Recombinant expression
The default first attempt: cheap, fast, grows to high density, and T7 systems can make a third of the cell's new protein your target. No glycosylation, a reducing cytoplasm that disfavours disulfides, and a strong tendency to put aggregation-prone eukaryotic proteins into inclusion bodies.
Our ruleSimulations prepare you for the bench but don't replace hands-on practice.
E. coli
The default first attempt: cheap, fast, grows to high density, and T7 systems can make a third of the cell's new protein your target. No glycosylation, a reducing cytoplasm that disfavours disulfides, and a strong tendency to put aggregation-prone eukaryotic proteins into inclusion bodies.
How it differs from the standard method
- T7 (pET) systems need a DE3 host carrying T7 RNA polymerase; pBAD and rhamnose systems are tunable instead of all-or-nothing.
- Rosetta and CodonPlus hosts supply rare-codon tRNAs; pLysS and pLysE hosts tighten repression of a toxic product.
- SHuffle and Origami strains have an oxidising cytoplasm for disulfide-containing proteins; periplasmic targeting with a pelB or DsbA signal is the other route.
- Autoinduction medium removes the need to watch the OD, at the cost of control over when induction happens.
1Making a protein in a host that never asked for it
Recombinant expression puts a coding sequence under a promoter the host can be told to switch on, and then asks the host's own ribosomes, chaperones and (sometimes) secretory machinery to build the protein for you. The construct is the easy part.
2Total protein is not the answer
The commonest mistake is to read a thick new band on a Coomassie gel as success. A T7 promoter induced with 1 mM IPTG at 37 °C can turn a third of the cell's new protein into your target, and still give you nothing: chains come off the ribosome faster than they can fold, partly folded intermediates find each other instead of folding, and the result is inclusion bodies.
Temperature
Dropping the induction to 18–20 °C overnight cuts the synthesis rate several-fold and gives chaperones time to keep up. Usually less total protein per hour, far more of it soluble.
Inducer
T7 saturates near 0.2–0.4 mM IPTG. Going to 0.01–0.05 mM is a deliberate under-induction: fewer transcripts, slower translation, better folding.
Fusion partners
SUMO, MBP, thioredoxin and GST hold the nascent chain soluble. You then have to cleave the partner off with a site-specific protease and purify again.
A different cell
If the protein needs disulfides, glycans, or a eukaryotic chaperone network, no amount of tuning in E. coli will fix it. Insect, mammalian or cell-free systems are the answer, not more IPTG.
3Codon usage and the tRNA pool
Translation runs at the speed of the slowest tRNA. Human genes are full of codons E. coli reads rarely (AGA and AGG for arginine, CGA, ATA, CTA, CCC, GGA) and a run of them makes ribosomes pause.
4Toxic products and tight repression
No promoter is completely off. A T7 system leaks a few percent of its induced level before you add anything, and if the product is toxic (a protease, a membrane protein, a nuclease, a kinase that phosphorylates the host) that leak selects against the cells carrying an intact plasmid all through the overnight starter.
5Reading the gel
Load four lanes from equal culture volumes: uninduced whole cells, induced whole cells, cleared lysate, washed pellet. A new band in the induced lane that is absent before induction is your protein.
Key terms
Practise these as flashcards →A dense intracellular aggregate of misfolded recombinant protein; nearly pure, usually inactive, recoverable only by solubilisation and refolding.
Product made before the inducer is added, because no promoter is fully repressed.
A codon the host decodes with a scarce tRNA, so ribosomes pause there and can drop off.
A fusion partner (SUMO, MBP, thioredoxin, GST) that keeps the nascent chain soluble and is cleaved off afterwards.
Product made per cell per unit time; distinct from titer, which also depends on how many cells there are and how long they live.
Medium formulated so that lactose takes over as the carbon source when glucose runs out, inducing the culture without you adding IPTG.
Product concentration in the harvest, usually mg/L: the number quoted for secreted expression.
When to use it
- Producing milligram to gram quantities of a protein for structural work, enzymology, screening or as an immunogen.
- Making a protein that cannot be purified from its natural source in useful amounts, or that carries engineered mutations and tags.
- Producing antibodies, antibody fragments or Fc fusions, where mammalian expression gives the right folding and glycosylation.
- Expressing a labelled protein (¹⁵N, ¹³C, selenomethionine) in defined medium for NMR or crystallography.
Compared with alternatives
The next step: a tag on the expressed protein is what lets you capture it out of the lysate.
Where the aggregate and truncated forms that expression left behind are separated out.
The soluble-versus-insoluble gel is the actual readout of an expression trial; learn to run and read it first.
If you only need RNA, or you are feeding a cell-free system, you may not need a host at all.
The delivery step behind transient mammalian expression, and the route to a stable producer line.
Scaling an expression strain from a shake flask into a controlled bioreactor, where oxygen and feeding are the limits.
Sub-techniques
Choosing one here switches the whole page, including the protocol and the lab, to it.
The default first attempt: cheap, fast, grows to high density, and T7 systems can make a third of the cell's new protein your target. No glycosylation, a reducing cytoplasm that disfavours disulfides, and a strong tendency to put aggregation-prone eukaryotic proteins into inclusion bodies.
- T7 (pET) systems need a DE3 host carrying T7 RNA polymerase; pBAD and rhamnose systems are tunable instead of all-or-nothing.
- Rosetta and CodonPlus hosts supply rare-codon tRNAs; pLysS and pLysE hosts tighten repression of a toxic product.
- SHuffle and Origami strains have an oxidising cytoplasm for disulfide-containing proteins; periplasmic targeting with a pelB or DsbA signal is the other route.
- Autoinduction medium removes the need to watch the OD, at the cost of control over when induction happens.