Amyloid fibrils, rope-like structures formed by linked protein molecules, are the common feature of these diseases and may well hold important clues to these diseases, said David Eisenberg, director of the UCLA-DOE Institute of Genomics and Proteomics, a Howard Hughes Medical Institute investigator, and a member of the research team.
Eisenberg and his colleagues report in Nature the three-dimensional structure of a small piece of a fibril-forming protein from yeast that behaves similarly to proteins involved in Alzheimer's and these other diseases. Knowledge of the structure of this small peptide -- known by the code of its amino acids, GNNQQNY -- reveals a surprising "molecular zipper" that Eisenberg described as "pathologically dry."
"Proteins live in water, but here all the water is squeezed out as the fibril is sealed and zipped up," Eisenberg said. "Our hypothesis is that this dry steric zipper forms in all of these diseases, and is universal in the fibrils. Once this steric zipper has formed, it's very difficult to reverse because it's so tight."
"Knowing the structure may provide a rational basis for developing drugs to fight these diseases," said Melinda Balbirnie, a UCLA postdoctoral scholar and a member of the research team.
Can scientists prevent the steric zipper from forming in the first place, or pry it open once it has formed?
Balbirnie is able to produce fibrils from the small peptide, and has developed a test, called an assay, to determine whether the fibrils break up.
"Her strategy is to add to this assay a wide variety of chemical compounds to see whether any will break up the fibrils," Eisenberg said. Balbirnie said she is "hopeful" her strategy will succeed in breaking up the fibrils.
Eisenberg and his colleagues also are investigating whether disease-forming proteins have similar structures. Their hypothesis is that Alzheimer's and other fatal "amyloid fibril" diseases have proteins containing the steric zipper.
"Our Nature paper presents the first atomic-level look at any of these structures," said Rebecca Nelson, a UCLA graduate student in biochemistry and molecular biology, and member of the team that determined the precise positions of all the atoms in the peptide.
The UCLA chemists and molecular biologists had difficulty analyzing tiny crystals from the small peptide using standards methods of X-ray crystallography. Nelson and coworker Robert Grothe, formerly of the Howard Hughes Medical Institute, worked indefatigably from 2000 to 2004 trying to develop new methods that would work.
"We wanted to learn which atomic-level interactions were giving the peptide the property to form fibrils of the type which the body cannot break down," Nelson said. "We tried many techniques with promising technologies that didn't work, but we never got discouraged. We thought if we could better understand the structure of the molecules inside the fibrils, we would understand more about why they have the properties they do, how they form, why they might be involved in disease and conceivably how to get rid of them or even prevent their formation."
A key breakthrough occurred when the UCLA team began working with a distinguished scientist in Grenoble, France, Christian Riekel, who conducts X-ray microcrystallography with an instrument designed to analyze very small crystals.
"We sent some of our crystals to Christian Riekel and his student, Anders Madsen, and we worked closely with them," said Michael Sawaya, a research scientist with UCLA and the Howard Hughes Medical Institute, and a member of the team. "Christian invented ways to get a fine beam of X-rays to bombard tiny crystals. He and Anders were able to collect diffraction data that allowed us to determine the structure of the peptide, as well as a second, related peptide that also contains the steric zipper."
"So many times I thought we were close," Nelson said, "but it didn't work until we tried this approach. When we solved the structure, I started dancing in the lab."
Nelson describes the proteins associated with Alzheimer's and other amyloid fibril diseases as "transformer" proteins that instead of doing their normal work, start forming pathological fibril structures.
"Like a transformer toy -- a car that changes its shape and turns into a robot -- the protein changes its shape, going from its normal function to a diseased state," Nelson said.
"Other proteins just do their jobs," Eisenberg said, "but these transformer proteins are different, and exceedingly strange. We believe we are now coming to grips with these proteins."
The researchers discovered that their measurements from the fibrils can all be characterized by what they describe as a "cross-beta diffraction pattern," Sawaya said. "They diffract in such a way that tells us there are many extended protein chains stacked like a spine or the rungs of a latter," he said. "That pattern is a common feature in these amyloid diseases."
Summarizing the connections, Eisenberg said, "All of these diseases have fibrils as their common feature; all of these fibrils have the same characteristic X-ray diffraction pattern, which is called cross-beta; our fibrils also have the cross-beta diffraction pattern in a small section of the protein that we call the spine. Because all of these diseased fibrils have a spine with the same diffraction pattern, and because diffraction patterns are characteristic of the arrangement of atoms, our hypothesis is that the two-dozen other diseases will each have a similar arrangement of atoms.
"Our hypothesis is that in all these diseases, a water-tight steric zipper has formed in the fibrils," Eisenberg said. "We have seen the teeth of the zipper in two related peptides."
The research was funded by the National Institutes of Health, the National Science Foundation and the Howard Hughes Medical Institute.
Balbirnie made the discovery that a small fragment of a protein -- a mere 1 percent of the protein -- can behave similarly to the entire protein, and is able to form fibrils. She and her colleagues reported this surprising finding in the journal Proceedings of the National Academy of Sciences in 2001.
"Like a detective, Melinda traced this fibril-forming property down to a little peptide," Eisenberg said. "Nobody expected that 1 percent of the protein could have the essence of the whole protein and could form fibrils on its own; I certainly didn't expect that. There were only seven amino acids in that fragment. Rebecca later found the peptide could be cut to only four amino acids and form fibrils."
"No one has known the details of these structures before, which we can now see," Balbirnie said. "The fibrils are stable in all of these diseases; we can account for that stability, which suggests this may be a common feature. We are learning how these biological machines work."