Several protein toxins, including the A chain of ricin (RTA), enter mammalian cells by endocytosis and subsequently reach their cytosolic substrates by translocation across the endoplasmic reticulum (ER) membrane. To achieve this export, such toxins exploit the ER-associated protein degradation (ERAD) pathway but must escape, at least in part, the normal degradative fate of ERAD substrates. Toxins that translocate from the ER have an unusually low lysine content. Since lysyl residues are potential ubiquitination sites, it has been proposed that this paucity of lysines reduces the chance of ubiquitination and subsequent ubiquitin-mediated proteasomal degradation [Hazes, B., and Read, R. J. (1997) Biochemistry 36, 11051-11054]. Here we provide experimental support for this hypothesis. The two lysyl residues within RTA were changed to arginyl residues. Their replacement in RTA did not have a significant stabilizing effect, suggesting that the endogenous lysyl residues are not the usual sites for ubiquitin attachment. However, when four additional lysines were introduced into RTA in a way that did not compromise the activity, structure, or stability of the toxin, degradation was significantly enhanced. Enhanced degradation resulted from ubiquitination that predisposed the toxin to proteasomal degradation. Treatment with the proteasome inhibitor clasto-lactacystin beta-lactone increased the cytotoxicity of the lysine-rich RTA to a level approaching that of wild-type ricin. The introduction of four additional lysyl residues into a second ribosome-inactivating protein, abrin A chain, also dramatically decreased the cytotoxicity of the holotoxin compared to wild-type abrin. This effect could also be reversed by proteasomal inhibition. Our data support the hypothesis that the evolution of a low lysine content is a degradation-avoidance strategy for toxins that retrotranslocate from the ER.