Increase CO₂ recycling of Escherichia coli containing CBB genes by enhancing solubility of multiple expressed proteins from an operon through temperature reduction

This study aims to assess the influence of culture temperature on IB formation, providing insights into enhancing CO₂ recycling in biological carbon fixation strategies. As a control group (MOCK), we used E. coli BL21(DE3) containing empty vectors pET-21a and pET-28b. Additionally, the CBB strain refers to E. coli BL21(DE3) transformed with pET-21a and pET-28b vectors containing the cbb _I operon and cbb _II operon, respectively (11).

Figure. 1A presents the cell growth and measured CO₂ release from the MOCK and CBB strains at 37°C and 30°C. As a result, at both temperatures, in contrast with the MOCK strain, the CBB strain showed greater bacterial growth, but rather reduced CO₂ release. Notably, at 30°C, the CBB strain exhibited approximately 5.76 times lower CO₂ release than at 37°C, highlighting the effectiveness of temperature control in enhancing CO₂ fixation activity. Next, we confirmed the presence of IB, characterized by dense refractile particles with smooth or irregular rough surfaces, mainly composed of expressed foreign proteins, through the transmission electron microscopy (TEM) images shown in Fig. 1B (12). Consistently, no clear IB was observed in the MOCK strain, whereas clear IB was observed in the CBB strain at 37°C due to heterologous protein expression. However, at 30°C, there was a notable reduction in IB formation in the CBB strain.

Low culture temperature for recombinant protein expression in E. coli is a well-established strategy to suppress IB formation (13). At low temperatures, protein solubility is enhanced, and protein aggregation is reduced (14). Therefore, based on these findings, it is evident that temperature control from 37°C to 30°C effectively suppresses IB formation and enhances the endogenous CO₂ recycling capacity of the CBB strain. On the other hand, at low temperatures, cellular processes slow down, reducing the rate of cell division (15). As a result, it can be confirmed that the growth of the MOCK strain was reduced by low culture temperature. However, the CBB strain showed rather increased growth. It is generally known that the production of recombinant proteins in recombinant strains inhibits cell growth (16), but we mentioned in a previous study that the biomass of the CBB strain was increased through enhanced glycolysis by CO₂ recycling (11). As such, it can be considered that the biomass of the CBB strain was further increased through increased CO₂ recycling due to low culture temperature.

Based on the growth results, we analyzed the levels of glyceraldehyde-3-phosphate (G3P) and pyruvate in both the MOCK and CBB strains under different culture temperatures, as shown in Fig. 1C. G3P serves as an intermediate in the CBB pathway and glycolysis (17), while pyruvate is produced through glycolysis from G3P (11). Therefore, when the glycolytic flux is stimulated, the synthesis of G3P and pyruvate can increase. However, lowering the temperature to 30°C did not significantly affect the levels of G3P and pyruvate in the MOCK strain, indicating that temperature reduction did not enhance glycolytic flux.

On the other hand, the CBB strain exhibited higher levels of G3P and pyruvate compared to the MOCK strain at both temperature conditions. Overall, the introduction of the CBB pathway increased the synthesis of organic acids such as pyruvate, which provides reducing power and ATP for carbon fixation (18). This indicates a beneficial cycle of CO₂ recycling and the synthesis of metabolic intermediates. Importantly, the levels of G3P and pyruvate in the CBB strain were higher at 30°C (5,700.17 ± 349.09 ppm, G3P; 30.04 ± 0.14 ppm, pyruvate) than at 37°C (1,962.20 ± 568.27 ppm, G3P; 13.52 ± 0.01 ppm, pyruvate). These changes in G3P and pyruvate levels were consistent with the CO₂ release results observed under each condition. Based on these findings, including the growth results of the CBB strain (Fig. 1A), we can predict that temperature control from 37°C to 30°C increased the metabolic rate of the CBB pathway (e.g., G3P synthesis) in the CBB strain, leading to enhanced glycolytic reactions (e.g., pyruvate synthesis).

Generally, the CBB pathway requires redox cofactors such as NAD(P)H and ATP to facilitate CO₂ fixation (19, 20). Figure 1D illustrates the changes in ATP concentration and the NAD⁺/NADH ratio in both the MOCK and CBB strains with respect to culture temperature. In comparison to the MOCK strain, the CBB strain exhibited lower levels of ATP and NADH at both temperature conditions. Notably, the reduction in ATP and NADH was more pronounced at 30°C compared to 37°C. Specifically, the ATP concentration in the CBB strain was 6.80 ± 0.14 nmol/g DCW (dry cell weight) at 30°C and 15.09 ± 0.47 nmol/g DCW at 37°C. Furthermore, the NAD⁺/NADH ratio was 16.90 ± 0.28 at 30°C and 10.92 ± 0.12 at 37°C. These changes in ATP and the NAD⁺/NADH ratio correlated with the CO₂ release results obtained under each temperature condition, indicating that temperature control from 37°C to 30°C enhanced the utilization of electrons from redox cofactors, thereby facilitating increased CO₂ recycling (11). However, it is important to note that the energy requirements for CO₂ fixation in the CBB cycle (approximately 9 ATP per carbon atom) exceed the energy provided by glycolysis (approximately 1.7 ATP per carbon atom from G3P to pyruvate). Hence, it should be considered that the CO₂ fixation in the CBB strain relies on an external energy source, such as glucose.

Next, we compared the electron consumption of the bacteria using bioelectrochemical techniques (Fig. 1E). The bioelectrochemical technique enables the biotransformation and biosynthesis of reduced products by supplying electrons through the cathode to provide reducing power to redox cofactors (21). A single-chamber reactor was used to monitor the current change as a function of voltage. Methylene blue (MB) was utilized to mediate electron transfer to the bacteria, minimizing toxicity and growth inhibition of microorganisms (22). Chronoamperometry and cyclic voltammetry were initiated after isopropyl β-D-1-thiogalactopyranoside (IPTG) induction. In M9 media supplemented with 0.5 mM MB, a peak reduction current was observed around −0.2 V. No significant redox peaks were observed in the MOCK and CBB strains without MB (data not shown). The cyclic voltammetry results demonstrate increased redox currents in the CBB strain compared to the MOCK strain at each temperature. Remarkably, the largest redox current was observed in the CBB strain at 30°C. These significant redox currents indicate improved electrotrophy of the CBB strain, suggesting enhanced electrochemical interactions, such as electron transport between the bacteria and the electrode surface (23). This suggests that more electrons were required for CO₂ fixation activation and the maintenance of metabolism (24).

The CBB strain contains the cbb _I operon (cbbF1-prkA-cfxA-cbbL-cbbS) and the cbb _II operon (fbpB-prkB-tklB-gapB-cfxB-rbpL) (25). Both operons contain genes encoding the enzymes fructose-1,6-bisphosphatase (cbbF1 and fbpB) (25), PRK (prkA and prkB) (26), aldolase (cfxA and cfxB) (27), and RubisCO (cbbL, cbbS, and rbpL) (28). Additionally, the cbb _II operon contains genes that encode the enzymes transketolase (tklB) (29) and G3P dehydrogenase (gapB) (25). The solubility of proteins expressed from CBB genes was predicted using the Protein–Sol web server (http://protein-sol.manchester.ac.uk) to provide clues as to which proteins may be the major contributor to IB formation (Fig. 2). Protein–Sol is a web server that predicts protein solubility through the calculation of 35 sequence-based properties using data for protein solubility of E. coli in a cell-free expression system (30). Predicted solubility is expressed as a value from 0 to 1, and an experimental data set of the average soluble E. coli protein represented by PopAvrSol is used as a control (31). Thus, a value higher than the control value (0.45) means predicted to have a higher solubility than the average soluble E. coli protein; conversely, a lower value means predicted to be less soluble (30). As a result, out of a total of 11 proteins, two proteins were predicted to have higher solubility (0.484, cbbS; 0.509, fbpB) and seven proteins to have lower solubility (0.311, cbbF1; 0.258, prkA; 0.444, cfxA; 0.328, prkB; 0.350, gapB; 0.420, cfxB; 0.179, rbpL) than the average soluble E. coli protein. The other two proteins’ solubility (cbbL and tklB) could not be predicted because they were predicted to have a transmembrane segment (30). Comprehensively, based on protein solubility prediction results, cfxB is expected to have the most significant effect on IB formation, followed by prkA, cbbF1, cfxA, prkB, gapB, and prkA.

We conducted this study to investigate the impact of culture temperature on CO₂ recycling in the CBB strain. By reducing the culture temperature from 37°C to 30°C, we successfully suppressed IB formation, leading to a significant reduction in CO₂ release by approximately 5.76 times. Moreover, we observed a simultaneous increase in pyruvate accumulation by approximately 2.3 times. These results demonstrate the effectiveness of temperature control in improving CO₂ recycling and organic acid synthesis. Based on these findings, we believe that temperature control holds promise as an approach to enhance chemical biosynthetic productivity and biological CO₂ fixation in integrated autotrophic biorefinery strategies. However, it is important to note that maximizing the carbon fixation capacity of the CBB strain is a complex challenge. In this study, we focused on temperature control, but it is worth considering that other environmental factors, such as IPTG concentration and pH, can influence IB formation in the recombinant strain. Additionally, the interplay of multiple environmental conditions can impact the overall system. Future research will explore the optimization of various environmental parameters to maximize carbon fixation activity. In conclusion, temperature control represents a valuable avenue for improving CO₂ recycling and organic acid synthesis in the CBB strain, offering potential benefits for advancing integrated autotrophic biorefinery strategies.

The bacteria were grown in 40 mL of M9 minimal medium (3 g L^–1 KH₂PO₄, 1 g L^–1 NH₄Cl, 12.8 g L^–1 Na₂HPO₄·7H₂O, 0.1 mM CaCl₂, 2 mM MgSO₄, 0.5 g L⁻¹ NaOH, and 4 g L⁻¹ glucose) with 50 ng L^–1 ampicillin and 50 ng L^–1 kanamycin. The bacteria were cultured in a 100 mL serum bottle at 180 rpm and each temperature under aerobic conditions. After 3 h of incubation (optical density at 600 nm [OD₆₀₀], ~0.4), 1 mM IPTG was added to induce heterogeneous expression. After IPTG induction, the cultures were tightly capped with a gas-tight rubber stopper for semi-anaerobic cultures. Unless otherwise noted, both strains were incubated for 24 h. The growth of the cells was monitored by measuring the optical density at 600 nm with sufficient dilution using an ultraviolet-visible (UV-Vis) spectrophotometer (Mecasys Co., LTD., Daejeon, Korea).

The amount of CO₂ release was estimated by YL6500 GC (YOUNGIN Chromass Co., Ltd., Korea) equipped with a 60/80 carboxen-1000 column (Supelco, Bellefonte, PA, USA) and a thermal conductivity detector. The oven was set at 150°C, and the temperature of the injector and detector was 70°C and 180°C, respectively. N₂ was used as the carrier gas. The CO₂ release was calculated as the amount of CO₂ release divided by the number of the colony-forming unit.

The methods from primary fixation to ethanol dehydration were performed according to a previously published protocol (32). The dehydrated samples were transitioned twice with 100% propylene oxide for 5 min at room temperature. For infiltration, the samples were added to a resin series (resin:propylene oxide = 1:1 for 2 h; resin:propylene oxide = 3:1 overnight; 100% resin for 2 h). The samples were centrifuged at 10,000 rpm for 1 min in each step. After infiltration, the samples were transferred to a mold and polymerized at 60°C for 48 h by adding 100% resin. The samples were cut into ultra-thin sections of ~70 nm using an ultramicrotome (Leica EM UC7/FC7; Leica Microsystems, Wetzlar, Germany). Next, the samples were double-strained with 2% uranyl acetate and 2% Reynold’s lead citrate for 7 min each on a 200-mesh grid. The samples were washed with distilled water seven times for 1 min each between each staining. The samples were observed by TEM (Hitachi H-7650; Hitachi High-Technologies, Canada) under an accelerating voltage of 100 kV.

According to a previously published protocol, LC/MS/MS analysis was performed (32). Pyruvate standard was purchased from Sigma-Aldrich (St. Louis, MO, USA). G3P standard was purchased from the Cayman Chemical Company (Ann Arbor, MI, USA). Standard solutions of pyruvate and G3P were prepared using analytical-grade ethanol.

NADH and NAD⁺ were detected using a NAD/NADH-Glo Assay kit (Promega Co., Ltd, Fitchburg, WI, USA). NADH and NAD⁺ were extracted using formulated extraction solutions (0.4 N HCl, 0.2 N NaOH/20% DTAB, 0.2 N HCl/0.25 M Trizma, and 0.5 M Trizma). The luminescence of a mixture of 100 μL of the NAD/NADH detection reagents and 100 μL of the sample was detected using the GloMax Explorer system (Promega Co.). ATP was detected using a BacTiter-Glo microbial cell viability assay kit (Promega Co.). An ATP standard curve was generated from 10^–12 to 10^–9 M. The luminescence of a mixture of 100 μL of reagents and 100 μL of the sample was detected using a GloMax Explorer system.

The method of bioelectrochemical analysis was based on previously published protocols with some modifications (11). Electrochemical measurements were performed after purging with Ar for 15 min to remove oxygen from the electrolyte (M9 minimal medium).

The solubility of proteins was predicted using the Protein–Sol web server. Amino acid sequences for solubility prediction were obtained through National Center for Biotechnology Information (NCBI)’s database (RSP_1285, cbbF1; RSP_1284, prkA; RSP_1283, cfxA; RSP_1282, cbbL; RSP_1281, cbbS; RSP_3266, fbpB; RSP_3267, prkB; RSP_3268, tklB; RSP_3269, gapB; RSP_3270, cfxB; RSP_3271, rbpL).

All experiments were performed independently in triplicate for error analysis. The results are shown along with the correlation and standard deviation for the experimental conditions. The experimental data were analyzed using Sigma Plot (Systat Software Inc., San Jose, CA, USA).