There is an amazing array of products available for protein expression and purification. Many of the larger companies produce complete systems for purification consisting of vectors for the addition of specific tags to the termini of cloned proteins (six-his, GST, MBP, CBP etc), matrices that specifically interact with tags, antibodies to recognize tags and so on. Despite the increasing number of systems available to purify a tagged protein of interest, bacteria still remain the predominant source of expressed protein. Stratagene, known for high efficiency competent cells, market perhaps the most comprehensive collection of E. coli for protein expression.
Despite marketing so many strains, I tend to use Stratagene's BL21(DE3) for my expression work. Perhaps Stratagene have a strain that would produce more of my proteins, but if it works, why change? BL21(DE3) and Stratagene's other competent cells for protein expression are intended for use with T7 RNA polymerase-based expression systems. The gene of interest is cloned into an expression cassette downstream of a T7 promoter, synthesis of T7 polymerase (not usually expressed in bacteria) leads to transcription of the RNA of interest and hopefully protein synthesis. This system is extremely popular for protein expression and is compatible with several commercially available vectors including pET (no tag) and pCAL (CBP tag) from Stratagene and pRSET (6-his tag) from Invitrogen. The DE3-derivatives of BL21 contains a l prophage that encodes the T7 RNA polymerase gene under the control of the lacUV5 promoter. As such, addition of IPTG (or lactose) to medium induces expression of T7 polymerase with consequent expression of the gene of interest.
High levels of protein synthesis are achieved by the efficient T7 polymerase system. Once the protein is generated, however, it must not be degraded. All BL21 cells are derived from the E. coli B strain and so naturally lack the Lon protease, a protease that degrades abnormal proteins. BL21 strains also lack the OmpT protease, an outer membrane protease that can digest proteins are cell lysis. I have never experienced proteolysis issues and I have never found it necessary to introduce protease inhibitor cocktails to my purification buffers.
Typically, I transform 50 ul of cells with nanogram quantities of plasmid, this gives me enough cells on a 10 cm plate for 2-3 full expressions. Cells come chemically competent and so electroporation is unnecessary. I generate a fresh plate by transformation every 7 to 10 days and never streak cells to generate fresh plates. For a protein expression, I start precultures from multiple colonies on the day of the induction, as opposed to the night before, and always include fairly high concentrations of antibiotic in the medium. Under these conditions, if I start a preculture early in the morning, then I can induce the main culture during log phase growth (OD600 ~ 0.6) on the same day. Using these guidelines and varying growth temperature (25-37C) and IPTG concentration (0.1-1 mM) have meant that I have always successfully expressed large amounts of my proteins. Further, once a method for induction is established with small scale experiments I have found that protocol scale-up is reproducible. I always freeze cells after harvesting and prior to cell lysis, this step may facilitate lysis, which I achieve with lysozyme and sonication without too much effort.
As stated, Stratagene offer many strains for expression work and it is worth considering the alternatives before starting an expression project. The proteins that I express are non-toxic and so I don't have to worry about leaky expression - expression without induction. If your proteins are toxic, as can be the case for membrane proteins, then strains with tight control of expression compared to BL21(DE3) exist. The tightest expression of proteins from T7 promoters is achieved in BL21 that are infected with lambda CE6 bacteriophage. Alternatively, BL21(DE3)pLysS contains a plasmid (pLysS with chloramphenicol resistance) that encodes T7 lysozyme to reduce background levels of T7 polymerase prior to induction. Other Stratagene strains are directed at alleviating problems associated with codon bias, their CodonPlus cells contain tRNA genes for codons that are frequent in higher organisms but less frequent in E. coli (specifically arg, ile, leu and pro). Interestingly, all my genes have eukaryotic codon bias and they reproducibly expressed at high yields in BL21(DE3). Stratagene also offer methionine auxotrophs to facilitate 35S labeling of proteins. Finally, although biased towards T7 polymerase systems BL21(DE3) has favorable characteristics as a general expression strains and I have produced proteins directly from Plac in BL21(DE3) by adding IPTG to medium.
Peter Haggie, Ph.D.
Post-Doctoral Fellow
University of California, San Francisco