RESEARCH NEWS & VIEWS
Sc
F
Sc–F bond
Figure 1 | Twisting vibration in the network structure of scandium trifluoride. Scandium atoms (Sc, grey) are at the centres and fluoride atoms (F, green) at the six shared corners of each of the eight ScF6 octahedra seen here in projection. When the Sc–Sc distances are fixed (dotted line), twisting vibration stretches the Sc–F bonds. If these bonds stretch harmonically, the potential energy of the motion varies with x4, where x is transverse fluoride displacement (as indicated by the arrow). Li et al.1 discovered that this form of vibration makes an important contribution to the phenomenon of negative thermal expansion in scandium trifluoride. (Modified from ref. 1.)
the authors demonstrated that the 25-meV vibration originates from a twisting mode of ScF6 octahedra (Fig. 1). A key question on structural NTE materials has been whether polyhedra such as the ScF6 octahedra distort during twisting vibrations. If the octahedra remain rigid during the vibrations, this ‘rigid unit mode’ has a harmonic potential: the potential energy of the motion varies with the square of the amplitude (x), like the stretching of a mechanical spring. This vibration would bring the scandium atoms at the centre of the octahedra together, shrinking the lattice. However, if the Sc–Sc distances instead remain fixed so that the Sc–F bonds have to stretch during the twisting motion, the potential energy has a quartic (x4) variation — an example of anharmonicity. Li and colleagues’ analysis1 of ScF3 reveals that the twisting vibration has an almost ideal quartic potential. They validate their analysis by showing that both the initial energy and the stiffening of the 25-meV phonon are consistent with the quartic potential for this twisting mode. After modifying the phonon calculations slightly by allowing the stretched Sc–F bonds to pull the scandium atoms together, the authors demonstrate that this anharmonic model predicts NTE that agrees well with the observed thermal behaviour for ScF3 (ref. 2).
This result shows that the non-rigidity of the ScF6 octahedra makes an important contribution to NTE. However, this does not signal the end of the rigid-unit-mode approach: Li et al.1 acknowledge that both harmonic rigid unit motions and quartic anharmonic effects are likely to be needed to provide a full explanation for NTE. With continuing improvements in inelastic neutron spectroscopy and in the computer power available for the supporting phonon calculations, the role of quartic anharmonicity in structural NTE is likely to be clarified further through future studies of more complex materials. Finally, I note that the discovery2 of NTE in ScF3 marks another step in the seemingly unstoppable rise of perovskite-type mater ials. Perovskites, named after the mineral form of the calcium titanium oxide CaTiO3, have structures like that of ScF3 but with an additional atom such as calcium that occupies the holes formed between groups of eight octahedra (Fig. 1). They are already
known to have excellent electronic and magnetic properties. Very large electronically driven NTE at ambient temperatures has been discovered in materials that are based on two perovskites, ZnNMn3 (ref. 4) and BiNiO3 (ref. 5); now the ‘empty perovskite’ ScF3 has been confirmed as a fundamental member of the structural NTE family. ■ J. Paul Attfield is in the Centre for Science at Extreme Conditions and the School of Chemistry, University of Edinburgh, Edinburgh EH9 3JZ, UK. e-mail:
[email protected] 1. Li, C. W. et al. Phys. Rev. Lett. 107, 195504 (2011). 2. Greve, B. K. et al. J. Am. Chem. Soc. 132, 15496–15498 (2010). 3. Mary, T. A., Evans, J. S. O., Vogt, T. & Sleight, A. W. Science 272, 90–92 (1996). 4. Takenaka, K. & Takagi, H. Appl. Phys. Lett. 87, 261902 (2005). 5. Azuma, M. et al. Nature Commun. 2, 347 (2011).
PH YSI O LO GY
On time metabolism In mammals, molecular clocks regulate transcription and glucose homeostasis. One way they do so is by controlling glucocorticoid-receptor signalling, which suggests that clocks are embedded in liver metabolism. See Letter p.552 JOSEPH BASS
T
he idea that genes program behaviour dates back to the 1970s, when researchers discovered that heritable mutations in fruitflies alter the daily, or circadian, sleep– wake cycle. At a molecular level, the clock machinery, which is composed of an autoregulatory transcription feedback loop1, is present in almost all cell types and generates oscillations in the expression of at least 10% of the genes transcribed2,3. Such observations led to an intensive search for the clock’s physiological functions. Lamia et al.4 describe one such function on page 552 of this issue. They report that the clock protein cryptochrome acts as an on/off switch for the nuclear receptor of glucocorticoids — hormones that are involved in glucose and lipid metabolism5. The finding expands our understanding of how metabolism is coordinated with geophysical time. The impact of circadian clocks on organismal energetics and survival has been investigated in photosynthetic organisms. In cyanobacteria6 and plants7, alignment of the period length with the environmental light cycle (a defining property of the clock) influences growth and reproduction. This suggests that ‘resonance’ of internal oscillators with the environment may provide a selective
4 6 6 | NAT U R E | VO L 4 8 0 | 2 2 / 2 9 D E C E M B E R 2 0 1 1
© 2011 Macmillan Publishers Limited. All rights reserved
advantage during evolution. In these organisms, the clock machinery also separates oxygen- and nitrogen-fixation processes, averting accumulation of cytotoxic free radicals and preventing futile energetic cycles8. In mammals, circadian rhythms operate at the behavioural, physiological and molecular levels. But studying the link between circadian cycles and metabolism in organisms with a nervous system is complicated. This is because the light response originates from the brain’s pacemaker neurons, whereas nutrient sensing occurs in other neurons and throughout many peripheral tissues. The discovery that fibroblasts maintained in culture exhibit selfsustaining circadian oscillations that show signature properties of a circadian clock9 was crucial because it hinted that peripheral clocks can function autonomously. Normally, however, brain and peripheral timepieces seem to communicate to give rise to coherent metabolic outputs. Indeed, genetic studies suggest that the brain and peripheral clocks participate in weight regulation and in glucose and lipid metabolism3. Circadian oscillations in hormone levels are certainly crucial for metabolic health, and disturbances in these have been used in clinical diagnosis for decades. For instance, high blood levels of the glucocorticoid hormone cortisol
NEWS & VIEWS RESEARCH at midnight are regarded as a sign of Cushing’s disease. Similarly, a ‘dawn phenomenon’ of elevated blood glucose levels arises in humans with diabetes owing to increased production of glucocorticoids in the early morning. At a molecular level, expression of the nuclear receptors for glucocorticoids shows strong circadian oscillations10. In fact, certain nuclear hormone receptors are integral components of the clock machinery11. These receptors’ activity results in a complex pattern of gene activation and repression, affecting metabolic and haematopoietic tissues in various ways. The clock protein PER2 interacts with several nuclear hormone receptors12, but the physiological impact of the clock machinery on the receptors’ function has not been established. Lamia and colleagues previously showed13 that the clock-repressor protein cryptochrome (Cry) is a sensor of cellular energy state. To further dissect the coupling between Cry and metabolic signals, some of the same authors now survey interactions of Cry with a broad range of nuclear hormone receptors in mice4. The authors report a strong and specific interaction between Cry and glucocorticoid receptors that shifts the balance towards gene repression after activation of these receptors. In addition, Cry interacted with two nuclear hormone receptors that activate transcription of the clock-activator protein Bmal1. This suggests that Cry may contribute to the oscillator’s robustness by participating in a short feedback loop controlling Bmal1 expression. Intriguingly, Cry interaction with glucocorticoid receptors was enhanced by dexamethasone4 — a drug that acts as a glucocorticoid. Because circadian levels of Cry vary, this observation directly links glucocorticoid signalling to circadian-clock oscillations (Fig. 1). Lamia and collaborators also note that Cry modulates patterns of gene activation and repression stimulated by glucocorticoid receptors. In Cry-deficient cells, glucocorticoid signalling led to a higher number of activated targets, and a vast decrease in the number of genes normally repressed by this clock protein. Because the main outcome of glucocorticoid-receptor activity is glucose generation in the liver by the process of gluconeogen esis, and because Cry-deficient animals display impaired glucose tolerance14, Lamia et al.4 focused on Cry function in the liver. The authors found that Cry directly represses the expression of Pck1 — a gene that encodes the rate-limiting enzyme in gluconeogenesis (Fig. 1). In Cry-deficient cells, dexamethasone increased Pck1 expression, but did not affect repression of genes mediating the inflammatory NF-κB signalling pathway. This observation indicates a separation of Cry functions in the gluconeogenic and inflammatory pathways. Cry deficiency also increased the sensitivity
a Non-autonomous circadian control Brain
HPA
Glucocorticoids Glucose homeostasis
Autonomous circadian control Liver clock
Glucocorticoidreceptor signalling
b
↓ Pck1
Bmal1 ↑Cry Clock
↓ GR
Gluconeogenesis OFF Light Dark Gluconeogenesis ON
Figure 1 | Biological clocks organize glucose metabolism. a, Glucose homeostasis is regulated by the interplay between non-autonomous circadian control — which affects glucocorticoid hormone levels through the brain and the hypothalamic–pituitary–adrenal (HPA) axis — and autonomous oscillatory regulation of these hormones’ signalling in peripheral tissues such as the liver. Lamia et al.4 report that the clock-repressor protein Cry plays a central part at both of these levels. b, In the liver, the clock-activator proteins Clock and Bmal1 increase Cry transcription, which downregulates gluconeogenesis both by interacting with — and so inhibiting — glucocorticoid receptors (GRs), and by reducing the expression of the Pck1 enzyme. As is typical of the clock machinery, high levels of Cry also downregulate Clock and Bmal1 transcription in a negative feedback loop, thus reducing its own levels and so promoting gluconeogenesis during the dark phase of the light–dark cycle.
of mice to glucocorticoid-induced diabetes4. This finding has therapeutic implications given the widespread use of steroids — of which glucocorticoids are an example — for the management of inflammation associated with cancer and conditions related to the immune system. Increasing Cry levels may enhance the anti-inflammatory effects of glucocorticoids by blocking the adverse effects of these hormones on glucose metabolism. Whereas previous work14 implicated Cry in regulating the signalling of glucagon (another hormone that mediates glucose homeostasis) in the cytoplasm, this paper pinpoints another role for Cry in the control of gluconeogenesis at the nuclear level. The modulation of
glucocorticoid signalling by Cry to control glucose homeostasis may be relevant for understanding glucose overproduction by the liver in human diabetes, not least because a genomic-association study15 indicated that CRY2 gene variants are associated with glucose homeostasis in humans. At the organismal level, glucocorticoid regulation in animals is controlled through a classical neuroendocrine feedback loop involving communication between the hypothalamus, pituitary and adrenal tissues (Fig. 1). Lamia et al. find that, in Cry-deficient mice, blood glucocorticoid levels remain high at all times of the day–night cycle, with the hormone’s overproduction arising at the level of the hypothalamus. How Cry deficiency might impair feedback inhibition of steroido genesis in the hypothalamus is a fascinating question. Lamia and colleagues’ findings may also be relevant to the spike seen in metabolic disorders among individuals subjected to shift work, sleep curtailment and other forms of circadian disruption. A feature of these disorders may be misalignment of oscillations in blood levels of glucocorticoids with oscillations in both the interaction between Cry and glucocorticoid receptors, and their signalling in the liver. This study provides an additional example of the extensive molecular integration and relatedness of temporal behavioural programmes with cell metabolism. ■ Joseph Bass is in the Division of Endocrinology, Metabolism and Molecular Medicine, Department of Medicine, Feinberg School of Medicine, Northwestern University, Chicago, Illinois 60611, USA. He is also at the Center for Sleep and Circadian Biology, Department of Neurobiology, Weinberg College of Arts and Sciences, Northwestern University. e-mail:
[email protected] 1. Rosbash, M. et al. Cold Spring Harb. Symp. Quant. Biol. 72, 75–83 (2007). 2. Panda, S. et al. Cell 109, 307–320 (2002). 3. Ramsey, K. M. & Bass, J. Cold Spring Harb. Symp. Quant. Biol. http://dx.doi.org/10.1101/ sqb.2011.76.010546 (2011). 4. Lamia, K. A. et al. Nature 480, 552–556 (2011). 5. Hollenberg, S. M. et al. Nature 318, 635–641 (1985). 6. Ouyang, Y., Andersson, C. R., Kondo, T., Golden, S. S. & Johnson, C. H. Proc. Natl Acad. Sci. USA 95, 8660–8664 (1998). 7. Dodd, A. N. et al. Science 309, 630–633 (2005). 8. Rutter, J., Reick, M. & McKnight, S. L. Annu. Rev. Biochem. 71, 307–331 (2002). 9. Balsalobre, A., Damiola, F. & Schibler, U. Cell 93, 929–937 (1998). 10. Yang, X. et al. Cell 126, 801–810 (2006). 11. Alenghat, T. et al. Nature 456, 997–1000 (2008). 12. Schmutz, I., Ripperger, J. A., Baeriswyl-Aebischer, S. & Albrecht, U. Genes Dev. 24, 345–357 (2010). 13. Lamia, K. A. et al. Science 326, 437–440 (2009). 14. Zhang, E. E. et al. Nature Med. 16, 1152–1156 (2010). 15. Dupuis, J. et al. Nature Genet. 42, 105–116 (2010).
2 2 / 2 9 D E C E M B E R 2 0 1 1 | VO L 4 8 0 | NAT U R E | 4 6 7
© 2011 Macmillan Publishers Limited. All rights reserved
