Professor Roderick MacKinnon set up his
Laboratory of Molecular Neurobiology and Biophysics at Rockefeller
with a specific goal in mind—to obtain a high-resolution three-dimensional
structure of a potassium ion channel protein. He fully expected
that his quest for the structure would be a long-term undertaking,
requiring at least a few years. But less than two years after his
arrival here, MacKinnon has already accomplished his mission, as
evidenced by the image of the X-ray crystallographic structure of
an intact potassium channel protein on the cover of the April 3,
1998 issue of Science.
MacKinnon, who was appointed an investigator with the Howard Hughes
Medical Institute last year, modestly attributes his success to
"some very lucky breaks and extremely hard work on the part of my
Outside experts, however, are far more forthcoming with their praise
for this achievement."A remarkable accomplishment," proclaimed Clay
Armstrong of the University of Pennsylvania, who reviewed MacKinnon’s
paper in the same issue ofScience. "It is a dream come true
Ever since their discovery in the 1950s, ion channels have been
the subject of intense interest to many scientists because of the
key role they play in maintaining some of the body’s most vital
functions. Seated in the oily layers of the cell membrane that preserves
the integrity of each cell, these proteins govern the flow of different
ions such as potassium, sodium and calcium into and out of cells.
The proper balance of these ions is essential for fundamental operations
such as the transmission of nerve impulses throughout the body and
Shaped like tiny doughnuts floating in oil, the ion channels perform
the dual functions of gateway and gatekeeper. The holes in the doughnut
form the gateway through which the ions flow. However, these holes,
or pores, are endowed with special properties that enable different
channel proteins to be selective as to which ions they allow passage.
Like security guards who check ID before allowing entrance, these
molecular gatekeepers only allow specific ions to pass through their
What allows the ion channel proteins to work as they do? This is
the question that has engaged MacKinnon for over a decade, beginning
with his observation of an experiment monitoring the electrical
activity of a potassium channel, when he was a medical student just
contemplating a switch to basic science. He joined the laboratory
of his undergraduate mentor Christopher Miller at Brandeis University,
where, as a postdoc, he began to work on biophysical aspects of
channel function. He chose to focus specifically on the protein
selective for potassium ions, because it happened to be the target
of Miller’s lab and also because, at the time, it was the least-studied
ion channel, says MacKinnon.
Beginning with electrophysiological and biochemical approaches,
MacKinnon studied the interaction of the potassium channel with
a specific toxin derived from scorpion venom and deduced that the
toxin inhibited the flow of ions by sitting directly on the pore
of the channel. This led to questions about which specific regions
of the channel proteins bind to ions and the toxins. Investigations
along those lines became possible after 1987, when scientists cloned
the genes for potassium channels in fruit flies.
In the wake of this development, MacKinnon turned to the field
of molecular biology. By systematically mutating the channel gene
at specific locations and observing the effects on the channel’s
ability to bind to the specific ions, he was able to pinpoint ion-binding
capacity to a single region of the protein.
"All potassium channel proteins contain a little signature stretch
of about 8 to 10 amino acids, which is specific for the ion," explains
Mackinnon. "The sequence is conserved across evolution from bacteria
to humans. In fact, some of these proteins have no features in common
except for this signature. It’s as if biology chose only one way
to select for potassium ions."
Meanwhile, genetic evidence suggested that a single potassium channel
consisted of multiple protein subunits. MacKinnon determined that
a functional channel has four identical subunits that join together
like the staves of a barrel around the central hole. Each of the
subunits contains the ion-specific signature sequence, "which forms
a loop extending into the center of the hole to create a pore selective
for potassium ions," he explains.
The pore-forming loops seemed to govern specific functions such
as ion selectivity, but without a clear idea of the 3-D structure
of the channel, MacKinnon says there was no way to test the truth
of this idea. So he came to Rockefeller where he hoped to master
crystallographic techniques and solve the structure of the potassium
The very nature of the problem led MacKinnon and others to anticipate
slow progress in their work. A basic requirement for protein crystallography
is that scientists have sufficient quantities of the starting material
or protein, in order to grow crystals for X-ray analysis. But although
potassium channels are present in virtually all cells, they are
usually produced in low quantities.
Fortunately, this hurdle was overcome when a group of scientists
discovered and cloned a potassium channel from the bacterium Streptomyces
lividans, using the very signature sequence that MacKinnon had discovered
a few years earlier.
"The bacterial system cleared the path by giving us a way to express
channel proteins in much larger quantities than had been possible,"
Armed with sufficient amounts of channel protein, the scientists
could turn their attention to tackling their other problem, namely
growing good crystals. Historically, membrane-bound proteins like
ion channels have been notoriously bad candidates for protein crystallography,
because the detergents used to separate such proteins from lipids—a
necessary step in making good protein crystals—often destroy the
proteins as well.
"To give an idea of the labor involved in determining the ideal
parameters for growing good crystals, we concocted a crystal screen
with a total of 900 different conditions at two different temperatures
and with 10 detergents," says MacKinnon. "Luckily we hit on the
right combination relatively early."
The structure of the potassium channel (see figure) confirms MacKinnon’s
theory about the pore–loop structure determining ion selectivity
and also offers an explanation as to why the channel prefers potassium
over sodium ions. "Basically the potassium selectivity filter is
a cylindrical cavity lined by slightly charged oxygen atoms which,
in turn, are held in place by structural elements in the portion
of the channel protein that traverses the membrane," explains MacKinnon.
This structure confers a certain rigidity to the cylinder and determines
the diameter of the oxygen-lined pore. A charged potassium ion fits
neatly into the narrowest configuration of this cylinder. Sodium
ions, which are smaller, do not fit properly in the cavity. "The
oxygens in the pore cannot get as close to the sodium as they do
to potassium. Consequently sodium ions are better stabilized by
oxygens in water molecules and do not enter the pore," MacKinnon
The channel structure also suggests an explanation for the seeming
paradox of a channel protein’s ability to reconcile high ion selectivity
with high throughput, or the rapid flow of ions through the channel.
"The two properties seem incompatible because high selectivity
implies that the interaction between the channel and ion must be
a strong one, which would not be conducive to letting the ions escape
past the protein into the cell," MacKinnon says.
But it appears as though the selectivity pore resolves this paradox
by placing two ions near each other where they experience mutual
repulsion by virtue of their electric charge. "This suggests that
the strong attractive forces between the potassium ions and channel
are counteracted locally by the repulsive forces between two positively
charged potassium ions, which allows each ion to fall through at
a rapid rate," he adds.
Having solved the elusive structure of the channel, MacKinnon plans
to obtain information about channel structure at still higher resolutions
and probe other aspects of their function. Ultimately, he says,
"My aim is to figure out how channel proteins function as the electrical
impulse generators in biology."
This article originally appeared in the Spring 1998 issue
of Search, a Rockefeller University publication.