Riddle of the folds
“Folded” 3 D proteins may hold the secret of how
human DNA works
 “Fold assignment”
may sound like a domestic chore (“You fold the sheets, I’ll get the towels”),
but to molecular biologist David Eisenberg, it signifies a frontier of
molecular medicine. It is a tool by which scientists can speed up the painstakingly
slow process of learning what all 80,000 proteins in the human body look
like -- a step toward developing knowledge that can be used to devise treatments
for the most stubborn diseases.
Eisenberg, director of the UCLA DOE Laboratory of Structural Biology
and Molecular Medicine at the UCLA Molecular Biology Institute, explains
that it’s not enough to know the DNA sequence — the genetic blueprint —
of a protein. That task, which has been taken up with great success by
the nationwide Human Genome Project, answers only part of the riddle of
how proteins, the engines that run our body’s cells, work.
“The Human Genome Project is fundamental research of titanic proportions,”
says Eisenberg. “But the next necessary stage is learning the three dimensional
structure of proteins.”
That’s where folds come in. The Genome Project, which should complete
its task of unraveling human DNA by the year 2002, provides a linear picture
of the amino acid molecules that form proteins. In order to function, however,
proteins must “fold” into complicated 3D structures. “Here, it’s just a
string of letters,” says Eisenberg, pointing to the alphabet soup that
makes up the genetic code. “Here,” he points to a 3D model, “we show how
it works.”
Molecular biologists know how to make proteins fold: Just put them in
water. (They also know how to make them unfold.) And they have discovered
how to photograph protein’s structure — through X ray crystallography,
which Eisenberg has been doing at UCLA since he came here in 1969. Using
several $.5 million machines in a lab generously funded by the Department
of Energy and the National Institutes of Health, he and his colleagues
shine X rays from a multitude of angles at a tiny (about 1/100th of an
inch in diameter) protein crystal. The crystal diffracts the rays and leaves
black spots of various densities on electronic film, which is read by a
laser and downloaded into a computer for analysis.
The problem is, with proteins — each of which typically contains anywhere
from 10,000 to 100,000 atoms — the process of determining what all those
spots mean can take anywhere from months to years. Postdoctoral fellow
Lesa Beamer, for example, recently spent three years determining the structure
of the protein BPI. In Eisenberg’s research group, about 30 proteins have
been determined, and throughout the country scientists are completing the
picture at the rate of 1,000 to 2,000 proteins a year. With 80,000 human
proteins in total to account for, this effort is running far behind the
Genome Project.
Enter fold assignment, which Eisenberg hopes will substantially speed
the process. In 1991, Eisenberg, postdoctoral fellow (now assistant professor)
Jim Bowie and postdoctoral fellow Roland Luthy developed a method for predicting
protein folds by computer. Given an amino acid sequence, the computer checks
its resemblance to 1,000 distinctive already known folds, and in one quarter
of the cases is able to make a match. UCLA has approved making Eisenberg’s
fold-recognition software available for non-commercial use on the lab’s
World Wide Web page at http://www.doe-mbi.ucla.edu/. Scientists at other
labs need only submit a sequence of amino acids, and the software will
identify as many folds as possible.
In scientific publications, schematic drawings of the folded 3D proteins,
representing only the connective backbone of the protein molecule, look
like colorful strands of confetti. To Eisenberg, however, these aren’t
just pretty models, but guideposts to medical breakthroughs. Consider,
for example, his model of diphtheria toxin (DT), a protein secreted by
the diphtheria bacterium. The DT kills by tricking the protein receptors
of healthy cells into treating it as one of their own. Eisenberg’s model
is currently being used by Ralf Landgraf to develop what he hopes will
be a “magic bullet” against breast cancer: If he can replace the receptor-binding
port of DT with a protein port that binds only to breast cancer cells,
maybe the diphtheria toxin will kill the cancer cells.
“The nice thing about UCLA,” says Eisenberg, “is that the Molecular
Biology Institute is right between the departments of chemistry and biochemistry
and the medical school. So we can easily talk to people about what they
need to know. Then we choose proteins we think will give some medical or
biochemical answer.
“The classical way of finding drugs has been screening them in trials,”
he concludes. “The newer way is to look at the protein. The two methods
together are very powerful.”
— Michele Kort
|