Skip to main content
Premium Trial:

Request an Annual Quote

Stanford Team Develops Enzymes for Improved In Vivo Proximity Labeling


NEW YORK (GenomeWeb) – A team led by researchers at Stanford University has developed two enzymes that could help improve proximity labeling of proteins in living organisms.

Described in a study published last month in Nature Biotechnology, the enzymes, called TurboID and miniTurbo, allow for rapid proximity labeling without the use of toxic reagents like hydrogen peroxide. Those reagents have limited the use of some proximity labeling approaches in living samples, said Tess Branon, first author on the paper and a graduate student in the lab of Stanford researcher Alice Ting, the study's senior author and an expert in the development of proximity labeling approaches.

Branon said Ting's lab has patented the enzymes through Stanford and has licensed them to three biotech companies that are using them for internal research. They are also in talks with a company about commercializing the enzymes as a kit that could be sold for research use.

Proximity labeling typically uses a target protein to tag other nearby proteins with a molecule, often biotin, that allows them to be extracted from a cell and analyzed. The method is increasingly popular with researchers who want to identify the protein components of specific subcellular compartments or identify interaction partners of proteins that are difficult to isolate via conventional immunoprecipitation mass spec experiments.

One of the most commonly used proximity labeling methods is APEX labeling, which was developed in Ting's lab and uses engineered ascorbic acid peroxidase (APEX) tags that are genetically inserted into proteins of interest. Upon stimulation with hydrogen peroxide, this tag releases biotin-phenoxyl radicals that tag nearby proteins in the cell, which can then be pulled out of the sample using streptavidin-based enrichment and analyzed using mass spec.

Capable of labeling proteins in under a minute and without notably disrupting the localization of the tagged proteins, APEX is highly effective for proximity labeling in cell cultures where the required hydrogen peroxide can be delivered to the cells quickly enough so measurements can be made before the cells' oxidative stress pathways are activated, Branon said. This, however, is not typically possible in living organisms, she noted.

"When you're trying to go into an actual animal, you can't really deliver hydrogen peroxide efficiently into tissues, because the time that it can take to diffuse and reach your target cells is probably going to be long enough to trigger stress pathways," she said. "And because of that kind of perturbation of the proteome, the data you get might not be physiologically relevant."

Another popular proximity labeling approach uses the enzyme BioID, a mutant form of the Escherichia coli biotin ligase that promiscuously labels proteins with biotin. As opposed to APEX, BioID requires only the addition of biotin, so it doesn't run into any cellular toxicity issues.

The BioID reaction is very slow, however, requiring around 18 to 24 hours, which makes it poorly suited for studying dynamic processes that occur on shorter timescales. Additionally, the authors noted, "the low catalytic activity makes BioID difficult or impossible to apply in some contexts, such as in worms, flies, or the endoplasmic reticulum lumen of cultured mammalian cells."

To address this limitation, Branon and her colleagues used adaptive evolution of BioID to generate a faster-acting version of the enzyme. Using error-prone PCR, the researchers generated roughly 107 mutants of BirAR118S, a form of the BioID enzyme, and displayed it on yeast fused to the Aga2p protein, adding biotin and ATP to allow for biotinylation. They then used FACS to select yeast cells showing high levels of self-biotinylation, repeating the process over 29 rounds of selection, during which they lowered the reaction time from 18 hours to 10 minutes to select for fast-acting BirA variants.

The process took several years, Branon said, noting that the researchers ran into several challenges along the way, including the fact that during the initial rounds of selection, the mutants had such low activity levels that they needed to apply an amplification process to boost the signal above the FACS assay's limit of detection, and the fact that in some cases, mutants that were effective at self-biotinylation were not necessarily good at biotinylating other proteins.

Ultimately, the effort produced two enzyme variants, TurboID and miniTurbo, a smaller version with the N-terminal domain deleted. They tested these in labeling experiments in mammalian mitochondria, nuclei, and ER-membrane-facing cytosol prepared from HEK293T cells, finding that the two enzymes labeled in 10 minutes essentially the same proteins as conventional BioID labeled in 18 hours. This, the authors wrote, indicates the enzymes have "a similar labeling radium despite much faster labeling kinetics."

The researchers also tested the two enzymes in model organisms, including Drosophila melanogaster and Caenorhabditis elegans, finding "extensive biotinylation" while detecting no signal in organisms labeled using BioID. They also found that TurboID decreased fly size and viability and caused developmental delay in worms, due, perhaps, to it using up the organisms' endogenous biotin supply. They suggested that this issue could be addressed by supplementing the animals with exogenous biotin.

Branon said APEX remains a better reagent for most cell culture work, with TurboID being a complementary technology "for in vivo applications where it's not possible to use APEX or BioID."

Another area where it could prove useful, she said, is plant research where BioID is poorly suited due to its low activity and APEX can't be used because of the presence of other peroxidases in plants.

The TurboID and miniTurbo enzymes have been used in "well over 100 labs at this point," she said. "So we're really excited to see where it goes."