Microbial community dynamics in traditionally fermented milk

Anneloes E. Groenenboom

Research output: Thesisinternal PhD, WU


Traditional fermented products are an important part of the diet in many African countries. These products can provide much of the needed proteins, vitamins and other micro-nutrients to its consumers, which include young children. Also, traditional fermented products are rooted in the local context and have high social and cultural value. Increasing our understanding of these spontaneously fermented products can be used to increase the value chain of these products in order to increase economic stability of producers as well as nutritional intake in rural and urban areas.

Most traditional fermented products are produced by spontaneous fermentation, meaning that there is no starter culture added to initiate fermentation. As a consequence these products contain a diverse microbial community, which are responsible for the characteristics of the product. Often a back-slopping method is used, where a portion of previously fermented product inoculates fresh raw material. Unknowingly, producers of these products have domesticated these bacterial communities and allowed the species to co-evolve within their community.

In this way, these bacterial communities form a model system for ecological and evolutionary research. Historically, experimental evolution was mainly performed with single organisms. In recent years the interest in how individuals evolve in the context of ecosystems is growing. This eco-evolutionary research needs model systems that can represent natural communities in their interactions and complexities, while still being a tangible model system.

In this thesis I investigated the microbial communities that are responsible for the fermentation and used the constituting bacteria to learn about bacterial community dynamics over time. This was done by combining three disciplines: 1) Food Microbiology, on the conversion of compounds by bacteria during spontaneous fermentation; 2) Evolution, on the changes in the fermenting community over time under selection pressure; 3) Ecology, on the roles and niches bacteria take within the fermenting community. This thesis had two aims. The first aim was to provide the first step towards the use of microbial communities of spontaneously fermented milk, such as Mabisi and Lait caillé, as model systems for studies on eco-evolutionary dynamics. For this, I performed a series of experiments with an increasing level of control and a decreasing level of complexity. These experiments started with producing Mabisi in traditional ways in the field to allow a study of bacterial dynamics, which resulted in a low level of control of variables. Later, I brought the microbial communities to the laboratory and used controlled environments without changing the diversity of the natural communities. The second aim was to provide ecological and evolutionary insights in Mabisi fermentations in the context of research for development. Here, the focus was on the practical application of bacterial dynamics for dairy fermentation. The potential use of the outcome for producers and consumers of spontaneously fermented foods by taking into account when constructing and interpreting the studies.  

This thesis starts with a perspective on the use of microbial communities of spontaneously fermented products as model system for eco-evolutionary dynamics (Chapter 2). We outline what model systems are used for experimental testing of evolutionary theory so far, ranging from simple microbial communities in the laboratory and, more recently, to complex (natural) communities. Microbial communities from spontaneously fermented products bear several intrinsic advantages for executing evolution experiments: short generation times, small size and ability to be stored frozen and defrosted to perform competition experiments (fitness tests) between evolved and ancestral lines. Moreover, these natural microbial communities have a limited number of players and form an island of microorganisms that does not have a lot of influx from outside the confined system boundaries. There are several research questions with an evolutionary background that can be addressed using these microbial communities from fermented foods. This includes questions on changes species frequency in space and time, the diversity-stability relationship, niche space, fluctuating environment and community coalescence. Hypotheses on the influence of these factors on community evolution are described as well as a short descriptions of the experimental approach of such studies when microbial communities of spontaneously fermented foods are used.

Natural communities of Mabisi were analysed in a field experiment. A method was developed using filter paper disks for the preservation of DNA from diverse microbial communities for later analyses (Chapter 3). The bacterial species composition obtained through DNA extraction via the filter paper method showed sufficient resemblance to the composition obtained via traditional DNA extraction from the liquid milk sample. This method could therefore successfully be used to analyse diverse microbial communities from Mabisi in our remote field sites in rural Zambia. Field experiments were conducted to determine the effect of variations in fermentation vessels and types of back-slopping procedures on acidification and bacterial community composition during fermentation (Chapter 4). Due to high costs and a reduced availability of the traditionally used calabashes, nowadays more and more plastic containers are used for Mabisi production. However, the effect of this change in production practice on the properties of the product is unknown. Together with the local community, I have performed 15 fermentations, using two types of fermentation vessels (calabashes and plastic buckets) and three levels of back-slopping (active back-slopping, passive back-slopping and no back-slopping). In passive back-slopping, the bacteria that start the fermentation are transferred from the previous fermentation round via the fermentation vessel. During active back-slopping, finished product is transferred to raw material to inoculate the fermentation. Overall, bacterial communities decreased in diversity over time, where the drop in pH correlates with a decreased diversity, although the rate of acidification showed variation. In case of active back-slopping, the pH drop started right after inoculation. In the ‘no back-slopping’ and ‘passive back-slopping’ fermentations, there was a clear lag phase before acidification started. No difference was found in bacterial diversity during and at the end of fermentation performed in plastic buckets or previously used calabashes. Besides small differences, all processing methods resulted in a microbial community dominated by Lactococcus lactis.

After the natural communities of traditional spontaneous fermentation were analysed, the next step was to bring the natural communities into the laboratory. Here, experiments were performed with lower complexity and an increased control. The species composition and metabolic profile of six different Mabisi samples were analysed before and after repeated propagation cycles (Chapter 5). These communities had similarity in their bacterial species composition and therefore had the potential to converge towards the same final species composition upon repeated propagation in a common environment. Species composition in all replicates propagated from the same original samples changed in a parallel way, yet that groups of communities derived from different original samples did not change in the same way. We observed that communities at the end of the repeated propagation cycles were either dominated by Lactobacillus helveticus or Lactobacillus delbrueckii. By modelling species compositions we tested the influence of four main factors on the species composition: initial species composition, selection imposed by the environment, selection caused by interaction between species and random processes in species dynamics. We found that the final species composition is mostly dependent on initial species composition, followed by random processes. The environment showed to have the least influence on the change in species composition.

We had the opportunity to work with a second spontaneously fermented dairy product, called Lait caillé. This traditional product originates from Senegal and is produced in wooden bowls, called lahals (Chapter 6). In terms of complexity and control, this experiment would position in between the field and laboratory experiments with Mabisi. The mode of propagation was traditional, meaning that the bacteria were transferred via the inside of the lahals that were used for repeated fermentations (passive back-slopping). This traditional production method allowed us to study the natural bacterial communities in the lahals and translate our findings towards production practices in Senegal. We analysed the species composition of Lait caillé over time and added a probiotic starter containing Lactobacillus rhamnosus yoba 2012 to the traditional process attempting to enrich the bacterial species community in the final product. We found detectable levels of L. rhamnosus in the final products, which were dominated by Lactobacillus helveticus and Acetobacter species. The abundance of L. rhamnosus ranged between 0.2 and 1 percent of the total bacterial population, which is comparable to the concentration found in probiotic yoghurt. Subsequent rounds of fermentation using passive back-slopping without the addition of new L. rhamnosus led to a loss of this strain from the community of fermenting bacteria. Just as Mabisi it is an important part of the daily diet of men, women and children in rural and urban areas in Senegal. The addition of a health-promoting, probiotic bacterium to these products ensures the delivery of that probiotic activity to consumers. The addition of probiotic strains at every fermentation cycle can enrich the existing complex communities of traditionally fermented Lait caillé while traditional bacterial strains remain dominant in the bacterial communities.

In conclusion, the bacterial communities found in the spontaneously fermented products show high diversity independent of production method. Our results show that even in a simple propagation environment and when confronted with new bacteria entering, these communities are able to stay diverse. The next step towards understanding these natural communities would be to construct simple communities based on combinations of individual species, while aiming to generate communities with similar characteristics, for example in terms of stability and aroma formation. The diversity we observed in the bacterial communities leads to functional stability as well as high levels product safety compared to fermented products with a lower species diversity. We found no differences in species diversity between Mabisi resulting from plastic buckets compared to calabashes, suggesting that other fermentation vessels are suitable to replace calabashes. Further, we found that the addition of a probiotic starter can enhance the nutritional value of the fermented product without losing the original fermenting community. These results will be used for improvements in production of traditionally fermented milk to ensure widespread availability of these products thereby improving public health.

Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • Wageningen University
  • Smid, Eddy, Promotor
  • Zwaan, Bas, Promotor
  • Schoustra, Sijmen, Co-promotor
  • Linnemann, Anita, Co-promotor
Award date21 Jun 2019
Place of PublicationWageningen
Print ISBNs9789463439664
Publication statusPublished - 2019


Dive into the research topics of 'Microbial community dynamics in traditionally fermented milk'. Together they form a unique fingerprint.

Cite this