Labeling surface proteins with high specificity: Intrinsic limitations of phosphopantetheinyl transferase systems

Fluorescent labeling of specific cell-surface proteins enables a manifold of techniques to study their function in health and disease. A frequently cited family of methods employs phosphopantetheinyl transferases (PPTases) to attach probes, provided as conjugates of Coenzyme A. This method appears attractive, as only short peptide tags genetically fused to the protein of interest are needed as conjugation sites. Here, we describe observations we made when evaluating such protocols for delicate single-molecule applications where we require a particular combination of dyes, low background binding or low labeling of other proteins, and a high degree of labeling.

When we tested a PPTase-acceptor peptide couple with several experimental protocols and various CoA conjugates for labeling of a protein on the cell surface, we noticed substantial non-specific labeling. For the first time, we provide here a quantification of the non-specific fraction of the signals obtained using appropriate controls. We further present evidence that this background is due to CoA-dye conjugates entering the cell, where they may be covalently attached to endogenous proteins. However, when studying cell-surface proteins, most fluorescent readouts require that labeling is strictly limited to the protein of interest located at the cell surface. While such data have so far been missing in the literature, they suggest that for applications where labeling of unwanted molecules would affect the conclusions, researchers need to be aware of this potential non-specificity of PPTase methods when selecting a labeling strategy. We show, again by quantitative comparison, that the HaloTag is a viable alternative.