Yeast, media, and cultivation

The industrial yeast for fuel ethanol production, Superstart™ active distillers dry yeast, Saccharomyces cerevisiae, was purchased from Lallemand Ethanol Technology (Milwaukee, WI, USA) and used in this study. Media for seed cultures and ethanol fermentation as well as the cultivation conditions were used as described in Wang et al. (2017). Briefly, yeast nitrogen base (YNB) medium was made of 50 g/L glucose (MilliporeSigma, Burlington, MA, USA) and 6.7 g/L yeast nitrogen base with amino acids (MilliporeSigma) in 0.1 M sodium phosphate buffer (\(NaH_2PO_4\cdot2H_2O/Na_2HPO_4\cdot2H_2O,pH\;6.0\), Thermo Fisher Scientific, Waltham, MA, USA), and the medium was sterilized by filtration (0.2 µm Sartolab™ P20 Pressure Filters, Sartorius, Göttingen, Germany). A glycerol stock of the yeast was thawed and steaked on a yeast-peptone-dextrose agar (YPDA) plate which was made of 50 g/L yeast-peptone-dextrose medium (BD Difco™, Bergen Country, NJ, USA) and 15 g/L agar (Thermo Fisher Scientific) sterilized by autoclaving (121 °C/15 min). The YPDA plates were incubated at 30 °C for 2–3 days in an incubator shaker (New Brunswick Innova 44). A single isolated colony was then transferred to 20 mL YNB medium (in a 50-mL Erlenmeyer flask) and grown at 30 °C/230 rpm for 18–20 h in the incubator shaker to make up the first seed culture. A sample from this first seed culture was transferred to 200 mL YNB medium (in a 500-mL Erlenmeyer flask) to reach an initial optical density at 600 nm (OD600) of ~ 0.2 and grown at 30 °C/250 rpm to make up the second seed culture. Finally, when an OD600 of ~ 0.5 was reached, the active yeast culture was transferred to the 5-L bioreactor as inoculum for fed-batch fermentation.

Fed-batch fermentationSettings and accessories for 5-L bioreactors

All fermentation experiments were performed in a 5-L stirred tank bioreactors (Infors-HT, Bottmingen, Switzerland), and the operation settings and accessories were used as described in Hung et al. (2023); Wang et al. (2020). Briefly, 160 mL from the second seed culture at OD600 ~ 0.5 was added to 1.84 L of YNB medium for an initial working volume of 2.00 L, where the inoculum was 8% (v/v). The bioreactors were operated anaerobically at 30 °C and 600 rpm. Dissolved oxygen and pH were monitored using a dissolved oxygen probe (VisiFerm DO ECS 325 mm H0, Hamilton Company, Reno, NV, USA) and a pH probe (EasyFerm Plus PHI K8 325 mm, Hamilton Company), respectively, via the software Iris 6.0 (Infors-HT). 2 N NaOH (Thermo Fisher Scientific) was used for pH control (set point at pH > 4.0) to mitigate contamination and reduce metabolic inhibition at low pH. A mass flow meter (M-200SCCM-D/5 M, Alicat Scientific Inc., Tucson, AZ, USA) was connected to the effluent of the bioreactor to monitor the evolved gas flow rate and cumulative gas production in real time. The evolved gas flow rate has been shown to be an effective monitoring parameter for ethanol fermentation (Wang et al. 2020).

Continuous feeding strategies

The feed medium added during fed-batch fermentation was a modified YNB medium: 500 g/L glucose (MilliporeSigma), 6.7 g/L yeast nitrogen base with amino acids (MilliporeSigma), and 47.7 g/L urea (MilliporeSigma) in 0.1 M sodium phosphate buffer (pH 6.0). The nitrogen-to-carbon ratio (N/C) in the feed medium was 0.95:10 (mol/mol), which is similar to the N/C ratio of 1:10 commonly used to support yeast growth and ethanol production in industrial settings (Kiran and Liu 2015).

Continuous feeding was performed using a peristaltic pump (MFLX78001-62, Masterflex Ismatec Reglo Independent Channel Control Peristaltic Pumps, Avantor, Mississauga, Ontario, Canada).

Evolved gas produced per reactor volume (E with units of mL/L), also known as evolved gas production—the amount of gas released from the metabolic activity of the cell culture per volume, during fermentation positively correlated with glucose consumed per reactor volume (G with units of g/L), as expressed in Eq. 1 (R2 = 0.895; Supplementary Fig. S1):

$$G=0.005\cdot E+0.302$$

(1)

Evolved gas from the bioreactor was used to control the glucose addition in adapted continuous feedings, and the real-time evolved gas flow rate was used as a monitoring parameter to indicate the end of the initial batch phase, at which point the fed-batch phase was initiated.

Details for fixed and adapted feeding are stated as follows.

(1)

In the fixed continuous feeding strategy, glucose was fed at a rate of 2.1 g/L/h. This glucose feed rate was set based on the average glucose consumption rate in batch fermentation of S. cerevisiae in previous works.

(2)

In the adapted continuous feeding strategy, glucose was fed at a rate of 2.1 g/L/h at the beginning of the fed-batch fermentation. Glucose consumed per reactor volume (g/L) was evaluated every 12 h based on the evolved gas production (using Eq. 1). The glucose feed rate (F) was adjusted according to Eq. 2:

$$F_t=F_\cdot\frac+\Delta E)}}$$

(2)

where Ft is the adapted glucose feed rate at time t (g/L/h), F(t−1) is the glucose feed rate in the previous time interval (g/L/h), E(t−1) is the total evolved gas at the previous sampling time (mL), and \(\Delta E\)  is the increase in evolved gas over the time interval (mL/h).

When the estimated glucose consumption rate was lower than the current feed rate, the glucose feed rate was maintained to ensure that ethanol fermentation was not limited due to insufficient substrate input. Overall, for both fixed and adapted feeding strategies, glucose was continuously fed until the total glucose added reached approximately 200 g/L, this ensured the ethanol contents could reach above 11% (v/v), a common target for the fuel industry for recovery in the downstream process.

To further investigate the application of adapted feeding strategies on ethanol production, the composition of the feed medium was modified to reduce its nitrogen content. As the addition of a nitrogen source is often required for high-density cell cultures, which can represent a substantial added cost, we tested whether a medium with reduced nitrogen content could sustain the same level of growth and ethanol production. The lower nitrogen-to-carbon ratio was obtained by removing urea from the medium. The modified feed medium thus contained 500 g/L glucose and 6.7 g/L YNB with amino acids in 0.1 M sodium phosphate buffer (N/C = 0.046:10, mol/mol). This feed medium composition was tested with both the adapted and fixed continuous feeding strategies, with all other fermentation conditions kept the same.

Analyses of fermentation samples

The biomass content of samples from the bioreactors was assessed by OD600 using a UV–Vis spectrophotometer (Ultrospec 4300 Pro, Amersham Biosciences, Mississauga, ON, Canada) and converted to cell dry weight (CDW, g/L) using a calibration curve (\(CDW=0.4412\cdot OD_+0.049;\;R^2=0.967\); Supplementary Fig. S2). Fermentation samples were also centrifuged at 10,100g for 10 min (Eppendorf® Centrifuge 5418, Eppendorf AG, Hamburg, Germany), and the supernatants were analyzed by high-performance liquid chromatography (HPLC) and gas chromatography (GC) for determining glucose and ethanol concentrations, respectively (Parashar et al. 2016).

Briefly, the HPLC (Agilent 1200 series, Agilent, Santa Clara, CA, USA) was equipped with an Aminex HPX-87H column (300 \(\times\) 7.8 mm; Bio-Rad Laboratory, Hercules, CA, USA) held at 60 °C and a refractive index detector. A 5 mM sulfuric acid was used as the mobile phase with a constant flow rate of 0.5 mL/min. Glucose concentration was determined based on the peak area on the chromatogram and a calibration curve relating it to glucose standards (0.188–60.0 g/L).

The GC system (Agilent 7890A series, Agilent Technologies, Mississauga, ON, Canada) was equipped with the Agilent 7693 series autosampler (injector temperature, 170 °C; pressure, 7.5 psi; septum purge flow, 3 mL/min; spill ratio, 10 to 1), a Restek™ Stabilwax-DA column (30 m \(\times\) 0.53 mm \(\times\) 0.5 µm; Restek, Bellefonte, PA, USA), and a flame ionization detector (temperature, 190 °C; airflow, 400 mL/min; hydrogen flow, 40 mL/min; nitrogen flow, 25 mL/min). A programmed temperature sequence was applied to the column (35 °C held for 3 min, heating by 20 °C/min, and then 190 °C held for 1 min) after injection. 1% (v/v) butanol was used as the internal standard added in the sample for ethanol quantification, and helium was used as the carrier gas (constant pressure: 51.710 kPa).

Statistical analysis

All fermentation experiments were run in independent triplicates. Data were expressed as means ± standard deviation. Statistical analysis for comparing multiple groups was performed by one-way analysis of variance (ANOVA) and Tukey’s honestly significant difference test, where p < 0.05 indicated significant differences.