Comparison of U. maydis K14 and U. cynodontis ITA MAX pH for itaconate production at low pH values
Despite the advantage of acidic pH values for organic acid production, previous fermentations of U. maydis MB215 ∆cyp3 ∆MEL ∆UA ∆dgat Pria1::Petef ∆fuz7 PetefmttA_K14, henceforth named strain K14 for ease of reference, were performed at a neutral pH of 6.5 due to the largely uninvestigated growth- and production characteristics of this strain at lower pH values. In shake flask cultivations, engineered U. maydis hyper-producers do not grow at acidic pH conditions, but still produce significant amounts of itaconate at lower pH values (28, 29). Hence, itaconate production of U. maydis K14 was assessed in fermentations with different pH values for growth and itaconate production phase (Fig. 1A). To gain a comprehensive comparison, U. cynodontis NBRC9727 ∆fuz7 ∆cyp3 PetefmttA Pria1ria1, henceforth named strain ITA MAX pH for ease of reference, was cultured under similar conditions (Fig. 1B). To study the effect of reduced ammonium concentrations on the product to substrate yield in U. cynodontis ITA MAX pH, an additional low-density fed-batch fermentation was performed (Fig. 1C).
Until 72 h, both high-density fermentations behaved similarly (Fig. 1A, B). The depletion of nitrogen was achieved and itaconate accumulated to approximately 30 g L− 1. However, upon reaching lower pH levels, difference became apparent between the two species. In the case of you U. maydis K14, production almost completely stopped once the pH reached values below 4.0. On the contrary, U. cynodontis ITA MAX pH produced an additional 30 g L− 1 itaconate after reaching acidic pH values. Remarkably, the U. maydis K14 culture did not even reach the final pH value 3.6, although the glucose concentration still declined until the end of the fermentation. This continued glucose uptake in the absence of further production indicates a very high metabolic energy demand for maintaining intracellular pH homeostasis. These results show that U. maydis K14 is not suitable to produce itaconate at lower pH values and is possibly more sensitive towards weak acid stress. Consequently, U. cynodontis ITA MAX pH was identified as a preferable candidate for subsequent characterization. If itaconic acid is to become a bulk chemical, yield is one of the most relevant production parameter because substrate cost is a decisive price-determining factor (14). The availability of nitrogen and the resulting C/N ratio offer a dimension to optimize the product to substrate yield by controlling the biomass density. Typically, lower nitrogen levels result in higher yields but also lower productivities (30). In previous fed-batch fermentations of U. maydis K14 with a reduced ammonium concentration, itaconate was produced at the maximal theoretical yield of 0.72 ± 0.02 gITA gGLC−1 during the production phase (15). A similar trend was observed for U. cynodontis ITA MAX pH during the low-density fermentation (Fig. 1C). This fermentation resulted in a similar itaconate titer of 67.8 ± 0.7 g L− 1 compared to the high-density fermentation. Interestingly, the 5-fold reduction in ammonium chloride as growth-limiting nutrient only resulted in an approximately 2-fold reduction of the maximum OD600 value as well as of the overall production rate (0.22 ± 0.01 g L− 1 h− 1) A similar phenomenon was observed for U. maydis (15). The lower substrate requirement for biomass production enabled a higher yield of 0.55 ± 0.02 gITA gGLC−1. When disregarding the glucose consumed during the first 24 hours in the growth phase, this fermentation achieved the theoretical maximal yield of 0.72 ± 0.01 gITA gGLC−1. This yield is the highest yield ever reported for U. cynodontis. Compared to previously published low-density pulsed fed-batch fermentation with a pH shift from 6.0 to 3.6, the fed-batch with continuous feed increased the titer by 62% and the yield by 41%, while the overall productivity remained nearly constant. These results clearly illustrate the benefit of a continuous glucose feed, preventing osmotic shocks caused by pulsed feeding. However, it is to note that the baseline glucose concentration during feeding was significantly higher than in the pulsed fed-batch fermentation. Although a higher osmotic stress due to elevated glucose concentration would be expected, it is also plausible that approximately 100 g L− 1 glucose represents a threshold concentration for achieving more efficient itaconate production while maintaining relatively low osmotic stress. These findings are in line with those reported for itaconate production with A. terreus, where the highest yields were obtained at glucose concentration between 120 to 200 g L− 1 (31). Almost the same is reported for citrate production in A. niger (32). This phenomenon should be further investigated for itaconate production with Ustilaginaceae.
Comparison of the itaconate production capabilities of U. cynodontis ITA MAX pH at neutral and acidic pH values
Previous research has demonstrated that U. cynodontis is also able to grow at the acidic pH value 3.6 (19). To investigate the impact of reduced pH values throughout the entire fermentation process, additional continuous fed-batch fermentations were conducted as described above. Growth and production capabilities achieved at pH 3.6 were compared to values obtained from fermentations at pH 6.5.
In the neutral pH fermentation, 64.7 ± 10.5 g L− 1 itaconate was produced within approximately 160 h with an overall productivity of 0.40 ± 0.06 g L− 1 h− 1 and the yield 0.42 ± 0.02 gITA gGLC−1 (Fig. 2B). Previous fed-batch fermentation with a pH-shift from 6.5 to 3.6 achieved similar KPIs (Fig. 1B). However, the low pH fermentation resulted in a higher itaconate yield of 0.49 ± 0.01 gITA gGLC−1 (Fig. 2A). In addition, the overall productivity was increased by 39% (0.57 ± 0.01 g L− 1 h− 1) and the titer by 22% (79.2 ± 1.3 g L− 1). These results show that U. cynodontis ITA MAX pH not only tolerates acidic conditions, it actually produces better at lower pH values. This result is promising, given that the acidic pH fermentation required approximately 3.5-fold less NOH compared to the neutral pH fermentation (112 mL and 398 mL 5 M NaOH solution). The KPIs are in good accordance with those previously achieved with this strain using a constant glucose feed controlled by an inline glucose sensor (78.6 g L− 1, 0.45 gITA gGLC−1, 0.42 g L− 1 h− 1) (19). Less optimal progenitor strains of U. cynodontis ITA MAX pH are capable of producing itaconate even at pH levels below 3.6 (19). Given the pKa values of itaconic acid of 3.84 and 5.55, further reduction of the fermentation pH is expected to still significantly reduce base consumption (33). Thereby, costs associated with pH adjusting reagents can be further reduced. In addition, less hydrochloric acid (HCl) is necessary for DSP. This leads to a lower amount of co-salt in the fermentation broth, which can therefore be further concentrated which increases the crystallization yields. As shown by Saur et al. (14), this increased yield in DSP is able to compensate for partial yield loss in fermentation (cf. introduction). As a result, this capability holds the potential to improve the economic viability of the itaconic acid production process with U. cynodontis.
Identification of the pH optimum for itaconate production with U. cynodontis ITA MAX pH
The pH plays a crucial role as it determines the itaconic acid species distribution during the fermentation process. On the one hand, protonated itaconic acid negatively impacts the efficiency of the fermentation as it leads to weak acid uncoupling, which increases maintenance demand through energydriven export of protons, and it possibly increases product inhibition by raising the intracellular itaconate concentration. On the other hand, the protonated acid greatly facilitates DSP as it avoids additional acid use and salt coproduction as described above. Therefore, it is crucial to carefully determine the optimum pH value in order to balance these effects and achieve the overall most efficient itaconate production. However, the pH optimum for itaconate production with U. cynodontis has so far only been determined with the sub-optimal production strain containing only the fuz7 deletion. To examine the pH optimum of the new itaconate hyper-producing strain, a series of pH controlled fed-batch fermentations were conducted in standardized conditions. To avoid growth defects due to pH values below 3.6, the initial biomass production phase was performed at pH 3.6 for all fermentations. Following the depletion of the nitrogen source, the pH was allowed to drop to the corresponding lower pH value. To adjust the pH value above values of 3.6, NaOH was added.
During the initial biomass production phase, a similar growth was observed across all experiments. However, after the pH was adjusted to the corresponding value being tested, differences in the cell densities became apparent. pH values below 3.6 resulted in decreased optical densities (Fig. 3D), indicating acid stress of the cells. The stress is most likely caused by weak acid uncoupling, which is more prominent at low pH values due to higher fractions of the double-protonated species (Fig. 3A). The increased weak acid uncoupling at these lower pH values is also reflected in reduced yields (Fig. 3B). The lowest yield was observed at the minimal pH value of 2.1, which was identified in a batch fermentation without pH control during the itaconate production phase. However, the NaOH consumption during this fermentation was 36-fold reduced compared to the fermentation at pH 5.5 (Fig. 3C, Additional file 1). The highest yield with moderate base addition was achieved at pH 3.6, similar to what was previously determined for the morphology-engineered strain. The double-protonated form remained relatively constant within the pH range of 2.8 and 3.4, suggesting a potential inhibitory threshold for the cells (Fig. 3A). At pH 3.6, there is a significant reduction in H2ITA. This reduction could explain the high KPIs observed at this pH value, indicating that H2ITA concentrations are key to achieving maximum KPIs. Considering only the yield, the additional genetic modification did not change the pH optimum. The morphology-engineered strain also achieved the highest itaconate titer at 3.6, and showed afterwards declining titer with increasing pH values. Regarding itaconate titers, it consequently appears that higher pH values negatively impact the regulation of the itaconate cluster genes of the fuz7 variant. The new itaconate hyper-producing strain however, showed increasing titers with increasing pH values until a pH of 5.0 (Fig. 3A). This may be due to the interference in the regulatory mechanisms of the itaconate cluster genes by the ria1 overexpression. The overexpression of ria1 may have contributed to an increased tolerance of the strain towards higher product concentrations. It may also be possible that the overexpression of ria1 reduced the pH dependency of the regulation of the itaconate cluster genes. One of the main potential benefits of production at more neutral pH values lies in the potentially higher titers (10, 33). In order to determine the maximum itaconate titer with this strain, an additional fed-batch fermentation was performed with a prolonged feeding phase.
In the fed-batch fermentation with a prolonged feeding phase, the itaconate titer kept linearly increasing up to 264 h up to approximately 92.3 ± 10.7 g L− 1 at a rate of 0.34 ± 0.01 g L− 1 h− 1 and a yield of 0.42 ± 0.01 gITA gGLC−1 (Fig. 4). During the remaining fermentation time, a linear increase of the itaconate titer could still be observed, however with a strongly reduced rate as the product inhibition became more and more prominent, taking another 434 h to produce only 32.9 ± 4.0 g L− 1 additional itaconate. This fermentation reached a final itaconate titer of 125.2 ± 14.6 g L− 1, one of the highest itaconate titers reported for Ustilaginaceae with NaOH titration. The itaconate concentration remained constant between 696 h and 720 h, indicating that the maximum titer was finally reached after approximately one month of fermentation time. In total, this fermentation resulted in an overall productivity of 0.17 ± 0.02 g L− 1 h− 1 and a yield of 0.36 ± 0.01 gITA gGLC−1. Although a very high titer could be achieved through extended feeding, this came at the major expense of a lower yield end rate. Despite the reduced weak acid stress at the pH value of 5.0 and the higher itaconate production per cell, this fermentation highlighted the strong inhibitory effect of elevated product titers on the over KPIs, indicating a major efficiency loss at titers above 92.3 ± 10.7 g L− 1 even at a higher pH value. This result highlights that not only elevated H2ITA concentrations are inhibitory for the cells, but also higher overall product titers due to increased osmotic stress.
In summary, the engineered U. cynodontis ITA MAX pH strain shows an extended operational range for itaconate production compared to the previous morphology-engineered strain, particularly in terms of itaconate titers. The carbon balance of all fermentations is shown in Additional file 3. Analogous to the morphology-engineered strain, the achieved KPIs for the hyper-producing strain exhibited a moderate decrease at pH values below 3.6, which became more prominent at pH values below 3.0 (Fig. 3B). However, the volumes of NaOH added during these low-pH fermentations were also significantly reduced (Fig. 3C) and associated reductions in acid consumption and saline waste production during DSP can be expected. To assess whether these can economically compensate for the losses in fermentation yield at pH values lower than 3.6, an operational cost analysis was performed. We also added the economic results for fed-batch fermentations at higher pH values than 3.6 to provide a full picture of the cost structure changes.
Identification of the pH optimum by operative cost analysis
The process KPIs from the fed-batch fermentations served as an input parameter for the simulation. Those consist of product titer, fermentation yield, and pH during product formation phase (Fig. 3A, B). The results of the cost analysis are displayed in Fig. 5.
As expected, the simulated costs associated to acid and base use and saline waste disposal decrease with lower fermentation pH. Reduced amounts of HCl and saline waste could also be confirmed in crystallization experiments using real cultivation supernatants from batch fermentations (Additional file 4). However, those costs comprise only a small fraction of the total operational costs. Independent of the selected pH, the total operational costs are most strongly influenced by overall substrate yield. The outstanding fermentation yield at pH 3.6 cuts costs significantly while specific operational costs are visibly larger at high and low pH values due to the poor substrate to product conversion. This cannot be outweighed by higher DSP yields as at lower pH as discussed in Saur et al. (14). The specific operational costs of approximately 1.04 EUR kg− 1 at pH 3.6 achieved in this work are approximately 0.40 EUR kg− 1 lower than those previously obtained for this strain (14). Nevertheless, to successfully compete with the petrochemical production of the counterparts acrylic acid and methacrylic acid, it is necessary to further reduce the production costs (34).