Spinal muscular atrophy (SMA) is a degenerative disorder of spinal motor neurons caused by homozygous mutations in the survival motor neuron (SMN1) gene. Because increased tissue levels of human SMN protein (hSMN) in transgenic mice reduce the motor neuron loss caused by murine SMN knockout, we engineered a recombinant SMN fusion protein to deliver exogenous hSMN to the cytosolic compartment of motor neurons. The fusion protein, SDT, is comprised of hSMN linked to the catalytic and transmembrane domains of diphtheria toxin (DTx) followed by fragment C of tetanus toxin (TTC). Following overexpression in Escherichia coli, SDT possessed a subunit molecular weight of approximately 130 kDa as revealed by both SDS-PAGE and immunoblot analyses with anti-SMN, anti-DTx, and anti-TTC antibodies. Like wild-type SMN, purified SDT showed specific binding in vitro to an RG peptide derived from Ewing's sarcoma protein. The fusion protein also bound to cultured primary neurons in amounts similar to those achieved by TTC. Unlike the case with TTC, however, immunolabeling of SDT-treated neurons with anti-TTC and anti-SMN antibodies showed staining restricted to the cell surface. Results from cytotoxicity studies in which the DTx catalytic domain of SDT was used as a reporter protein for internalization and membrane translocation activity suggest that the SMN moiety of the fusion protein is interfering with one or both of these processes. While these studies indicate that SDT may not be useful for SMA therapy, the use of the TTC:DTx fusion construct to deliver other passenger proteins to the neuronal cytosol should not be ruled out.