In situ adsorption of itaconic acid from fermentations of Ustilago cynodontis improves bioprocess efficiency

Abstract

Background: Reducing the costs of biorefinery processes is a crucial step in replacing petrochemical products by sustainable, biotechnological alternatives. Substrate costs and downstream processing present large potential for improvement of cost efficiency. The implementation of in situ adsorption as an energy-efficient product recovery method can reduce costs in both areas. While selective product separation is possible at ambient conditions, yield-limiting effects, as for example product inhibition, can be reduced in an integrated process.

Results: An in situ adsorption process was integrated into the production of itaconic acid with Ustilago cynodontis IA_max, as an example of a promising biorefinery process. A suitable feed strategy was developed to enable efficient production and selective recovery of itaconic acid by maintaining optimal glucose concentrations. Online monitoring via Raman spectroscopy was implemented to enable a first process control and understand the interactions of metabolites with the adsorbent. In the final, integrated bioprocess, yield, titre, and space-time yield of the fermentation process were increased to values of 0.41 g_IA/g_Glucose, 126.5 g_IA/L and 0.52 g_IA/L/h. This corresponds to an increase of up to 30% in comparison to the first extended batch experiment without in situ product removal. Itaconic acid was recovered with a purity of at least 95% and high concentrations above 300 g/L in the eluate.

Conclusion: Integration of product separation via adsorption into the bioprocess was successfully conducted and improved the efficiency of itaconic acid production. Raman spectroscopy was proven to be a reliable tool for online monitoring of various metabolites and facilitated design and validation of the complex separation and feed process. The general process concept can be transferred to the production of various similar bioproducts, expanding the tool kit for design of innovative biorefinery processes.

Keywords: Adsorption; Biorefinery processes; Downstream processing; Integrated bioprocesses; Itaconic acid; Raman spectroscopy.

Fig. 1

Extended batch cultivation of U. cynodontis Δfuz7^r Δcyp3^r p_etefmtta p_ria1ria1 in a 2-L fermenter with 100 g/L initial glucose and sequentially decreasing glucose feed rates. Depicted are A oxygen transfer rate (OTR) and concentrations of glucose and itaconic acid, B volumetric aeration rate and filling volume, C pH, optical density (OD₆₀₀), cell dry weight (CDW) and volume of titrated base solution and D respiratory quotient (line at RQ = 1) and erythritol concentration over time. Cultivation was performed at T = 30 ℃, n = 300–800 rpm, pH_initial = 5.92, pH_control = 3.6 (10 M NaOH), OD_600,initial = 0.5 and V_L,initial = 1250 mL in adapted Verduyn medium in a 2-L fermenter. Aeration rate adjusted at beginning of each feed phase to reach a mean volumetric aeration rate of q_in = 60 SL/L/h. Glucose was fed from a 517.3 g/L solution, resulting in feed rates of I: 3.6 g/h, II: 3.0 g/h, III: 2.4 g/h, IV: 1.9 g/h and V: 1.3 g/h. Feed rates were calculated from glucose accumulation in the experiment shown in Fig. SI 7. For concentrations, mean values of two replicates are shown

Fig. 2

Adsorption characteristics for itaconic acid and glucose on activated carbons. A Relation between polarity and affinity to hydrophobic activated carbons for the different itaconic acid species and glucose. Metabolites depicted in order of decreasing affinity. H₂IA: fully protonated itaconic acid, HIA⁻: single protonated itaconic acid, IA²⁻: fully dissociated itaconic acid. B pH-dependent distribution of itaconic acid species as described in [45]. C pH-dependent adsorption capacity of itaconic acid and glucose from a 1:1 mixture (30 g/L) on Blücher 100562 activated carbon at T = 30 ℃ after 1 h of batch adsorption. D Relation of outlet to inlet concentration for itaconic acid and glucose over the relation of accumulated inlet volume to adsorbent volume for a continuous adsorption on an adsorption column. The column was filled with 4 mL of activated carbon. The adsorption experiment was performed with a 1:1 mixture (60 g/L) at T = 30 ℃ and an inlet flow rate of V_F = 1 mL/min. E Illustration of the distribution of metabolites adsorbed to the column for the three different phases from the experiment depicted in D

Fig. 3

Simplified illustration of liquid flows for in situ adsorption and Raman analysis for the different phases of product separation cycles. A schematic drawing of the complete experimental setup with all pumps and valves is shown in Additional file 1: Figure S7. Tanks are filled with DI-water (dark blue), waste (green), ethanol (light blue) and product solution (red). Flow directions during different steps: A fermentation: bioreactor → Raman sensor 1 → bioreactor. B Washing: water tank → adsorption column → Raman sensor 2 → waste tank (fermentation step is run simultaneously). C Water displacement: bioreactor → Raman sensor 1 → adsorption column → Raman sensor 2 → waste tank. D Adsorption: bioreactor → Raman sensor 1 → adsorption column → Raman sensor 2 → bioreactor. E Broth displacement: ethanol tank → adsorption column → Raman sensor 2 → bioreactor. F Desorption: ethanol tank → adsorption column → Raman sensor 2 → product tank (fermentation step is run simultaneously)

Fig. 4

Proof of concept cultivation of U. cynodontis Δfuz7^r Δcyp3^r p_etefmtta p_ria1ria1 in a 2 L fermenter with 100 g/L initial glucose, a linear glucose feed rate and in situ separation of itaconic acid by adsorption. Depicted are oxygen transfer rate (OTR) and concentrations of glucose, itaconic acid and ethanol over time. Cultivation was performed at T = 30 ℃, n = 300–650 rpm, q_in = 60 SL/L/h, pH_initial = 5.78, pH_control = 3.6 (10 M NaOH), OD_600,initial = 0.5 and V_L,initial = 1250 mL in adapted Verduyn medium in a 2-L fermenter. DOT was controlled at 30% by adjusting stirrer speed. Glucose was fed from a 440.7 g/L solution. Column was filled with 11.3 mL of Blücher 100562 activated carbon. The adsorption steps of product separation cycles were performed from 92 to 96 h and 116–118 h. For concentrations, mean values of two replicates are shown

Fig. 5

Proof of concept cultivation of U. cynodontis Δfuz7^r Δcyp3^r p_etefmtta p_ria1ria1 in a 2 L fermenter with 100 g/L initial glucose, sequentially decreasing glucose feed rates and in situ separation of itaconic acid by adsorption. Depicted are A oxygen transfer rate (OTR) and concentrations of glucose, itaconic acid and ethanol over time. B, C Are enlarged representations of concentrations of itaconic acid during adsorption cycles over time. Cultivation was performed at T = 30 ℃, n = 300–600 rpm, q_in = 60 SL/L/h, pH_initial = 5.75, pH_control = 3.6 (10 M NaOH), OD_600,initial = 0.5 and V_L,initial = 1250 mL in adapted Verduyn medium in a 2-L fermenter. DOT was controlled at 30% by adjusting stirrer speed. Glucose was fed from a 500.3 g/L solution, resulting in feed rates of I: 4.0 g/h, II: 3.3 g/h, III: 2.6 g/h, IV: 1.9 g/h and V: 1.2 g/h. Column was filled with 11.3 mL of Blücher 100562 activated carbon. The adsorption steps of product separation cycles were performed from 47.7 to 48.1 h and 70.6 to 70.9 h. For concentrations, mean values of two replicates are shown

Fig. 6

Mass fractions of different metabolites calculated from the Raman signal of flow cell 2 (see Additional file 1: Fig. S7) behind the adsorption column. Depicted are mass fractions of the analysed Raman-active substances over time and the different phases of the product separation cycle according to Fig. 3. Start and end points of different phases in the experiment were adjusted according to the measured Raman signals. The displayed product separation cycle is the second cycle from the experiment displayed in Fig. 7 (87.4–99.5 h). The column was filled with 47 mL of Blücher 100562 activated carbon

Fig. 7

Cultivation of U. cynodontis Δfuz7^r Δcyp3^r p_etefmtta p_ria1ria1 in a 2-L fermenter with 100 g/L initial glucose, sequentially decreasing glucose feed, in situ separation of itaconic acid by adsorption and concentration monitoring by Raman spectroscopy. Depicted are A oxygen transfer rate (OTR) and concentrations of glucose and itaconic acid, B volumetric aeration rate and filling volume, C pH, optical density (OD₆₀₀), cell dry weight (CDW) and volume of titrated base solution and D respiratory quotient (line at RQ = 1) and erythritol concentration over time. Cultivation was performed at T = 30 ℃, n = 300–800 rpm, Q_in = 60 SL/h, pH_initial = 5.67, pH_control = 3.6 (10 M NaOH), OD_600,initial = 0.5 and V_L,initial = 1800 mL in adapted Verduyn medium in a 2 L fermenter. DOT was controlled at 30% by adjusting stirrer speed. Glucose was fed from a 510.2 g/L solution resulting in feed rates of I: 3.4 g/h, II: 2.8 g/h, III: 2.0 g/h and IV: 1.6 g/h. Column was filled with 47 mL of Blücher 100562 activated carbon. The adsorption steps of product separation cycles were performed from 68.9 to 73.3 h, 92.2 to 95.7 h, 113.4 to 116.9 h and 136.4 to 140.4 h. For concentrations from HPLC measurements, mean values of two replicates are shown