 
Revised
model of protein-drug interactions could make job of drug designers
a little easier
14
Feb 2000
By
Robert Sanders, Public Affairs
BERKELEY--
While most people may have difficulty distinguishing a person
from his or her mirror image, proteins in cells have no such
problem. They are exquisitely selective, able to latch on tightly
to one molecule but reject its mirror image.
Now
scientists claim that a 50-year-old theory explaining how proteins
and enzymes discriminate so precisely, considered gospel by
most researchers, needs revision.
The
finding could significantly help drug designers, who may needlessly
be discarding good drug candidates because of this misconception.
"The
reigning theory of how enzymes distinguish between very slight
differences in molecules is the three-point attachment model
found in essentially all biochemistry textbooks," said
Daniel E. Koshland Jr., professor of molecular and cell biology
at the University of California, Berkeley, and a researcher
in the Center for Advanced Materials at Lawrence Berkeley National
Laboratory.
"This
is very important in pharmacology, where drug designers rely
on the theory to design a particular mirror image, or enantiomer,
to fit exactly into the active site of an enzyme or receptor.
Well, the classic explanation needs correction."
In
a paper in the Feb. 10 issue of Nature, Koshland and Andrew
D. Mesecar, a former postdoctoral fellow at UC Berkeley and
now an assistant professor in the Department of Medicinal Chemistry
and Pharmacognosy and the Center for Pharmaceutical Biotechnology
at the University of Illinois, Chicago, argue that the so-called
Ogsdon three-point attachment model must be replaced by a new
four-point location model.
These
models explain how proteins bind to "chiral" molecules,
that is, molecules that cannot be superimposed on their mirror
image. Typically the mirror-image versions of chiral molecules
act very differently in the body. Some bacteria can degrade
one version of a pollutant but not its mirror image; a key receptor
in the brain is turned on by an amino acid but not its mirror
image.
The
reigning assumption is that the business end of a molecule homes
in on the "active site" of a protein - an enzyme,
for example, or a receptor sitting on the surface of a cell
- and makes a three-point landing. That is, three groups arranged
around a tetrahedral carbon atom at the business end interact
with three locations in the active site, usually with such precision
that no other alignment would work.
In
fact, Koshland and Mesecar say, a three-point landing is not
sufficient to allow a protein to discriminate between mirror
images of the incoming molecule. While one molecule could fit
snugly in the active site from the front, for example, the mirror
image might fit just as well coming in from the rear.
Some
other attribute of the active site determines whether it specifically
binds one or the other mirror image. Sometimes it is a constraint,
such as an obstacle preventing binding from one direction. Often
it is a fourth interaction in the active site, for example with
a metal ion, that turns the encounter into a four-point landing.
"We
usually thought there was only one binding mode at an active
site of a protein," said Mesecar. "This shows you
can have at least two. Selection is dependent on the active
site environment of the protein."
This
new way of looking at protein interactions would significantly
affect drug designers, who, Koshland says, may needlessly be
discarding good drug candidates because of this misunderstanding.
Juggling
three-dimensional computer models of enzymes or receptors, drug
designers typically engineer small molecules to fit just so
into a particular active site, with such precision that even
the mirror image wouldn't fit. The goal is to switch a receptor
on or off or disable an enzyme in hopes of stopping some disease
process.
However,
assuming only one of two mirror images of a drug will bind strongly,
when in fact both may bind to the active site, can lead to confusing
assays and possible rejection of a promising drug candidate.
"What
we are saying is, you never really know whether one or both
mirror images are capable of binding in the active site,"
Mesecar said. "People are going to have to revise their
thinking. This may change the way they interpret their binding
data for enantiomers."
"To
really do it right with designing drugs, you must think in terms
of four points, not three," Koshland said.
In
their paper they detail one specific example of how this works.
The enzyme isocitrate dehydrogenase has the sole purpose of
grabbing hold of isocitrate and breaking it down as one link
in the citric acid cycle, or Krebs cycle, that generates energy
for the cell.
Until
now, it was thought that only one form of isocitrate (the D
isomer) was able to make a three-point landing in the active
site of IDH and hold on tightly.
Koshland
and Mesecar found that in the absence of magnesium ions, the
enzyme latches on preferentially to the other, inactive form
of isocitrate (the L isomer). Only when magnesium is present
does it bind the active isomer, and go on to generate the energy
of the cell.
In
fact, when the enzyme binds to L-isocitrate the cycle stops
dead in its tracks and generates no energy. This implies that
the interaction between the enzyme and D-isocitrate evolved
in the presence of magnesium ions.
"This
work shows how important metals are in molecular discrimination
between mirror-image molecules," Mesecar said, and is further
evidence that metals like magnesium, zinc, chromium and copper
- always known to be critical in the diet, if in minuscule amounts
- play a major role in the molecular machinery of the cell.
Koshland
notes that other enzymes have been shown recently to bind equally
well to both mirror image molecules. In some cases, one enantiomer
is the actual substrate - the molecule the enzyme is designed
to react with while the other acts as an inhibitor.
The
work was funded by the National Science Foundation.
###
|
