An end-to-end pipeline for succinic acid production at an industrially relevant scale using Issatchenkia orientalis

The study addresses the challenge of producing succinic acid (SA), a top bio-based building block, via microbial conversion of renewable biomass as an alternative to fossil-based processes. Traditional bacterial SA fermentations require neutral pH maintained by bases (e.g., CaCO3 or NaOH) and post-fermentation reacidification with strong acids (e.g., H2SO4), generating gypsum and increasing downstream processing costs and environmental burdens. Producing SA at low pH can reduce operating costs and environmental impacts. Yeasts, being more acid tolerant, are promising hosts. The authors aim to engineer the acid-tolerant yeast Issatchenkia orientalis to produce SA efficiently at low pH, demonstrate scalable production using industrially relevant sugarcane juice, and evaluate end-to-end process economics and environmental impacts.

Prior work has engineered bacteria such as Escherichia coli, Corynebacterium glutamicum, and Mannheimia succiniciproducens to high SA performances, but the neutral pH requirement limits sustainability due to gypsum formation during reacidification. Yeast platforms (e.g., Yarrowia lipolytica) have achieved high titers, often using complex media and near-neutral pH or non-sugar substrates (glycerol); truly low-pH, sugar-based processes are less developed. Transport engineering (e.g., MAE transporters) and pathway modifications (oTCA versus rTCA) have been explored, with rTCA offering carbon fixation advantages over oTCA which includes decarboxylation steps. The authors build on prior demonstrations of SA production in I. orientalis with rTCA to improve low-pH performance and industrial relevance.

Engineering strategy and strain construction: The reductive TCA (rTCA) pathway previously installed in I. orientalis (strain SA) was augmented by integrating a codon-optimized dicarboxylic acid exporter SpMAE1 from Schizosaccharomyces pombe, creating SA/MAE1. To reduce byproducts and improve redox balance, alcohol and glycerol formation pathways were targeted: PDC (pyruvate decarboxylase) and GPD (glycerol-3-phosphate dehydrogenase) were deleted (strain SA/MAE1/pdcA/gpdA). Based on genome-scale model insights (ADH mitochondrial localization), cytosolic NADH limitation was identified via 13C metabolic flux analysis (MFA). To limit SA reuptake, JEN-family dicarboxylate importers (annotated as g3473 and g3068) were disrupted; g3473 deletion improved titer, while dual deletion reduced performance. External mitochondrial NADH dehydrogenase (NDE) was deleted to conserve cytosolic NADH (strain g3473Δ/ndeΔ). To enhance glycerol utilization and cytosolic NADH generation, a codon-optimized glycerol dehydrogenase from Pichia angusta (PaGDH) and endogenous dihydroxyacetone kinase (DAK) were overexpressed (strains g3473Δ/PaGDH-DAK and g3473Δ/ndeΔ/PaGDH-DAK). To alleviate glucose catabolite repression on glycerol co-consumption, hexokinase candidates (g1398, g2945, g3837) were evaluated; only g3837 deletion enabled simultaneous glucose–glycerol use in the engineered background, though at reduced consumption rates. Fermentations: Shake-flask fermentations employed SC-URA minimal medium with glucose (50–70 g/L) and optionally glycerol (20–30 g/L), under oxygen-limited or aerobic conditions. Bench-top bioreactor studies (0.35 L vessel; 0.1 L working volume) performed batch and fed-batch fermentations at pH 3 with continuous O2 and CO2 sparging; feeding strategies included glucose/glycerol or concentrated sugarcane juice. Fed-batch conditions were optimized for titer, yield, and productivity. For industrial relevance, invertase (ScSUC2) was expressed to enable sucrose utilization from sugarcane juice. Scale-up: A pilot-scale bioreactor (75 L vessel, 30 L working volume) was operated in batch mode at pH 3 using sugarcane juice medium. Scale-up maintained similar power input per volume (P/V) and Reynolds number compared to bench-scale systems; DO was controlled at 8% via airflow. Downstream processing (DSP): SA was recovered directly from low-pH fermentation broth via two-stage vacuum distillation concentration followed by direct crystallization at 0 °C with optimized seeding (1% w/v), temperature (10 °C seeding), time (4 h), and agitation. Stage-wise and overall recoveries and product purities were quantified by HPLC. Analytical and systems analyses: Extracellular metabolites were quantified by HPLC. 13C-MFA used targeted LC-MS of intracellular metabolites and amino acids with INCA to resolve fluxes and redox balance. Techno-economic analysis (TEA) and life cycle assessment (LCA) were performed in BioSTEAM at a 26,800 t/y SA capacity under laboratory batch, laboratory fed-batch, and pilot batch scenarios. Uncertainty was assessed via 2000-run Monte Carlo simulations; metrics included minimum product selling price (MPSP, 2016$), cradle-to-grave 100-year global warming potential (GWP100), and cradle-to-gate fossil energy consumption (FEC). Sensitivity analyses used Spearman’s rank correlations. Performance landscapes across yield–titer–productivity combinations were modeled for low-pH versus neutral fermentations.

- Transport and byproduct pathway engineering: Introducing SpMAE1 increased SA titer in shake flasks from 6.8 to 24.1 g/L under oxygen-limited conditions. Deleting PDC and GPD did not raise SA titer due to cytosolic NADH limitation, instead causing pyruvate accumulation (~19.8 g/L). 13C-MFA confirmed cytosolic NADH scarcity limits rTCA flux. - Co-substrate strategy and deletions: Co-fermenting glucose (50 g/L) with glycerol (20 g/L) increased SA to 38.6 g/L under aerobic conditions versus 30.5 g/L under oxygen-limited. Deleting a dicarboxylate importer (g3473) improved titer to 42.0 g/L; deleting both importers (g3473 and g3068) reduced titer to 34.5 g/L. Deleting NDE (g3473Δ/ndeΔ) further raised titer to 46.4 g/L but slowed glucose uptake. - Enhanced glycerol utilization: Overexpressing PaGDH and DAK improved substrate consumption rates and productivities (e.g., from 0.29 to 0.44 g/L/h in g3473Δ/PaGDH-DAK) without significantly increasing titer. Using 50 g/L glucose + 20 g/L glycerol achieved a higher yield (0.60 g/g glucose equivalent) than 50–70 g/L glucose alone (0.50–0.51 g/g). Optimal glycerol with 50 g/L glucose was 20 g/L. - Bioreactor performance: In bench-top reactors at pH 3, deleting hexokinase g3837 (in g3473Δ/PaGDH-DAK background) improved SA formation during glycerol utilization, likely by lowering expression of TCA cycle genes (CIT, ACO, IDH) while maintaining rTCA expression. Fed-batch in minimal medium achieved 109.5 g/L SA, 0.65 g/g glucose equivalent yield, and 0.54 g/L/h productivity. Using sugarcane juice with ScSUC2 in fed-batch reached 104.6 g/L, 0.63 g/g, and 1.25 g/L/h. - Pilot scale (30 L working volume, pH 3): Batch fermentation on sugarcane juice reached 63.1 g/L SA, 0.50 g/g, and 0.66 g/L/h. - Downstream crystallization: Direct crystallization from low-pH broth (no re-acidification) achieved an overall 64.0% SA recovery in two stages (stage 1: 31.0%; stage 2: 47.7% of the filtrate) with crystal purities of 88.9% and 86.23% for stages 1 and 2, respectively. - TEA/LCA: Estimated MPSP (2016$) was $1.70/kg (lab batch), $1.06/kg (lab fed-batch), and $1.37/kg (pilot batch). GWP100 was 1.95, 0.93, and 1.67 kg CO2-eq/kg, respectively; FEC was −3.74, −5.36, and −0.21 MJ/kg. The pilot scenario’s GHG emissions were 34–90% lower than fossil-based SA (3.27–12.1 kg CO2-eq/kg). Sensitivities: MPSP most sensitive to fermentation yield; GWP100 and FEC more sensitive to titer. Low-pH fermentation consistently outperformed neutral fermentation in MPSP, GWP100, and FEC across performance landscapes.

Engineering I. orientalis for low-pH (pH 3) SA production enabled high titers on sugar-based media, addressing downstream cost and environmental issues associated with neutral pH bacterial processes. The primary bottleneck was cytosolic NADH supply for the rTCA pathway; co-utilizing glycerol (higher reducing power) with glucose improved both yield and titer. Preventing SA reuptake (g3473 deletion) and conserving cytosolic NADH (NDE deletion) further enhanced titers, though NDE deletion reduced substrate consumption rates and productivity, likely due to impacts on respiration and ATP synthesis. Attempts to increase NADH via transhydrogenase conversion of NADPH were deemed unlikely to help given low PPP flux and NADPH demand for biosynthesis. Compared to bacteria and Yarrowia lipolytica, the engineered I. orientalis achieves competitive titers at low pH in minimal media and leverages the carbon-fixing rTCA pathway (potentially more sustainable than oTCA). Demonstrations at bench and pilot scales and direct crystallization from acidic broth highlight process practicality. TEA and LCA confirm financial viability at or below market prices and substantial environmental benefits versus fossil routes, with low-pH fermentation offering consistent advantages over neutral processes. Sensitivity analysis indicates improving yield most strongly reduces MPSP, while increasing titer most reduces GWP100 and FEC due to lower separation energy demand and enhanced cogeneration benefits.

The authors established an end-to-end low-pH SA production pipeline using engineered Issatchenkia orientalis, achieving >100 g/L titers in fed-batch on minimal and sugarcane juice media and 63.1 g/L at pilot scale, with direct crystallization enabling 64% overall recovery without re-acidification. TEA/LCA under uncertainty demonstrate financial viability (MPSP down to $1.06–1.37/kg) and markedly reduced GHG emissions compared to fossil-based SA, with low-pH processes outperforming neutral ones. Future work will target higher yields and titers by coupling the rTCA with the glyoxylate shunt (theoretical 1.71 mol/mol glucose) and integrating crude glycerol supplementation to boost reducing power. The pipeline may extend to other organic acids (e.g., muconic acid, 3-hydroxypropionic acid) at low pH.

- Redox limitation: Cytosolic NADH availability limits rTCA flux and SA yield; NDE deletion improves titer but lowers substrate consumption and productivity. - Glycerol phase performance: In bioreactors with higher aeration, more flux to the TCA cycle reduced SA formation during glycerol utilization; improvements required careful regulation (e.g., g3837 deletion) at the cost of slower sugar uptake. - Downstream purity: Direct crystallization from fermentation broth achieved ~86–89% purity; coloring impurities remain and pre-polishing steps may be needed to reach commercial-grade purity and increase recovery yield beyond 64%. - Scale-up scope: Pilot trials were batch-only due to volume constraints; fed-batch at pilot scale remains to be demonstrated. - Media and substrate co-utilization: Hexokinase deletion enabling co-consumption reduced uptake rates, not improving productivity; broader strategies for robust co-utilization without trade-offs may be needed.