Pharmaceuticals and fermentation: from molecule to medicine: Fermentation: microbes as chemical plants

4.1 How an industrial fermentation works

Shreve and Dryden give the same list of what a fermentation needs: an organism that makes the desired product reliably, cheap raw materials, acceptable yields, a fast fermentation, and a product that is easy to recover. The work divides into stages.

The strain comes first. Specialists select and mutate organisms for the highest yield and fewest by-products. A stock culture is grown up through flasks and seed tanks to an inoculum, typically a few per cent of the main fermenter volume.

The medium, or broth, supplies carbon (sugar, molasses, starch), nitrogen (corn steep liquor, ammonium salts, urea) and minerals. It is sterilised by heat, usually steam at about 115 to 120°C, and cooled. Air is sterilised too; Dryden describes compressing it adiabatically to 200°C and filtering out dust and oil.

Most industrial fermentations are aerobic, so the fermenter is a gas-liquid contactor. Kent gives airflows of 0.1 to more than 1 volume of air per volume of broth per minute (vvm) and mixing power of under 0.5 to over 5 kW/m3. Dryden notes that foaming worsens as tanks grow, so height-to-diameter ratios end up between 2:1 and 3:1.

Heat removal is often the real limit on scale. Kent puts the metabolic heat of a vigorous aerobic fermentation above 35 kW/m3, much more than the stirring power, and hard to remove through coils at 37°C with cooling water at 18°C.

Fermenters run batch, fed-batch (nutrients added during the run) or continuous. Kent and Dryden agree that batch and fed-batch dominate for antibiotics, organic acids, amino acids and enzymes, because long runs risk contamination and takeover by faster-growing, less productive mutants.

• Shreve (1984), Ch 31 Fermentation Industries · pdf pp. 598-599
• Dryden (1997), IIID Carbohydrates and Fermentation Industries · pdf pp. 337-339
• Kent (2012), Ch 30 Industrial Biotechnology: Discovery to Delivery · pdf pp. 1139-1143
• Pollak (2011), Fine Chemicals · pdf p. 65

4.2 Penicillin: from mould to crystals

Penicillin is made by the mould Penicillium chrysogenum in deep, aerated tanks. To steer the mould towards penicillin G, the plant feeds a precursor, phenylacetic acid or a derivative, which the mould builds into the side chain of the molecule. Feeding phenoxyacetic acid instead gives penicillin V.

Kent's typical medium contains, per litre, 30 g corn steep liquor, 30 g lactose, 5 g glucose, 3 g sodium nitrate, 3 g calcium carbonate, trace salts and 0.05 g phenyl acetamide as precursor. It is sterilised, cooled to 24°C and seeded from a 24-hour seed culture. Sterile air is sparged at about 1 vvm, and production runs 60 to 200 hours. Shreve gives 100 to 150 h at 25 to 27°C and pH kept below about 6.5 to 7.

Recovery exploits the fact that penicillin is an acid. At low pH it is uncharged and moves into an organic solvent; at neutral pH it is a salt and moves back into water. The mycelium is filtered off on rotary vacuum filters, the filtrate is chilled (Dryden says to 1 to 2°C, since penicillin is unstable) and acidified to pH 2 to 2.5, and the penicillin is extracted into amyl acetate or methyl isobutyl ketone in a Podbielniak centrifugal contactor. It is then back-extracted into a buffer at pH 6.5 to 7.5, with about 90% recovery at this step. Shreve notes that purity rises from 5 to 10% in the broth to 75 to 80% after the first extraction. Further purification and sterile filtration give the final salt.

Shreve's typical 45,000 L batch yields about 300 kg of procaine penicillin G, or about 6.7 g per litre of fermenter, consistent with his broth concentration of 5 to 10 mg/mL. Kent's later figure of 20 to 30 g/L shows how much strain improvement has added since.

• Shreve (1984), Ch 40 Pharmaceutical Industry · pdf pp. 836-840
• Kent (2012), Ch 30 Industrial Biotechnology: Discovery to Delivery · pdf pp. 1178-1179
• Dryden (1997), IVH Pharmaceutical Industry · pdf pp. 567-568, 574-577

4.3 Bulk fermentation: ethanol from molasses

Ethanol is the largest fermentation product and shows the economics of a cheap one. Yeast, Saccharomyces cerevisiae, converts sugar to ethanol and carbon dioxide without air: C₆H₁₂O₆ → 2 C₂H₅OH + 2 CO₂. By weight that is 51% ethanol on glucose; Kent notes that because some sugar feeds yeast growth, real yields are 42 to 46%.

Dryden's Indian process starts from cane molasses, the syrup left after sugar crystallisation, with 50 to 55% sugar. It is diluted to 10 to 15% sugar and brought to pH 4 to 5, which favours yeast over bacteria. Nutrients such as ammonium sulphate are added if lacking. Yeast grown on sterile mash is added, and fermentation runs 30 to 70 hours at 20 to 30°C, rising to 35°C at the end, with cooling coils removing heat.

The product, called beer, is only 8 to 10% ethanol, so most of the energy goes into distillation. A beer still takes off 50 to 60% alcohol; an aldehyde column removes light impurities; a rectifying column brings the alcohol to 95%, the azeotrope, a mixture that boils without changing composition. Anhydrous fuel ethanol needs a further step: azeotropic distillation with benzene or extractive distillation with ethylene glycol. The still bottoms, slops, become cattle feed or a high-strength effluent.

Shreve quantifies the energy. Conventional corn plants used 13 kg of steam per litre of anhydrous alcohol, newer designs claimed 2 kg. Some processes consumed up to 42 MJ per litre, against 23 MJ released by burning it. Kent adds that continuous fermentation and yeast recycle cut capital but barely move product cost, because raw materials dominate. The "Bioethanol" tile is this process at fuel scale.

• Dryden (1997), IIID Carbohydrates and Fermentation Industries · pdf pp. 341-344
• Kent (2012), Ch 30 Industrial Biotechnology: Discovery to Delivery · pdf p. 1166
• Shreve (1984), Ch 31 Fermentation Industries · pdf pp. 601-604, 608

4.4 Citric acid and lysine: an organic acid and an amino acid

Citric acid and lysine are bulk fermentation products sold on price, and they show how strain, aeration and recovery chemistry set cost.

"Citric acid" is made by the mould Aspergillus niger. Kent explains that one mole of glucose gives one mole of citric acid with no net oxygen consumed, so the theoretical yield is 112% by weight on sucrose; growth is minimal during production. In the modern submerged process the medium holds about 200 g/L sucrose with carefully balanced iron, zinc, copper, magnesium, manganese and phosphate; the medium recipe is the single most critical factor. Air is bubbled through at 0.5 to 1.5 vvm, without mechanical stirring, at 25 to 27°C for 6 to 9 days. The mould is coaxed into pellets 1 to 2 mm across, which keeps the broth thin and eases oxygen transfer and filtration.

Recovery uses lime and sulphuric acid. The mycelium is filtered off. Heating with a little lime first drops out oxalic acid, a by-product. More lime at about 95°C precipitates calcium citrate, which is filtered and treated with dilute sulphuric acid, freeing citric acid and leaving gypsum. Carbon treatment, ion exchange, vacuum evaporation and crystallisation give USP-grade monohydrate. World output is about one million tonnes a year, and Bharat Starch Industries of India is among the major producers Kent names.

Lysine, in the "Amino acids (lysine, methionine)" tile, is made by mutants of Corynebacterium glutamicum or Brevibacterium flavum, at 28 to 33°C and pH 6 to 8 with high aeration. Biotin above 30 mg/L must be supplied. After 48 to 72 hours the broth holds 100 to 140 g/L lysine, a yield of 40 to 50% on carbohydrate. Ion exchange isolates it, and it is sold mostly as 98% lysine hydrochloride for animal feed.

• Kent (2012), Ch 30 Industrial Biotechnology: Discovery to Delivery · pdf pp. 1157-1158, 1173-1174
• Shreve (1984), Ch 31 Fermentation Industries · pdf pp. 615-617

See it in the table

Check your understanding

In penicillin recovery, why is the filtered broth acidified to about pH 2 before solvent extraction?

Kent says continuous fermentation and yeast recycle in ethanol plants cut capital but barely lower product cost. Why?

What is the main reason A. niger is grown as small pellets in citric acid fermentation?