Lipid-loving Toxins

Peptide toxins found in spider venom block the opening and closing of membrane-bound ion-channel proteins. It has been assumed that they do so via complementary binding surfaces on peptide and protein, but two new studies suggest that these toxins' ability to partition into the waxy lipid membrane is the real key.

Philip A. Gottlieb of State University of New York, Buffalo; Olaf S. Andersen of Cornell University's Weill Medical College; and coworkers studied how the tarantula venom toxin GsMTx4 inhibits an ion channel that opens and closes in response to stretching of the membrane [Nature, 430, 235 (2004)].

To their surprise, they found that GsMTx4's mirror image blocks channel gating just as GsMTx4 does. Furthermore, both GsMTx4 and its enantiomer have the identical effect on a completely unrelated stretch-activated channel.

"These results violate the traditional lock-and-key model of ligand-protein interactions," Gottlieb says. Unlike normal ligand-protein interactions, the toxin doesn't bind its target via a binding surface that is sterically and electrostatically complementary to a surface on the channel. Instead, "it achieves its effect by modulating the lipids that surround the ion channel," perhaps even without contacting the channel, he says.

A second, independent study suggests that related venom peptide toxins--all of which sport a highly hydrophobic face that could drive partitioning into the membrane--might also access their targets through the membrane.

Seok-Yong Lee and Roderick MacKinnon of Howard Hughes Medical Institute and Rockefeller University probed the interaction of VSTX1--a related toxin from the same tarantula--with an ion channel that opens and closes in response to voltage changes across the membrane [Nature, 430, 232 (2004)]. They report that VSTX1 binds to membrane-bound channels with high affinity but has only low affinity for channels that have been removed from the membrane with detergent. Lee and MacKinnon conclude that VSTX1 tends to accumulate in the membrane and only reaches its weak binding site on the channel via the lipid bilayer.

"Much of the binding energy of VSTX1 and GsMTx4 is derived from the nonspecific free energy of membrane partitioning, whereas the actual peptide-channel interaction is rather weak," notes Maria L. Garcia of Merck Research Laboratories in a commentary that accompanies the two papers. She suggests this could make rational drug design of mimics of these toxins difficult.

There is a bright side, however: Gottlieb notes that GsMTx4 has been shown to prevent atrial fibrillation in rabbit hearts and may also hold promise as a treatment for muscular dystrophy and incontinence. Its equally effective all-D enantiomer is likely to be more resistant to breakdown and to be orally administrable.