ingredient for encapsulation, to replace fat, to create new textures, and to improve sensorial properties of foods. In order to explore the possible functionalities of microbubbles in food systems, a good understanding is required regarding the formation of protein stabilized microbubbles as well as their stability in environments and at conditions encountered in food products. The aim of this research was to investigate the key parameters for applications of microbubbles in food systems. In Chapter 1 an introduction to this topic is given. In Chapter 2, the effect of the microbubble preparation parameters on the microbubble characteristics, like the microbubble yield, size and stability, was investigated. The protein Bovine Serum Albumin (BSA) and the method sonication was used to manufacture the microbubbles. The manufactured number and stability of microbubbles was highest when they were prepared at a pH around 5 to 6, just above the isoelectric point, and at an ionic strength of 1.0 M. This can be related to the protein coverage at the air/water interface of air bubbles formed during sonication. At a pH close to the isoelectric point the BSA molecules is in its native configuration. Also the repulsion between the proteins is minimized at these pH values and ionic strength. Both the native configuration and the limited repulsion between the proteins result in an optimal protein coverage during the first part of sonication. Also a high protein concentration contributes to a higher surface coverage. The surface coverage is proportional to the protein concentration up to a concentration of 7.5% after which an increase in protein concentration did not lead to a substantial increase in the number of microbubble . In the second part of sonication the protein layer around the air bubble becomes thicker and stronger by heat induced protein-protein interactions. We found that and at a preheating temperature of 55-60°C, about 5 °C below the BSA denaturation temperature, and a final solution temperature of 60-65°C most microbubbles were obtained, while at higher temperatures mainly protein aggregates and (almost) no microbubbles are formed. This suggests that at temperature of around 60°C to 65°C protein aggregated mostly at the air-water interface creating a multi-layered shell, while at higher temperature, they also aggregated in bulk. These aggregates cannot form microbubbles. We found that optimal preparation parameters strongly depend on the protein batch. We hypothesize that the differences in microbubble formation between the protein batches is due to (small) differences in the protein molecular and denaturation properties that determine the temperature at which the molecules start to interact at the air-water interface. Microbubbles made with different protein concentration and preheating temperatures shrunk in time to a radius between 300 nm and 350 nm, after which the size remained constant during further storage. We argue that the driving force for the shrinkage was the Laplace pressure, resulting in an air flux from the bubbles to the solution. We argue that the constant final size can be explained by a thickening of the microbubble shell as a result of the microbubble shrinkage, thereby withstanding the Laplace pressure. In Chapter 3 and Chapter 4, microbubble stability at environments and conditions representative for food products were studies. In Chapter 3 we investigated the stability upon addition of surfactants and acid, When surfactants or acid were added, the microbubbles disappeared in three subsequent steps. The release of air from the microbubble can be well described with the two-parameter Weibull process. This suggests two processes are responsible for the release of air: 1) a shell-weakening process and 2) a random fracture of the weakened shell. After the air has been released from the microbubble the third process is identified in the microbubble disintegration: 3) the shell disintegrated completely into nanometer-sized particles. The probability of fracture was exponentially proportional to the concentration of acid and surfactant, meaning that a lower average breaking time and a higher decay rate were observed at higher surfactant or acid concentrations. For different surfactants, different decay rates were found. The disintegration of the shell into monomeric proteins upon addition of acid or surfactants shows that the interactions in the shell are non-covalent and most probably hydrophobic. After surfactant addition, there was a significant time gap between complete microbubble decay (release of air) and complete shell disintegration, while after acid addition the time at which the complete disintegration of the shell was observed coincided with the time of complete microbubble decay. In Chapter 4 the stability of the microbubbles upon pressure treatment, upon fast cooling after heating and at different storage temperatures was studied. The microbubble stability significantly decreased when microbubbles were pressurized above 1 bar overpressure for 15 seconds or heated above 50°C for 2 minutes. Above those pressures the microbubbles became unstable by buckling. Buckling occurred above a critical pressure. This critical pressure is determined by the shell elastic modulus, the thickness of the shell, and the size of the microbubble. Addition of crosslinkers like glutaraldehyde and tannic acid increased the shell elastic modulus. It was shown that microbubbles were stable against all tested temperatures (up to 120°C) and overpressures (4.7 bar) after they were reinforced by crosslinkers. From the average breaking time at different storage temperatures, we deduced that the activation energy to rupture molecular bonds in the microbubbles shell is 27 kT. In Chapter 5, we investigated the effect of microbubbles on the rheological, tribological sensorial properties of model food systems and we compared this effect to the effect on food systems with emulsion droplets and without an added colloid. We investigated the effect in three model food systems, namely fluids with and without added thickener and a mixed gelatine-agar gel. In a sensory test panellists were asked whether they could discriminate between samples containing microbubbles, emulsion droplets or no added colloid. Emulsions could be sensorially well distinguished from the other two samples, while the microbubble dispersion could not be discriminated from the protein solution. Thus, we concluded that at a volume fraction of 5% of these BSA covered microbubbles were not comparable to oil-in-water emulsions. The good discrimination of emulsion might be ascribed to the fact that emulsion had a lower friction force (measured at shear rates form 10 mm/s to 80 mm/s) than that microbubbles dispersions and protein solutions. Upon mixing emulsions and microbubble dispersions the friction value approximated that of emulsions. This effect was already noticed at only 1.25% (v/v) oil, indicating that microbubbles had not a significant contributions to the friction of these samples. Also microbubble dispersions with and without protein aggregates were compared. The microbubble dispersions with and without thickener containing protein aggregates had a higher viscosity than the those samples without protein aggregates. Protein aggregates in the gelled microbubble sample yielded a higher Young’s modulus and fracture stress. The differences between the gelled samples could be well perceived by the panellists. We attribute this mainly to the fracture properties of the gel. In general we concluded that microbubbles, given their size of ~ 1 mm and volume fraction of 5%, did not contribute to a specific mouthfeel. Finally in Chapter 6, the results presented in the previous chapters are discussed and put in perspective of the general knowledge on microbubbles production, stability, and applications in food. We described the main mechanisms leading to microbubble formation and stability. We showed that the production parameters significantly influence the interactions in the microbubble shell, and the those interactions highly determine the stability of the microbubbles under several conditions. We reported about limitations of sonication as a method to produce microbubbles suitable for food applications and we provided some ways to overcome these limitations. The use of microbubbles in food systems has been explored and we clearly see possible applications for microbubbles in food. We reported about directions for possible further research. In this work we made significant progress in understanding the interactions in the microbubble shell and their relation to microbubble stability. We also advanced in comprehension towards possible applications of microbubbles in food.

Melk speelt wereldwijd een belangrijke rol bij het voorzien in de nutriëntenbehoeften van de mens. De koe zet voor mensen niet-eetbaar eiwit uit voer om in eiwit van hoge kwaliteit. Steeds vaker rijst de vraag of deze eiwitconversie voldoende toegevoegde waarde heeft en hoe dit zich verhoudt tot de milieudruk van de zuivelproductie. Professor Toon van Hooijdonk van Wageningen Universiteit ontwikkelt samen met collega's een model om dit inzichtelijk te maken.

In 2012 is een proef met melkkoeien uitgevoerd met als doel de waarde van grasvezel te onderzoeken. In het rantsoen van de koeien werd een deel van de graskuil vervangen door grasvezel. De grasvezel kwam beschikbaar na de raffinage van gras. De resultaten van de proef vielen tegen. De voeropname van de koeien die grasvezel kregen was namelijk lager dan van de (controle)koeien die het gangbare rantsoen kregen. De melkgift was eveneens lager op het rantsoen met grasvezel. De oorzaak van de lagere voeropname heeft waarschijnlijk te maken met de versheid en daarmee de smakelijkheid van grasvezel. Daarom is in 2013 besloten om eerst te kijken naar de mogelijkheid om grasvezel te conserveren (in te kuilen), waardoor de kwaliteit en de houdbaarheid mogelijk werd verbeterd. In 2012 is eveneens een oriënterend onderzoek gedaan met graseiwit verstrekt aan kalveren. Graseiwit is het eiwit dat gewonnen wordt uit het grassap en in de proef werd het in gelvorm verstrekt. De resultaten van de proef met kalveren waren uitermate positief. De dieren namen het graseiwit graag op. De groei van de kalveren was vergelijkbaar met de controlegroep.

Melk speelt wereldwijd een belangrijke rol bij het voorzien in de nutriëntenbehoeften van de mens. De koe zet voor mensen niet-eetbaar eiwit uit voer om in eiwit van hoge kwaliteit. Steeds vaker rijst de vraag of deze eiwitconversie voldoende toegevoegde waarde heeft en hoe dit zich verhoudt tot de milieudruk van de zuivelproductie. Professor Toon van Hooijdonk van Wageningen Universiteit ontwikkelt samen met collega's een model om dit inzichtelijk te maken.

Summary In this thesis we present the design and characterization of bio-inspired hybrid protein polymers. All polymers are composed of two distinct types of building blocks. The first type is a silk-inspired block that is pH-responsive and can fold and self-assemble into highly ordered structures. The basic structure of this building block is an octapeptide (GAGAGAGX), denoted SX. We include multiple repeats n of this octapeptide (SXn) in our protein polymers. The amino acid X is always an acidic or basic one. This way, the pH of the solvent determines the charge of this amino acid; in the charged state the silk-like blocks repel each other and the proteins are molecularly dissolved. When charge neutralized, the silk-like blocks become hydrophobic and can fold and stack. The second building block is very hydrophilic and acts as a random coil under a wide range of aqueous solvent conditions. The basic structure of this block is a 99 amino acids long sequence of mostly hydrophilic amino acids, that we included either as a single block or as multimers, denoted Cn. The combination of these two blocks in one molecule leads to a pH-responsive protein polymer that switches from hydrophilic to amphiphilic due to solvent pH. The amphiphilic nature of the neutralized protein polymer leads to microscopic phase separation. All protein polymers were designed by genetic engineering and produced by genetically modified Pichia pastoris in a batch fermentation. A simple ammonium sulfate precipitation was sufficient for all types of proteins to acquire highly pure samples. The types of hybrid protein polymers we produced and characterized differ in three aspects. Firstly, we designed different silk-like blocks in which the amino acid X had three varieties. We included the acidic amino acid glutamic acid (E) with a pKa ~ 4, to obtain a block that is charged at high pH values and neutral under acidic conditions. We also designed this silk-like block with basic amino acids lysine (K) or histidine (H). These blocks are positively charged under acidic conditions, while being neutral at higher pH. Lysine with a pKa ~ 10 remains charged under slightly alkaline conditions, and is only neutralized at rather extreme pH values. The pKa of histidine (~ 6) means this is the only pH-responsive amino acid that’s almost completely neutral under physiological conditions (a pH of 7.4). This makes histidine the most interesting residue from a biomedical point of view. The second variety in design is the relative sizes of the two blocks. For any amphiphile, the relative sizes of the two blocks determine the structure that is formed upon self-assembly. The size of the silk-like block was chosen to be 8, 16, 24 or 48 repeats of the octapeptide SXN. The random coil block was included as monomer C1, dimer C2 or tetramer C4. Our third variation in molecular design is the order of the two building blocks. We constructed two diblock structured protein polymers: C1SH48 and C2SH48. All other protein polymers had a triblock structure. We constructed C2SXNC2 protein polymers, where the self-assembling silk-like domain was the central block, and SX24C4SX24 protein polymers, with telechelic end blocks. All types of protein polymers that we studied are presented in Table 6.1. Table 6.1: Overview of different protein polymers studied in this thesis. Telechelic (Ch. 2) Varying Silk-Like Block (Ch. 3 & 4) Varying Random Coil Block (Ch. 5) SE24C4SE24 C2SH8C2 C1SH48 SH24C4SH24 C2SH16C2 C2SH48 SK24C4SK24 C2SH24C2 C2SH48C2 C2SH48C2 In chapter 2 we report on the self-assembly behavior of triblock structured telechelic protein polymers SX24C4SX24. We analyzed the pH-dependent self-assembly into fibrillar structures of three different protein polymers. These proteins only differ in the amino acid X in the silk-like block. We found that all proteins self-assemble into fibrils under solvent conditions at which the amino acid X is uncharged. This self-assembly is completely reversible; changing the solvent pH to a value at which the amino acid X is fully charged, leads to immediate disassembly of the fibrils. The secondary structures of the fibrils are comparable, and are a combination of a random coil corona and a crystalline folded and self-assembled core. The self-assembly process is a pseudo-first order one. Initial fast (heterogeneous) nucleation is followed by elongation of existing fibrils, without the formation of new fibrils. These kinetics lead to monodisperse samples of fibrils. Existing fibrils have at least one living end: addition of new proteins in solution leads to further growth of these fibrils. In chapter 3 we use super resolution fluorescence microscopy and atomic force microscopy to analyze the self-assembly mechanism of the protein polymer C2SH48C2. Surprisingly, we found that self-assembly of these fibrils is an asymmetric process. The fibrils grow in only one direction with one living end, although the protein polymer that is the building block of these fibrils is highly symmetric. We therefore conclude that nucleation is a heterogeneous process. We observed that once a protein molecule is part of a fibril, it is kinetically trapped. In a timeframe of 3 days, we don’t observe exchange of protein molecules inside fibrils, with proteins in solution. The interactions of uncharged folded protein polymers inside a fibril are simply too strong to overcome to be released into solution. We also report that self-assembly of these fibrils is a process that involves continuous nucleation; elongation of existing fibrils is accompanied by the genesis of new ones. This leads to samples that contain fibrils with a wide variety of sizes, quite different from the populations found for the protein polymers with the inverted block sequence that we presented in chapter 2. In chapter 4 and 5 we present our findings on several protein polymers in which we varied the relative sizes of the silk-like block and of the random coil blocks. In chapter 4 we present the characterization of protein polymers that have differently sized silk-like domains. We studied 4 protein polymers with the general structure C2SHnC2; the series consisted of n = 8, 16, 24 or 48. The two smallest protein polymers form micelles when charge neutralized (pH 8). The two largest protein polymers form fibrils under these conditions. At low pH, when the silk-like block is highly charged, this block behaves as (extended) random coil, according to circular dichroism measurements. This behavior is consistent for all block sizes that we studied. In the self-assembled state, there is a distinct difference in secondary structure of the micelles and the fibrils. The silk-like core of the micelles has a secondary structure that differs only slightly from the structure in the charged state. It merely acts as a collapsed coil. The secondary structures of the fibril forming protein polymers are very different in neutralized state. Their structures are mutually nearly identical, similar to that of a betaroll. We observed that the size of the silk-like block has a strong effect on the kinetics of self-assembly. The largest protein polymer C2SH48C2 self-assembles into fibrils at a rate that is over a decade faster than the protein polymer with the smaller silk-like block C2SH24C2. Both fibril forming protein polymers can form hydrogels. There is however a great difference in rigidity of the gels at similar concentrations. The gel that consists of fibrils of C2SH48C2 is a decade stiffer than the one consisting of C2SH24C2. This stronger fibril-fibril interaction due to the more exposed silk-like core of C2SH48C2 clearly has a strong effect on macroscopic gel properties. We used partial enzymatic degradation of the random coil block to determine the influence of decreasing the hydrophilic block on self-assembly behavior. Both micelle forming protein polymers are able to form fibrils after up to 80% of the random coil blocks has been cleaved off. This shows the intrinsic capacity of the silk-like block to form fibrils even at a size as small as 64 amino acids. The fibrils of C2SH24C2 show an increase in interaction after partial cleavage of the random coil block. Individual fibrils start to associate laterally. This is a strong indication that fibrils with smaller random coil blocks can form more rigid hydrogels, based on their increased core-core interaction. Based on these findings, we designed two new protein polymers with smaller random coil blocks: C1SH48 and C2SH48. With these protein polymers we systematically probe the role of a more exposed silk-like core in gel properties, presented in chapter 5. Both proteins self-assemble into fibrils when neutralized. Fibrils of C1SH48 differ from those of C2SH48 and C2SH48C2 as they start to laterally associate. Surprisingly, the rates at which self-assembling fibrils are formed, are identical for these protein polymers, and also equal to the rate of C2SH48C2. Apparently, for these sizes of the blocks, the size of the silk-like block is what determines the rate of self-assembly. These two protein polymers attain secondary structures that are very similar to that of C2SH48C2. When looking at macroscopic properties of hydrogels formed by these protein polymers, we do observe a very clear difference. Both C2SH48C2 and C2SH48 form fibril based hydrogels that act as gels with very few (weak) crosslinks. These two gels show similar scaling behavior of modulus with concentration (exponents of 1.52 and 1.66). The attractive interaction of fibrils of C1SH48 leads to a different type of gel. The modulus of this hydrogel scales strongly with concentration (exponent of 2.8), typical for a (physically) cross-linked gel. The latter gels can have much greater moduli than gels of C2SH48C2 and C2SH48, but are also slightly more brittle. The porosity of the gels (an important parameter for biomedical applications) increases when decreasing the size of the random coil domain. However, for this series of protein polymers, the porosity is in the order of 10-100 nm, which makes these gels not suitable for using them to grow 3D cell cultures.

Gelation is a ubiquitous process in the preparation of foods. As most foods are multi constituent mixtures, understanding gelation in mixtures is an important goal in food science. Here we presented a systematic investigation on the influence of molecular interactions on the gelation in protein mixtures. Gelatin gels with added globular protein and globular protein gels with added gelatin were analyzed for their gel microstructure and rheological properties. Mixed gels with altered microstructure (compared to single gels) also differed in modulus from single gels. Mixed gels with microstructures similar to single gels were rheologically similar to single gels. Alterations in microstructure were attributed to segregative phase separation between proteins which occurred during gelation. Gelation was treated as a growth process from macromolecule to space spanning network. At conditions where electrostatic interactions were screened the occurrence of phase separation was attributed to the molecular size ratio between gelling and non-gelling proteins before gelation and changes of this size ratio during gelation. Here only mixtures that during gelation passed a region of high compatibility (similar molecular sizes) before entering a region of decreasing solubility phase separated. For applications this implies that whenever the gelling molecule is larger than the non-gelling molecule phase separation during gelation is unlikely while reversely, if the gelling molecules is smaller than the non-gelling molecule phase separation during gelation typically does occur

Aeration of food is considered to be a good method to create a texture and mouthfeel of food products that is liked by the consumer. However, traditional foams are not stable for a prolonged time. Microbubbles are air bubbles covered with a shell that slows down disproportionation significantly and arrests coalescence. Protein stabilized microbubbles are seen as a promising new food ingredient for encapsulation, to replace fat, to create new textures, and to improve sensorial properties of foods. In order to explore the possible functionalities of microbubbles in food systems, a good understanding is required regarding the formation of protein stabilized microbubbles as well as their stability in environments and at conditions encountered in food products. The aim of this research was to investigate the key parameters for applications of microbubbles in food systems. In Chapter 1 an introduction to this topic is given. In Chapter 2, the effect of the microbubble preparation parameters on the microbubble characteristics, like the microbubble yield, size and stability, was investigated. The protein Bovine Serum Albumin (BSA) and the method sonication was used to manufacture the microbubbles. The manufactured number and stability of microbubbles was highest when they were prepared at a pH around 5 to 6, just above the isoelectric point, and at an ionic strength of 1.0 M. This can be related to the protein coverage at the air/water interface of air bubbles formed during sonication. At a pH close to the isoelectric point the BSA molecules is in its native configuration. Also the repulsion between the proteins is minimized at these pH values and ionic strength. Both the native configuration and the limited repulsion between the proteins result in an optimal protein coverage during the first part of sonication. Also a high protein concentration contributes to a higher surface coverage. The surface coverage is proportional to the protein concentration up to a concentration of 7.5% after which an increase in protein concentration did not lead to a substantial increase in the number of microbubble . In the second part of sonication the protein layer around the air bubble becomes thicker and stronger by heat induced protein-protein interactions. We found that and at a preheating temperature of 55-60°C, about 5 °C below the BSA denaturation temperature, and a final solution temperature of 60-65°C most microbubbles were obtained, while at higher temperatures mainly protein aggregates and (almost) no microbubbles are formed. This suggests that at temperature of around 60°C to 65°C protein aggregated mostly at the air-water interface creating a multi-layered shell, while at higher temperature, they also aggregated in bulk. These aggregates cannot form microbubbles. We found that optimal preparation parameters strongly depend on the protein batch. We hypothesize that the differences in microbubble formation between the protein batches is due to (small) differences in the protein molecular and denaturation properties that determine the temperature at which the molecules start to interact at the air-water interface. Microbubbles made with different protein concentration and preheating temperatures shrunk in time to a radius between 300 nm and 350 nm, after which the size remained constant during further storage. We argue that the driving force for the shrinkage was the Laplace pressure, resulting in an air flux from the bubbles to the solution. We argue that the constant final size can be explained by a thickening of the microbubble shell as a result of the microbubble shrinkage, thereby withstanding the Laplace pressure. In Chapter 3 and Chapter 4, microbubble stability at environments and conditions representative for food products were studies. In Chapter 3 we investigated the stability upon addition of surfactants and acid, When surfactants or acid were added, the microbubbles disappeared in three subsequent steps. The release of air from the microbubble can be well described with the two-parameter Weibull process. This suggests two processes are responsible for the release of air: 1) a shell-weakening process and 2) a random fracture of the weakened shell. After the air has been released from the microbubble the third process is identified in the microbubble disintegration: 3) the shell disintegrated completely into nanometer-sized particles. The probability of fracture was exponentially proportional to the concentration of acid and surfactant, meaning that a lower average breaking time and a higher decay rate were observed at higher surfactant or acid concentrations. For different surfactants, different decay rates were found. The disintegration of the shell into monomeric proteins upon addition of acid or surfactants shows that the interactions in the shell are non-covalent and most probably hydrophobic. After surfactant addition, there was a significant time gap between complete microbubble decay (release of air) and complete shell disintegration, while after acid addition the time at which the complete disintegration of the shell was observed coincided with the time of complete microbubble decay. In Chapter 4 the stability of the microbubbles upon pressure treatment, upon fast cooling after heating and at different storage temperatures was studied. The microbubble stability significantly decreased when microbubbles were pressurized above 1 bar overpressure for 15 seconds or heated above 50°C for 2 minutes. Above those pressures the microbubbles became unstable by buckling. Buckling occurred above a critical pressure. This critical pressure is determined by the shell elastic modulus, the thickness of the shell, and the size of the microbubble. Addition of crosslinkers like glutaraldehyde and tannic acid increased the shell elastic modulus. It was shown that microbubbles were stable against all tested temperatures (up to 120°C) and overpressures (4.7 bar) after they were reinforced by crosslinkers. From the average breaking time at different storage temperatures, we deduced that the activation energy to rupture molecular bonds in the microbubbles shell is 27 kT. In Chapter 5, we investigated the effect of microbubbles on the rheological, tribological sensorial properties of model food systems and we compared this effect to the effect on food systems with emulsion droplets and without an added colloid. We investigated the effect in three model food systems, namely fluids with and without added thickener and a mixed gelatine-agar gel. In a sensory test panellists were asked whether they could discriminate between samples containing microbubbles, emulsion droplets or no added colloid. Emulsions could be sensorially well distinguished from the other two samples, while the microbubble dispersion could not be discriminated from the protein solution. Thus, we concluded that at a volume fraction of 5% of these BSA covered microbubbles were not comparable to oil-in-water emulsions. The good discrimination of emulsion might be ascribed to the fact that emulsion had a lower friction force (measured at shear rates form 10 mm/s to 80 mm/s) than that microbubbles dispersions and protein solutions. Upon mixing emulsions and microbubble dispersions the friction value approximated that of emulsions. This effect was already noticed at only 1.25% (v/v) oil, indicating that microbubbles had not a significant contributions to the friction of these samples. Also microbubble dispersions with and without protein aggregates were compared. The microbubble dispersions with and without thickener containing protein aggregates had a higher viscosity than the those samples without protein aggregates. Protein aggregates in the gelled microbubble sample yielded a higher Young’s modulus and fracture stress. The differences between the gelled samples could be well perceived by the panellists. We attribute this mainly to the fracture properties of the gel. In general we concluded that microbubbles, given their size of ~ 1 mm and volume fraction of 5%, did not contribute to a specific mouthfeel. Finally in Chapter 6, the results presented in the previous chapters are discussed and put in perspective of the general knowledge on microbubbles production, stability, and applications in food. We described the main mechanisms leading to microbubble formation and stability. We showed that the production parameters significantly influence the interactions in the microbubble shell, and the those interactions highly determine the stability of the microbubbles under several conditions. We reported about limitations of sonication as a method to produce microbubbles suitable for food applications and we provided some ways to overcome these limitations. The use of microbubbles in food systems has been explored and we clearly see possible applications for microbubbles in food. We reported about directions for possible further research. In this work we made significant progress in understanding the interactions in the microbubble shell and their relation to microbubble stability. We also advanced in comprehension towards possible applications of microbubbles in food.

In 2012 is een proef met melkkoeien uitgevoerd met als doel de waarde van grasvezel te onderzoeken. In het rantsoen van de koeien werd een deel van de graskuil vervangen door grasvezel. De grasvezel kwam beschikbaar na de raffinage van gras. De resultaten van de proef vielen tegen. De voeropname van de koeien die grasvezel kregen was namelijk lager dan van de (controle)koeien die het gangbare rantsoen kregen. De melkgift was eveneens lager op het rantsoen met grasvezel. De oorzaak van de lagere voeropname heeft waarschijnlijk te maken met de versheid en daarmee de smakelijkheid van grasvezel. Daarom is in 2013 besloten om eerst te kijken naar de mogelijkheid om grasvezel te conserveren (in te kuilen), waardoor de kwaliteit en de houdbaarheid mogelijk werd verbeterd. In 2012 is eveneens een oriënterend onderzoek gedaan met graseiwit verstrekt aan kalveren. Graseiwit is het eiwit dat gewonnen wordt uit het grassap en in de proef werd het in gelvorm verstrekt. De resultaten van de proef met kalveren waren uitermate positief. De dieren namen het graseiwit graag op. De groei van de kalveren was vergelijkbaar met de controlegroep.