A pressing need exists for long-acting, non-addictive medicines to treat chronic pain, a major societal burden. Botulinum neurotoxin type A (BoNT/A) complex – a potent, specific and prolonged inhibitor of neuro-exocytosis – gives some relief in several pain disorders, but not for all patients. Our study objective was to modify BoNT/A to overcome its inability to block transmitter release elicited by high [Ca2+]i and increase its limited analgesic effects. This was achieved by fusing a BoNT/A gene to that for the light chain (LC) of type/E. The resultant purified protein, LC/E-BoNT/A, entered cultured sensory neurons and, unlike BoNT/A, inhibited release of calcitonin gene-related peptide evoked by capsaicin. Western blotting revealed that this improvement could be due to a more extensive truncation by LC/E of synaptosomal-associated protein of Mr = 25 k, essential for neuro-exocytosis. When tested in a rat spared nerve injury (SNI) model, a single intra-plantar (IPL) injection of LC/E-BoNT/A alleviated for ˜2 weeks mechanical and cold hyper-sensitivities, in a dose-dependent manner. The highest non-paralytic dose (75 U/Kg, IPL) proved significantly more efficacious than BoNT/A (15 U/Kg, IPL) or repeated systemic pregabalin (10 mg/Kg, intraperitoneal), a clinically-used pain modulator. Effects of repeated or delayed injections of this fusion protein highlighted its analgesic potential. Attenuation of mechanical hyperalgesia was extended by a second administration when the effect of the first had diminished. When injected 5 weeks after injury, LC/E-BoNT/A also reversed fully-established mechanical and cold hyper-sensitivity. Thus, combining advantageous features of BoNT/E and/A yields an efficacious, locally-applied and long-acting anti-hyperalgesic.