November 24, 2003
Volume 81, Number 47
CENEAR 81 47 p. 11
ISSN 0009-2347


Approach yields amazingly accurate, nativelike structure with novel shape


A research team imagined a large protein having a shape that's never been seen before, tailored it computationally, synthesized it, determined its structure, and found it to correspond extremely closely with the original conception [Science, 302, 1364 (2003)]. The work--by David Baker, Howard Hughes Medical Institute investigator and associate professor of biochemistry at the University of Washington, Seattle, and coworkers--opens up the possibility of custom-designing potentially useful proteins that nature hasn't gotten around to devising yet.

TWINS Computationally designed model of Top7 (blue) is an extraordinarily close match with solved X-ray crystal structure of the
synthesized protein (red).

The new protein, called Top7, is the first globular protein with a new fold that was designed from scratch and validated by determination of its 2.5-Å-resolution X-ray crystal structure. Globular proteins are roughly spherical structures with multiple long-range interactions--contacts between amino acids far removed from each other in the protein's sequence.

Previous attempts to design globular proteins have been based on folds already seen in nature. In contrast, Top7's 93 amino acid residues, two -helices, and five -strands adopt a new shape that has not been observed previously.

Baker and coworkers designed and synthesized Top7 by using protein sequence-design algorithms and the group's structure-prediction program, Rosetta (C&EN, Aug. 4, page 26). "Starting with an initial backbone conformation," Baker says, "a low-energy sequence was found using the design algorithms. Keeping this sequence fixed, a low-energy structure was then optimized using Rosetta." Ten iterations of this process yielded the Top7 sequence and structure. Baker believes such an approach will now prove applicable to the design and creation of many other new types of proteins.

"This is groundbreaking work--an enormous step forward," comments chemistry professor Michael H. Hecht of Princeton University, who identifies de novo proteins a different way--by plucking them out of combinatorial libraries (C&EN, Nov. 10, page 14). "The ability to design novel protein folds is clearly a major goal in the field of protein design, but I expect many scientists will be surprised it has succeeded so soon. Moreover, it is surprising and impressive that the experimentally determined structure agrees so precisely with the structure specified by the design."

Biophysicists Chao Tang and Ned S. Wingreen of NEC Laboratories America, in Princeton, N.J., who specialize in protein folding and design, call the work a milestone. "This is the first demonstration that a fold significantly different from known folds can be designed," they note. "The iterative method developed in this study is an impressive technical achievement. It has been demonstrated beautifully in this example and will have broad applications."

However, the researchers add that "one should not conclude that any arbitrary folds are designable, as the paper implies. Top7 differs from natural folds only by the topological connections of the secondary-structural elements--the order in which the helices and strands are linked to each other. One could argue that the work demonstrates a valuable iterative method for the design of a new fold but may also rely on the ingenious ability of Baker and coworkers to have selected a 'good' fold for design."

Biochemistry professor Anna Tramontano of the University of Rome "La Sapienza," who specializes in protein structure prediction and analysis, says she would ideally like to have seen one thing that was missing in the Baker paper: some element of functionality in the protein, such as the ability to bind a metal or catalyze a reaction. However, she notes that de novo-designed proteins created this way could indeed ultimately turn out to have interesting functions and that the study "is quite an impressive achievement."

Rob B. Russell of the European Molecular Biology Laboratory, Heidelberg, Germany, who specializes in protein structure studies, agrees that "the next step will probably be to engineer a designed protein with a specific function. Anyway, what can I say--Baker is impressive. He's doing things that people only dreamed about a few years ago."


TOP7 De novo-designed protein has 93 amino acid residues, two -helices, and five -strands.



Chemical & Engineering News
Copyright © 2003 American Chemical Society