Gajdusek’s “odd, slow viruses,” as he called them, didn’t fit this description: Heat, formalin and disinfectants did not harm them. Even stranger, radiation did not diminish their virulence. But the agent seemed to have strains—some sheep with scrapie had a lot of trembling; others had an unrelenting itch. These facts added up to a very strange infectious agent: an indestructible strain-specific particle that didn’t die under conditions that would kill a virus and provoked no immune system response.

What was this particle? Gajdusek did not go as far as other researchers who posited that it was no virus at all. Experiments with centrifuges had turned up a lot of protein but no nucleic acid. Was it because there wasn’t any in the infectious agent? Could an infectious agent be entirely made of protein?

The conventional wisdom said no. Proteins are building blocks and engines and messengers of the body, but they are not alive. And to suggest, as some researchers began to, that a protein could make copies of itself in the victim’s body—that a nonliving agent was somehow replicating itself or being replicated in the victim’s body without DNA and causing disease—pushed up against the impossible.

During the late 1960s and 1970s, Gajdusek gradually committed his prestige to the protein-only point of view. In 1976 he was honored with a Nobel Prize in Physiology or Medicine for discovering “a completely new type of infectious agent.” Certainly he was being rewarded for doing the key experiments in the field (showing the agent was transmissible) and for the expert framing of a question (Was infection possible without a living agent?) rather than for answering it. All that could be said with certainty was that there was a group of diseases out there that were not like conventional ones.

Stanley Prusiner, a professor of neurology at the University of California at San Francisco (UCSF), entered what Gajdusek named the “slow virus” field in 1974, when it was stuck somewhere between theory and particle. Prusiner had trained as a chemist and thought the only way to be sure about the nature of the scrapie particle was to pull it out of the muck of the scrapie or kuru homogenate. He set to work and by the mid-1980s had separated out an agent 5,000 times more concentrated than in nature. Prusiner’s measurements indicated that it was small even for a protein and much smaller than the smallest known viruses. He mixed the agent with enzymes that dissolve proteins and with enzymes that dissolve nucleic acid, and found that chemicals that destroy proteins destroyed the particle’s infectivity, whereas chemicals that destroyed nucleic acids actually increased the infectivity, strongly hinting that the infection resided in a protein and not in a virus.

In 1982 Prusiner coined the word prion, an acronym (for “proteinaceous infectious particle”) that didn’t silence the skeptics. To do that, Prusiner had to synthesize a prion in a test tube and show it was sufficient to cause a prion disease. Then no one could say there was a virus hiding behind the protein. But prions had another surprise in store: Once Prusiner, working with other labs, purified the prion enough to determine part of its amino acid sequence, he found that prions were ordinary proteins manufactured by a healthy gene in the host’s own body. In other words, prions were not something that entered the victim—they were something the body itself produced.

Working over the next decade, Prusiner developed a theory to explain the apparent impossibility of an infection caused by the host itself. The body manufactures proteins as ribbons, which then fold into three-dimensional shapes that allow them to fulfill whatever functions they perform. Most proteins, because of the arrangement of their atoms, have only one such shape, but prions—Prusiner’s theory goes—have two: one normal, the second infectious and lethal. When a disease-causing prion is placed beside a normal prion in a test tube, the abnormal one can sometimes even convert the normal one to the former’s shape.

After one prion becomes altered, its new shape allows it to bond with adjoining healthy prions, and the bonding causes those prions, in turn, to fold wrongly and bond with others, making them fold wrongly too. The process is called conformational influence. Infectious prion disease, then, would result from the introduction of a misshapen prion from outside, and inherited prion disease from faulty manufacture by the genes. A third kind of prion disease, sporadic, would come from the chance misfolding of a healthy protein in the body.