Engineering ethanologenic bacterium Zymomonas mobilis for economic cellulosic bioethanol production using agro-industrial waste corncob residue

Abstract

Cellulosic ethanol is an important biofuel derived from lignocellulosic biomass, which has the potential to mitigate environmental pollution and contribute to achieving carbon neutralization for a sustainable circular economy. However, the economic production of cellulosic ethanol is hindered by lack of robust microbial cell factories capable of overcoming the challenges of efficient utilization of C5/C6 sugar mixture and inhibitors in the cellulosic hydrolysate. The ethanologenic bacterium Zymomonas mobilis with excellent industrial characteristics was engineered in this study, which was then applied for bioethanol production at different scales of 5 L, 50 L, and 30 m³ fermentors using the hydrolysate of the agro-industrial waste of corncob residues. The results demonstrated that the recombinant Z. mobilis had a consistent fermentation performance from lab-scale flasks to pilot-scale fermentors up to 30 m³ with the ethanol titer, yield, and productivity above 60 gꞏL⁻¹, 0.47 gꞏg⁻¹, and 3.0 gꞏL⁻¹ꞏh⁻¹, respectively. Techno-economic analysis (TEA) and life cycle assessment (LCA) were performed to evaluate the commercialization potential and environmental impacts of this process. The TEA indicated that the minimum selling price of ethanol with an annual capacity of 16,000 tons ranged from $0.58 kg⁻¹ to $0.79 kg⁻¹ with a 10 % internal rate of return. The LCA suggested that this process could reduce greenhouse gas emissions by 57 % compared to fossil fuels. The consistent performance of recombinant strain of Z. mobilis at different scales and the economic feasibility for cellulosic bioethanol production highlight the potential of Z. mobilis as a suitable cell factory for commercial biomanufacturing of lignocellulosic biochemicals using sustainable agro-industrial feedstocks.

Highlights

• •Z. mobilis produced cellulosic ethanol using acidic corncob residue without pH adjustment and sterilization.
• •Recombinant Z. mobilis produced cellulosic ethanol consistently in different scales.
• •Cellulosic ethanol at 30-m³ scale achieved >60 gꞏL⁻¹, ∼0.47 gꞏg⁻¹, and 3.0 gꞏL⁻¹ꞏh⁻¹.
• •Cellulosic ethanol is economic with a minimum selling price of $0.58–0.79 kg⁻¹.
• •Cellulosic ethanol from corncob residues can reduce 57 % GHG emissions.

1. Introduction

Bioethanol, an environmentally friendly renewable liquid biofuel, has high expectations as one of the most promising alternatives to fossil fuels [1,2]. Currently, food crops like corn and economic plants such as sugarcane are the major feedstocks of microbial fermentation, yielding over 90 million tons of bioethanol annually around the world. However, this practice is unsustainable due to the increasing population and consequent demand across the world [3]. Lignocellulosic biomass has the advantages of being the most abundant, low-cost, and no competing for food with people [4], which is a suitable renewable feedstock for the production of biofuels and commodity biochemicals. Therefore, cellulosic ethanol produced from non-food agricultural resources of lignocellulosic biomass is attractive to meet the needs of environmental protection, carbon neutralization, energy security, and the development of bioeconomy [5].

Lignocellulosic materials are naturally recalcitrant, which require pretreatment and subsequent enzymatic hydrolysis to release fermentable sugars for microbial fermentation [6]. The acidic pH of lignocellulosic hydrolysates and the accumulation of inhibitory biochemicals such as acetic acid, furfural, 5-hydroxymethyl-2-furaldehyde (HMF), and phenolic compounds during pretreatment and enzymatic hydrolysis as well as the ethanol product itself, can inhibit substrate utilization, cell growth, and bioethanol production [[7], [8], [9]]. Robust and efficient industrial microorganisms, capable of maintaining consistent ethanol fermentation performance under various environmental conditions and process parameters, are essential for economically viable cellulosic ethanol production [10]. This could be the major obstacle that commercial cellulosic ethanol production is still severely lacking, despite the historical efforts of industry pioneers such as Abengoa, Beta Renewables, DuPont, Granbio, and POET/DSM world widely [11,12].

A promising ethanol producing strain derived from wild-type Zymomonas mobilis, jointly developed by DuPont and National Renewable Energy Laboratory (NREL), demonstrated the ability to utilize both the C6 sugar glucose and C5 sugar xylose for cellulosic ethanol production using the corn stover [4]. This strain serves as the primary biocatalyst in the largest commercial cellulosic ethanol plant, which has an annual capacity of 30 million gallons of cellulosic ethanol [4]. Z. mobilis is a natural ethanologen with industrial advantages for cellulosic ethanol production. For instance, Z. mobilis possesses a unique anaerobic Entner-Doudoroff (ED) pathway that enables high sugar uptake, high specific ethanol productivity, and ethanol yield. It is also a Generally Recognized As Safe (GRAS) facultative anaerobic bacterium without the risk of phage contamination for continuous economic fermentation [13]. Its ethanol fermentation process does not require oxygen or complex oxygen control devices. The anaerobic fermentation can save 30–50 % cost compared to those products produced by aerobic fermentation [14]. In addition, Z. mobilis is capable of directly fermenting acidic lignocellulosic hydrolysates without pH adjustment and control, reducing contamination risk, chemical usage, and cost [9]. In anaerobic conditions, Z. mobilis converts nearly 98 % of sugars into ethanol, producing minimal by-products due to its small genome (2 Mb) and simple metabolic pathways, which reduces the cost of downstream product purification [6,15].

However, Z. mobilis is a non-model polyploid bacterium facing extreme challenges of effective and efficient genome engineering [16], despite that the polyploidy aids in genome stability under stressful conditions for consistent fermentation performance. With the accumulation of omics datasets and biological components from systems biology and synthetic biology [17,18], efficient genome-editing tools have been developed, including those based on heterologous CRISPR-Cas12a [19], endogenous Type I–F CRISPR-Cas [20] and associated repair pathways such as microhomology-mediated end joining (MMEJ) [21], as well as a GW-ICE (Genome-wide Iterative Continuous Editing) system [22]. The utilization of these tools has significantly enhanced the efficiency of polyploid genome editing for the construction of efficient cellulosic ethanol or d-lactate producer [23], thus advancing industrial implications.

In this study, recombinant strains of Z. mobilis were engineered by integrating biological parts related to ethanol tolerance, resulting in selected recombinant strains that demonstrated improved ethanol fermentation performance. Lab-scale fermentations utilizing corncob residue hydrolysate (CRH) were initially tested in shake flasks as well as 5-L and 50-L fermentors. Subsequently, a pilot-scale fermentation using the engineered ethanol-tolerant strain was carried out at a 30-m³ fermentor. The economic feasibility and environmental impacts of CRH-based ethanol production were then evaluated through techno-economic analysis (TEA) and life cycle assessment (LCA), respectively.

2. Materials and methods

2.1. Strains, plasmids, and media

E. coli DH5α was used for plasmid construction. Z. mobilis RJK01 derived from Z. mobilis wild-type strain ZM4 was used to construct the ethanol-tolerant recombinant strains. Primers, plasmids, and strains used in this study are listed in Table S1. E. coli strains were cultured in Luria-Bertani (LB) medium (10 gꞏL⁻¹ tryptone, 5 gꞏL⁻¹ yeast extract, 10 gꞏL⁻¹ NaCl) at 37 °C with shaking at 250 rpm. Z. mobilis strains were cultured in Rich medium RMG5 (50 gꞏL⁻¹ glucose, 10 gꞏL⁻¹ yeast extract, 2 gꞏL⁻¹ KH₂PO₄) at 33 °C with shaking at 100 rpm. RMG5E7.5 is RMG5 supplemented with 75 gꞏL⁻¹ ethanol. For pilot-scale fermentation, Z. mobilis strains recovered from the glycerol stocks were inoculated into 5 mL RMG8 medium (80 gꞏL⁻¹ glucose, 10 gꞏL⁻¹ yeast extract, 2 gꞏL⁻¹ KH₂PO₄), and then transferred sequentially to 5-L, 50-L, and finally 30-m³ fermentors with an 80 % working volume. The initial inoculation volume was set as 10 % v/v for the final fermentation volume.

2.2. DNA manipulation and recombinant strain construction

The ethanol-tolerant genes groELS, grpE, hfq flanking 20 bp overlap regions with the linearized vector were PCR amplified by the primer pair of groELS-F/R, grpE-F/R, and Hfq-F/R, respectively. And the linearized vector pEZ15A amplified by 15 A-FK-F/15 A-Peno-FK-R was ligated with the ethanol-tolerant gene DNA fragment using the T5 exonuclease-dependent assembly approach [24] to construct the over-expression plasmids driven by Peno. The sequences of genes and promoters can be found in the one-stop database ZymOmics [22].

Genome editing of Z. mobilis was carried out using the native Type I–F CRISPR-Cas system to replace the chromosomal genes with the ethanol-tolerant genes that were constructed into the editing plasmid pL2R [20]. For editing plasmid construction, pL2R was digested with Bsa I (NEB, WA, USA) at 37 °C for 3 h to generate a linearized vector containing 4 bp protruding sequences at both ends. The spacer DNAs were designed to bear 32 bp following a 5′CCC-3′ PAM. Double-stranded spacer DNAs were prepared by annealing two spacer oligonucleotides through heating at 95 °C for 5 min followed by cooling down gradually to room temperature. The spacer fragment carrying 4 bp protruding ends and linearized vector pL2R were ligated with T4 DNA ligase (Thermo Scientific, USA) at 22 °C for 2 h [20], and then transformed into E. coli DH5α. The donor fragments were designed to cover 15–20 bp overlap regions with adjacent DNA fragments. The upstream and downstream donor sequences were PCR amplified by primer pair “gene”-US-F/R and “gene”-DS-F/R, separately. T5 exonuclease-dependent assembly method was used to construct upstream fragment, downstream fragment, and ethanol tolerant gene fragment to be inserted into the editing plasmids [24].

All over-expression plasmids or editing plasmids were then transformed into Z. mobilis RKJ01 by electroporation as described [25]. Briefly, plasmids were transformed into the competent cells via electroporation (0.1 cm cuvette, 1.6 kV, 200 Ω, 25 μF) using a Gene Pulser® (Bio-Rad, CA, USA), the electroporated cells were immediately transferred to RMG5 for recovering, and subsequently spread on RMG5 plates containing appropriate antibiotics [26]. Candidate recombinant strains were screened by colony PCR with the primer pair located in both over-expression plasmid and either side of the target gene. The colony PCR products with correct size were further confirmed through Sanger sequencing (TsingKe Biological Technology, Wuhan, China). The recombinant strains obtained by genome editing were then cultivated in RMG5 without antibiotics at 30 °C and passaged for five generations to cure the editing plasmid.

2.3. Enzymatic hydrolysis of corncob residue

The corncob residue was purchased from a local company, which uses the corncob as raw materials for xylose production. The residue after xylose extraction was composed of cellulose (59.8–65.6 %), hemicellulose (5∼6.66 %), and lignin (21–26 %), as quantified by NREL standard methods NREL/TP-510-42618 [27]. Cellic CTec3 HS used in this study was purchased from Novozymes (Copenhagen, Denmark). The enzymatic activity was quantified as 373.16 FPU/mL (filter paper units per millilitre) using the standardized protocol outlined in QB/T 2583-2023, while its protein concentration reached 128 mg mL⁻¹ based on Bradford method. The hydrolysis reaction was carried out with a fed-batch mode to reach 22–25 % of solids loading in different scale reactors, while supplying 4 % of enzyme of dry feedstock, which means 40 mg liquid enzyme/g dry feedstock. The hydrolysis system was conducted in a water system with pH maintained at 4.8–5.2 by automated fed-batch addition of 25∼28 % (v/v) aqueous ammonia during feeding corncob residue.

The supernatant of the enzymatic hydrolysate was obtained through plate-frame pressure filtration and then directly used for fermentation without sterilization. Nitrogen source of 2 gꞏL⁻¹ urea was added during fermentation. All fermentations were carried out at least two times.

2.4. Analytical methods

The concentration of glucose, xylose, and ethanol were analyzed by high-performance liquid chromatography (HPLC) system (LC-20 AD, Shimadzu, Japan) equipped with a refractive index detector (RID-20 A) and an Aminex HPX-87H ion exclusion column (300 mm × 7.8 mm, Bio-Rad, USA). Sulfuric acid (5 mM) was used as the mobile phase with a flow rate of 0.5 mLꞏmin⁻¹. The column temperature was set to 60 °C and injection volume of sample was 20 μL. All samples were centrifuged at 4000 rpm for 5 min and then the supernatants were filtered through a 0.2 μm syringe filter.

2.5. Process design

Schematics of the ethanol production from corncob residues process are presented in Fig. 1. In this study, the ethanol production process was designed using Aspen Plus® software, which was also facilitated the calculation of mass and energy balance. This process encompasses five areas: glucose production (A100), ethanol production (A200), ethanol purification (A300), wastewater treatment (A400), and utilities (A500) (Supplementary method 1). Within TEA, the inside battery limit (ISBL) typically covers essential costs related to warehouse investments, site development, and additional piping. Specifically, in this study, the ISBL includes the costs of equipment and installation for areas A100 to A300.

Fig. 1.

Flow diagram of the simplified process of ethanol production using the agro-industrial waste of corncob residues. This process encompasses five areas: glucose production (A100), ethanol production (A200), ethanol purification (A300), wastewater treatment (A400), and utilities (A500).

The major assumptions and details of the process designed in Aspen Plus are listed in Table S2. The assumptions concerning production for process design are based on our experimental methodologies and findings, including enzyme load, cellulose yield, glucose yield, ethanol yield, productivity, and titer (Table S2). These data was processed in an Excel spreadsheet using established formulas to calculate the CAPEX and OPEX of the proposed processes.

2.6. Techno-economic analysis (TEA)

TEA represents an integrated process that comprehensively evaluates the economic viability and technical feasibility of ethanol production utilizing corncob residues as a feedstock. The CAPEX, OPEX, and MSP were estimated using an in-house model developed by NREL [28]. The mass balance of the ethanol production process is presented in Fig. S1, enabling the calculation of CAPEX and OPEX. An annual ethanol production capacity of 16,000 tonnes for the biorefinery was assumed, considering factors such as equipment size, raw material expenditures, and other associated costs.

The TEA is based on an “nth-plant” model, which estimates the investment necessary for the pre-commercial process as well. The key assumptions and basic parameters for the scale-up process are shown in Table S3, which are sourced from NREL's reports and studies [[28], [29], [30]]. Equipment and raw materials costs for CAPEX and OPEX estimation were obtained from published literature and official reports. Equipment costs vary with implementation size and follow a power law based on original prices [31], while raw material costs are summarized in Table S4.

Using the discounted cash flow method, the MSP of ethanol is determined at the point where the NPV equals zero and the IRR reaches 10 %. All costs are converted to 2023 US dollars to account for inflation, according to the Plant Cost Index from Chemical Engineering Magazine, the Industrial Inorganic Chemical Index from SRI Consulting, and the US Department of Labor Bureau of Labor Statistics [30].

2.7. Life cycle assessment (LCA)

The primary objective of the LCA was to quantify and compare the environmental impacts of producing ethanol from corncob residues. The system boundary is defined as “from cradle to gate” and is illustrated in Fig. S2, encompassing raw material consumption, the conversion process, and waste disposal, as well as the associated raw material, energy inputs, products, and waste streams. The functional unit of LCA studies related to ethanol is based on calorific value (MJ). To facilitate comparison with similar research, this study selected 1 MJ of ethanol as the functional unit, providing a reference for inputs and outputs. The heating value of ethanol is around 27 MJꞏkg⁻¹ ethanol.

Raw materials and energy inputs for the ethanol production pathway in this study were derived from Aspen Plus, with the corncob residues disposal capacity set at 8000 kgꞏh⁻¹. Table S5 provides detailed material and energy consumption data for ethanol production. The life cycle inventory (LCI) process database details, including input flows, corresponding GWP values, and data sources, are summarized in Table S6. The tool for reduction and assessment of chemicals and other environmental impacts (TRACI) version 2.1, developed by the environmental protection agency (EPA), was used to provide characterization factors for quantifying the potential impacts of inputs and releases from the production processes. This study assessed the global warming potential (GWP, kg CO₂-eq) of ethanol production process. To ensure consistency with the TEA system boundary, the environmental credit of lignin as a by-product was also included in the LCA. Economic allocation was applied based on the market prices of ethanol (MSP from this study) and lignin (Table S4). This method is extensively utilized in biomanufacturing LCA and is considered more suitable for this analysis than mass or system expansion allocation approaches, which are more applicable to products with analogous physical characteristics or situations involving the displacement of particular substituted products, respectively [[32], [33], [34], [35]].

In addition, the greenhouse gas reduction was quantified by directly comparing the GWP of ethanol production with that of fossil gasoline (Table S7) [36]. The percentage reduction was calculated according to the following equation:

GHG reduction (%) = (fossil gasoline – this study) / fossil gasoline × 100%

3. Results and discussion

3.1. Construction of ethanol-tolerant strains of Z. mobilis

Increasing ethanol concentration in the fermentation broth can greatly decrease the separation energy cost by distillation [37]. However, high ethanol titer accumulated during fermentation leads to ethanol stress, which affects cell growth and ethanol production. To evaluate the effect of ethanol accumulation during fermentation on cell growth, ethanol of different concentrations (0, 50, 75, and 100 gꞏL⁻¹) was supplemented and the results demonstrated the adverse effect of ethanol supplementation on cell growth (Fig. S3). Therefore, improving the ethanol tolerance of producing strains is pivotal for economic bioethanol production with the maximum possible ethanol concentration at industrial scales.

Prior researches have demonstrated that some chaperones, including Hfq (RNA chaperone) [8], GroELS (chaperonin Cpn10) [38], GrpE [39], and ClpB (ATP-dependent chaperone) [39], can enhance ethanol tolerance. Additionally, the sigma factor RpoD (sigma factor 70) [40,41], regulatory factor IrrE [42], and the uncharacterized protein ZMO0994 [43] have been shown to boost growth rate, glucose consumption rate, or final ethanol titer in media with high ethanol concentration. Meanwhile, the knockout of ZMO0128, which encodes a TonB-dependent receptor, has been shown to enhance acetic acid tolerance and may also confer potential ethanol tolerance [44]. TonB-dependent receptor encoded by ZMO0128 mediates substrate-specific transport such as siderophores, vitamin B12, and saccharides across the outer membrane by energy utilizing [45]. The downregulation of ZMO0128 may reduce energy needs for stress responses, which was a potential genetic target for future robust strain construction.

A series of recombinant Z. mobilis strains were then constructed using strain RJK01 as the parent strain to improve ethanol production, which included the ZMO0128 knockout mutant (ZET1), the hfq overexpression mutant (ZET2), and the groELS_grpE overexpression mutant (ZET3). To test the additive effect of multiple genetic modifications on ethanol production improvement, recombinant strains of ZET4 and ZET5 were subsequently constructed by overexpressing hfq or groELS-grpE in ZET1 strain, respectively.

The results showed that in the presence of 75 gꞏL⁻¹ (9.5 % v/v) ethanol supplementation, cell growth of RJK01 was effectively improved by deleting ZMO0128 (ZET1), over-expressing either hfq (ZET2) or groELS_grpE (ZET3), with final OD_600nm achieved 2.27 ± 0.03, 2.99 ± 0.07, and 3.12 ± 0.06, respectively, compared to that of the control strain RJK01 with an OD_{600 nm} value of 2.07 ± 0.01 (Fig. 2A). Consistent with cell growth, these recombinant strains exhibited faster glucose consumption and ethanol production than those of the control strain (Fig. 2B). In addition to the overexpression of hfq, groELS-grpE, ZMO0128 knockout also showed enhanced ethanol tolerance, which was consistent with previous work [8,38,39].

Fig. 2.

Cell growth, glucose consumption, and ethanol production of the recombinant strains compared with the control strain in RMG5 with the supplementation of 75 gꞏL⁻¹ ethanol. Cell growth (A) and fermentation performance (B) of ZMO0128 deletion mutant (ZET1) and recombinant strains with hfq (ZET2) or groELS-grpE (ZET3) overexpressed on plasmid in RMG5 supplemented with 75 gꞏL⁻¹ ethanol. Cell growth (C) and fermentation performance (D) of recombinant strains with hfq (ZET4) or groELS-grpE (ZET5) overexpressed on genome of ZET1 in RMG5 supplemented with 75 gꞏL⁻¹ ethanol. Three replicates were performed for the experiment.

The effects of overexpressing either hfq or groELS-grpE in the ZET1 strain with deletion of ZMO0128 were then investigated. The results demonstrated that the control strain RJK01 consumed approximately half of the available glucose (21.67 gꞏL⁻¹) and yielded 9.43 gꞏL⁻¹ ethanol. As expected, when hfq was overexpressed and ZMO0128 was knocked out, the resultant strain ZET4 exhibited an increased glucose consumed of 9.83 gꞏL⁻¹ and an enhanced ethanol production of 5.06 gꞏL⁻¹ compared to RJK01 (Fig. 2C and D). The recombinant strain ZET5 with overexpressing GroELS_GrpE and deleting ZMO0128 achieved the highest final biomass of 2.45 OD_{600 nm} (Fig. 2C), and consumed almost all glucose with 20.09 gꞏL⁻¹ ethanol produced (Fig. 2D).

In the absence of 75 gꞏL⁻¹ (9.5 % v/v) ethanol supplementation, the ZMO0128 knockout strain ZET1 had similar cell growth as its parental strain RJK01 although glucose consumption and ethanol production were slightly improved. While both cell growth and fermentation performance of the hfq over-expressing strain ZET2 improved, those of the groELS_grpE over-expressing strain ZET3 were adversely affected. In contrast, genomic integration of hfq or groELS-grpE in ZMO0128 knockout strain ZET1 had no significant effect on cell growth, but glucose utilization and ethanol production were reduced (Fig. S4).

Previous studies suggest that ZMO0128 knockout improves acetic acid tolerance [44], while groESL and grpE genes have been implicated in stress responses to various environmental inhibitors including temperature, heat, salinity, solvent, and toxic metabolites [[46], [47], [48], [49], [50]]. And hfq play important roles in resisting multiple lignocellulosic hydrolysate inhibitors [8,51]. Based on these synergistic mechanisms, we hypothesize that the multiple tolerance elements in ZET4 and ZET5 may act cooperatively to confer robustness in environments containing multiple inhibitors, such as corncob residue hydrolysate. Therefore, we evaluated the fermentation performance of these two strains first in media with ethanol supplementation and then examined them in corncob residue hydrolysate.

The assessment using CRH indicated that ZET4 and ZET5 with hfq or groELS-grpE overexpression in ZMO0128 deletion strain had higher ethanol titer of 85.70 ± 2.97 gꞏL⁻¹ and 86.51 ± 2.45 gꞏL⁻¹than that of the control strain RJK01 at 76.16 ± 5.62 gꞏL⁻¹ (Table 1). As a result, ZET5 with the highest ethanol titer and productivity was selected as the production strain for further pilot-scale fermentation.

Table 1.

Glucose consumption as well as ethanol titer, yield, and productivity of different strains using corncob residue hydrolysate.

Strain	Glucose consumption (gꞏL−1)	Ethanol Titer (gꞏL−1)	Yield (gꞏg−1)	Productivity (gꞏL−1ꞏh−1)
RJK01	149.42 ± 8.51	76.16 ± 5.62	0.51 ± 0.01	0.53 ± 0.04
ZET4	170.15 ± 9.80	85.70 ± 2.97	0.50 ± 0.01	0.60 ± 0.02
ZET5	177.71 ± 0.98	86.51 ± 2.45	0.49 ± 0.01	0.60 ± 0.02

The enhanced ethanol tolerance and fermentation performance in ZET5 may stem from a synergistic “resource re-allocation and protection” strategy. The knockout of ZMO0128, encoding a TonB-dependent receptor, likely reduces the energy expenditure (ATP) dedicated to the active transport of specific substrates across the outer membrane under stress conditions. This “energy-saving” effect could allow the cell to reallocate limited resources towards essential stress defense mechanisms. Concurrently, the overexpression of the chaperone system GroELS-GrpE directly bolsters protein homeostasis by preventing aggregation and facilitating the refolding of denatured proteins, a common consequence of ethanol and inhibitor stress. The combination may thus create a virtuous cycle: the ZMO0128 knockout potentially conserves cellular energy, which can be more efficiently utilized by the enhanced chaperone capacity to maintain membrane integrity, key enzyme activity, and overall proteostasis under the dual pressures of high ethanol and hydrolysate inhibitors.

The ZET5 strain, constructed by knocking out ZMO0128 and overexpressing groESL_grpE, demonstrated a biomass increase of approximately 1.4-fold (OD600 nm) and a 122 % improvement in ethanol production compared to the control RJK01 in RMG5 medium supplemented with 75 gꞏL⁻¹ (9.5 % v/v) ethanol (Fig. 2C and D). Previous studies in Z. mobilis have reported that overexpression of groESL enhanced ethanol production by approximately 7.4 % and 11 % at 37 °C and 40 °C, respectively, during fermentation using sweet sorghum juice [38]. Under 5 % v/v ethanol stress, groESL overexpression alone increased viable cells (CFU) by 1.9-fold, while co-expression of grpE and groESL improved cell viability (CFUꞏmL⁻¹) by 2.1-fold [52]. However, there have been no reports on the effect of ZMO0128 knockout on ethanol tolerance. This study further confirms the function of these tolerance-related elements. Through integration of multiple tolerance determinants, ZET5 achieved a significant greater improvement in ethanol yield compared to previously reported values, indicating its potential for ethanol production under high sugar concentrations. The relatively lower biomass increase compared to results reported under 5 % v/v ethanol reflects the severe cellular damage caused by 9.5 % v/v ethanol, highlighting the need for further enhancement of ethanol tolerance in future work.

3.2. Enzymatic hydrolysis and batch fermentation of corncob residue hydrolysate

The corncob residue is the agro-industrial waste of xylose production industry, which uses the corncob as raw materials for xylose production. The raw corncob primarily consisted of cellulose (32–36 %), hemicellulose (35–40 %), and lignin (17–20 %) [53]. After sulfuric acid pretreatment for xylose production, the recalcitrant cell wall structure was deconstructed, altering the composition to cellulose (59.8–65.6 %), hemicellulose (5∼6.66 %), and lignin (21–26 %). Due to this structural modification, the resulting corncob residues can be directly utilized as feedstock for enzymatic hydrolysis following pH adjustment. In this study, enzymatic hydrolysis of the corncob residue was carried out in 5-L and 50-L fermentor firstly, and then scaled up to a 70-m³ fermentor.

The results suggested that the glucose concentration was gradually increased to 140–150 gꞏL⁻¹ in all reactors of different scales within 72 h with a reaction condition of 22∼25 % solid loading, 4 % liquid CTec3 cellulase product, and 50 °C (Fig. 3A). Eighteen batches of enzymatic hydrolysis were performed at different scales to evaluate the stability and consistency of enzymatic hydrolysis. The results demonstrated that the main sugar of glucose concentration can reach a consistent level around 140–150 gꞏL⁻¹ in different scale reactors of 5 L, 50 L, and 70 m³ when 22–25 % of solid loading and 4 % liquid enzyme loading were used (Fig. 3B), which was higher than 120–140 gꞏL⁻¹ reported using similar substrate [53]. Fermentation was then carried out using fermentors of different scales of 5 L, 50 L, and 30 m³ using the hydrolysate supernatant with 10 % seed culture inocula.

Fig. 3.

The reproducibility of enzymatic hydrolysis and fermentation performance at different scales. Change of glucose concentrations within enzymatic hydrolysis of corncob residue in different scale reactors (A) and the final glucose concentration of 18 batches enzymatic hydrolysis (B) using 5-L, 50-L, and 70-m³ reactors. Initial glucose concentrations of corncob residue hydrolysates (CRH) as well as the corresponding ethanol titers and yields of hydrolysate fermentation using different scale of 5-L, 50-L, and 30-m³ reactors (C). The condition of enzymatic hydrolysis was 22∼25 % solid loading, 4 % liquid CTec3 cellulase, and 50 °C for 72 h. The presented data points are derived from duplicate fermentation runs conducted under identical conditions.

Since the acidic hydrolysate with pH 4.5 contains multiple inhibitors, including lactic acid (0.11–1.88 gꞏL⁻¹), acetic acid (0.32–1.22 gꞏL⁻¹), and furfural (0∼0.22 gꞏL⁻¹), the growth of most microorganisms may be significantly inhibited under this condition. However, the recombinant Z. mobilis exhibits excellent robustness against acidic pH above pH 3.5 and inhibitory compounds. Additionally, the glucose was rapidly utilized by Z. mobilis before competing microbes could establish significant growth, and the ethanol it produces can further inhibit the growth of other microorganisms in the system that could cause contamination. Therefore, the acidic hydrolysate was used directly without sterilization in our fermentation.

The fermentation results demonstrated that ethanol titers kept stable around 60–70 gꞏL⁻¹, and ethanol yields were all around 0.47 gꞏg⁻¹ in fermentors of different scales (Fig. 3C–Table S8). The initial pH of fermentation was started with 4.5–5.0. Then, the pH was decreased to finial pH with 4.0–4.2, which remained within the pH tolerance limits of the cells (Fig. S5). Notably, ethanol yield was calculate using the stoichiometric equation (C₆H₁₂O₆ → 2C₂H₅OH + 2CO₂). To date, lignocellulosic ethanol production at pilot-scale are mostly using the recombinant S. cerevisiae strains [54] except for that of the former DuPont Facility in Nevada, Iowa, USA using the xylose-utilizing recombinant Z. mobilis [55]. Additionally, ethanol concentrations achieved in these pilot-scale fermentations were typically around 50 gꞏL⁻¹ [12]. Although the consistent performance of the recombinant strain of Z. mobilis in fermenting CRH across different scale fermentors with ethanol titers exceeding 60 gꞏL⁻¹ represents its potential for commercial-scale cellulosic ethanol production, the strain can be further engineered and evolved to achieve higher titer and productivity than those of the current recombinant strain. Furthermore, the consistent fermentation performance at different scales indicated that the results at small scale can reflect future large-scale pilot and even commercial scale fermentation results.

To evaluate the performance of the process at pilot scale, mass balance was applied based on the data of 30 m³ scale with 22 % solid loading (Fig. 4). For enzymatic hydrolysis with 22 % solid loading, 0.65 ton water and 25.18 L ammonia were supplied per dry ton corncob residue with 37 kg Ctec3 cellulase to release sugars. The conversion of cellulose to glucose in the obtained hydrolysate supernatant was 76.4 %. The result of fermentation showed that total 1.29 ton ethanol was produced from 5.4 ton dry corncob residue with 90 % sugar conversion at 30 m³ scale, which means 4.18 ton materials can produce 1 ton ethanol. The fermentation was completed within 16 h with ethanol concentration at 60.7 gꞏL⁻¹ (Tabel S7), which was faster than other previous results with similar solid loading [56,57].

Fig. 4.

Mass balance to evaluate the performance at 30-m³ scale with 22 % solid loading. The conversion rate means the ratio of the actual ethanol conversion yield to the theoretical maximum yield.

3.3. Economic feasibility and environmental impacts of ethanol production from corncob residues

The Aspen Plus simulations were conducted by using the experimental data collected from our pilot fermentation and assumptions based on other literature. According to simulation results, an ethanol production facility with an annual capacity of 16,000 tons would require a capital expenditure (CAPEX) of $14.72 million, with 94.29 % accounting for fixed capital investment (FCI), 4.76 % for working capital, and 0.95 % for land (Table S9). Key components of the total indirect cost (TIC) including office and construction fees, project contingency, field expenses, and prorated expenses, and other costs account for 35.46 % of the total. Equipment and installation expenses constitute 50.27 % of the CAPEX, which is consistent with a previous NREL design report on lignocellulosic ethanol production [58]. Significant investments in equipment and installation for glucose production (A100) and ethanol production (A200) are mainly attributed to the costs of pretreatment reactors and fermentors, directly impacting glucose extraction yield and ethanol production parameters (Fig. S6). The total Operating expenditure (OPEX) of this process amounts to $10.08 million (Table S10), with raw materials representing the largest operating expense at 89.76 %. Corncob residues and commercial CTec3 cellulase are the primary sources of raw material expenditure, totaling $10.91 million annually, which is in line with other studies on bioethanol production [58,59]. Additionally, annual revenues of $2.42 million can be generated from selling biomass residues from the glucose production area to offset production costs.

Once the CAPEX and OPEX have been determined, a discounted cash flow rate of return analysis can be used to determine the minimum selling price (MSP) of ethanol until the net present value (NPV) of the project is zero. The MSP for ethanol is $0.79 kg⁻¹ at a 10 % internal rate of return (IRR), which is lower than the current market price of $1.08 kg⁻¹ [60], while the MSP for cellulosic ethanol with an average of $0.89 kg⁻¹ [61]. To evaluate the impact of cost-driving factors on the MSP of ethanol, a single-point sensitivity analysis was conducted by considering major raw material costs, by-product credit, glucose yield, and other crucial parameters related to ethanol production as depicted in Fig. 5. The analysis revealed that the MSP fluctuates ranging from $0.58 kg⁻¹ to $0.98 kg⁻¹, with the notable sensitivity to the costs of corncob residues and CTec3 cellulase, resulting in a variation of over 20 %. To further assess the economic impact of enzyme dosage, we performed an additional sensitivity analysis (Fig. S7), which showed that reducing the cellulase dosage from 40 kg per ton corncob residue to 10 kg per ton led to a 27.85 % decrease in the MSP. These findings highlight the significant cost-reduction potential through improved enzymatic efficiency and lower cellulase consumption, underscoring the importance of continued optimization of the hydrolysis step. Moreover, by-product credits are another important cost-driving force impacting the MSP. A combustion-grade value was assigned to lignin based on its realistic end use in current biomass conversion systems, where lignin is commonly used for on-site energy generation or sold as a low-grade solid fuel [[62], [63], [64]]. This assumption reflects the impurity profile of the enzymatic hydrolysis residue and aligns with estimates reported in prior TEA studies [[58], [62],65]. This revenue reflects a conservative and technically realistic estimate within the base-case scenario, not requiring additional purification or market development. For example, the conversion of lignin biomass into fulvic acid for application as a biofertilizer could potentially increase annual revenue to $5.52 million, thereby reducing the MSP by 25.3 %. However, this valorization process requires additional capital investments, which will neutralize the MSP reduction. Subsequent analysis further confirmed the diminishing effect of productivity and titer on the MSP beyond 3 gꞏL⁻¹ꞏh⁻¹ and 50 gꞏL⁻¹, respectively (Fig. S8).

Fig. 5.

Single-point sensitivity analysis of the MSP to produce 16,000 tons/yr of ethanol.

A preliminary LCA was performed to evaluate potential environmental impacts of this production process, specifically the GWP of producing 1 MJ of ethanol (Table S7). As shown in Fig. 6A, more than 34.54 g CO₂-eq will be emitted during the production of 1 MJ of ethanol using CRH. The use of economic allocation, based on product market values, slightly influences the distribution of environmental burdens between ethanol and lignin (Table S7). An analysis of the cradle-to-gate life cycle GWP identified cellulolytic enzymes as the predominant environmental impact factor, responsible for 86.12 % of the total GWP, comparable findings have been reported in the previous studies (Table S7) [35,[66], [67], [68]]. This is mainly attributable to the high emission intensity of commercial CTec3 (5.5 kg CO₂-eq/kg enzyme) [66], which surpasses other renewable or locally sourced inputs (Table S4–S5).

Fig. 6.

Contribute analysis of the cradle-to-gate GWP to produce 1 MJ of ethanol (A), effect of CTec3 enzyme loading and enzyme recycling rate on GWP (B), and GWP comparisons of producing 1 MJ of ethanol using different strategies to that of fossil-based gasoline (C).

A detailed scenario analysis was conducted to explore the implications of varying enzyme-related parameters on GWP. The analysis was based on our quantitative model that links enzyme loading and recycling to GWP (Fig. 6B). Specifically, reducing enzyme loading by 50 % would lower the GWP by approximately 43 %, and increasing the enzyme recycling rate to 60 % would reduce the GWP to about 16.69 g CO₂-eq/MJ ethanol. On-site enzyme production could reduce the GWP by approximately 53∼56 % [35,69], bringing it down to approximately 17 g CO₂-eq/MJ ethanol. Looking forward, advances in enzyme technology, such as the development of bio-based enzymes and more efficient recycling methods, could lead to further GWP reductions. The integration of these technologies holds significant promise for enhancing the sustainability of ethanol production. While the GWP reductions associated with enzyme management are significant, practical considerations must be addressed. Scaling up these processes could present challenges related to operational costs, energy demands, and the feasibility of enzyme recycling in industrial systems. A thorough evaluation of these factors is crucial for understanding the broader implications of enzyme management on both the environment and the economy.

The ethanol production from corncob residues by recombinant Z. mobilis demonstrates significant promise for further greenhouse gas (GHG) emission reduction by reducing enzyme usage, improving sugar yield, and utilizing renewable energy. Furthermore, a comparative analysis of GHG emissions from molasses [70], spruce chips [71], sweet sorghum [66], and switchgrass [66] for ethanol production versus fossil gasoline [36] indicates that biorefineries can achieve substantial GWP reductions (Fig. 6C). The LCA results suggest that our proposed route of bioethanol production could potentially reduce GHG emissions by up to 57 % compared to fossil-based gasoline (87.4 g CO₂-eq/MJ) [36], which is attributed to the perceived neutral carbon footprint of bio-based feedstocks [72].

Reducing GHG emissions can potentially decrease the MSP and enhance economic viability of bio-based ethanol production. The European Union's Emission Trading System (EU ETS) permits participating facilities to engage in the trading of CO₂ allowances [73]. Although biorefineries are currently not included in that system, the revised EU ETS II scheduled for implementation in 2027, is expected to expand coverage to additional options. As of 2023, the market price of a European Emission Allowance stands at €100/ton CO₂ [73]. According to our LCA findings, the reduced GHG emissions associated with bio-based ethanol in comparison to fossil-based gasoline, could potentially yield $130 in revenue per ton of ethanol produced.

4. Conclusion

Our results demonstrated consistent performance with glucose concentration of ∼150 gꞏL⁻¹ and ethanol titer of >60 gꞏL⁻¹ at different scales of enzyme hydrolysis and fermentation, suggesting the stability of the entire process using engineered Z. mobilis strain. TEA and LCA findings revealed that the proposed process could be economically competitive with reduced environmental impacts, achieving a lower MSP and GWP compared to current market price and fossil-based gasoline, respectively. The utilization of corncob residues and cellulase showed a positive correlated with both the MSP and GWP, indicating that improving the utilization rate of raw materials and enzymes could enhance the economic and environmental performance of ethanol production in the future. In addition to corncob residue, other agro-industrial wastes such as corn stover, wheat straw, and bagasse could also be utilized as feedstocks for large-scale economic production of ethanol using engineered Z. mobilis strain. Therefore, ethanol production from corncob residues using the engineered Z. mobilis not only offers a feasible approach for reusing biomass on an industrial scale, but also has the potential to mitigate local GHG emissions, thereby further enhancing the economic viability of the product.

CRediT authorship contribution statement

Yongfu Yang: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation. Chenyue Zhang: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Investigation, Formal analysis. Binan Geng: Methodology, Investigation. Ying Tang: Validation, Investigation. Haixia Yi: Validation, Investigation. Jun Du: Supervision, Resources, Project administration. Qiang Fei: Writing – review & editing, Supervision, Software, Resources, Project administration, Methodology, Formal analysis. Shihui Yang: Writing – review & editing, Writing – original draft, Visualization, Supervision, Resources, Project administration, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Funding

This work was supported by the National Key Research and Development Program of China (No. 2022YFA0911800), the International Science and Technology Cooperation base of Hubei Province (No. SH2318), the Technological innovation Plan Project of Hubei Province (No. 2024BCB025), Natural Science Basic Research Program of Shaanxi (2025JC-QYXQ-004), the Technology Achievement Transformation Project of Wuhan Science and Technology Innovation Bureau (No. 2024030803010187), and the Special Zone Project of Wuhan Natural Science Foundation (No. 2024040701010048).

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Jun Du is associated with Beijing Tsingke Biotech Co., Ltd., and Shihui Yang is the founder of Wuhan ZymoBiotech Inc.

Abbreviations

CAPEX

capital expenditures

CRH

corncob residue hydrolysate

ED

Entner-Doudoroff

EPA

environmental protection agency

ETS

emission trading system

EU

European union

FCI

fixed capital investment

GHG

greenhouse gas

GRAS

generally recognized as safe

GW-ICE

genome-wide iterative continuous editing

GWP

global warming potential

HMF

5-hydroxymethyl-2-furaldehyde

HPLC

high-performance liquid chromatography

IRR

internal rate of return

ISBL:

inside battery limit

LB

Luria-Bertani

LCA

life cycle assessment

LCI

life cycle inventory

MMEJ

microhomology-mediated end joining

MSP

minimum selling price

NPV

net present value

NREL:

National Renewable Energy Laboratory

OPEX

operating expenditures

TEA

techno-economic analysis

TIC

total indirect cost

TRACI

tool for reduction and assessment of chemicals and other environmental impacts

Appendix A. Supplementary data

The following is the Supplementary data to this article: