Botulinum neurotoxin D-C uses synaptotagmin I and II as receptors, and human synaptotagmin II is not an effective receptor for type B, D-C and G toxins

Abstract

Botulinum neurotoxins (BoNTs) are classified into seven types (A-G), but multiple subtype and mosaic toxins exist. These subtype and mosaic toxins share a high sequence identity, and presumably the same receptors and substrates with their parental toxins. Here, we report that a mosaic toxin, type D-C (BoNT/D-C), uses different receptors from its parental toxin BoNT/C. BoNT/D-C, but not BoNT/C, binds directly to the luminal domains of synaptic vesicle proteins synaptotagmin (Syt) I and II, and requires expression of SytI/II to enter neurons. The SytII luminal fragment containing the toxin-binding site can block the entry of BoNT/D-C into neurons and reduce its toxicity in vivo in mice. We also found that gangliosides increase binding of BoNT/D-C to SytI/II and enhance the ability of the SytII luminal fragment to block BoNT/D-C entry into neurons. These data establish SytI/II, in conjunction with gangliosides, as the receptors for BoNT/D-C, and indicate that BoNT/D-C is functionally distinct from BoNT/C. We further found that BoNT/D-C recognizes the same binding site on SytI/II where BoNT/B and G also bind, but utilizes a receptor-binding interface that is distinct from BoNT/B and G. Finally, we also report that human and chimpanzee SytII has diminished binding and function as the receptor for BoNT/B, D-C and G owing to a single residue change from rodent SytII within the toxin binding site, potentially reducing the potency of these BoNTs in humans and chimpanzees.

Fig. 1.

BoNT/D-C does not share the same receptor with BoNT/C. (A) Schematic diagrams showing the protein sequence identity between each functional domain of BoNT/D-C with the corresponding regions in its parental toxins, BoNT/D and BoNT/C. (B) Recombinant C/H_C, D-C/H_C, and D/H_C proteins (150 ng) were subjected to immunoblot analysis, using either a rabbit polyclonal BoNT/C antibody generated using the full-length BoNT/C (upper panel), or a human monoclonal BoNT/D-C antibody (clone 8DC2) raised against the full-length BoNT/D-C (lower panel). The polyclonal BoNT/C antibody recognized C/H_C, but not D-C/H_C and D/H_C. The monoclonal BoNT/D-C antibody recognized D-C/H_C, but not C/H_C and D/H_C. (C) Cultured rat hippocampal neurons were exposed to BoNT/C (10 nM) and BoNT/D-C (0.3 nM) simultaneously for 5 minutes in high K⁺ buffer, with or without (No H_C) the presence of 1 µM of recombinant D-C/H_C or C/H_C. Cells were washed and further incubated in toxin-free media for 6 hours. Cell lysates were subjected to immunoblot analysis using antibodies against the proteins that are indicated. Cells that were not exposed to toxins served as a control (No toxin). Actin serves as an internal loading control. Cleavage of Syb by BoNT/D-C results in the loss of its immunoblot signals. BoNT/C cleaves both Syx and SNAP-25. Cleavage of Syx by BoNT/C results in the loss of its immunoblot signals, whereas cleavage of SNAP-25 by BoNT/C generates a smaller fragment indicated by an asterisk. D-C/H_C reduced the cleavage of Syb by BoNT/D-C, but it did not affect Syx and SNAP-25 cleavage by BoNT/C. Similarly, C/H_C reduced entry of BoNT/C, but had no effect on BoNT/D-C entry. (D) Neurons were exposed to BoNT/D-C (0.3 nM) for 5 minutes in conditions that either stimulate synaptic vesicle exocytosis (High K⁺) or reduce synaptic vesicle exocytosis (PBS plus TTX). Using high K⁺ buffer increased cleavage of Syb. Unless indicated in the figure, high K⁺ buffer was used to load toxins into neurons in all toxin-loading experiments. (E) Experiments were carried out as described in panel D, except that neurons were exposed to the concentrations of BoNT/C that are indicated. Similar levels of Syx and SNAP-25 cleavage were observed in high K⁺ buffer and in PBS plus TTX conditions.

Fig. 2.

D-C/H_C binds directly to the luminal domains of SytI and II. (A) Recombinant D-C/H_C, C/H_C and B/H_C were purified as GST fusion proteins and incubated with detergent (Triton X-100) extracts from rat brain. GST protein serves as a control. Pellets were analyzed by immunoblot. (B) Upper panel: schematic diagram showing the topology of SytI/II on vesicles. Lower panel: truncations of SytI, II and IX containing the luminal domain and the transmembrane domain were used to pull down soluble D-C/H_C, C/H_C or D/H_C, with (+) or without (−) gangliosides (Gangl). Bound materials were analyzed by immunoblot. (C) Pull-down assays were carried out as described in panel B, with indicated concentrations (µM) of D-C/H_C. Bound materials were subjected to SDS-PAGE and visualized by staining with Coomassie blue. The asterisk indicates a degradation band from GST-SytII 1–87. Binding of D-C/H_C to SytII 1–87 is saturated at high concentrations, suggesting a stoichiometric interaction.

Fig. 3.

Expression of SytI/II is required for BoNT/D-C entry into neurons. (A) Cultured rat hippocampal neurons were infected with (+) or without (−) SytI KD lentiviruses. Cells were harvested 7 days later and cell lysates were subjected to immunoblot analysis. SytI expression was greatly reduced in SytI KD neurons, demonstrating a high KD efficiency. Expression of Syb was not affected in SytI KD neurons, indicating the specificity of SytI KD. Actin serves as a loading control. (B) Control and SytI KD neurons were exposed to BoNT/A (20 nM) for 5 minutes in high K⁺ buffer. Cells were washed, fixed and subjected to immunostaining analysis detecting BoNT/A and Syb. GFP was co-expressed with SytI KD shRNA to label infected neurons. The infection efficiency of lentiviruses is close to 100%. Syb serves as a marker for nerve terminals. SytI KD did not affect binding and entry of BoNT/A into nerve terminals. (C) Control neurons, SytI KD neurons and KD rescue neurons with the expression of KD resistant SytI were exposed to BoNT/B (20 nM) for 5 minutes in high K⁺ buffer. Cells were washed, fixed and subjected to immunostaining analysis. Binding and entry of BoNT/B into nerve terminals was abolished in SytI KD neurons, and was restored by KD resistant SytI. (D) Similar levels of SNAP-25 cleavage by BoNT/A (10 nM, 5 minutes exposure, 6 hours incubation) were observed between SytI KD (+) and control neurons (−). Cleavage of SNAP-25 by BoNT/A generates a smaller fragment, indicated by an asterisk. (E) Control neurons, SytI KD neurons and KD rescue neurons with the expression of KD resistant SytI were exposed to BoNT/B (20 nM, 5 minutes exposure, 24 hours incubation). Cleavage of Syb was detected through immunoblot analysis. Entry of BoNT/B was blocked in SytI KD neurons and was restored by KD resistant SytI, as evidenced by Syb cleavage. (F) SytI KD (+) and control neurons (−) were exposed to indicated concentrations of BoNT/C (5 minutes exposure, 24 hours incubation). Cleavage of SNAP-25 by BoNT/C is similar at all toxin concentrations between control and SytI KD neurons, indicating that SytI is not required for BoNT/C entry into neurons. (G) Neurons infected with (+) or without (−) lentiviruses expressing SytI KD shRNA were exposed to BoNT/C (3 nM) and BoNT/D-C (0.3 nM) simultaneously for 5 minutes. Cells were washed and further incubated in toxin-free media for 24 hours. SytI KD blocked entry of BoNT/D-C, but did not affect BoNT/C entry. (H) Rescue experiments were carried out for SytI KD neurons, using either KD-resistant SytI or SytII. Neurons were exposed to BoNT/D-C (0.3 nM) and BoNT/E (0.3 nM) simultaneously for 5 minutes. Cells were washed and incubated further for 6 hours. Entry of BoNT/D-C into SytI KD neurons was restored by KD resistant SytI and SytII. Cleavage of SNAP-25 by BoNT/E, which generates a smaller fragment (indicated by an asterisk) is not affected by the expression levels of SytI/II, and serves as an internal control.

Fig. 4.

The luminal domain of SytII can inhibit BoNT/D-C entry into neurons in vitro and reduce the toxicity of BoNT/D-C in vivo. (A) BoNT/D-C (0.3 nM) was pre-incubated for 30 minutes with indicated concentrations (µM) of soluble GST-tagged SytII 1–87 or soluble GST protein, with (−) or without (+) gangliosides, in high K⁺ buffers. Neurons were then exposed to the mixtures for 5 minutes, washed and incubated further in toxin-free media for 6 hours. SytII 1–87 inhibited entry of BoNT/D-C into neurons in a dose-dependent manner. Gangliosides enhanced the ability of SytII 1–87 to inhibit BoNT/D-C entry. (B) The same amount of BoNT/D-C was pre-mixed with either soluble GST or GST–SytII 1–87 proteins, with or without gangliosides for 30 minutes at 4°C. The mixtures were injected into mice and time-to-death of each mouse is listed. The effective toxicity of the mixture in vivo is estimated from an established standard curve reflecting the linear relationship between the time-to-death and the log value of the toxicity (in LD₅₀/ml units) (Boroff and Fleck, 1966; Dong et al., 2003).

Fig. 5.

BoNT/D-C shares the same binding site on SytII with BoNT/B. (A) Upper panel: sequence alignment of the membrane adjacent regions and (partial) transmembrane domains (TMD) of SytI (rat) and SytII (mouse). The BoNT/B-binding site is labeled. Pull-down assays were carried out as described in Fig. 2B, using either immobilized SytII 1–61 and 1–87 (middle panel), or the SytII luminal domain truncations indicated (lower panel). To facilitate protein purification, all SytII luminal domain truncations end at residue 267. (B) Immobilized SytII 1–87 was used to pull down BoNT/B (100 nM), in the presence of indicated concentrations (µM) of D-C/H_C or C/H_C. (C) Neurons were exposed to BoNT/B (20 nM) and BoNT/A (20 nM) simultaneously, with the presence of 3 µM D-C/H_C or C/H_C for 5 minutes. Control cells were exposed to BoNT/B and A without H_Cs. Cells were washed and fixed. Binding and entry of BoNT/B and BoNT/A into nerve terminals were detected using their specific antibodies through immunostaining. Syb was detected as a marker for nerve terminals.

Fig. 6.

Defining the binding interface between BoNT/D-C and SytI/II through targeted mutagenesis. (A) Pull-down assays were carried out as described in Fig. 2B, using the indicated SytII point mutations to pull down D-C/H_C or BoNT/B. (B) The following SytI/II mutants were expressed in SytI KD neurons through lentiviral transduction: SytI mutation (SytI mut.: F46A, M47A, E49K), SytI deletion (SytI del.: Δ39–49 of rat SytI), SytII mutation (F54A of mouse SytII), and SytII deletion (Δ63–65 of mouse SytII). SytI mutants were based on a KD-resistant SytI sequence. Neurons were then exposed to either BoNT/D-C (0.3 nM) or BoNT/B (20 nM) for 5 minutes. Cells were washed and further incubated for either 6 hours (BoNT/D-C, upper panel) or 24 hours (BoNT/B, lower panel). Cell lysates were analyzed by immunoblot. (C) Upper panel: superimposed structures of D-C/H_CC (orange) versus the BoNT/B (blue)–SytII (light blue helix) complex. Lower panel: comparison of D-C/H_CC (orange) versus C/H_CC (gray). (D) Targeted mutagenesis sites in D-C/H_CC are marked as sticks (magenta, loss of Syt binding; black, no binding difference, see panel E). The black ellipse highlights the region corresponding to where SytI/II bind BoNT/B. The magenta ellipse indicates the proposed SytI/II binding site in BoNT/D-C. (E) Wild type (WT) and indicated D-C/H_C mutants were purified as GST fusion proteins and incubated with detergent extracts from rat brain. Bound materials were subjected to immunoblot analysis detecting SytI and SytII.

Fig. 7.

Human SytII is not an effective receptor for BoNT/D-C, B and G. (A) Upper panel: human SytII differs from mouse SytII by one residue within the toxin-binding site (residue 54 in mouse SytII, 51 in human SytII). Lower panel: immobilized mouse SytII 1–87 and a mutant SytII 1–87 mimicking human SytII (F54L) were used to pull down D-C/H_C, BoNT/B or G/H_C. (B) Full-length mouse SytII WT and SytII (F54L) were expressed in SytI KD neurons. Binding and entry of BoNT/B into neurons was detected through immunostaining as described in Fig. 5C. (C) Full-length mouse SytII WT and SytII (F54L) were expressed in SytI KD neurons. Neurons were exposed to BoNT/D-C (0.3 nM, 5 minutes exposure, 6 hours incubation), BoNT/B (20 nM, 5 minutes exposure, 24 hours incubation) or BoNT/G (40 nM, 5 minutes exposure, 24 hours incubation). SytII (F54L) is less efficient than WT SytII in mediating entry of these three toxins. (D) Rat SytI 1-83 and human SytI 1–80 were purified as GST fusion proteins and used to pull down soluble D-C/H_C, BoNT/B and G/H_C. For all three toxins, human SytI mediated similar levels of toxin binding to rat SytI.