Biophysics. Lonely voltage sensor seeks protons for permeation.

Full text

Page 1

28 APRIL 2006VOL 312 SCIENCEwww.sciencemag.org
534
ILLUSTRATION: K. SUTLIFF AND A. STONEBRAKER/SCIENCE
T
time when ion channels were known only by the
electrical properties they display in electrophysio-
logical experiments. Today, nearly all these ion
channels are also known as proper membrane pro-
teins, their sequences assigned to molecular fami-
lies, the detailed workings of many well under-
stood. Over the past 25 years, these ion channels
fell one by one to the awesome powers of recom-
binant DNA, membrane biochemistry, and in a
few cases x-ray crystallography: channels gated
by voltage, neurotransmitters, or intracellular lig-
ands; channels modulated by G proteins, Ca2+,
phosphorylation, or all of these together; channels
sensitive to heat and cold, jalapeno and menthol,
mechanical stretch, even light.
But in this all-electronic age, there is one
manila folder that I still dip into. It is labeled
“Funny Little Channels,” and its reprints deal
with a stubbornly defiant hold-out against molec-
ular identification: the voltage-dependent proton
channel originally discovered in snail neurons (1)
and meticulously characterized in mammalian
cells by DeCoursey (2, 3). Cells use this channel
for proton extrusion in response to membrane
voltage changes, in contexts as diverse as bacter-
ial killing by macrophages and bone resorption
by osteoclasts (2). The channel’s quirky func-
tional personality seems to place it outside all
known ion channel families, but the gene encod-
ing this protein has eluded discovery. Two recent
papers (4, 5) unveil the molecular identity of the
voltage-gated proton channel. The surprising
results match the channel’s functional idiosyn-
crasies and open the door to a deeper understand-
ing of voltage sensing in membrane proteins.
To appreciate the force of the discovery, it is
useful to review a group of ion channels called
the S4 family. These proteins—gated transmem-
brane pores specific for Na+, K+, or Ca2+—fulfill
a range of biological needs, the best known of
which is the charge flow required to generate
electrical signals in nerve cells. S4 proteins sense
and respond to transmembrane voltage changes.
For example, at a typical resting potential of –70
mV within the cell, Na+channels are closed, but
when the voltage moves positive, they open and
trigger electrical activity.
This conformational switching can be under-
he file-cabinet in my office houses detritus
of a pre-electronic past: manila folders
stuffed with long-neglected reprints from a
stood by looking at the structures of S4 proteins
(see the figure, top panels). In all these proteins,
a membrane-spanning module is repeated four
times around the pore axis. Each module con-
sists of six transmembrane helices. The last two
helices, S5 and S6, form the ion-selective pore in
the center. The first four, S1 to S4, fold together
on the pore periphery into a separate voltage-
sensing domain (VSD), a molecular voltmeter
that tells the pore when to open or close.
The physical basis of this key action resides
in the unusual sequence of helix S4. This other-
wise hydrophobic sequence is peppered with
positively charged residues (usually arginines
separated by three or four residues). At negative
voltage, the helix is electrostatically drawn
toward the intracellular side of the membrane;
when the voltage becomes positive, the helix is
pushed outward. These movements tug on the
pore domain to close or open the channel. The
recent crystal structure of an open S4 channel (6)
reveals the four-leaf-clover arrangement of
VSDs around the central pore.
VSDs were long thought to occur only in the
S4 family. But in 2005, in a sea-
squirt genome project, Okamura
and co-workers discovered a VSD-
containing protein that is not part
of an ion channel (7). Instead, the
VSD, complete with an arginine-
laden S4 helix, is attached to the
catalytic domain of a lipid phos-
phatase (see the figure, bottom
panels). This protein is the first
known voltage-controlled enzyme.
Okamura and co-workers then used
the sea-squirt VSD sequence to find
mammalian homologs, but no such
phosphatases were pulled out.
Instead, the mouse genome
coughed up the VSD-containing
protein described by Sasaki et al.
on page 589 of this issue (4). Using
a slightly different approach,
Ramsey et al. obtained a human
homolog (5). The big surprise is
that this VSD is not attached to
anything at all: The protein is a
free-floating VSD.
To find out what the VSD does,
both groups used a heterologous
system (one that does not normally
express the protein) to produce the
protein for electrophysiological
analysis. Robust voltage-dependent
proton currents were observed,
with properties similar to those that
appear naturally in many cell types.
Similarities in the heterologous ver-
sus natural systems include gating
that depends on voltage and trans-
membrane pH gradient, high tem-
perature sensitivity, and inhibition by Zn2+.
Cherny and DeCoursey (8) had predicted from
detailed mechanistic analysis that the voltage-
dependent H+channel would carry a conserved
pair of extracellular-facing histidines mediating
Zn2+inhibition; such a pair is observed in (4, 5).
Moreover, mutation of the S4 arginines pro-
duces changes in gating kinetics and softening
of voltage dependence, as expected for a VSD.
These correspondences led both groups to con-
PERSPECTIVES
The proton channel that ushers H+ions across
cell membranes to control cellular pH resembles
an isolated voltage-sensing domain with no
apparent pore.
Lonely Voltage Sensor Seeks
Protons for Permeation
Christopher Miller
BIOPHYSICS
The author is in the Department of Biochemistry, Howard
Hughes Medical Institute, Brandeis University, Waltham,
MA 02454, USA. E-mail: cmiller@brandeis.edu
+ + +
Positive voltage
Positive voltage
Negative voltage
Negative voltage
+ + +
S4
S4
S4
Open channel
Closed channel
Inactive enzyme
Active enzyme
Active site “open”
Active site “closed”
+ + +
S4
+ + +
S4
+ + +
S4
+ + +
– – –
– – –
Intracellular space
Voltage-sensing domains in two contexts. (Top two panels) A
typical voltage-gated K+channel, with voltage-sensing domain
(yellow and brown rods) attached to a gated pore. (Bottom two
panels) Voltage-sensing domain attached to lipid phosphatase to
form a voltage-gated enzyme. [Adapted from (7)]
Published by AAAS

Page 2

clude that the VSD alone forms the classical
voltage-gated proton channel, the founding
member of the new Hvchannel family (4, 5).
That is a satisfying conclusion. But here is a
gentle caution from someone once burned by
applying exactly the same criteria—similarity of
electrophysiological properties, alteration of funct-
ional behavior by mutation, appropriate tissue dis-
tribution—to infer that a quirky K+channel pro-
tein, minK, is identical to a particular K+current
of human cardiac tissue (9). That conclusion was
later shown to be wrong: minK is an auxiliary sub-
unit of a conventional K+channel (10). Never-
theless, the idea of an ion channel formed by a
solitary VSD is so chemically intriguing, biologi-
cally rich, and aesthetically pleasing that I will
refrain from demanding the tight proof normally
expected for family founders: functional reconsti-
tution of the purified protein.
With the arrival of the Hvfamily, I can almost
hear the patch-pipettes pulling and PCR tubes
popping as biophysicists rush to attack new ques-
tions. Why does the channel not have a pore
domain? Maybe because protons, unlike metal
cations, do not need an aqueous pathway to move
through proteins (3). Where do the protons go?
Probably not along the S4 helix itself, despite the
fact that a proton leak can be engineered into a K+
channel’s S4 helix (11). How many VSDs associ-
ate to form the channel? Two or more, surely,
because at least six charges move across the
membrane upon opening(3). What does the VSD
look like? Probably similar to the known struc-
ture of an isolated, though nonfunctional, VSD of
a K+channel (12). And how will knocking out
this gene affect the health of a mouse?
I reckon that I will be saving the many future
papers addressing these questions as PDFs.
References
1. R. C. Thomas, R. W. Meech, Nature 299, 826 (1982).
2. T. E. DeCoursey, Biophys. J. 60, 1243 (1991).
3. T. E. DeCoursey, Physiol. Rev. 83, 475 (2003).
4. M. Sasaki, M. Takagi, Y. Okamura, Science 312, 589
(2006); published online 23 March 2006 (10.1126/
science.1122352).
5. L. S. Ramsey, M. M. Moran, J. A. Chong, D. E. Clapham,
Nature, 10.1038/nature.04700 (22 March 2006).
6. S. B. Long, E. B. Campbell, R. Mackinnon, Science 309,
903 (2005).
7. Y. Murata, H. Iwasaki, M. Sasaki, K. Inaba, Y. Okamura,
Nature 435, 1239 ( 2005).
8. V. V. Cherny, T. E. DeCoursey, J. Gen. Physiol. 114, 819
(1999).
9. S. A. N. Goldstein, C. Miller, Neuron 7, 403 (1991).
10. M. C. Sanguinetti et al., Nature 384, 80 (1996).
11. D. M. Starace, F. Bezanilla, Nature 427, 548 (2004).
12. Y. Jiang et al., Nature 423, 33 (2003).
10.1126/science.1127186
www.sciencemag.orgSCIENCEVOL 312
28 APRIL 2006
535
PERSPECTIVES
A
about the Sun. Long-period comets, so named
because they have orbital periods of more than
200 years, originate from the Oort Cloud of
comets surrounding the Sun and stretching at least
10% of the way to the nearest star (1). The second
group are known as the Jupiter-family comets,
with orbital periods near 20 years, whose dynam-
ical evolution is controlled by gravitational
encounters with the giant planet. Theoretical work
pinpointed the source of this second group to a
comet belt beyond the planet Neptune (2–4); this
was dramatically proven by the discovery of the
first such object in 1992 (5). Now, in a report on
page 561 of this issue, Hsieh and Jewitt (6) have
issued a shock to the system by demonstrating the
existence of a third dynamical class of comets,
orbiting much closer to the Sun and lying entirely
within the main asteroid belt.
The story starts in 1996 with the discovery
that an asteroid first seen 17 years earlier was in
fact a comet, henceforth named 133P/Elst-
Pizzaro (7). Observationally, all but the largest
asteroids are optically unresolved and appear as
point sources, whereas active comets are recog-
nizable when near the Sun from the surrounding
atmosphere of sublimated ices and dust parti-
cles. Each year, several objects classified as
asteroids but lying in elongated comet-like orbits
stronomers have known for more than
two centuries that comets can be split into
two groups as defined by their orbits
are found to exhibit a coma and/or tail and hence
are reclassified as comets. The surprising fact
about 133P/Elst-Pizarro was that its orbit was
unlike that of any other comet, as it lay com-
pletely within the asteroid belt between Mars and
Jupiter. Another comet on a similar orbit was dis-
covered late last year, and Hsieh and Jewitt
report finding a third in a dedicated survey for
such objects. All three objects are relatively sta-
ble against strong gravitational perturbations
from Jupiter, which implies that they exist where
they formed.
Hsieh and Jewitt show that the detected atmo-
sphere of dust particles cannot be caused by weak
processes such as electrostatic levitation, nor can
it be the debris cloud from an impact by a smaller
body, and hence it must result from the steady sub-
limation of ices as with other comets. In most
walks of life, two is company but
three is a crowd, and there is no
escaping the recognition that we now
have a third dynamical class of active
comets identified in the solar system,
which Hsieh and Jewitt have labeled
main-belt comets.
Like all good discoveries, this
throws up a number of questions.
Perhaps the most important is how
they can exist in the first place.
Comets are ephemeral bodies, as
each time they pass the Sun they
lose a small fraction of their mass
via sublimation of the surface ices.
For example, the lifetime of Mark
Twain’s nemesis, Halley’s comet,
has been estimated as less than
100,000 years (8). The comets we
see today disappear on these time
scales, to be replenished by new
comets from the Oort Cloud and the
trans-Neptunian reservoirs. But the
main-belt comets are still in their
source regions, where continuous
solar heating would have seen them
vanish very soon after formation.
Two groups of comets are known: those with
orbital periods of hundreds of years or greater,
and those with decade-long periods. Athird class
appears to be orbiting within the asteroid belt.
Ice Among the Rocks
Alan Fitzsimmons
PLANETARY SCIENCE
The author is at the Astrophysics Research Center, School of
Mathematics and Physics, Queen’s University Belfast, Belfast
BT7 1NN, Northern Ireland, UK. E-mail: a.fitzsimmons@
qub.ac.uk
Closer to home. Orbits of the three main-belt comets discussed by
Hsieh and Jewitt (blue) and the planets Mars and Jupiter (red). The
green points are the first 5000 asteroids numbered, showing the
position of the main asteroid belt.
Published by AAAS