E. coli Strains for Protein Expression

Many challenges can arise when over-expressing a foreign protein in E. coli. We will review the potential pitfalls of recombinant protein expression and some of the most popular commercial strains designed to avoid them.

Why do I need an expression strain?

Protein expression from high-copy number plasmids and powerful promoters will greatly exceed that of any native host protein, using up valuable resources in the cell thus leading to slowed growth. Additionally, some protein products may be toxic to the host when expressed, particularly those that are insoluble, act on DNA, or are enzymatically active. For this reason, recombinant proteins are typically expressed in E. coli engineered to accomodate high protein loads using inducible promoter systems (which will be discussed later). In addition to the basic genotypes outlined below, certain specialized strains are available to confer greater transcriptional control, assist with proper protein folding, and deal with sub-optimal codon usage (Table 1)

A few mutations are common to all or most expression strains to accomodate high protein levels including: 

• ompT: Strains harboring this mutation are deficient in outer membrane protease VII, which reduces proteolysis of the expressed recombinant proteins.
• lon protease: Strains where this is completely deleted (designated lon or Δlon) similary reduce proteolysis of the expressed proteins.
• hsdS_B (r_B^- m_B^-): These strains have an inactivated native restriction/methylation system. This means the strain can neither restrict nor methylate DNA.
• dcm: Similarly, strains with this mutation are unable to methylate cytosine within a particular sequence.

Table 1: E. coli Expression Strains 

Note: All strains are derived from the E. coli B strain, except ** which are K12

Strain	Resistance	Key Features	Genotype	Use
BL21 (DE3)		Basic IPTG-inducible strain containing T7 RNAP (DE3)	F- ompT lon hsdSB(rB- mB-) gal dcm (DE3)	General protein expression
BL21 (DE3) pLysS*	Chloramphenicol (pLysS)	pLysS expresses T7 lysozyme to reduce basal expression levels; expression vector cannot have p15A origin of replication	F- ompT lon hsdSB(rB- mB-) gal dcm(DE3) pLysS (CamR)	Expression of toxic proteins
BL21 (DE3) pLysE*	Chloramphenicol (pLysE)	pLysE has higher T7 lysozyme expression than pLysS; expression vector cannot have p15A origin of replication	F- ompT lon hsdSB(rB- mB-) gal dcm(DE3) pLysE (CamR)	Expression of toxic proteins
BL21 star (DE3)		Lacks functional RNaseE which results in longer transcript half-life	F- ompT lon hsdSB(rB- mB-) gal dcm rne131 (DE3)	General expression; not recommended for toxic proteins
BL21-A1	Tetracycline	Arabinose-inducible expression of T7 RNAP; IPTG may still be required for expression	F- ompT lon hsdSB(rB- mB-) gal dcm araB::T7RNAP-tetA	General protein expression
BLR (DE3)	Tetracycline	RecA-deficient; best for plasmids with repetative sequences.	F- ompT lon hsdSB(rB- mB-) gal dcm(DE3) Δ(srl-recA)306::Tn10 (TetR)	Expression of unstable proteins
HMS174 (DE3)**	Rifampicin	RecA-deficient; allows for cloning and expression in same strain	F- recA1 hsdR(rK12- mK12+) (DE3) (RifR)	Expression of unstable proteins
Tuner (DE3)		Contains mutated lac permease whch allows for linear control of expression	F- ompT lon hsdSB(rB- mB-) gal dcm lacY1(DE3)	Expression of toxic or insoluble proteins
Origami2 (DE3)**	Streptomycin and Tetracycline	Contains highly active thioredoxin reductase and glutathione reductase to faciliate proper folding; may increase multimer formation	Δ(ara-leu)7697 ΔlacX74 ΔphoA PvuII phoR araD139 ahpC galE galK rpsL F′[lac+ lacIq pro] (DE3) gor522::Tn10 trxB (StrR, TetR)	Expression of insoluble proteins
Rosetta2 (DE3)*	Chloramphenicol (pRARE)	Good for “universal” translation; contains 7 additional tRNAs for rare codons not normally used in E. coli.Expression vector cannot have p15A origin of replication	F- ompT hsdSB(rB- mB-) gal dcm (DE3) pRARE2 (CamR)	Expression of eukaryotic proteins
Lemo21 (DE3)*	Chloramphenicol (pLemo)	Rhamnose-tunable T7 RNAP expression alleviates inclusion body formation. Expression vector cannot have p15A origin of replication	fhuA2 [lon] ompT gal (λ DE3) [dcm] ∆hsdS/ pLemo (CamR)	Expression of toxic, insoluble, or membrane proteins
T7 Express		IPTG-inducible expression of T7 RNAP from the genome; does not restrict methylated DNA	fhuA2 lacZ::T7 gene1 [lon] ompT gal sulA11 R(mcr-73::miniTn10--TetS)2 [dcm] R(zgb-210::Tn10--TetS)	General protein expression
m15 pREP4*, **	Kanamycin (pREP4)	Cis-repression of the E. coli T5 promoter (found on vectors such as pQE or similar), inducible under IPTG (lac repressor on the pREP4 plasmid). Expression vector cannot have p15A origin of replication	F-, Φ80ΔlacM15, thi, lac-, mtl-, recA+, KmR	Expression of toxic proteins

* Denotes the presence of an additional plasmid-- make sure to maintain this by growing on appropriate media. Note: Purifying your expression plasmid from these strains is not recommended as these auxillary plasmids may be isolated during the prepping process.

How does inducible expression work?

As mentioned above, many expression plasmids utilize inducible promoters, which are 'inactive' until an inducer such as IPTG is added to the growth medium. Induction timing is important, as you typically want to make sure your cells have first reached an appropriate density. Cells in the exponential growth phase are alive and healthy, which makes them ideal for protein expression. If you wait too long to induce, your culture will start collecting dead cells, and, conversely, you cannot induce too early as there are not enough cells in the culture to make protein. 

The DE3 lysogen/T7 promoter combination is the most popular induction system. The DE3 lysogen expresses T7 RNA polymerase (RNAP) from the bacterial genome under control of the lac repressor, which is inducible by the addition of IPTG. T7 RNAP is then available to transcribe the gene of interest from a T7 promoter on the plasmid. Many commercial strains carry the DE3 lysogen, as indicated by the name of the strain. Conversely, other strains such as M15(pREP4) use a lac repressor to act directly on the expression plasmid in order to repress transcription from a hybrid promoter.

Although the DE3/T7 RNAP system works well for most experiments, the lac promoter can “leak,” meaning that a low level of expression exists even without the addition of IPTG. This is mostly a problem for toxic protein products, which can prevent the culture from reaching the desired density within a reasonable time-frame. For these cases, some strains carry an additional measure of control such as the pLys plasmid, which suppresses basal T7 expression. The pLys plasmid contains a chloramphenicol resistance cassette for positive selection and a p15A origin of replication, making it incompatible with other p15A plasmids. pLys comes in two flavors—pLysS and pLysE—the difference being that the latter provides tighter control of basal expression.

What if I don't see protein overexpression?

The strains described above should generate sufficient expression levels for most purposes, but what do you do when you’ve tried a common strain and don’t get the desired level (or any) protein expression? Low expression outcomes can result from variety of sources, so fear not—there are a few simple troubleshooting measures that can help get you back on track:

• Compatibility: Double-check your plasmid backbone and expression strain to make sure they are compatible. An arabinose-inducible plasmid will not express in an IPTG induction strain for example, nor will a p15 plasmid be compatible with a pLys strain. Your strain may require additional antibiotic selection or a special growth media, or if your plasmid is low-copy, consider reducing the antibiotic concentration.
• Growth Tempurature: Analyze your expression conditions by setting up a small-scale expression experiment to test variables such as temperature, time, and media conditions. Many recombinant proteins express better at 30°C or room-temperature, which is accomplished by growing your culture to the desired density at 37°C and reducing the temperature or moving it to a bench-top shaker 10-20 minutes before adding the inducer.
• Growth Media: Changing media is tricky, because there can be a trade-off between growth rate and protein quality. For many proteins, a rich media such as TB or 2XYT is optimal because of the high cell-density they support; however, minimal media supplemented with M9 salts may be preferable if the protein product is secreted to the medium or if slow expression is required due to solubility concerns.
• Insoluble and Secreted Proteins: The most common purification protocols are designed for soluble, cystosolic protein products, but this is not always achievable. Proteins which contain hydrophobic regions or multiple disulfide bonds may aggregate and become insoluble. These insoluble globs of misfolded protein are known as inclusion bodies, and can be recovered and purified using a special protocol. Alternatively, reducing the concentration of inducer or adding anaffinity tag such as GST may help with solubility issues.