On the significance of the C-terminal primary amide in cholecystokinin