Bacillus subtilis is the most widely used Gram-positive bacterium in industrial fermentation, responsible for the production of proteases, amylases, lipases, and biosurfactants at scales exceeding 100,000 litres. Its GRAS (Generally Recognized As Safe) status, natural protein secretion capacity of up to 20 g/L, and well-characterized genetics make B. subtilis fermentation the backbone of the global industrial enzyme market. Yet three interlinked challenges define every B. subtilis bioprocess: sporulation that diverts cellular resources, autolysis that destroys biomass, and acetate overflow that poisons high-cell-density cultures. This guide covers how to control each of these during Bacillus subtilis fermentation, from strain selection through to 3,000 L scale-up.
Why B. subtilis Dominates Industrial Enzyme Production
Bacillus subtilis secretes proteins directly into the culture supernatant through the Sec and Tat pathways, eliminating the cell lysis and refolding steps required for E. coli inclusion-body products. This secretory capacity, combined with a doubling time of 20 to 40 minutes at 37 °C and established genetic tools (CRISPR-Cas9, integrative vectors, inducible promoters P43, Pgrac, PxylA), has made B. subtilis the preferred host for industrial enzyme manufacturing since the 1960s.
| Parameter | B. subtilis | E. coli | P. pastoris | A. niger |
|---|---|---|---|---|
| Secretion capacity | Up to 20 g/L | <1 g/L (periplasm) | 1 to 10 g/L | Up to 30 g/L |
| Doubling time | 20 to 40 min | 20 to 30 min | 90 to 120 min | 2 to 4 h |
| GRAS status | Yes | No (K-12 exempt) | Yes | Yes |
| Glycosylation | None | None | High mannose | High mannose |
| Inclusion bodies | Rare | Common | Rare | None |
| Sporulation risk | Yes | No | No | Yes (conidiation) |
| Methanol/flammable feed | No | No | Yes (AOX1 strains) | No |
| Typical products | Proteases, amylases, lipases | Insulin, cytokines | Antibody fragments | Glucoamylase, cellulase |
The global enzyme market exceeds $12 billion annually (as of 2025), with Bacillus species producing the majority of detergent proteases (subtilisin), food-grade amylases, and technical lipases. Companies including Novozymes, BASF, and DuPont rely on high-cell-density B. subtilis fermentation at 50,000 to 200,000 L scale.
Fermentation Strategy by Product Type
The correct Bacillus subtilis fermentation strategy depends on the target product: secreted enzymes, biosurfactants, or spore-based biocontrol agents each require different strain backgrounds, media, dissolved oxygen setpoints, and downstream processing.
Sporulation Control: Genetic and Chemical Approaches
Sporulation is the primary metabolic drain during Bacillus subtilis fermentation for enzyme production because the sporulation programme diverts 20 to 30% of cellular resources (amino acids, energy, regulatory proteins) toward building a dormant endospore instead of secreting the target enzyme. Controlling sporulation is therefore essential for maximizing enzyme yields in any B. subtilis bioprocess.
The sporulation cascade in B. subtilis is triggered by nutrient depletion and controlled by the master regulator Spo0A. When phosphorylated Spo0A~P accumulates above a threshold, it activates hundreds of genes for asymmetric septation, forespore engulfment, and cortex/coat synthesis. Two approaches suppress this cascade:
Genetic sporulation knockouts
- Δspo0A deletion eliminates sporulation entirely and is the industry-standard approach for secreted enzyme production. Spo0A-null strains produce 15 to 40% more enzyme because the diverted metabolic flux stays in protein synthesis and secretion pathways.
- ΔspoIIE or ΔspoIIG block sporulation at later stages, preserving some Spo0A-dependent gene regulation that can benefit protease expression in certain strain backgrounds.
- Combining sporulation and autolysis knockouts (Δspo0A ΔlytC) provides the most robust industrial host, eliminating both sporulation diversion and cell lysis during extended fermentations.
Chemical sporulation suppression
- Excess glucose (>5 g/L) represses sporulation through carbon catabolite repression (CcpA-mediated). However, this conflicts with the need for glucose limitation to avoid acetate overflow (see Section 5).
- Manganese chelation with EDTA (0.1 to 0.5 mM) depletes the Mn2+ required for sporulation-specific enzymes. This is effective for enzyme fermentations but must not be used when sporulation is the desired outcome (biocontrol products).
- Continuous ammonium supply suppresses sporulation through nitrogen sensing, making ammonium-limited fed-batch feeding a dual-purpose strategy that controls both growth rate and sporulation.
For media and feeding optimisation of B. subtilis enzyme fermentations, a Design of Experiments approach is the most efficient way to screen carbon source, nitrogen source, Mn2+ concentration, and pH simultaneously, especially when balancing sporulation suppression against growth rate.
Preventing Autolysis During Stationary Phase
Autolysis destroys 30 to 50% of B. subtilis biomass within hours of growth arrest, making it the single largest cause of product loss in industrial Bacillus subtilis fermentation. The major autolysins LytC (N-acetylmuramoyl-L-alanine amidase) and LytD (endo-β-N-acetylglucosaminidase) degrade the peptidoglycan cell wall, releasing intracellular proteases that can degrade the target enzyme in the supernatant.
Autolysis is not random. It is a genetically programmed response triggered through three parallel pathways:
- Nutrient depletion signalling. Carbon or nitrogen exhaustion activates sigma factors σH and σE, which upregulate autolysin expression.
- Cannibalism factors. SdpC (sporulation delay protein) and SkfA (sporulation killing factor) are toxins that lyse sibling cells to release nutrients for the surviving population. This cannibalism can account for 10 to 15% of biomass loss before sporulation even begins.
- D-alanine ester depletion. As cell-wall teichoic acids lose D-alanine esters during stationary phase, the wall becomes more accessible to autolysins, accelerating lysis.
| Strategy | Target | Biomass Retention | Product Yield Impact | Notes |
|---|---|---|---|---|
| ΔlytC deletion | Major amidase | +25 to 40% | +20 to 35% enzyme | Most effective single knockout |
| ΔlytC ΔlytD double | Both autolysins | +35 to 50% | +30 to 45% enzyme | Standard industrial strain |
| ΔsdpC ΔskfA | Cannibalism | +10 to 15% | +5 to 10% enzyme | Additive with autolysin KOs |
| Fed-batch feeding | Nutrient depletion | +15 to 25% | +10 to 20% enzyme | Avoids feast-famine trigger |
| Early harvest | Timing | N/A | Variable | Sacrifices late-phase enzyme |
The most robust industrial approach combines ΔlytC ΔlytD genetic knockouts with carbon-limited fed-batch feeding to avoid the nutrient-depletion trigger. This extends the productive culture window by 12 to 24 hours compared to wild-type batch fermentation, giving more time for enzyme secretion and accumulation in the supernatant.
Acetate Overflow and Fed-Batch Feeding Strategies
Acetate overflow in Bacillus subtilis fermentation begins when the specific glucose uptake rate exceeds the TCA cycle capacity, at approximately 1.0 g glucose per g DCW per hour. Unlike E. coli, B. subtilis lacks a native glyoxylate shunt, so it cannot reassimilate acetate through the anaplerotic pathway. Once acetate accumulates above 5 g/L, it inhibits growth and protein secretion.
The feeding strategy must balance two conflicting goals: maintain enough glucose to sustain growth and enzyme secretion, but not so much that overflow metabolism produces inhibitory acetate concentrations. Three fed-batch approaches address this:
- Exponential feeding at μ = 0.15 to 0.25 h-1 (below the critical μ that triggers overflow, typically 0.3 to 0.4 h-1 for B. subtilis). This keeps residual glucose below 0.5 g/L and acetate below 2 g/L.
- DO-stat feeding links feed rate to dissolved oxygen demand. When the culture consumes the current feed pulse, DO rises and triggers the next pulse. This is simpler to implement than exponential feeding and self-corrects for metabolic shifts.
- pH-stat feeding exploits the pH rise that occurs when glucose is exhausted and the culture begins consuming amino acids. The controller adds glucose feed when pH rises above the setpoint.
Worked Example: Exponential Feed Profile for B. subtilis Protease Fermentation
Given: 50 L working volume, initial biomass X0 = 2.0 g/L DCW, target μ = 0.20 h-1, YX/S = 0.40 g DCW / g glucose, maintenance coefficient ms = 0.02 g glucose / g DCW / h.
Feed rate equation:
F(t) = (μ/YX/S + ms) × X0 × V0 × e(μ × t) / Sf
At t = 0 h: F(0) = (0.20/0.40 + 0.02) × 2.0 × 50 / 500 = 0.104 L/h (using 500 g/L glucose stock)
At t = 10 h: F(10) = 0.104 × e(0.20 × 10) = 0.104 × 7.39 = 0.769 L/h
Expected biomass at 10 h: X(10) = 2.0 × e(0.20 × 10) = 14.8 g/L DCW
Expected acetate: <2 g/L (below inhibitory threshold)
For complex optimisation problems involving multiple feeding parameters, temperatures, and medium components in Bacillus subtilis fermentation, the DOE Generator helps screen feeding rate, μ setpoint, carbon source, and nitrogen source in a statistically efficient design.
Dissolved Oxygen Requirements by Product
Dissolved oxygen requirements differ sharply between B. subtilis enzyme and biosurfactant fermentations, making DO the single most important parameter to get right at the outset. Enzyme production requires aerobic conditions (20 to 40% air saturation) because the TCA cycle and oxidative phosphorylation provide the ATP for high-rate protein secretion. Surfactin production, by contrast, requires microaerobic conditions (1 to 5% DO) because oxygen limitation activates the srfA biosynthetic gene cluster.
| Parameter | Protease | Amylase | Surfactin | Spores |
|---|---|---|---|---|
| DO setpoint (% air sat.) | 20 to 40 | 20 to 30 | 1 to 5 | 30 to 50 |
| Aeration (VVM) | 0.5 to 1.0 | 0.5 to 0.8 | 0.1 to 0.3 | 0.5 to 1.0 |
| Agitation (rpm, bench) | 300 to 500 | 200 to 400 | 100 to 200 | 200 to 400 |
| kLa target (h-1) | 100 to 300 | 80 to 200 | 20 to 60 | 100 to 250 |
| Foaming risk | Moderate | Low | Severe | Moderate |
Foaming is the defining challenge for surfactin B. subtilis fermentation. Surfactin is a powerful lipopeptide biosurfactant that reduces surface tension to below 27 mN/m, causing uncontrolled foaming that can fill headspace, block exhaust filters, and contaminate the fermentation. The solution is to turn this problem into an advantage: integrated foam fractionation collects surfactin-enriched foam in an external column, recovering over 98% of produced surfactin with enrichment factors of 4 to 48 times. This approach achieves surfactin concentrations of up to 7.7 g/L in the foamate.
Scale-Up to 3,000 L: Strategy Comparison
Scaling Bacillus subtilis fermentation from bench to 3,000 L introduces oxygen transfer limitations, mixing heterogeneity, and foaming control challenges that do not exist at small scale. Four scale-up strategies have been tested industrially, with different trade-offs for enzyme versus biosurfactant production.
A study by Abuhena et al. (2022) demonstrated successful optimisation of B. subtilis CW-S at 3,000 L for agricultural biocontrol, reporting biomass densities of 109 CFU/mL with optimised media composition. At this scale, the choice of scale-up criterion determines both enzyme activity and operational challenges.
Constant kLa with a mechanical foam breaker delivers the best results for enzyme production at 3,000 L: it preserves oxygen transfer rates matched to the bench-scale process while the foam breaker prevents headspace fouling. For surfactin production, cascade continuous fermentation with integrated foam fractionation achieves the highest volumetric productivity because the product is continuously removed in the foam, avoiding feedback inhibition.
Fed-Batch Performance: Wild-Type vs Engineered Strains
The combined effect of sporulation, autolysis, and overflow knockouts on Bacillus subtilis fermentation performance is dramatic. An engineered Δspo0A ΔlytC strain maintains biomass for 12 to 24 hours longer than a wild-type strain, producing up to 90% more total enzyme in the same vessel.
The key differences visible in the fed-batch profile are:
- Autolysis onset at 32 h (wild-type) causes a rapid OD600 drop from 25 to 12, releasing intracellular proteases that degrade the secreted enzyme. The ΔlytC strain maintains OD600 above 22 through 48 h.
- Protease activity peaks at 1,200 U/mL (wild-type) versus 2,100 U/mL (Δspo0A ΔlytC) because the engineered strain has more time for secretion and no sporulation diversion.
- Acetate stays below 2 g/L in both strains when glucose-limited exponential feeding at μ = 0.20 h-1 is used, confirming that the feed strategy is effective regardless of genetic background.
Fed-Batch Calculator
Model exponential and constant-rate feeding profiles for B. subtilis enzyme fermentations. Predicts biomass, substrate, and product trajectories.
OTR/kLa Estimator
Calculate kLa and oxygen transfer rates for B. subtilis fermentations. Ensures adequate DO supply at bench and production scale.
Frequently Asked Questions
What is the optimal temperature and pH for Bacillus subtilis fermentation?
The optimal temperature for most B. subtilis enzyme fermentations is 30 to 37 °C, with 37 °C preferred for protease production and 30 °C for thermosensitive recombinant proteins. The optimal pH range is 6.5 to 7.5 for growth, though alkaline protease production benefits from pH 7.0 to 8.0 during the stationary phase. Surfactin production typically uses 30 °C and pH 7.0.
How do you prevent sporulation during Bacillus subtilis fermentation?
Sporulation can be prevented through genetic deletion of spo0A (the master sporulation regulator) or chemical suppression using EDTA at 0.1 to 0.5 mM to chelate Mn2+. The genetic approach is preferred industrially because it is permanent, does not require medium additives, and eliminates sporulation entirely rather than delaying it. Maintaining exponential growth through continuous feeding also delays sporulation onset.
Why does Bacillus subtilis autolyse during fermentation and how can it be prevented?
Autolysis is triggered by nutrient depletion, sporulation signals, and activation of cell-wall hydrolases LytC and LytD. These enzymes degrade peptidoglycan, causing cell lysis that can reduce biomass by 30 to 50% within hours. Prevention strategies include deleting autolysin genes (ΔlytC, ΔlytD), maintaining carbon-limited fed-batch feeding, and harvesting before late stationary phase. The cannibalism factors SdpC and SkfA also contribute and can be knocked out additively.
What dissolved oxygen level is needed for B. subtilis fermentation?
For enzyme production, dissolved oxygen should be maintained at 20 to 40% of air saturation. Below 20%, specific enzyme productivity drops and overflow metabolism increases acetate. For surfactin biosurfactant production, microaerobic conditions at 1 to 5% DO are preferred because oxygen limitation activates the surfactin biosynthetic gene cluster srfA. Typical bench-scale aeration is 0.5 to 1.0 VVM at 200 to 500 rpm.
What cell density can Bacillus subtilis reach in fed-batch fermentation?
In optimised fed-batch fermentations, cell densities of 20 to 75 g/L DCW are achievable. Nattokinase-producing strains have reached 73 g/L DCW with exponential glucose feeding. For enzyme production, a practical target is 15 to 30 g/L DCW, balancing biomass with volumetric enzyme activity. Glucose-limited feeding at μ = 0.15 to 0.25 h-1 avoids acetate overflow while sustaining growth.
Related Tools
- Fed-Batch Calculator — model exponential feeding profiles and predict biomass, substrate, and product trajectories for B. subtilis
- OTR/kLa Estimator — calculate oxygen transfer rates for aerobic B. subtilis fermentation at any scale
- Scale-Up Calculator — compare constant P/V, constant tip speed, and constant kLa strategies for scaling B. subtilis to production
References
- Sun B, Zou K, Zhao Y, et al. The fermentation optimization for alkaline protease production by Bacillus subtilis BS-QR-052. Frontiers in Microbiology. 2023;14:1301065. doi:10.3389/fmicb.2023.1301065
- Liu ZY, Yu XZ. Engineering Bacillus subtilis for high-value bioproduction: recent advances and applications. Microbial Cell Factories. 2025;24(1). doi:10.1186/s12934-025-02818-6
- Abuhena M, Al-Rashid J, Azim MF, et al. Optimization of industrial (3000 L) production of Bacillus subtilis CW-S and its novel application for minituber and industrial-grade potato cultivation. Scientific Reports. 2022;12:11006. doi:10.1038/s41598-022-15366-5
- Ren K, Wang Q, Hu M, et al. Research progress on the effect of autolysis to Bacillus subtilis fermentation bioprocess. Fermentation. 2022;8(12):685. doi:10.3390/fermentation8120685
- Kabisch J, Pratzka I, Meyer H, et al. Metabolic engineering of Bacillus subtilis for growth on overflow metabolites. Microbial Cell Factories. 2013;12:72. doi:10.1186/1475-2859-12-72