We developed modified RBCs to serve as carriers for systemic delivery of a wide array of payloads. These RBCs contain modified proteins on their plasma membrane, which can be labeled in a sortase-catalyzed reaction under native conditions without inflicting damage to the target membrane or cell. Sortase accommodates a wide range of natural and synthetic payloads that allow modification of RBCs with substituents that cannot be encoded genetically. 2 with a favorable surface-to-volume ratio; and (vi) the absence of a nucleus, mitochondria, and any DNA. Thus, any modification made to the DNA of RBC precursors is eliminated upon their enucleation and cannot lead to abnormal growth or tumorigenicity after their transfusion into a recipient.Engineered RBCs have been generated using encapsulation (2-4), by noncovalent attachment of foreign peptides, or through installation of proteins by fusion to a monoclonal antibody specific for an RBC surface protein (5, 6).However, modified RBCs have limitations if intended for application in vivo. Encapsulation allows the entrapment of sizable quantities of material but does so at the expense of disrupting plasma membrane integrity, with a concomitant reduction in circulatory half life of the modified RBCs. Osmosis-driven entrapment limits the chemical nature of materials that can be encapsulated successfully, the site of release is difficult to control, and encapsulated enzymes are functional only at the final destination, compromising reusability at other sites (5, 6). Targeting of cargo to RBCs by fusion to an RBC-specific antibody, (e.g., anti-glycophorin antibody), has limitations because this mode of attachment to the RBC is noncovalent and dissociates readily, thus reducing circulatory half life and mass of cargo available for delivery (5, 6). Other developments that exploit RBCs for targeted delivery include nanoparticles enveloped by an RBC-mimicking membrane and RBC-shaped polymers (1). The short in vivo survival rate of these RBC-inspired carriers (∼7 d maximum) may limit their therapeutic utility.There is a need to develop new methodology for engineering RBCs so that they can carry a wide variety of useful cargoes to specific locations in the body. We describe an approach that involves minimal modification of the RBCs, with preservation of plasma membrane integrity. The method involves sortase-mediated site-specific covalent attachment of payloads to specific RBC surface proteins.Bacterial sortases are transpeptidases capable of modifying suitably modified proteins in a covalent and site-specific manner (7,8). Sortase A from Staphylococcus aureus recognizes an LPXTG motif positioned close to the substrate's C terminus and cleaves between T and G to form a covalent acyl-enzyme intermediate. This intermediate is resolved by a nucleophilic N-terminal glycine residue on an appropriately designed probe (9) with concomitant formation of a peptide bond between substrate and probe. Conversely, a protein may be labeled at its N terminus by extending it with suitably exposed g...