Enhanced malic acid production from glycerol with high-cell density Ustilago trichophora TZ1 cultivations

Bioreactors enable higher cell density resulting in higher volumetric production rates

The potential of Ustilaginaceae as production organisms of different industrially relevant compounds, such as organic acids, lipids, or polyols, has been discussed and demonstrated consistently over the last years [17, 19–25]. Recently, U.
trichophora was found to produce malic acid naturally from glycerol at high titers. By adaptive laboratory evolution and medium optimization, the production rate of this strain in shake flask could be improved to around 0.4 g L^?1 h^?1 reaching titers near 200 g L^?1 [18]. All cultivations ended either upon glycerol depletion or by oxygen limitations due to the viscosity of the cultures. These viscosity issues resulted mainly from the buffering agent, CaCO₃, reacting with produced malate, forming insoluble calcium malate. Although this precipitation might be beneficial for alleviation of product inhibition, it greatly hinders oxygenation of the culture broth in shaking flasks [26].

To overcome handling issues with insoluble components and to avoid glycerol depletion, here we investigate the production process with U.
trichophora TZ1 in bioreactors, in which the pH was kept constant by titration with NaOH. By this, effects of insoluble buffer components on production can be minimized. Further, by feeding additional glycerol prior to depletion, malate titers might be further increased. Additionally, better oxygenation through sparging and stirring, which has a strong influence on microbial organic acid production processes [27], also enables higher cell densities.

Initially, U.
trichophora TZ1 was cultured in pH controlled bioreactors (pH 6.5, NaOH titration) in MTM containing 0.8 g L^?1 NH₄Cl and 200 g L^?1 initial glycerol. An additional 160 g glycerol was fed when the concentration dropped below 50 g L^?1. This results in a slight drop in the measured malate concentrations due to the dilution of the culture broth. The resulting titer (119.9 ± 0.9 g L^?1) and rate (0.13 ± 0.00 g L^?1 h^?1) (Fig. 1b) were significantly lower than those reached in shake flasks with CaCO₃ [18]. Likely, these reductions can be attributed to product inhibition caused by the drastically increased dissolved malate concentration in NaOH-titrated cultures. To improve the production rate, the cell density was increased by using higher concentrations of the growth-limiting nutrient NH₄Cl (1.6, 3.2, and 6.4 g L^?1). Dependent on the initial NH₄Cl concentration, a delay in the onset of malate production could be observed, which can be attributed to a longer growth phase. Maximal OD₆₀₀, however, could be increased from 42 ± 2 with 0.8 g L^?1 NH₄Cl to 80 ± 0 and 115 ± 1 using 1.6 and 3.2 g L^?1 NH₄Cl, respectively (Fig. 1a). As expected, also the overall volumetric malic acid production rate (from the beginning of cultivation until the end) increased to 0.46 ± 0.02 and 0.54 ± 0.07 g L^?1 h^?1 with 1.6 and 3.2 g L^?1 NH₄Cl, respectively (Fig. 1b). 6.4 g L^?1 NH₄Cl, however, did not lead to increased biomass and subsequently production, but had the opposite effect (data not shown). In these cultures, NH₄Cl was no longer depleted during the fermentation. A similar effect was observed for itaconate producing Ustilago maydis MB215 in MTM with NH₄Cl concentrations above 4 g L^?1 [19]. This likely explains the reduced productivity, since nitrogen limitation is the most efficient trigger for organic acid production with Ustilaginaceae [28]. To compensate for this effect, all medium components except for glycerol were doubled in combination with 6.4 g L^?1 NH₄Cl in a subsequent fermentation, resulting in an overall volumetric production rate of 0.54 ± 0.00 g L^?1 h^?1, with a maximal production rate of 1.99 ± 0.04 g L^?1 h^?1 between 45 and 69 h (Fig. 1b).

As expected, an increase in the growth-limiting nutrient led to more biomass formation and consequently to a higher volumetric production rate. There is a good correlation between the maximum malate production rate and the initial NH₄Cl concentration, indicating that the production rate could be further increased as long as secondary limitations are excluded. However, further increases will strongly impact the product yield, since more glycerol is used for biomass formation. Assuming no CO₂ co-consumption, the maximum theoretical yield would be 0.75 mol malate per mole glycerol. However, the glycerol needed for biomass production reduces this maximum, and this reduction is proportional to the initial ammonium concentration. Based on the glycerol consumption during the growth phase (Fig. 1a), approximately 11.5 g of glycerol are needed for biomass formation per gram NH₄Cl. Thus, taking into account the total amount of glycerol consumed by each culture, biomass formation reduces the maximum theoretical yield to 0.73, 0.71, 0.68, and 0.62 mol mol^?1, for 0.8, 1.6, 3.2, and 6.4 g L^?1 NH₄Cl, respectively. This in part explains the reduction in the observed yields in the cultures with higher NH₄Cl concentrations, although in general the yields are only 30–55 % of these theoretical maxima, suggesting that the impact of biomass formation is at the moment relatively low. Improvement in the product yield should be the main focus of future optimization, possibly by the reduction in by-product formation through the disruption of competing pathways. The improvement in specificity for the production of one organic acid is generally considered a promising approach to improve microbial organic acid production. For U.
trichophora TZ1, however, besides 5–10 g L^?1 succinate, no significant amounts of other organic acids were found in HPLC analysis. Additionally, CO₂ and extra- and intracellular lipids are most likely the main by-products. The formation of lipids under organic acid production conditions and their effect on the cells have been described extensively [28, 29]. These by-products can be reduced by knock-out of single genes in the responsive gene clusters [30–32].

Since a significant influence of the starting glycerol concentration on the malic acid production rate has been observed in shake flasks [18], this relation was also studied in bioreactors. Concentration steps of 50 g L^?1 between 150 and 300 g L^?1 were investigated in MTM containing 3.2 g L^?1 NH₄Cl. Additional 160 g glycerol was fed to the cultures one time (300 g L^?1 initial glycerol), two times (150 and 200 g L^?1 initial glycerol), and four times (250 g L^?1 initial glycerol), when the concentration became lower than 50–100 g L^?1 (150 and 200 g L^?1 initial glycerol) or 200 g L^?1 (250 and 300 g L^?1 initial glycerol). Thus, after the consumption of the initial glycerol, its concentrations generally ranged between 50 and 150 g L^?1 (150 and 200 g L^?1 initial glycerol) and 100 and 250 g L^?1 (250 and 300 g L^?1 initial glycerol). Just as in shake flasks, increasing initial glycerol concentrations between 150 and 300 g L^?1 decreased growth rates, final OD₆₀₀ and malic acid production rates (Fig. 2). Possibly, higher glycerol concentrations impose a stress upon the cells. This is also known in other organisms, such as S.
cerevisiae, even though lower glycerol concentrations are generally known to contribute to osmotolerance in different yeast, such as Zygosaccharomyces
rouxii and S.
cerevisiae [33, 34].