In the late 1970’s, the KGB inducted ricin into the “cloak and dagger” hall-of-fame by turning a more or less innocuous umbrella tip into their assassin’s weapon of choice. In recent popular imagination, Walter White immortalized this toxin from the castor oil plant by deploying it against his opponents in the TV show Breaking Bad. Ricin’s remarkable efficacy – just a few grains of it in salt form are enough to kill a human – belies its torturous journey into the cell. Starting with endocytosis, ricin moves using vesicles through a series of compartments of the secretory system in reverse until it reaches the endoplasmic reticulum (ER). Normally, secreted proteins depend on a protein translocation channel in the ER membrane to escape from the cell. By going through this channel in reverse, even a single molecule of ricin can proceed to kill the cell by rapidly inactivating one ribosome after another. Mercifully, a small molecule was discovered in 2010 with the ability to fully protect mice from lethal doses of ricin. This compound, aptly named Retro-2, rendered ricin innocuous by halting its retrograde progress through the secretory pathway. In the years that followed, Retro-2 was shown to also confer protection against a broad variety of toxins and viruses that take advantage of the host cell’s vesicle transport machinery, as well as against certain pathogens with distinct etiologies, including the Ebola virus. Despite rapidly acquiring its sensational reputation, the cellular target of Retro-2 and its mechanism of action have remained a mystery.
In our paper, done in collaboration with the Bassik (Stanford University) and Sello labs (then at Brown University), we take a CRISPR-based approach for target identification by finding genes that when knocked down result in a Retro-2-like phenotypic signature. Our analysis identifies as its top hits several components of the transmembrane domain recognition complex (TRC) pathway, which mediates the post-translational ER targeting and insertion of tail-anchored (TA) proteins. This class of membrane proteins is defined by a single C-terminal transmembrane domain and many of its members get sorted out of the ER to other compartments by vesicle transport.
So does Retro-2 actually inhibit the TRC pathway as predicted by genetic interaction analysis? We validated three key expectations for such a small-molecule mechanism of action. First, we use fluorescent reporters for TA protein insertion to show that cells treated with Retro-2 have reduced post-translational substrate flux through the TRC pathway to the ER. By re-wiring one TA protein for recognition by the signal recognition particle, we render its co-translational flux to the ER insensitive to Retro-2. Second, we show using purified proteins that Retro-2 inhibits a specific TRC pathway step during which substrates are loaded onto ASNA1, the factor that takes newly-synthesized TA proteins to the ER membrane for insertion. Third, we use a CRISPR base editing tool to isolate a point mutant of ASNA1 that abolishes Retro-2’s activity as measured by three different assays: protection from ricin, disruption of substrate flux through the TRC pathway in cells, disruption of substrate loading onto ASNA1 in vitro.
But how does inhibition of the TRC pathway enable ricin to become arrested in endosomes en route to the ER? Our work argues for a mechanism of action whereby Retro-2 depletes the secretory and endocytic system of its TRC pathway clients. Among these, tail-anchored SNARE proteins are a particularly promising set of effector molecules. These proteins are capable of zippering up with one another on opposing vesicle and target compartment membranes, which overcomes the otherwise high energy barrier to spontaneous lipid bilayer fusion imposed by intervening water molecules. One particular SNARE protein – Stx5 – has a Retro-2-like phenotypic signature and is depleted by Retro-2 from its Golgi residence where it normally mediates fusion with retrograde vesicles derived from endosomes. Thus, engineering Stx5 ER targeting to be co-translational might be a promising future way to abolish Retro-2’s protective effect against ricin by targeting a critical TA protein effector. More broadly, depletion of other SNARE proteins might be at the heart of Retro-2’s panacean abilities to halt other toxins and pathogens en route to their final destinations.