Enhancing aroma production by lactic acid bacteria at near-zero growth rates: a retentostat approach

Lactic acid bacteria (LAB) are extensively used for the production of fermented foods from both animal and plant origin, such as cheese, yoghurt, kimchi and sauerkraut. Their predominant function in these processes is to produce organic acids, mainly lactic acid, thereby lowering the pH, which contributes to the safety and shelf life of the product. Nowadays, many fermented foods are made using starter cultures for a consistently safe product of constant quality. In Dutch-type cheeses, the starter cultures consist of various strains of mainly two species: Lactococcus lactis and Leuconostoc mesenteroides. L. lactis subsp. cremoris and L. lactis subsp. lactis are considered to be the main acid producers that dominate during the early stages of cheese manufacturing when nutrients are abundant. During cheese ripening, nutrients are scarce and LAB encounter long periods of nutrient limitation leading to slow growth. In particular, L. lactis biovar diacetylactis and Lc. mesenteroides survive these periods of extreme nutrient limitation and still contribute to aroma formation. In this thesis, the dairy-isolated strains Lactococcus lactis biovar diacetylactis FM03P and Leuconostoc mesenteroides FM06 were studied focussing on the effect of near-zero growth rates on their quantitative physiology and aroma formation capacity. Moreover, possible strategies to enhance aroma formation at near-zero growth rates are described.

The genomes of both lactic acid bacteria strains were sequenced, including the plasmids they harboured, (Chapter 2) revealing that their genomes carried clear signatures of adaptation to the dairy environment.

The plasmids of L. lactis FM03P and the genes they carried were further characterised in Chapter 3. Plasmid-encoded functions that were identified include lactose utilisation, citrate uptake, oligopeptide uptake and peptide degradation, bacteriophage resistance, uptake of cations (magnesium and manganese), exopolysaccharide production, and stress resistance. Acquisition of the plasmids most likely facilitated the adaption of these strains to the dairy environment. Interestingly, some plasmids were already lost in a single propagation step, signifying their instability in the absence of a selection pressure and demonstrating that propagation should be minimised when studying dairy isolates of L. lactis.

In L. lactis, citrate utilisation has been linked to the production of the important aroma compounds diacetyl and acetoin. In L. lactis biovar diacetylactis, the limiting step in citrate utilisation is the transport of citrate across the cell membrane, which is facilitated by the plasmid-encoded citrate permease (CitP). Although it is known citP expression is induced at low pH, the effect of citrate is still under debate. Moreover, the role of the plasmid copy number (number of plasmid copies per chromosome) has never been taken into account. In Chapter 4, we systematically analysed regulation of citrate utilisation by pH, nutrient limitation and presence of citrate at four different levels: i) plasmid copy number, ii) citP transcription, iii) citP mRNA processing and iv) citrate utilisation capacity. Induction of the citP gene at low pH was confirmed, but citP expression increased even more in the presence of citrate. In cells grown at low pH or under amino acid limitation, also the copy number of the citP-containing plasmid slightly increased. No significant effects on citP mRNA processing were found. Due to the increased citP expression, the citrate utilisation rate increased from approximately 1 to 65 μmol.min-1.gDW-1 and also the production of acetoin increased significantly. This knowledge was used in other conditions to select conditions that improved citrate-driven flavour formation.

In cheese manufacturing, aroma compounds are mainly formed by the bacteria in the ripening phase in which nutrients are limited resulting in severely reduced growth rates. To investigate if reduced nutrient availabilities and low growth rates are required for the formation of aroma compounds, L. lactis FM03-V1 was grown for 2 weeks in batch and retentostat cultures and in a milli-cheese model system. Subsequently, aroma compounds were analysed by headspace solid phase microextraction gas chromatography mass spectrometry (Chapter 5). While batch cultures were performed with milk, hydrolysed micellar casein isolate (MCI) and a chemically defined medium (CDM), retentostat cultures were only performed with CDM. Despite the use of CDM, aroma production in retentostat cultivations had the biggest qualitative overlap with aroma production in the milli-cheese model system demonstrating that low growth rates are required to produce cheese-related compounds. In total, 52 known cheese compounds were produced in retentostat cultures. In cultures with CDM and MCI, free fatty acids and their corresponding degradation products were underrepresented compared to the milli-cheeses. Therefore, addition of free fatty acids to MCI and CDM might help to enhance flavour formation, thereby better resembling flavour formation in cheese.

To further study the effect of near-zero growth rates on the aroma formation capacity and quantitative physiology, L. lactis FM03-V1 was grown for 5 weeks in retentostat cultures (Chapter 6). Growth rates decreased to less than 0.001 h-1, while the viability remained above 80%. Interestingly, a large fraction of the cells lost the ability to grow in M17 plates, indicating that they entered a viable but non-culturable state. Dynamic modelling revealed that the maintenance coefficient of this dairy strain decreased 7-fold at near-zero growth rates compared to high growth rates. Most likely, the bacteria saved energy by limiting energy used for protein turnover. Changes in aroma formation in the retentostat culture resembled biochemical changes occurring during cheese ripening, such as amino acid and fatty acid catabolism. Analysis of complete and cell-free samples revealed that lipophilic compounds accumulated in the cells, most likely in the cell membranes.

Starter cultures for cheese production often consist of multiple strains and species, which could result in a higher aroma complexity. To investigate if retentostat co-cultures could further enhance aroma formation, L. lactis FM03-V1 and Lc. mesenteroides FM06 were grown in retentostat mono- and co-cultures (Chapter 7). Both species had similar physiological responses to near-zero growth rates, including that a large fraction was viable but not culturable and that the maintenance coefficient decreased approximately 7-fold. However, L. lactis reached a higher biomass concentration in retentostat mono-cultures due to a higher ATP yield on substrate, a higher biomass yield on ATP and a lower maintenance requirement. A dynamic model was developed to predict biomass accumulation in retentostat co-cultures. This model predicted the biomass accumulation very well with L. lactis dominating in the retentostat co-cultures (ratio 100:1). Aroma compounds specific for both species were identified in retentostat mono-cultures. No additional unique aroma compounds were identified in the co-cultures, although 2 out of 5 compounds specific for Lc. mesenteroides were found in the co-cultures despite its low abundance. This indicates that a similar or even higher aroma complexity could be obtained by mixing two retentostat mono-cultures.

Lactic acid bacteria can carry multiple plasmids affecting their performance in dairy fermentations. The expression of plasmid-borne genes and the activity of the corresponding proteins is severely affected by changes in the number of plasmid copies. Therefore, we analysed plasmid copy numbers in a wide variety of dairy-related conditions, including near-zero growth rates, typical for cheese ripening (Chapter 8). The copy numbers of theta-type replicating plasmids were negatively correlated with the size of the plasmids. Plasmid copy numbers were remarkably stable in the extremely wide range in growth rates (0.0003 h-1 to 0.6 h-1), suggesting strict control of the plasmid copy number even at extreme nutrient restriction. Copy numbers were also hardly affected by varying the pH value, the nutrient limitation or the presence of citrate, signifying the stability in copy number of the plasmids.

Although retentostat cultivation is a unique cultivation method to study microorganisms at near-zero growth rates, it has some disadvantages as production platform in particular due to the time dependency. Partial cell recycling chemostats overcome this problem, while near-zero growth rates can still be obtained as demonstrated in Chapter 9. Using a partial cell recycling chemostat cultivation, the decreased maintenance requirement of L. lactis at near-zero growth rates was confirmed. Moreover, it was demonstrated that particular aroma compounds were affected by the growth rate with the studied range (0.0025 to 0.025 h-1. Finally, the potential of partial cell recycling chemostat cultivation as production platform and as unique research tool are discussed.

The results of the experimental chapters (Chapter 2-9) were further discussed in Chapter 10, including a comparison with transcriptomics data of retentostat cultures of the plant-associated L. lactis KF147. This showed that the decreased maintenance requirement at near-zero growth rates might be explained by switching from degradation and re-synthesis of misfolded proteins to disaggregation and refolding of misfolded proteins with the ClpB-DnaK system. The function of aroma formation in lactic acid bacteria was also discussed, in particular in relation to their possible role as release valve for the excess of branched-chain amino acids.

In conclusion, this thesis described and revealed new physiological, genetic and metabolic adaptations of lactic acid bacteria to near-zero growth rates. This helps to understand how microbes adapt in food fermentation processes as well as in natural environments in which nutrients are often scarce. Moreover, this work demonstrates the importance of slow growth for the production of aroma compounds by lactic acid bacteria and provides new opportunities for enhanced production of aroma compounds.