BIOCHEMISTRY: Balancing Cellular Energy D. Grahame Hardie, et al. Science 315, 1671 (2007); DOI: 10.1126/science.

1140737

The following resources related to this article are available online at www.sciencemag.org (this information is current as of March 23, 2007 ):
Updated information and services, including high-resolution figures, can be found in the online version of this article at: http://www.sciencemag.org/cgi/content/full/315/5819/1671 Downloaded from www.sciencemag.org on March 23, 2007 A list of selected additional articles on the Science Web sites related to this article can be found at: http://www.sciencemag.org/cgi/content/full/315/5819/1671#related-content This article appears in the following subject collections: Biochemistry http://www.sciencemag.org/cgi/collection/biochem Information about obtaining reprints of this article or about obtaining permission to reproduce this article in whole or in part can be found at: http://www.sciencemag.org/about/permissions.dtl

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright c 2007 by the American Association for the Advancement of Science; all rights reserved. The title SCIENCE is a registered trademark of AAAS.

PERSPECTIVES
BIOCHEMISTRY

Balancing Cellular Energy

D. Grahame Hardie

The structure of the core of an enzyme complex that senses and responds to changes in the cells energy balance has been solved.

besity and type 2 diabetes are disorders of energy balance that are reaching epidemic proportions worldwide. A key target for drug development in this area is the adenosine monophosphateactivated protein kinase (AMPK) complex, which regulates cellular energy balance. For example, AMPK is switched on in muscle during exercise as a result of increased energy demand, triggering metabolic changes such as the switch from storing to consuming fats and carbohydrate. AMPK activation during regular exercise helps to protect against obesity and diabetes, and this protein complex is also the target for antidiabetic drugs such as the biguanides and thiazolidinediones (1). On page 1726 of this issue, Townley and Shapiro (2) describe crystal structures for the core of the AMPK complex from fission yeast. The study provides insights into binding of AMP and adenosine triphosphate (ATP), which may be very helpful in the design of new drugs. Mammalian AMPK regulates energy balance by monitoring changes in the cellular concentrations of the nucleotides AMP and ATP. An increase in AMP concentration indicating an energy deficitswitches the kinase on, causing it to attach phosphate to downstream target proteins, thereby switching on proteins involved in ATP-generating pathways and switching off those involved in ATP consumption. Conversely, high ATP concentrationsindicating energy sufficiency prevent AMP from switching on the kinase by competing for binding at the same site. AMPK and its relatives in nonmammalian species are composed of a catalytic subunit () and two regulatory subunits ( and ). The nucleotide-binding sites are located in two Bateman domains, each of which is formed by two repeats of a pattern of amino acids termed a CBS motif (3). The nucleotidebinding site has been modeled on the basis of the structure of other Bateman domains and on the locations of mutations causing human heart disease that are known to interfere with nucleotide binding (3, 4). Townley and Shapiro now show how the subunits interact in, and how the nucleotides are bound to, the core of the complex from fission yeast.

The author is in the Division of Molecular Physiology, College of Life Sciences, University of Dundee, Dundee DD1 5EH, UK. E-mail: d.g.hardie@dundee.ac.uk

www.sciencemag.org

SCIENCE

VOL 315

23 MARCH 2007

1671

Published by AAAS

Downloaded from www.sciencemag.org on March 23, 2007

The ability of mammalian AMPK to phosphorylate downstream target proteins is increased >100-fold by attachment of phosphate to its subunit by another protein (5, 6). This protein seems to phosphorylate AMPK constantly, but the phosphate is normally rapidly removed by yet another enzyme, so that AMPK immediately returns to the inactive state. Binding of AMP to the subunit inhibits this dephosphorylation (7, 8), providing a mechanism for converting the kinase to its active state during times of energy deficit. AMP binding also induces a change in structure that causes an additional activation; the combination of the two effects produces >1000-fold activation (9). However, neither effect of AMP appears to occur with the related protein from budding yeast (8), even though the AMP-binding sites are conserved. Large multiprotein complexes are often difficult to crystallize because of their complex shape and flexibility. The strategy adopted by Townley and Shapiro was to treat the

human complex with protein-degrading enzymes to generate the more compact and rigid core of the complex. This core contained the entire subunit but only the C-terminal domains of the and subunits, which were known to be required for formation of the complex. Bacteria were then programmed to make this core complex by genetic manipulation techniques, but suitable crystals were obtained only with the protein from fission yeast and not with that from humans. This is unfortunate, because little is known about the fission yeast enzyme, although its sequence is more similar to that of the protein from budding yeast than to the one from humans. In their structure, the C-terminal domain of the subunit has a compact shape, with the C-terminal domain of the subunit snaking around it (see the figure). There is minimal contact between and but extensive contact between and . The amino acids involved in the - interaction tend to be similar in AMPK relatives in all species, which suggests that the subunit interactions will also be similar. subunit C-terminal In the structures reported domain N by Townley and Shapiro, the N- and C-terminal Bateman subunit C-terminal domain domains (domains A and B) N face each other like mirror images. In crystals grown in the presence of AMP or ATP, a molC C Central sheet ecule of each nucleotide was bound in a very similar manner, Vacant AMP site? CBS1 but only in domain B. Although Bateman crystallized in the presence of CBS2 domain A N-terminal Mg2+, the ATP complex con+ domain + tained no metal. This agrees subunit + with findings that ATP binding Bateman domain B to the Bateman domains did not CBS3 N require Mg2+ (3). It also helps to explain how AMP can compete CBS4 C effectively for binding with AMP ATP, even though the latter is The core of the complex. The structures of Townley and Shapiro (2) normally present at 100-fold contain the C-terminal domains of the (yellow) and subunits (red) higher concentrations: Most shown at the top, and the entire subunit. The and subunits inter- ATP in the cell is present as act mainly via the central sheet. Bateman domains A (formed from the MgATP2 complex, whereCBS1 and CBS2) and B (CBS3 and CBS4) face each other like mirror as the concentration of free images, with domain B to the front in this view. A single AMP molecule is bound in domain B, with the negatively charged phosphate ATP is much lower and closer group (red oxygen atoms) interacting with a positively charged side to that of AMP. Bateman domain A was unchain on CBS4. The approximate positions of two equivalent positive side chains in human CBS1 and CBS2 are also indicated; these are not occupied by nucleotide, despite conserved in the yeast enzymes. high AMP and ATP concentra-

PERSPECTIVES
tions during crystallization. This was unexpected, because the isolated Bateman domains from the mammalian subunits bind two molecules of AMP or ATP (3). This is likely to be a genuine difference between the mammalian and yeast enzymes. In the new structures, a positively charged side chain in domain B (lowest + symbol in the figure) interacts with negatively charged phosphate(s) on AMP or ATP. A mutation of the equivalent side chain in the human enzyme (2 variant) that causes severe heart disease also greatly reduces binding of AMP (10). In the human enzyme, mutations in positively charged side chains occupying similar positions in CBS1 and CBS2 (upper + symbols in the figure) cause similar effects (3), supporting the idea that domain A also binds AMP in humans. Recently, my laboratory has provided evidence for a mechanism of activation of the human enzyme by AMP (11) that involves binding of AMP to these side chains. These residues are not conserved in the fission or budding yeast enzymes, which might explain why the latter is not activated by AMP. Resolving these remaining uncertainties and anomalies will require structures of mammalian complexes in the presence of AMP or ATP, together with other methods to study domain interactions, especially of those domains not present in the structures reported by Townley and Shapiro. The effort will be very worthwhile if it facilitates development of new drugs aimed at treatment of the epidemic of obesity and diabetes.
References
1. D. G. Hardie, Annu. Rev. Pharmacol. Toxicol. 47, 185 (2007). 2. R. Townley, L. Shapiro, Science 315, 1726 (2007). 3. J. W. Scott et al., J. Clin. Invest. 113, 274 (2004). 4. J. Adams et al., Protein Sci. 13, 155 (2004). 5. S. A. Hawley et al., J. Biol. 2, 28 (2003). 6. A. Woods et al., Curr. Biol. 13, 2004 (2003). 7. S. P. Davies, N. R. Helps, P. T. W. Cohen, D. G. Hardie, FEBS Lett. 377, 421 (1995). 8. M. J. Sanders, P. O. Grondin, B. D. Hegarty, M. A. Snowden, D. Carling, Biochem. J. 10.1042/BJ20061520 (2006). 9. M. Suter et al., J. Biol. Chem. 281, 32207 (2006). 10. B. Burwinkel et al., Am. J. Hum. Genet. 76, 1034 (2005). 11. J. W. Scott, F. A. Ross, J. K. Liu, D. G. Hardie, EMBO J. 26, 806 (2007). 10.1126/science/1140737

PHYSICS

How Does the Proton Spin?

Steven D. Bass

Protons are made of quarks and gluons, but their spins dont add up. New experiments may help resolve this discrepancy.

any particles, such as electrons, protons, and neutrons, behave like spinning tops. Unlike classical tops, however, the spin of these particles is an intrinsic quantum mechanical phenomenon. This spin is responsible for many fundamental properties of matter, including the protons magnetic moment, the different phases of matter in low-temperature physics, the properties of neutron stars, and the stability of the known universe. In recent experiments, a number of research groups have been seeking to shed some light on the puzzling origin of spin and how this might resolve some large discrepancies between theory and experiment. Particles such as the proton are actually combinations of more basic entities called quarks and gluons (which bind the quarks together). One of the challenges to physicists over the past 20 years has been to understand how the protons spin is built up from its quark and gluon constituents. Models of the proton generally predict that about 60% of the protons spin should be carried by the intrinsic spin of its three quarks, with the rest carried by orbital angular momentum (that is, the quarks flying around inside the proton). However, experiments at CERN (European Organization for Nuclear Research), DESY (Deutsches Elektronen-Synchrotron), and SLAC (Stanford

'

Spin story. Physicists use Feynman diagrams such as this to express the sequence of events in a highenergy particle collision. In one type of experiment, a polarized muon () and a polarized proton (p) approach each other on the left hand side. As they interact, the muon exchanges a polarized photon (). Pairs of charm-anticharm quark particles (c-c )are produced; the precise number of these particles created depends on the spin of the gluons (G) in the polarized proton, which allows the spin of the gluons to be reconstructed.

The author is at the Institute for Theoretical Physics, Universitt Innsbruck, A6020 Innsbruck, Austria. E-mail: sbass@mail.cern.ch

Linear Accelerator Center) have taught us that the contribution from the spin of the quarks inside is small, only about 30% (14). This shortfall offers a substantial challenge to our understanding about the structure of the proton. To sort this out, a vigorous global program has produced about 1000 theoretical papers, and dedicated spin experiments are under way at CERN, DESY, Jefferson Laboratory, and RHIC (Relativistic Heavy Ion Collider) to map individual quark and gluon angular momentum contributions to the protons spin. These experiments are now yielding exciting results (5). The proton is described by quantum chromodynamics (QCD, the theory of quarks and gluons) as a bound state of three confined valence quarks (6). The quarks have spin 1/2 and interact through the exchange of gluVOL 315 SCIENCE

ons, which have a spin of 1 (where spin is quoted in units of Plancks constant divided by ). When we probe deep inside the proton, the strength of quark-gluon and gluon-gluon interactions is small because of asymptotic freedom. This unusual idea means that, unlike some interactions, such as electrostatic forces, the force between quarks and gluons weakens as they get closer together. If a quark tries to escape, though, the force becomes strongerso strong, in fact, that the quarks and gluons are always bound inside nuclear particles such as the proton; they are never observed by themselves as free particles. In low-energy experiments, the proton behaves like a system of three massive constituent quarks carrying about 1/3 each of the mass of the proton. When we look deeper inside in high-energy experiments, these constituent quarks dissolve into near massless current quarks and a sea of quark-antiquark pairs and gluons. The spin experiments at CERN, DESY,

1672

23 MARCH 2007

www.sciencemag.org

Published by AAAS

Downloaded from www.sciencemag.org on March 23, 2007