Evidence that the A2 fragment of Shiga-like toxin type I is required for holotoxin integrity