Why use P. putida instead of P. aeruginosa for rhamnolipid production?

P. aeruginosa is the native high-producing host but it's an opportunistic human pathogen (BSL-2). That rules it out for most cosmetic, personal-care, and food-contact applications because of regulatory exposure. Engineered P. putida KT2440 (BSL-1, non-pathogenic) with heterologous rhlAB genes from P. aeruginosa has become the industrial standard. Titers are lower than native P. aeruginosa (30-50 g/L vs 60-110 g/L) but the regulatory path is much cleaner.

Why is foam fractionation used for rhamnolipids but not sophorolipids?

Rhamnolipids are strong surfactants that foam aggressively in stirred-tank bioreactors, and that foam is actually enriched in rhamnolipid compared to the bulk broth. Foam fractionation leverages this: a foam-overhead column collects the foam, breaks it, and concentrates the rhamnolipid 5-15x relative to broth. Sophorolipids don't foam like this because they're less surface-active in the bulk phase (they aggregate into a dense immiscible phase instead), so foam fractionation doesn't work for them.

What's the main scale-up challenge for P. putida rhamnolipid?

Oxygen transfer under heavy foaming. The organism is strictly aerobic and the rhamnolipid product aggressively foams during stirring. Conventional antifoam addition knocks down foam but also reduces kLa, so you end up oxygen-limited and titer caps out at 15-25 g/L. The mitigations are mechanical foam breakers (destroy foam without adding chemicals), inline foam fractionation (harvest the foam as product instead of suppressing it), and oxygen enrichment (compensate for reduced kLa). Getting to 30-50 g/L at 20,000 L requires at least two of these.

What feedstock options does the platform support for rhamnolipid?

P. putida can use glucose, glycerol, vegetable oil, or waste oil feedstocks. Glycerol is most common industrially because it's a cheap byproduct of biodiesel production and gives good rhamnolipid yields (0.2-0.3 g/g). Glucose gives slightly lower yield but is more consistent. Oil feedstocks (rapeseed, palm, waste cooking oil) can push yield up to 0.3-0.4 g/g because rhamnolipids have fatty-acid tails and oil is a direct precursor. The scenario builder lets you pick the feedstock and the platform adjusts the expected yield accordingly.

What COGS target is realistic for rhamnolipid at industrial scale?

At 40 g/L titer, 20,000 L, foam fractionation DSP, glycerol feedstock, and 25 batches per year: $18-32/kg COGS. High-titer strains (50-60 g/L) with waste cooking oil feedstock can reach $12-20/kg. The $6-10/kg target that some producers are pursuing requires both a titer step-change (to 80+ g/L) and fully-optimized foam fractionation DSP. Below $6/kg is very difficult for any microbial biosurfactant without a major break in the production technology.

Rhamnolipid production in P. putida: 20,000 L scale-up and COGS

Rhamnolipids are the largest microbial biosurfactant category by volume produced globally and the one with the most diverse commercial applications: industrial cleaners, enhanced oil recovery, agricultural adjuvants, personal care. Almost all new production today uses engineered Pseudomonas putida rather than the native high-producer P. aeruginosa, because the latter is an opportunistic pathogen and regulatory-unfriendly. This post walks through a 20,000 L scale-up for P. putida rhamnolipid, covers the foam fractionation DSP that defines the category, and shows where COGS lands.

Why engineered P. putida

P. aeruginosa natively produces rhamnolipids at 60-110 g/L in optimized processes. That's roughly 2x what the best engineered P. putida strains hit. On paper it's the obvious choice. In practice it's BSL-2, the product has to undergo rigorous endotoxin and pathogenicity clearance, and the regulatory burden adds both cost and years to product approval for any personal-care, food, or cosmetic application.

P. putida KT2440 is non-pathogenic (BSL-1) and has been engineered with heterologous rhlAB (or rhlABC) genes from P. aeruginosa to produce the same mono- and di-rhamnolipid products. Typical industrial titers: 30-50 g/L. Some academic groups (RWTH Aachen and others) have pushed engineered strains beyond 60 g/L. The tradeoff is titer vs regulatory cleanliness, and the market has mostly decided that clean wins for anything consumer-facing.

Lab-scale baseline

Representative 5 L fed-batch (30°C, pH 7.0, glycerol feed):

Duration: 80-120 h
Titer: 25-40 g/L at harvest
Biomass: 10-20 g/L dry weight (lower than S. bombicola; product is dominant)
Rhamnolipid yield on glycerol: 0.2-0.3 g/g
Ratio of mono- to di-rhamnolipid: varies by strain, typically 30:70 to 60:40
Foaming: heavy from 24 h onward, significant management required

The foaming is the defining operational characteristic. Any rhamnolipid fermentation at any scale has to actively manage foam or the process becomes unworkable. This constraint rules scale-up considerations more than kLa or mixing.

Scale-up physics

The classical scale-up concerns (kLa, mixing time, tip speed) apply, but for P. putida rhamnolipid they're dominated by the foam-management problem.

kLa at 20,000 L

Aerobic process, moderate biomass (under 25 g/L DCW), moderate viscosity (low because rhamnolipid doesn't thicken the broth like sophorolipid does). Expected kLa at 20,000 L: 80-150 h^-1. That's adequate for the organism's oxygen demand at typical biomass.

Foam at 20,000 L

The problem scales unfavorably. At 5 L you can suppress foam with a few mL of silicone antifoam per batch and accept the kLa hit. At 20,000 L, the same strategy requires 10-50 L of antifoam per batch, drops kLa by 30-50%, and coats the foam trap and downstream piping with silicone residue that's hard to clean.

Three industrial-scale foam management strategies:

Mechanical foam breaker. A high-speed impeller at the top of the reactor that physically destroys foam. No chemical addition. Capital cost: $80-200k for a 20,000 L reactor. Works but doesn't recover rhamnolipid in the foam.
Inline foam fractionation. Vent foam continuously to an overhead column, collapse it with mild heat or mechanical agitation, and collect the collapsed liquid as a rhamnolipid-enriched concentrate (5-15x enrichment over broth). This converts the foam problem into a DSP advantage.
Modified antifoam protocol. Use a high-performance silicone antifoam at very low addition rates (sub-mL per kg of product) with tight automation, minimizing the kLa hit. Compromise strategy.

Foam fractionation as primary DSP

For anyone going above 10,000 L, inline foam fractionation is the default industrial approach because the economics are compelling. The foam comes out of the reactor at 5-15x the rhamnolipid concentration of the broth, which means the first DSP step has effectively already done a 5-15x concentration for free.

Typical foam fractionation + polishing DSP chain:

Inline foam fractionation: 80-92% primary yield, concentrates to 100-200 g/L rhamnolipid in the collapsed foam
Centrifugation of residual broth to recover non-foamed rhamnolipid: 80-90% incremental yield
UF/DF of combined streams: 90-95% yield, concentrates to 400-600 g/L
Solvent extraction (optional, for polishing): 85-92% yield
Spray drying or crystallization to final product

Overall DSP yield for foam-fractionation-primary processes: 60-75%. Compare to 50-60% for conventional stirred-tank + solvent extraction of the whole broth. Foam fractionation also dramatically reduces solvent consumption, which is the second-biggest DSP cost driver.

COGS at 20,000 L

Representative scenarios at 20,000 L, 30 batches per year, glycerol feedstock:

Conventional (antifoam + solvent extraction):25 g/L titer, 50% DSP yield. 250 kg product per batch. Total COGS: $50-80/kg.
Foam fractionation primary: 40 g/L titer, 65% DSP yield. 520 kg product per batch. Total COGS: $18-32/kg.
High-titer strain + foam fractionation: 55 g/L titer, 70% DSP yield. 770 kg product per batch. Total COGS: $12-20/kg.
Waste oil feedstock + high titer: 50 g/L titer on waste cooking oil, 68% DSP yield. 680 kg product per batch. Total COGS: $8-14/kg.

The jump from conventional to foam fractionation cuts COGS roughly in half. The jump to a higher-titer strain adds another 30-40% reduction. Waste oil feedstock provides another 15-25% on top. Together these three moves push the category from $50+/kg (not competitive) to $8-14/kg (commercially viable for industrial cleaners, agricultural adjuvants, and some personal-care applications).

What Augur predicts for your strain

Upload 5-10 lab runs from your engineered P. putidastrain. Select P. putida as the organism; the DSP template defaults to biosurfactant with foam fractionation as the first unit operation. Override the foam fractionation parameters if you have measured enrichment factors for your strain (the platform default is 8x enrichment, which is mid-range for engineered strains).

Pilot customers working on P. putida rhamnolipid have seen titer prediction intervals of 20-35% at 20,000 L. The wider intervals compared to sophorolipid reflect the operational variability that foam management introduces. The platform surfaces oxygen-limitation warnings when the scaled-up kLa is predicted to fall below the strain's requirement and suggests foam fractionation as an alternative to antifoam-heavy strategies.

If you're developing a rhamnolipid process and want to see production-scale predictions with foam fractionation DSP modeling,request access. We can configure foam fractionation parameters for your specific strain and feedstock during pilot onboarding.