The problem is that the most potent material—brain extract—is a mess. A misfolded version of the endogenous PrP protein emerged decades ago as the putative infectious agent. But in infected brains it is clumped in an insoluble amyloid form that is difficult to purify and impossible to crystallize. Thus, the exact nature of the most infectious form is a mystery. “The only measurable thing we have right now is the amount of β-sheet,” says Legname. “We do not know the actual structure.”
His approach was to start with the recombinant version of a truncated form, known to be predisposed to forming amyloid, and select conditions that favored β-sheet content. This preparation was injected into the brains of mice overexpressing a similarly truncated protein. Between 380 and 660 d later, the mice showed disease symptoms, which could be serially transferred to other mice.
Some researchers believe that the test system is poised to become infectious, and the injection may be just one of many neurotoxic stresses that would eventually get it going (Couzin, 2004). Legname does acknowledge that “you may well need accessory factors in vivo”—factors that would only be transferred from brain extract not recombinant material. And even the most infectious conformation may be lost once the protein is injected into a host.
But Legname reminds the doubters that “yeast prion systems have established that the protein-only hypothesis is right,” although of course they lack the brain pathology of mammals. He remains confident that the “hundreds of different conditions” that the team is testing to increase infectivity will bring down the presymptomatic time and eventually allow disease initiation by recombinant protein in wild-type mice. ▪