Site-specific modification of proteins remains an important goal to endow them with desirable properties, in basic research as well as for practical applications. To achieve these goals, genetic fusions are limited to modifications that are template-encoded or that involve the installation of enzymatically active modules of considerable size. Chemical modification of proteins, while extremely versatile, often exploits the reactivity of free thiols or primary amines to install payloads of interest, but such methods do not always afford the requisite site-specificity.

Recent work in my lab has focused on sortase-mediated modification of proteins and cells. Sortases are transacylases expressed by Gram-positive bacteria;  they covalently link proteins to the peptidoglycan cell wall. In Staphylococcus aureus >30 proteins of unrelated sequence are attached via a conserved mechanism that exploits the presence of an LPXTG motif near the C-terminus of the substrate protein. Sortase cleaves the substrate between the Thr and Gly residues, with concomitant formation of a thioacyl enzyme intermediate that is then resolved by nucleophilic attack involving the (Gly)5 sidechain of a peptidoglycan precursor. We and others have exploited the fact that this LPXTG motif is portable and can be installed on a protein of choice, to serve in a “sortagging” reaction that involves synthetic (Gly)n derivatives, which can be modified with any payload of choice. These  include not only other proteins, but especially peptides carrying chemical modifications that are not template-encoded (lipids, carbohydrates, “click” handles, fluorophores, radioisotopes). All of the harsh chemistry required to attain the desired non-template-encoded modifications is carried out off-line mostly by solid phase synthesis, whereas the sortase reaction itself is then conducted at physiological pH and temperature. This strategy provides unprecedented access to protein modifications that are genetically inaccessible, including N-N and C-C fusions of polypeptides.

For fluorescent labeling we find sortagging particularly useful in situations where  standard GFP fusions fail, such as in the creation of fluorescent type II membrane proteins  (N-terminus in, C-terminus out), or for proteins that do not tolerate the attachment of GFP as a fusion partner without loss of function, as will be illustrated for the influenza glycoproteins. We have begun the construction of large sets of single domain antibodies derived from alpacas and shrunk them to their single variable region domains (VHHs), in conjunction with sortase reactions to create new tools for the engineering of mammalian cell surfaces to achieve  modulation of the mammalian immune response, several examples of which will be discussed.