Calcein permeability of liposomes mediated by type A botulinum neurotoxin and its light and heavy chains

The mode of botulinum neurotoxin action involves binding of its heavy chain for internalization into the presynaptic end of a nerve cell through endocytosis. The low-pH conditions of endosomes trigger translocation of the light chain across the endosomal membrane to the cytosol, where the light chain cleaves specific target proteins involved in the docking and fusion of synaptic vesicles for acetylcholine release. In an effort to model the interaction of botulinum neurotoxin and its subunit chains with lipid bilayer at low pH during the translocation process, we have examined type A botulinum neurotoxin-mediated calcein release from asolectin liposomes. At equimolar concentration (0.1 M), the neurotoxin and its heavy and light chains evoked 23%, 58%, and 28% calcein release, respectively. Calcein release was observed only when the cis-side (the side to which neurotoxin samples were added) pH was lowered to 4. Calcein release activity of the heavy chain was mostly blocked (76%) by a polyclonal antibody raised against the neurotoxin. Additionally, two peptide-specific polyclonal antibodies derived from the N-terminal and C-terminal halves of the heavy chain were also able to block the calcein release activity by 15–20%. In summary, these results suggest that calcein release from liposomes is specifically mediated by the heavy chain, and the light chain also integrates into the membrane. Implications of these results for the molecular mode of neurotoxin light-chain translocation across the endosomal membrane are discussed.     

Botulinum neurotoxins are zinc proteins.

The available amino acid sequences of 150-kDa botulinum and tetanus neurotoxins show the presence of a closely homologous segment in the middle of the light chain (NH2-terminal 50 kDa), which is the intracellularly active portion of the toxin. This segment contains the zinc binding motif of metalloendopeptidases, HEXXH. Atomic adsorption analysis of botulinum neurotoxins (serotypes A, B, and E) made on the basis of this observation demonstrated the presence of one zinc atom/molecule of 150-kDa neurotoxin. Conditions were found for the removal of the zinc ion with chelating agents and for the restoration of the normal metal content. The conserved segment, which includes the zinc binding motif, was synthesized and shown to bind [65Zn]2+. Chemical modification experiments indicated that two histidines and no cysteines are involved in Zn2+ coordination in agreement with a probable catalytic role for the zinc ion. The present findings suggest the possibility that botulinum neurotoxins are zinc proteases.     

A new, smaller actin-activatable myosin subfragment 1 which lacks the 20-kDa, SH1 and SH2 peptide

The actin-dependent ATPase activity of myosin is retained in the separated heads (S1) which contain the NH2-terminal 95-kDa heavy chain fragment and one or two light chains. The S1 heavy chain can be degraded further by limited trypsin treatment into characteristic 25-, 50-, and 20-kDa peptides, in this order from the NH2-terminal end. The 20-kDa peptide contains an actin-binding site and SH1 and SH2, two thiols whose modification dramatically affects ATPase activity. By treating myosin filaments with trypsin at 4 degrees C in the presence of 2 mM MgCl2, we have now obtained preferential cleavage at the 50-20-kDa heavy chain site without any cleavage at the head-rod junction and hinge region in the rod. Incubation of these trypsinized filaments at 37 degrees C in the presence of MgATP released a new S1 fraction which lacked the COOH-terminal 20-kDa heavy chain peptide region. This fraction, termed S1'(75K), has more than 50% of the actin-activated Mg2+-ATPase activity of S1 and the characteristic Ca2+-ATPase and K+-EDTA ATPase activities of myosin. These results show that SH1 and SH2 are not essential for ATPase activity and that binding of actin to the 20-kDa region is not essential for the enhancement of the Mg2+-ATPase activity.     

Isolated light chains of botulinum neurotoxins inhibit exocytosis. Studies in digitonin-permeabilized chromaffin cells

The effects of botulinum neurotoxins or their light and heavy chain subunits were investigated in digitonin-permeabilized adrenal chromaffin cells. Because these cells are permeable to proteins, the toxin had direct access to the cell interior. Botulinum type A neurotoxin and its light chain subunit inhibited Ca2+-dependent catecholamine secretion in a dose-dependent manner. The heavy chain subunit had no effect. Inhibition required introduction of the neurotoxin or light chain into the cell and was not seen when intact cells were incubated with these proteins. The inhibition of secretion by type A neurotoxin and light chain was incomplete, the maximal response being 65%. The inhibition was not overcome by increasing Ca2+ concentrations. The action of the light chain was irreversible and rapid. Botulinum type E neurotoxin also inhibited secretion in a dose-dependent manner. Its potency was increased 30-fold following mild trypsinization, which nicked the single chain protein to the dichain form. In contrast to the results seen with types A and E, botulinum type B neurotoxin did not inhibit secretion, while its light chain totally abolished secretion. Trypsinization of the neurotoxin produced the dichain form, which did not inhibit secretion. Reduction of the trypsinized neurotoxin with dithiothreitol produced inhibition equivalent to that seen with the purified light chain subunit. Isolated type A heavy chain had no effect on the inhibitory action of type A or B light chains. The data demonstrate that the ability of botulinum neurotoxins to inhibit secretion is confined to the light chain region of these proteins. Furthermore, while the botulinum neurotoxin types A, B, and E have similar macrostructures, they are not identical with respect to their biological activities.     

Is formation of visible channels in a phospholipid bilayer by botulinum neurotoxin type B sensitive to its disulfide?

Botulinum neurotoxin, produced by Clostridium botulinum as a approximately 150-kDa single-chain protein, is nicked proteolytically either endogenously or exogenously. The approximately 50- and approximately 100-kDa chains of the dichain molecule remain held together by an interchain disulfide bridge and noncovalent interactions. The neurotoxin binds to receptors of the target cell and is internalized by endocytosis. Thereafter, a portion of the neurotoxin, the approximately 50-kDa chain, escapes to the cytosol, where it blocks neurotransmitter release. Botulinum neurotoxin serotype B is released by the bacteria primarily as an unnicked single chain. We reduced this unnicked protein and used its binding to ganglioside in a lipid layer to produce helical tubular crystals of unnicked botulinum neurotoxin type B in its disulfide-reduced state. The helical arrangement of the neurotoxin allowed determination of the structure of the molecule using cryo-electron microscopy and image processing. The resulting model reveals that neurotoxin molecules formed loops extending out from the surface of the bilayer and bending toward a neighboring loop. Although channels have been seen with disulfide-linked neurotoxin (Schmid, Robinson, and DasGupta (1993) Direct visualization of botulinum neurotoxin-induced channels in phospholipid vesicles, Nature 364, 827-830), no channels were seen here, a finding which suggests that the reduced, unnicked neurotoxin is incapable of forming a visible channel.     

Inhibition of Acanthamoeba myosin I heavy chain kinase by Ca(2+)-calmodulin.

The actin-activated Mg(2+)-ATPase activity of Acanthamoeba myosins I depends on phosphorylation of their single heavy chains by myosin I heavy chain kinase. Kinase activity is enhanced > 50-fold by autophosphorylation at multiple sites. The rate of kinase autophosphorylation is increased approximately 20-fold by acidic phospholipids independent of the presence of Ca2+ and diglycerides. We show in this paper that Ca(2+)-calmodulin inhibits phospholipid-stimulated autophosphorylation of myosin I heavy chain kinase and hence also inhibits the catalytic activity of unphosphorylated kinase in the presence of phospholipid. Ca(2+)-calmodulin does not inhibit kinase activity in the absence of phospholipid. Micromolar Ca(2+)-calmodulin also inhibits binding of myosin I heavy chain kinase to phospholipid vesicles and purified plasma membranes. Proteolytic removal of a 7-kDa NH2-terminal segment from the 97-kDa kinase prevents binding of both calmodulin and phospholipid; therefore, we propose that they bind to the same or overlapping sites. These data provide a mechanism by which Ca2+ could inhibit the actin-activated Mg(2+)-ATPase activity of the myosin I isozymes in vivo and thus regulate myosin I-dependent motile activities.     

Interactions of synapsin I with small synaptic vesicles: distinct sites in synapsin I bind to vesicle phospholipids and vesicle proteins

Synapsin I is a major neuron-specific phosphoprotein that is specifically localized to the cytoplasmic surface of small synaptic vesicles. In the present study, the binding of synapsin I to small synaptic vesicles was characterized in detail. The binding of synapsin I was preserved when synaptic vesicles were solubilized and reconstituted in phosphatidylcholine. After separation of the protein and lipid components of synaptic vesicles under nondenaturing conditions, synapsin I bound to both components. The use of hydrophobic labeling procedures allowed the assessment of interactions between phospholipids and synapsin I in intact synaptic vesicles. Hydrophobic photolabeling followed by cysteine-specific cleavage of synapsin I demonstrated that the head domain of synapsin I penetrates into the hydrophobic core of the bilayer. The purified NH2-terminal fragment, derived from the head domain by cysteine-specific cleavage, bound to synaptic vesicles with high affinity confirming the results obtained from hydrophobic photolabeling. Synapsin I binding to synaptic vesicles could be inhibited by the entire molecule or by the combined presence of the NH2-terminal and tail fragments, but not by an excess of either NH2-terminal or tail fragment alone. The purified tail fragment bound with relatively high affinity to synaptic vesicles, though it did not significantly interact with phospholipids. Binding of the tail fragment was competed by holosynapsin I; was greatly decreased by phosphorylation; and was abolished by high ionic strength conditions or protease treatment of synaptic vesicles. The data suggest the existence of two sites of interaction between synapsin I and small synaptic vesicles: binding of the head domain to vesicle phospholipids and of the tail domain to a protein component of the vesicle membrane. The latter interaction is apparently responsible for the salt and phosphorylation dependency of synapsin I binding to small synaptic vesicles.     

Limited tryptic digestion of Acanthamoeba myosin IA abolishes regulation of actin-activated ATPase activity by heavy chain phosphorylation

Acanthamoeba myosin IA is a globular protein composed of a 140-kDa heavy chain and a 17-kDa light chain. It expresses high actin-activated Mg2+-ATPase activity when one serine on the heavy chain is phosphorylated. We previously showed that chymotrypsin cleaves the heavy chain into a COOH-terminal 27-kDa peptide that can bind to F-actin but has no ATPase activity and a complex containing the NH2-terminal 112-kDa peptide and the light chain. The complex also binds F-actin and has full actin-activated Mg2+-ATPase activity when the regulatory site is phosphorylated. We have now localized the ATP binding site to within 27 kDa of the NH2 terminus and the regulatory phosphorylatable serine to a 20-kDa region between 38 and 58 kDa of the NH2 terminus. Under controlled conditions, trypsin cleaves the heavy chain at two sites, 38 and 112 kDa from the NH2 terminus, producing a COOH-terminal 27-kDa peptide similar to that produced by chymotrypsin and a complex consisting of an NH2-terminal kDa peptide, a central 74-kDa peptide, and the light chain. This complex is similar to the chymotryptic complex but for the cleavage which separates the 38- and 74-kDa peptides. The tryptic complex has full (K+, EDTA)-ATPase activity (the catalytic site is functional) and normal ATP-sensitive actin-binding properties. However, the actin-activated Mg2+-ATPase activity and the F-actin-binding characteristics of the tryptic complex are no longer sensitive to phosphorylation of the regulatory serine. Therefore, cleavage between the phosphorylation site and the ATP-binding site inhibits the effects of phosphorylation on actin binding and actin-activated Mg2+-ATPase activity without abolishing the interactions between the ATP- and actin-binding sites.     

Distinct steps in the penetration of adenylate cyclase toxin of Bordetella pertussis into sheep erythrocytes. Translocation of the toxin across the membrane.

Adenylate cyclase (AC) toxin from Bordetella pertussis penetrates eukaryotic cells and upon activation by calmodulin generates unregulated levels of intracellular cAMP. The process of toxin penetration into sheep erythrocytes was resolved into three consecutive steps including insertion, translocation, and intracellular cleavage. Insertion of the toxin into the cell membrane occurred over a wide temperature range (4-36 degrees C). In contrast, translocation of the toxin, i.e. transfer of the NH2-terminal catalytically active fragment across the membrane, occurred only above 20 degrees C and was highly temperature-dependent. While a single exposure of the toxin to Ca2+ was sufficient for its insertion into the plasma membrane, toxin translocation required exogenous Ca2+ at mM concentrations. Translocation was not affected by pretreatment of cells with trypsin, N-ethylmaleimide, and sodium carbonate at alkaline pH. The NH2-terminal fragment of the toxin was cleaved in the cell releasing the 45-kDa active AC into the cytosol. The cleavage was blocked by treatment of cells with N-ethylmaleimide. It is hypothesized that the COOH-terminal portion of the toxin creates in the membrane a channel through which the NH2-terminal fragment is translocated.     

Effect of pH on the interaction of botulinum neurotoxins A, B and E with liposomes.

The interaction of botulinum neurotoxin serotypes A, B and E with membranes of different lipid compositions was examined by photolabelling with two photoreactive phosphatidylcholine analogues that monitor the polar region and the hydrophobic core of the lipid bilayer. At neutral pH the neurotoxins interacted both with the polar head groups and with fatty acid chains of phospholipids. At acidic pHs the neurotoxins underwent structural changes characterized by a more extensive interaction with lipids. Both the heavy and light chain subunits of the neurotoxins were involved in the process. The change in the nature and extent of toxin-lipid interaction occurred in the pH range 4-6 and was not influenced by the presence of polysialogangliosides. The present data are in agreement with the idea that botulinum neurotoxins enter into nerve cells from a low pH intracellular compartment.     

Double anchorage to the membrane and intact inter-chain disulfide bond are required for the low pH induced entry of tetanus and botulinum neurotoxins into neurons

Tetanus and botulinum neurotoxins are di-chain proteins that cause paralysis by inhibiting neuroexocytosis. These neurotoxins enter into nerve terminals via endocytosis inside synaptic vesicles, whose acidic pH induces a structural change of the neurotoxin molecule that becomes capable of translocating its L chain into the cytosol, via a transmembrane protein-conducting channel made by the H chain. This is the least understood step of the intoxication process primarily because it takes place inside vesicles within the cytosol. In the present study, we describe how this passage was made accessible to investigation by making it to occur at the surface of neurons. The neurotoxin, bound to the plasma membrane in the cold, was exposed to a warm low pH extracellular medium and the entry of the L chain was monitored by measuring its specific metalloprotease activity with a ratiometric method. We found that the neurotoxin has to be bound to the membrane via at least two anchorage sites in order for a productive low-pH induced structural change to take place. In addition, this process can only occur if the single inter-chain disulfide bond is intact. The pH dependence of the conformational change of tetanus neurotoxin and botulinum neurotoxin B, C and D is similar and take places in the same slightly acidic range, which comprises that present inside synaptic vesicles. Based on these and previous findings, we propose a stepwise sequence of molecular events that lead from toxin binding to membrane insertion.     

NH2-terminal calcium-binding domain of human complement C1s- mediates the interaction of C1r- with C1q

The assembly of C1, the first component of human complement, involves interactions between various domains of each of its three subcomponents, C1q, C1r, and C1s. The isolation, assignment of function, and structural characterization of the individual domains of C1r and C1s are critical for a thorough understanding of this complex assembly. The present study describes a 27-kDa plasmin-generated fragment derived from the NH2-terminal half of the heavy A chain of C1s-, the activated form of C1s. This fragment, C1s-alpha, was shown in the presence of Ca2+ to mimic the ability of whole C1s- to self-associate, bind to C1r-, and facilitate the binding of C1r to C1q. These results directly prove that the Ca2(+)-binding sites of C1s as well as all of the determinants necessary for binding of C1s- to C1r- and C1q are located in the NH2-terminal 27-kDa alpha region of the A chain.     

Chemical identification of serine 181 at the ATP-binding site of myosin as a residue esterified selectively by the fluorescent reagent 9-anthroylnitrile

The esterification reagent 9-anthroylnitrile (ANN) reacts with a serine residue in the NH2-terminal 23-kDa peptide segment of myosin subfragment-1 heavy chain to yield a fluorescent S1 derivative labeled by the anthroyl group (Hiratsuka, T. (1989) J. Biol. Chem. 264, 18188-18194). The labeling was highly selective and accelerated by nucleotides. In the present study, to determine the exact location of the labeled serine residue, the labeled 23-kDa peptide fragment was isolated. The subsequent extensive proteolytic digestion of the peptide fragment yielded two labeled peptides, a pentapeptide and its precursor nonapeptide. Amino acid sequence and composition analyses of both labeled peptides revealed that the anthroyl group is attached to Ser-181 involved in the phosphate binding loop for ATP (Smith, C. A., and Rayment, I. (1996) Biochemistry 35, 5404-5417). We concluded that ANN can esterify Ser-181 selectively out of over 40 serine residues in the subfragment 1 heavy chain. Thus ANN is proved to be a valuable fluorescent tool to identify peptides containing the phosphate binding loop of S1 and to detect the conformational changes around this loop.     

Reductive chain separation of botulinum A toxin--a prerequisite to its inhibitory action on exocytosis in chromaffin cells

Cleavage of the disulfide bond linking the heavy and the light chains of tetanus toxin is necessary for its inhibitory action on exocytotic release of catecholamines from permeabilized chromaffin cells [(1989) FEBS Lett. 242, 245-248; (1989) J. Neurochem., in press]. The related botulinum A toxin also consists of a heavy and a light chain linked by a disulfide bond. The actions of both neurotoxins on exocytosis were presently compared using streptolysin O-permeabilized bovine adrenal chromaffin cells. Botulinum A toxin inhibited Ca2+-stimulated catecholamine release from these cells. Addition of dithiothreitol lowered the effective doses to values below 5 nM. Under the same conditions, the effective doses of tetanus toxin were decreased by a factor of five. This indicates that the interchain S-S bond of botulinum A toxin must also be split before the neurotoxin can exert its effect on exocytosis.     

Tetanus toxin is a zinc protein and its inhibition of neurotransmitter release and protease activity depend on zinc.

Tetanus and botulinum neurotoxins are the most potent toxins known. They bind to nerve cells, penetrate the cytosol and block neurotransmitter release. Comparison of their predicted amino acid sequences reveals a highly conserved segment that contains the HexxH zinc binding motif of metalloendopeptidases. The metal content of tetanus toxin was then measured and it was found that one atom of zinc is bound to the light chain of tetanus toxin. Zinc could be reversibly removed by incubation with heavy metal chelators. Zn2+ is coordinated by two histidines with no involvement in cysteines, suggesting that it plays a catalytic rather than a structural role. Bound Zn2+ was found to be essential for the tetanus toxin inhibition of neurotransmitter release in Aplysia neurons injected with the light chain. The intracellular activity of the toxin was blocked by phosphoramidon, a very specific inhibitor of zinc endopeptidases. Purified preparations of light chain showed a highly specific proteolytic activity against synaptobrevin, an integral membrane protein of small synaptic vesicles. The present findings indicate that tetanus toxin, and possibly also the botulinum neurotoxins, are metalloproteases and that they block neurotransmitter release via this protease activity.     

Detection and localization of a cytoplasmic domain on the beta-subunit of Na+,K+-ATPase. A monoclonal antibody study

A monoclonal antibody directed against the beta-subunit of dog kidney Na+,K+-ATPase was generated. Immunoblots demonstrate that monoclonal antibody III 18A binds exclusively to the denaturated beta-subunit. Binding experiments with membranes and whole cells reveal that III 18A binds to membranes but not to whole cells, indicating that the antibody binds to a cytoplasmic domain on the native beta-subunit. To localize the antibody-binding epitope, purified membrane-bound enzyme was fragmented by protease treatment. Tryptic digestion yields a 30-kDa fragment of the beta-subunit, which still retains the binding capacity for the antibody. Thus III 18A probably does not bind to the NH2-terminal segment of the protein. On the other hand, fragmentation of the beta-subunit with low concentrations of papain, which is known to yield a 40-kDa NH2-terminal and a 16-kDa COOH-terminal fragment, results in a complete loss of III 18A binding. These results suggest that the antibody-binding epitope is localized at or near a papain cleavage site on the COOH-terminal part of the beta-subunit. This is inconsistent with a structure model of the beta-subunit containing only a single transmembrane hydrophobic segment with a cytoplasmic NH2-terminal portion, but agrees quite well with a hypothetical structure with four intramembrane segments.     

Purification and characterization of a protease from Clostridium botulinum type A that nicks single-chain type A botulinum neurotoxin into the di-chain form.

A protease that nicks the approximately 150-kilodalton (kDa) single-chain type A botulinum neurotoxin into the approximately 150-kDa di-chain form in vitro was isolated from Clostridium botulinum type A (Hall strain) cultures. The di-chain neurotoxin generated in vitro is composed of an approximately 50-kDa light chain and an approximately 100-kDa heavy chain which are disulfide linked and is indistinguishable from the di-chain neurotoxin that forms in vivo and is routinely isolated (M.L. Dekleva and B.R. DasGupta, Biochem. Biophys. Res. Commun. 162:767-772, 1989). This enzyme was purified greater than 1,000-fold by ammonium sulfate precipitation, QAE-Sephadex Q-50, Sephadex G-100, and CM-Sephadex C-50 chromatography steps with the synthetic substrate N-benzoyl-DL-arginine-p-nitroanilide. The approximately 62-kDa amidase (protease) is a complex of 15.5- and 48-kDa polypeptides (determined by polyacrylamide gel electrophoresis) that could not be separated without sodium dodecyl sulfate. The enzyme has an isoelectric point of pH 5.73, a pH optimum of 6.2 to 6.4, an absolute requirement for a thiol-reducing agent as well as a divalent metallic cation (probably Ca2+) for activity, and a temperature optimum of 70 degrees C. Tests with several synthetic substrates indicated the high specificity of the enzyme for arginyl amide bonds.     

Localization of the actin-binding sites of Acanthamoeba myosin IB and effect of limited proteolysis on its actin-activated Mg2+-ATPase activity

Acanthamoeba myosin IB contains a 125-kDa heavy chain that has high actin-activated Mg2+-ATPase activity when 1 serine residue is phosphorylated. The heavy chain contains two F-actin-binding sites, one associated with the catalytic site and a second which allows myosin IB to cross-link actin filaments but has no direct effect on catalytic activity. Tryptic digestion of the heavy chain initially produces an NH2-terminal 62-kDa peptide that contains the ATP-binding site and the regulatory phosphorylation site, and a COOH-terminal 68-kDa peptide. F-actin, in the absence of ATP, protects this site and tryptic cleavage then produces an NH2-terminal 80-kDa peptide. Both the 62- and the 80-kDa peptides retain the (NH+4,EDTA)-ATPase activity of native myosin IB and both bind to F-actin in an ATP-sensitive manner. However, only the 80-kDa peptide retains a major portion of the actin-activated Mg2+-ATPase activity. This activity requires phosphorylation of the 80-kDa peptide by myosin I heavy chain kinase but, in contrast to the activity of intact myosin IB, it has a simple, hyperbolic dependence on the concentration of F-actin. Also unlike myosin IB, the 80-kDa peptide cannot cross-link F-actin filaments indicating the presence of only a single actin-binding site. These results allow the assignment of the actin-binding site involved in catalytic activity to the region near, and possibly on both sides of, the tryptic cleavage site 62 kDa from the NH2 terminus, and the second actin-binding site to the COOH-terminal 45-kDa domain. Thus, the NH2-terminal 80 kDa of the myosin IB heavy chain is functionally similar to the 93-kDa subfragment 1 of muscle myosin and most likely has a similar organization of functional domains.     

Localization of distinct functional domains on prekallikrein for interaction with both high molecular weight kininogen and activated factor XII in a 28-kDa fragment (amino acids 141-371).

The predominant autolytic form of human kallikrein, beta-kallikrein, was used to localize the high molecular weight kininogen (HK) binding site on kallikrein as well as the substrate recognition site for activated factor XII on prekallikrein. beta-Kallikrein is formed by autolysis of the kallikrein heavy chain to give two fragments of approximately 18 and 28 kDa by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. A ligand binding technique established that the HK binding site on kallikrein residues on the 28-kDa fragment of the heavy chain. Limited NH2-terminal sequencing of this portion of beta-kallikrein showed that this fragment of the heavy chain consists of the COOH-terminal 231 amino acids of the heavy chain. A panel of five murine monoclonal antibodies to human prekallikrein (PK) were found to have epitopes on this same fragment of the heavy chain. None of the monoclonal antibodies were able to block binding of HK to PK. Three of the monoclonal antibodies (13G11, 13H11, and 6A6) were able to inhibit the activation of PK to kallikrein in both a plasma system and a purified system. The 28-kDa fragment of the PK heavy chain was purified and was able to compete with HK for binding to PK. The HK binding site and the site of recognition of factor XII are separate and distinct on PK, and both are contained in the COOH-terminal 231 amino acids of the PK heavy chain.     

The C-terminal transmembrane region of synaptobrevin binds synaptophysin from adult synaptic vesicles

Synaptophysin and synaptobrevin are abundant membrane proteins of neuronal small synaptic vesicles. In mature, differentiated neurons they form the synaptophysin/synaptobrevin (Syp/Syb) complex. Synaptobrevin also interacts with the plasma membrane-associated proteins syntaxin and SNAP25, thereby forming the SNARE complex necessary for exocytotic membrane fusion. The two complexes are mutually exclusive. Synaptobrevin is a C-terminally membrane-anchored protein with one transmembrane domain. While its interaction with its SNARE partners is mediated exclusively by its N-terminal cytosolic region it has been unclear so far how binding to synaptophysin is accomplished. Here, we show that synaptobrevin can be cleaved in its synaptophysin-bound form by tetanus toxin and botulinum neurotoxin B, or by botulinum neurotoxin D, leaving shorter or longer C-terminal peptide chains bound to synaptophysin, respectively. A recombinant, C-terminally His-tagged synaptobrevin fragment bound to nickel beads specifically bound synaptophysin, syntaxin and SNAP25 from vesicular detergent extracts. After cleavage by tetanus toxin or botulinum toxin D light chain, the remaining C-terminal fragment no longer interacted with syntaxin or SNAP 25. In contrast, synaptophysin was still able to bind to the residual C-terminal synaptobrevin cleavage product. In addition, the His-tagged C-terminal synaptobrevin peptide 68-116 was also able to bind synaptophysin in detergent extracts from adult brain membranes. These data suggest that synaptophysin interacts with the C-terminal transmembrane part of synaptobrevin, thereby allowing the N-terminal cytosolic chain to interact freely with the plasma membrane-associated SNARE proteins. Thus, by binding synaptobrevin, synaptophysin may positively modulate neurotransmission.