Biotinylation of proteins in vivo and in vitro using small peptide tags

The 15-residue AviTag (GLNDIFEAQKIEWHE) is biotinylated by E. coli BirA at a rate slightly better than that of the natural protein substrate. Of three BirA substrate peptide sequences examined, this sequence is biotinylated with substantially faster kinetics whether fused to the N terminus or the C terminus of maltose-binding protein (MBP).

A chemically synthesized 14-residue peptide, GLNDIFEAQKIEWH, is the minimal AviTag substrate for E. coli BirA and is biotinylated at a rate within twofold of the natural protein substrate. The 15-mer GLNDIFEAQKIEWHE is recommended over the 14-mer because it is consistently biotinylated at a slightly better rate than the natural substrate.

The AviTag works at either the N terminus or the C terminus of a target protein. An N-terminal AviTag needs an ATG initiation codon, which for E. coli expression is followed by an Ala or Ser codon to confer proteolytic stability if methionine aminopeptidase removes the N-terminal Met, while a C-terminal AviTag is generally joined to the protein through a short Gly-Gly linker, although the linker may not be necessary. An AviTag placed away from the termini gives extremely variable biotinylation, depending on whether the 15-mer folds so that BirA can recognize it.

In vitro biotinylation of an AviTag fusion by purified E. coli BirA reaches more than 95% completion in 30 min at 30 °C under these conditions: 38 µM protein substrate (stored in 10 mM Tris-HCl, pH 7.5), 50 mM Bicine buffer (pH 8.3), 10 mM magnesium acetate, 10 mM ATP, 40 µM biotin and 25 µg/ml BirA. These conditions were optimal in reaction volumes from 20 µl to 10 ml for the peptide LCDIFESQKIEWHSA and for the tested AviTag fusion proteins.

In vitro biotinylation rate scales with substrate concentration and BirA amount: 2.5 µg of E. coli BirA biotinylates more than 95% of 10 nmol of substrate at 40 µM in 30-40 min at 30 °C, the same amount of BirA needs about 5 hr at 4 µM substrate, and adding 10 times more BirA completes the reaction in 30 min instead. A suboptimal peptide sequence requires more BirA.

NaCl, glycerol and ammonium sulfate inhibit E. coli BirA activity in in vitro biotinylation reactions, whereas 0.1% Tween 20, 5 mM mercaptoethanol and 100 µM PMSF do not affect the enzyme. The AviTag substrate protein is best biotinylated in a low-salt, mildly buffered solution such as 10 mM Tris pH 7.5.

Proteases that degrade the AviTag during purification are the most common problem in in vitro biotinylation of purified AviTag fusions, and 0.1 mM PMSF with 1 µg/ml leupeptin and 0.7 µg/ml pepstatin can be effective against them. Biotinylating after purification suits AviTag fusions that form E. coli inclusion bodies, such as AviTag-MHC molecules, and expression systems not developed for efficient in vivo biotinylation, such as yeast or mammalian cells.

In E. coli with endogenous BirA, AviTag proteins expressed at low or moderate levels are typically biotinylated at about 10-30% efficiency, and highly overexpressed AviTag proteins are biotinylated at only 2-6% in strain AR1814, an E. coli B-derived strain with endogenous BirA. Coexpressing BirA, either from a dicistronic message, from its own promoter on the same plasmid as the AviTag fusion, or from a ColE1-compatible low-copy plasmid (pBirAcm, also called pJS169) that expresses BirA from a Tac promoter, raises the biotinylation efficiency.

A C-terminal MBP-AviTag fusion expressed from the pMAL-c vector in E. coli strain AVB101, which carries the BirA-overexpressing plasmid pBirAcm, is 70-80% biotinylated when induced with 1.5 mM IPTG for 3 hr in biotin-supplemented TYH medium. Amylose-column purification from a 1 liter culture yields more than 50 mg of MBP, and MBP is more than 60% of total cellular protein.

Two alternatives to E. coli inclusion-body expression of AviTag fusions are increasing solubility in E. coli by fusing the AviTag and the protein to thioredoxin, and using baculovirus vectors that coexpress the fusion protein and BirA in Spodoptera frugiperda insect cells, which avoids E. coli expression altogether.

The chromogenic biotin reaction timed assay (BRTA) measures biotinylation of an AviTag protein with an error of less than 20% using only picomole amounts of protein and tolerates free biotin in the sample. The HABA assay is linear from 0 to 7 nmol of biotin but needs up to 300 µg of MBP-AviTag protein for a reasonable signal and requires removal of all free biotin, and the nondenaturing PAGE gel-shift with avidin or streptavidin is only semiquantitative.