Choosing an expression system

The host sets the ceiling on what recombinant protein production can deliver, so match it to the protein's size, disulfide content and glycosylation needs before anything else. Four hosts cover most therapeutic and reagent proteins.

A practical rule: no glycans and few disulfide bonds, start in E. coli; mammalian glycans or Fc-mediated functions, start in CHO; use Pichia when secretion simplifies purification and the glycan pattern does not matter; use HEK293 when speed matters more than titer. ICH Q5B and CTD section 3.2.S.2.3 require the host and expression construct to be described and justified.

Upstream basics: media, feeding, harvest timing

Titer is set by three upstream choices, the medium, the feeding strategy and the harvest point, and each also changes the impurity load that purification must remove. Recombinant protein production is therefore planned with purification in view from the first flask.

Media. Mammalian processes use chemically defined, serum-free media screened against the specific clone; microbial processes use defined or semi-defined media with carbon, nitrogen and trace metals. Screening compares growth, titer and quality attributes such as glycan profile or charge variants, not titer alone.

Feeding. Fed-batch is the default in both systems. In CHO, concentrated feeds from about day 3 extend viability and raise titer severalfold over batch culture. In E. coli, exponential or DO-stat glucose or glycerol feeding builds cell density before induction with IPTG or a temperature shift; in Pichia, a glycerol batch phase precedes methanol induction.

Harvest timing. Harvest late enough to capture titer and early enough to avoid the proteases, host cell proteins and aggregates that accumulate as viability falls. Mammalian cultures are harvested at a viability floor; microbial cultures at a fixed time after induction.

Harvest and clarification

Clarification must deliver a particle-free, low-turbidity feed to the first column, because a fouled column costs yield and resin lifetime. The route depends on where recombinant protein production has left the product.

• Secreted products (CHO, Pichia). Cells are removed by centrifugation and depth filtration, or depth filtration alone, then 0.2 µm filtration. Turbidity and filter capacity per square meter set the sizing.
• Intracellular soluble products (E. coli). Cells are concentrated by centrifugation, resuspended in lysis buffer and disrupted by high-pressure homogenization, typically two passes at 600–1,000 bar, then clarified by centrifugation and depth filtration.
• Inclusion body products (E. coli). After homogenization, the dense inclusion bodies are recovered by centrifugation and washed with detergent or dilute denaturant to strip membrane proteins before solubilization.

Nucleic acid released during microbial lysis raises viscosity; nuclease addition or dilution keeps the lysate filterable. In every route, bioburden and endotoxin are measured on the clarified feed.

Chromatography strategy: affinity, ion exchange, HIC, SEC

A purification train uses orthogonal separation principles in sequence: capture on the most selective interaction available, then polish with one or two steps that exploit a different property. Three chromatography steps are typical in recombinant protein production; four is common for difficult proteins.

Resin screening starts in 1 mL pre-packed columns, moves to laboratory columns for loading and elution optimization, then to pilot columns where bed height, residence time and linear velocity are fixed. Column diameter grows with scale; bed height and residence time stay constant.

Inclusion bodies and refolding in microbial systems

Inclusion bodies are not a failure; in E. coli recombinant protein production they are often the purification strategy, because the dense aggregates are already 60–90% target protein after washing. The cost is a refolding step that must be developed and controlled.

The sequence is solubilization, refolding, then capture. Washed inclusion bodies are dissolved in 6–8 M guanidine hydrochloride or urea with a reducing agent. Refolding lowers the denaturant by rapid dilution, stepwise dialysis or on-column exchange, at protein concentrations usually below 1 mg/mL to favor folding over aggregation. Redox pairs such as reduced and oxidized glutathione drive disulfide formation; additives such as arginine suppress aggregation.

Refolding yield is the dominant economic variable, so the parameters to screen are protein concentration, denaturant endpoint, redox ratio, pH, temperature and time. Correct folding is confirmed by RP-HPLC, SEC and a functional assay; misfolded species are removed in the chromatography steps that follow, with HIC the most discriminating.

UF/DF and formulation

Ultrafiltration/diafiltration is the step that converts a purified chromatography pool into drug substance: it concentrates the protein to target and exchanges it into the formulation buffer in one tangential-flow operation. It is the last unit operation of recombinant protein production proper.

Membrane cutoff is chosen at roughly one third to one fifth of the protein's molecular weight (10 kDa for a 30 kDa protein, 30 kDa for a 150 kDa antibody). Loading per square meter, transmembrane pressure and crossflow rate balance flux against shear. Diafiltration of 7–10 diavolumes brings buffer components to within a few percent of target, confirmed by pH and conductivity. For high-concentration products, a second concentration step follows diafiltration.

Formulation development runs in parallel: pH, buffer species, tonicity agent, surfactant and, where needed, a stabilizer such as sucrose or arginine are selected from stress studies read by SEC, DLS and DSC. Proteins unstable in solution are lyophilized, which adds cycle development but extends shelf life.

Analytics that confirm purity and identity

Recombinant protein production is finished when an orthogonal analytical panel, not a single gel, shows that the protein meets its specification. ICH Q6B groups the tests into identity, purity and impurities, potency and quantity.

• Identity: peptide mapping by LC-MS, intact mass, N-terminal sequencing, isoelectric focusing.
• Purity and product-related variants: SEC-HPLC for aggregates, CE-SDS for fragments, ion exchange or icIEF for charge variants, RP-HPLC for oxidized and misfolded forms.
• Process-related impurities: host cell protein ELISA, host cell DNA qPCR, residual affinity ligand, residual detergents or refolding additives.
• Safety: endotoxin by LAL, bioburden, and for mammalian products the viral safety package.
• Potency and quantity: binding by ELISA or SPR, cell-based activity, concentration by A280.

Each method is developed alongside the process, qualified in late development and validated before GMP release use. Programs that leave analytics until the end routinely discover a variant they cannot remove without redeveloping a chromatography step.