Characterisation of Lactobacillus plantarum single and multi-strain biofilms

link to thesis

Research described in this thesis provides new insights in single and multi-strain biofilm formation of Lactobacillus plantarum WCFS1 and L. plantarum strains isolated from different environments including fermented and spoiled foods. Studies on biofilms in relation to L. plantarum strain diversity and mechanisms of biofilm formation are highly relevant for the development of biofilm control and prevention strategies.

Chapter 1 describes an overview of the role of L. plantarum in food spoilage in relation to its nutrient requirements and colonisation of abiotic surfaces. Additionally, general concepts of biofilm formation with single and multiple microbial species are provided, including a description of tools used to study biofilm formation and a description of cellular factors that influence L. plantarum biofilm development.

Chapter 2 describes the analysis of the biofilm forming capacity of L. plantarum WCFS1 and six L. plantarum food spoilage isolates. Biofilm formation as quantified by crystal violet (CV) staining and colony forming units was largely affected by the medium composition, growth temperature and maturation time as well as by strain specific features. All strains showed highest biofilm formation in Brain Heart Infusion medium supplemented with manganese and glucose. For L. plantarum biofilms, the CV assay, that is routinely used to quantify total biofilm formation, correlates poorly with the number of culturable cells in the biofilm. This can in part be explained by cell death and lysis resulting in CV stainable material, conceivably extracellular DNA (eDNA), contributing to the extracellular matrix. For example, increasing biofilm maturation times from 24 h up to 72 h, and increasing temperature from 20 to 37°C, resulted in decreased culturable cell numbers with a concomitant increase of CV staining. The strain to strain variation may in part be explained by differences in release of eDNA, likely as result of differences in lysis behaviour. In line with this, biofilms of all strains tested, except for one spoilage isolate, were sensitive to DNase treatment. In addition, biofilms were highly sensitive to treatment with Proteinase K suggesting a role for proteins and/or proteinaceous material in surface colonisation. This study shows the impact of a range of environmental factors and enzyme treatments on biofilm formation capacity for selected L. plantarum isolates associated with food spoilage, and may provide clues for disinfection strategies in food industry. Chapter 3 provides new insights into biofilm development by L. plantarum WCFS1 through comparative analysis of wild type and mutants affected in cell surface composition, including a mutant deficient in the production of Sortase A involved in the covalent attachment of 27 predicted surface proteins to the cell wall peptidoglycan (ΔsrtA) and mutants deficient in the production of capsular polysaccharides (CPS1-4, (Δcps1-4). Surface adhesion and biofilm formation studies revealed none of the imposed cell surface modifications to affect the initial attachment of cells to polystyrene while CV-stainable biofilm was severely reduced in the ΔsrtA mutant and significantly increased in mutants lacking the cps1 cluster, compared to the wild-type strain. Fluorescence microscopy analysis of biofilm samples pointed to a higher presence of eDNA in cps1 mutants and this corresponded with increased autolysis activity. Subsequent studies using Δacm2 and ΔlytA mutants, deficient for specific peptidoglycan hydrolases and affected in lytic behaviour, revealed reduced CV staining of biofilms, confirming the relevance of lysis for the build-up of the biofilm matrix with eDNA. Additionally, performance of L. plantarum strains obtained from different origins was determined in single and in competitive multi-strain biofilm formation models.

Two different approaches were used to monitor the population dynamics of multi-strainbiofilm formation: quantitative PCR (qPCR) (Chapter 4) and next generation sequencing (NGS) (Chapter 5). In Chapter 4, six spoilage related L. plantarum strains (FBR1-FBR6) and L. plantarum WCFS1 were characterised in single and multiple strain competition models. A quantitative PCR approach was used with added propidium monoazide (PMA) enabling quantification of cells in the biofilm without membrane damage, representing the viable cell fraction that determines the food spoilage risk. The results show that the performance of individual strains in multi-strain cultures generally correlates with their performance in pure culture, and relative strain abundance in multi-strain biofilms positively correlated with the relative strain abundance in suspended (planktonic) cultures. The total biofilm quantified by CV staining of the multi-strain biofilms showed a positive correlation with CV values of the dominant strain obtained in single strain studies. However, the combination of FBR5 and WCFS1 showed significantly higher CV values compared to the individual performances indicating that total biofilm formation was higher in this specific condition. Notably, L. plantarum FBR5 was able to outcompete all other strains and showed the highest relative abundance in multi-strain biofilms. All the multi-strain biofilms contained a considerable number of viable cells, representing a potential source of contamination. As a follow up, in Chapter 5, the population dynamics of a mixture of 12 L. plantarum isolates of different origins was studied in competitive planktonic and static and dynamic flow biofilm growth models as a function of different parameters, namely maturation time, temperature and medium composition (BHIGlucose containing low or high Mn(II), in presence and absence of haem and vitamin K2). A NGS approach based on detection of strain specific alleles was used to determine the relative abundance of each strain in the different conditions. Data were obtained in the presence and absence of PMA, thus allowing for identification and quantification of relative contributions of each individual L. plantarum strain to the fraction of viable cells in planktonic and biofilm phase and the fraction of dead cells (compromised membrane) and levels of eDNA in the biofilm matrix, respectively. This approach revealed that the relative abundance of each strain in the static biofilm positively correlates with its performance in static planktonic conditions. The genome content of the two groups of strains that dominated the static and dynamic flow biofilms was explored to identify genetic factors that potentially contribute to biofilm forming capacity under static and dynamic flow conditions, respectively.

The results presented in this thesis contribute to the understanding of mechanisms of biofilm formation and matrix composition of L. plantarum resulting in the design of a more detailed model for static biofilm formation on abiotic surfaces (Chapter 6). This model was built based on the analysis of biofilm mass (including matrix components) and culturable cells retained on the surface after multiple washing steps. Combining all observations, it is proposed that sedimentation of planktonic cells to surface is notonly required for the initial attachment, but also for biofilm development. Both, eDNA and protein-protein interactions and/or proteinaceous material contribute to for cellcell/cell-surface interactions and biofilm matrix. Based on the model, a role for DNase and/or proteinase enzyme-mediated prevention of biofilm formation and dispersal has been suggested. The newly obtained insights on the population dynamics of multi-strainL. plantarum isolates obtained from different origins in static and dynamic flow biofilm models, provide further leads for research into underlying mechanisms involved and may further add to development of tools and strategies to prevent (re)contamination from biofilms in food processing environments.