TULLIO POZZAN BACTERIAL TOXINS Two routes to neurotoxicity? Tetanus and botulinum toxins cause death by preventing small synaptic vesicle release. Recent results suggest mechanisms for these toxins and may provide new insights into the cell biology of synapses. Bacterial toxins still provide one of the most formidable threats to life. Diphtheria, cholera, tetanus and botulism are the most familiar names of the deadly diseases caused solely by these bacterial weapons. Despite the introduc- tion in the first half of this century of mass vaccination with inactivated toxins, each year bacterial toxins still account for millions oF human deaths, mostly in devel- oping countries, The toxins responsible for the diseases mentioned above act intracellularly and they all consist of two subunits: one subunit binds a receptor on the sur- face of target cells and translocates the other, catalytically active, subunit into the: cytosol [ 1,2]. Although progress has been made in understanding the catalytic activities of diphtheria and cholera toxins, the catalytic activities of the tetanus (TeTx) and botulinum (BoTx) toxins, made by Clostridium t-etani and Clostridium botulinurn respectively, have been mysterious. At the subcellular level, the effects of TeTx and BoTx are the same - both block exocytosis of small synaptic vesi- cles As, however, TeTx acts on inhibitory spinal cord in- terneurons, whereas BoTx acts on spinal cord motorneu- rons, the two toxins have opposite effects on skeletal muscle, causing spastic and flaccid paralysis, respectively. Determining the molecular mechanisms by which TeTx and BoTx produce their toxic effects is not only impor- tant from a medical and social point of view, it is also of interest for basic science as it is likely to increase our understanding of the affected cellular function - neuro- transmitter release. Thus, recent results leading to two new suggested mechanisms by which TeTx and BoTx may inhibit small synaptic vesicle exocytosis represent an exciting development in neurobiology [3-51. The new results were (obtained by two groups who both began by analyzing the neurotoxin sequences. Facchiano and Luini noticed that the sequences of both TeTx chains include short stretches with similarities to motifs that oc- cur in the sequences of a number of substrates of cellular transglutaminase [ 31. Transglutaminases are ubiquitous enzymes that catalyse the irreversible crosslinking of proteins [6]. Facchiano and Luini found that TeTx is a stimulater of transglutaminase, increasing the activity of the enzyme, assayed in the presence of physiologi- cal Ca2+ and GTP concentrations, more than ten-fold. This appears to be a stoichiometric stimulation, depend- ing on a 1: 1 physic:al interaction between TeTx and transglutaminase. The next step was to identify the important target pro- tein(s) of TeTx-stimulated transglutaminase. Synapsin I, a major protein of synaptic vesicle membranes [7], was Volume 2 Number 11 1992 shown to be a preferred substrate of transglutaminase, which in synaptic terminals is localized both in the cy tosol and on synaptic vesicle membranes (Facchiano F, Benfenati F, Valtorta F, Luini A personal communica- tion). Synapsin I is known to provide a dynamic link be- tween synaptic vesicles and the cytoskeleton, and thereby to be an indirect regulator of the interaction between synaptic vesicles and the plasma membrane. Thus, Fac- chiano et al. suggest that TeTx acts by stimulating transg- lutaminase, leading to the irreversible, covalent crosslink- ing of synapsin I to microfilaments. This crosslinking is suggested to ‘freeze’ the synaptic vesicles in the cytosol, preventing them from fusing with the plasma membrane. when Schiavo et al. analysed the neurotoxin sequences, however, they spotted a diierent homology - they noticed that a histidine-rich conserved segment of the TeTx and BoTx light chains contains the sequence His-Glu-X-X-His, where X is any amino acid, character- istic of the E-i”+-binding site of metalloproteases [4,5]. Schiavo et al. then proceeded to test the suggestion that TeTx and BoTx have Zn2+ -dependent protease activity crucial to their toxic effects on neurons. In support of this suggestion they found, first, that highly purified TeTx and BoTx preparations contain Zn2+ ions reversibly bound to their light chains in a 1: 1 stoichiometric ratio. Second, chemical modification and 6?Zn2+ -binding experiments indicated that the Zn2+ atom is indeed coordinated by the histidine-rich segments of the neurotoxins. Third, the authors carried out functional experiments in ~pZysia neurons, a well-characterized model system for studying the intracellular action of TeTx and BoTx, which proved particularly telling. Injection of TeTx or BoTx into 4Zysia neurons inhibits neurotransmitter release, and Schiavo et al. found that with Zn2+-depleted toxins (apotoxins) this effect is greatly delayed. The presence of intracellu- lar heavy-metal ion chelators, or chemical modification of the putative Zn2+ -binding his&dine residues of TeTx and BoTx, were found to prevent completely the in vivo reactivation of the apotoxins. And a known inhibitor of Zn2+ -dependent proteases was found to suppress com- pletely the inhibitory effect of TeTx on neurotransmitter release. If TeTx and BoTx are indeed metalloproteases then pre- sumably they act by cleaving some intracellular protein(s) involved in synaptic vesicle exocytosis. Gel electrophore- sis of proteins from highly puriiied synaptic vesicles revealed only one protein, of molecular weight lgk~, cleaved in the presence of the TeTx light chain [4,8]. As expected, only the Zn2+ -containing toxin could cleave the 19 kD protein and cleavage was prevented by Zn2+ 621