Biochemists have detected a subtle new gyration that protein molecules undergo in the intricate, squirming dance that influences their activity in the cell.
The researchers have also created a realistic geometrical model of the twisting "backrub" motion that could help scientists understand the basics of protein function and design proteins for medical use.
Also, they said, the backrub motion could have implications for understanding how proteins can accommodate locally to some mutations that occur during evolution, without altering their global structure or function.
Lead author on the paper was graduate student Ian Davis, and the other co-author was Bryan Arendall. The research was supported by the National Institutes of Health and a Howard Hughes Medical Institute predoctoral fellowship to Davis.
Proteins comprise strings of amino acids whose links form a "backbone." Each kind of amino acid sprouts a characteristic molecular "side chain," and together the backbone and side chains determine a protein's structure and function.
The Duke researchers suspected the presence of backrub motions for other reasons, but their reality could be conclusively shown only by studying proteins frozen in crystalline form for structural analysis by x-ray crystallography.
"We were pleased but surprised that these crystal structures at liquid nitrogen temperatures actually could show us something really interesting about dynamics," said Jane Richardson.
In the highest resolution such crystal structures, in which individual atoms are directly visible, it is quite common to see a side chain that "dances", or flips back and forth between two different conformations.
The researchers traced the consequences of this motion back into the backbone and deduced that the local backbone structure must twist slightly in a particular way to accommodate the larger side-chain movement.
According to Richardson, this motion, which they dubbed a "backrub", is a subtle, concerted shift of the two backbone units on either side of the dancing side chain.
"Nobody has described this particular kind of motion before," said Richardson. "And that's because it's down in the noise, in terms of what the backbone is doing."
"You have to get ultra-high-resolution, clean maps that really show you exactly where the side chain atoms are. And then you can work backwards to figure out what the backbone must have done." The researchers created a geometrical "backrub" software tool to model this motion.
"Investigators had theorized that the backbone moved, but it has been rather difficult to prove what's really going on," said Richardson.
"There have been other previous models, but these were not as successful as one would like, presumably because they were not based on the kind of empirical data that we've now developed."
To understand how frequently the backrub motion occurred in proteins, Davis undertook an analysis of crystallographic data on 19 proteins.
"We took nineteen of the highest-resolution structures available - the best data we could get our hands on," said Davis.
"We then went through each of those structures one residue at a time, looking for evidence of some sort of motion that extended beyond just a side chain spinning."
Davis distinguished backrub motions by detecting movement of the first side chain atom attached to the backbone - a shift that could only occur if the backbone had moved, said Davis.
In analyzing some 4,000 amino acid units, Davis found backrub motion for 3 percent of the total and 75 percent of all the local backbone shifts.
"We were expecting to see some examples of this motion, but we were very surprised that it was so dominant over any other very local backbone motions," he said.
The fact that the backbone motions are common in proteins frozen in crystals suggests that they are even more prevalent in proteins in the liquid environment of the cell, said Richardson.
"The backrub motions in these frozen crystals have to be a subset of what goes on in solution or in the body," she said. "It's got to be more common when you have more flexibility."
Such ubiquity means that the backrub model developed by the researchers could have useful applications, said Richardson.
"The backrub model can be used by people doing homology modeling, in which they are starting with a known protein structure and trying to model the structure of a protein with a related but different sequence," she said.
"One knows that the backbone will shift in such cases, but previous methods of modeling those shifts have usually produced results farther from reality rather than closer."
"We think our backrub shifts can help predict how, either in natural evolution or in protein engineering, the local structure would accommodate the substitution of an amino acid with a different shape or size of side chain," said Richardson.