Identification of Fibroblast Growth Factor Receptor 3 (FGFR3) as a Protein Receptor for Botulinum Neurotoxin Serotype A (BoNT/A)
(A) Schematic presentation of the eight deletion mutant FGFR3 peptides. FGFR3b Loop 1,2,3 and FGFR3c Loop 1,2,3 (long and short version); FGFR3b Loop 2,3 and FGFR3c Loop 2,3; and FGFR3b Loop 3 and FGFR3c Loop 3. AB: Acid Box; HBS: Heparin Binding Site. The purple area highlights the region where subtype b and c differ in sequence (Figure modified from  (See also Figure S3)). (B–C) Pre-incubation of BoNT/A with eight deletion mutant FGFR3 peptides, corresponding to the extra-cellular loops of FGFR3b/c (A) inhibit BoNT/A uptake in Neuro-2a cells. BoNT/A at 1 nM was incubated for 20 min with increasing concentrations of individual FGFR3 peptides before treatment of cells. In parallel, BoNT/A was pre-incubated with antibodies to HC/A (positive control), or Syt II1–20 (negative control). Data are shown as percentage BoNT/A uptake relative to the uptake after pre-incubation with the negative control Syt II1–20. SNAP25 cleavage, as a measure of BoNT/A uptake, was decreased when BoNT/A was pre-incubated with Anti-HC/A or with any of the eight deletion mutant FGFR3 peptides. The averages of three or more experiments were included. The data was fitted to a non-linear exponential decay model; Y = 100*e- IC*log(concentration). As a measure for inhibition, the Inhibition Constant (IC) and standard error (SE) for each inhibitor are shown. The IC value for Syt II1–20 was significantly lower than all the other IC values, p≤0.0096 and p≤0.0003 respectively. The IC value for Anti-HC/A was significantly higher than all the other IC values, p≤0.0125 and p≤0.0068 respectively. The IC value for FGFR3b Loop 2,3 was significantly higher than the IC values for the other FGFR3b peptides, p-value≤0.0489. There was no significant difference in IC value among the FGFR3c peptides. (D) Results from testing the eight FGFR3 deletion mutant peptides in a BIAcore SPR binding assay. Increasing concentrations, 0–4000 nM of each of the deletion mutant FGFR3 peptides were flowed across a CM5 sensor chip covered with rHC/A. The curves were fitted to a 1∶1 kinetic-binding model (A + B ↔ AB) with the BIAevaluation 3.0 software. The table shows the calculated KD's ± SEM of all eight FGFR3 deletion mutant peptides upon binding to rHC/A. The average ka's (association constants (1/Ms)) and kd's (dissociation constants (1/s)) are shown in figure S3B. The KD values for the FGFR3b Loop 1,2,3 peptides are significantly lower than the KD values for the FGFR3c Loop 1,2,3 peptides, P-value≤0.0005 and the KD values for FGFR3b/c Loop 3 and FGFR3c Loop 2,3 are significantly higher than the KD values for the longer FGFR3b/c peptides, P-value≤0.0024. Loop 2,3 of FGFR3 can therefore be identified as the minimal optimal binding region of FGFR3 (See also Figure S3). (E) Comparison of FGFR3b loop 2,3 and FGFR3b Loop 3 association (or on-rate) after normalization. The 1000 nM curves of three different runs were normalized using BIAcore evaluation software version 4.1. The on-rate for FGFR3b loop 2,3 was faster than the on-rate for FGFR3b Loop 3. The average ka (1/Ms), kd (1/s), and KD (KD = kd/ka (M)) are shown. ka for FGFR3b Loop 2,3 is ∼10-fold higher than ka for FGFR3b Loop 3 (See also Figure S3).