Glaciology: Ice–sheet advance in Antarctica
2
that this domain interacts with the C domain of the associated TCRβ chain. Within the crystal lattice, two pre-TCR molecules adopt a head-to-tail dimeric arrangement in which each pTα C domain is sandwiched between the C domain of its partner TCRβ chain and the V domain of the TCRβ chain of the other pre-TCR molecule (Fig. 1b). This quaternary arrangement, in which a large, surface-exposed hydrophobic patch on the V domain of a TCRβ is shielded by the C domain of pTα, provides an immediate explanation for the ability of the pre-TCR to dimerize autonomously. Pang and colleagues’ ligand-independent dimerization model also implies that the seg- ments that connect the extracellular domains of pre-TCR molecules to the cell membrane have sufficient rotational freedom to allow interactions between the pTα chain of one pre- TCR with the V domain of the TCRβ chain of another pre-TCR on the surface of the same cell. This surmise is based primarily on the authors’ analysis of the crystal lattice, and on their biophysical analysis of soluble pre-TCR. To validate it, Pang et al. tested the effect of mutations in amino-acid residues predicted to affect pre-TCR dimerization. They found that these residues are essential for dimerization of pre-TCR in solution and for its punctate distribution in cells that were not in contact with each other; this suggests that the pre-TCRs dimerize and probably func- tion even in the absence of pre-TCR expressed on the surface of another cell. Moreover, given that the residues within the V domain of the TCRβ chain that contact pTα are evolution- arily conserved in many Vβ genes, Pang and colleagues’ model of pre-TCR assembly also explains the ability of pTα to pair with any Vβ domain 4 . On the basis of these data 1 , the role of TCRβ in the pre-TCR complex does not involve its antigen-binding ability and differs from its contribution to a second checkpoint that occurs later in the life of a developing T cell — the TCRαβ selection. (At this point, pTα chains are replaced by TCRα chains and the result- ing αβTCRs are checked for their antigenic specificity.) Instead, the partnership between pTα and TCRβ at the level of the pre-TCR may have evolved through co-option of a proto- pTα chain, the primordial function of which was limited to allowing interaction between TCRβ-like chains expressed on opposite cells as hinted by a crystal structure 5 (Fig. 1c). But proteins related to pTα are found only in mammals, suggesting that αβTCR develop- ment in other vertebrates does not rely on pre- TCR signalling 6 . So what bonus was conferred by the emergence of a pTα chain? It has been argued 7 that, if rearrangements of the TCRα and the TCRβ genes were to happen concomitantly, and if their outcome were to be checked in a sin- gle step, developing T cells would not meet the stringent specificity requirements that charac- terize TCRαβ selection in mammals. Therefore, GLACIOLOGY Ice-sheet advance in Antarctica Reliable forecasts of sea-level rise depend on accurately modelling the dynamics of polar ice sheets. A numerical framework that better reflects ice-sheet basal drag adds greater realism to such models. FABIEN GILLET-CHAULET & GAËL DURAND S ea levels are rising, with about one-third of the increase being attributable to the release of ice from coastal glaciers in Greenland and Antarctica. A decade ago came news that these glaciers were accelerating, and they have continued to do so. But our poor understanding of the processes that are driving ice discharge casts doubt on the reliability of sea-level forecasts for the coming centuries 1 . In a paper in Geophysical Research Letters, Mor- lighem et al. 2 provide a leap forward in the modelling of ice-sheet dynamics that promises to improve matters considerably. Confident predictions of the evolution of a system obviously require a proper descrip- tion of its current state. For polar ice sheets, the advent of remote-sensing techniques has been a great help, and it was with such data that their significant mass loss was diagnosed. A particular example of rapid change is that two- thirds of the mass loss from the West Antarctic Ice Sheet can be attributed to the Pine Island Bay sector, where ice release has doubled over the past decade 3 . From a broader perspective, we know the surface elevation and surface velocities of a fair portion of both the Green- land and Antarctic ice sheets, and are starting to have a reasonable view of the changes that have occurred over the past ten years. But other crucial factors operating at the base of the ice sheets, such as the drag exerted by the underlying bedrock on the ice, are much less well known. Although description of the ice sheets’ current state is reaching maturity, forecasts require a numerical model that adequately describes the thermo-mechanical processes involved. Given the continental scale of the ice masses concerned, and the limited com- puting resources, most ice-sheet models use approximations of the ice-flow equations 4 . These approximations are fully justified when applied to the part of an ice sheet that is grounded (resting on bedrock), or the float- ing extensions (the ice shelves). But they fail to properly describe the flow at the transition zone in between, called the grounding line 5 (Fig. 1). We are therefore faced with deficient models for the very location at which most changes are now occurring. Moreover, ice- sheet models were mainly designed to study the long-term variations of ice masses, and so were focused on palaeoclimatic simulations and reconstruction of the area covered by ice in the past. So modelling of the configuration of present-day ice sheets arose more as a by- product — a final step — of such simulations, rather than as an objective in itself. These it seems likely that the pre-TCR maximizes the generation of a large collection of T-cell clones, each expressing a TCR endowed with a unique antigenic specificity. This allows mammals to successfully cope with any antigen. Pang and co-workers’ structure 1 is of sig- nificance not just because it elucidates the unique mechanism by which assembly of the pre-TCR is directly coupled to its signalling role. The ligand-independent dimerization that the authors describe for pre-TCR might occur more generally during the development of adaptive immunity. Developing B cells also use a pre-B-cell receptor that, although built up differently to pre-TCR 8 , monitors successful rearrangement at the level of the gene encoding the immunoglobulin heavy chains. Whether, akin to the pre-TCR, the pre-B-cell receptor also functions through ligand-independent dimerization remains to be seen. ■ Bernard Malissen and Hervé Luche are at the Centre d’Immunologie de Marseille Luminy, INSERM, CNRS, Université de la Méditerranée, Parc Scientifique de Marseille Luminy, Case 906, 13288 Marseille Cédex 9, France. e-mail: [email protected] 1. Pang, S. S. et al. Nature 467, 844–848 (2010). 2. Saint-Ruf, C. et al. Science 266, 1208–1212 (1994). 3. Yamasaki, S. et al. Nature Immunol. 7, 67–75 (2005). 4. Wilson, A., Maréchal, C. & MacDonald, H. R. J. Immunol. 166, 51–57 (2001). 5. Bentley, G. A., Boulot, G., Karjalainen, K. & Mariuzza, R. A. Science 267, 1984–1987 (1995). 6. Kreslavsky, T., Gleimer, M. & von Boehmer, H. Curr. Opin. Immunol. 22, 185–192 (2010). 7. Malissen, B., Ardouin, L., Lin, S. Y., Gillet, A. & Malissen, M. Adv. Immunol. 72, 103–148 (1999). 8. Bankovich, A. J. et al. Science 316, 291–294 (2007). 794 | NATURE | VOL 467 | 14 OCTOBER 2010 NEWS & VIEWS RESEARCH © 20 Macmillan Publishers Limited. All rights reserved 10
Transcript of Glaciology: Ice–sheet advance in Antarctica
that this domain interacts with the C domain of the associated TCRβ chain. Within the crystal lattice, two pre-TCR molecules adopt a head-to-tail dimeric arrangement in which each pTα C domain is sandwiched between the C domain of its partner TCRβ chain and the V domain of the TCRβ chain of the other pre-TCR molecule (Fig. 1b). This quaternary arrangement, in which a large, surface-exposed hydrophobic patch on the V domain of a TCRβ is shielded by the C domain of pTα, provides an immediate explanation for the ability of the pre-TCR to dimerize autonomously.
Pang and colleagues’ ligand-independent dimerization model also implies that the seg-ments that connect the extracellular domains of pre-TCR molecules to the cell membrane have sufficient rotational freedom to allow interactions between the pTα chain of one pre-TCR with the V domain of the TCRβ chain of another pre-TCR on the surface of the same cell. This surmise is based primarily on the authors’ analysis of the crystal lattice, and on their biophysical analysis of soluble pre-TCR. To validate it, Pang et al. tested the effect of mutations in amino-acid residues predicted to affect pre-TCR dimerization.
They found that these residues are essential for dimerization of pre-TCR in solution and for its punctate distribution in cells that were not in contact with each other; this suggests that the pre-TCRs dimerize and probably func-tion even in the absence of pre-TCR expressed on the surface of another cell. Moreover, given that the residues within the V domain of the TCRβ chain that contact pTα are evolution-arily conserved in many Vβ genes, Pang and colleagues’ model of pre-TCR assembly also explains the ability of pTα to pair with any Vβ domain4.
On the basis of these data1, the role of TCRβ in the pre-TCR complex does not involve its antigen-binding ability and differs from its contribution to a second checkpoint that occurs later in the life of a developing T cell — the TCRαβ selection. (At this point, pTα chains are replaced by TCRα chains and the result-ing αβTCRs are checked for their antigenic specificity.) Instead, the partnership between pTα and TCRβ at the level of the pre-TCR may have evolved through co-option of a proto-pTα chain, the primordial function of which was limited to allowing interaction between TCRβ-like chains expressed on opposite cells as hinted by a crystal structure5 (Fig. 1c).
But proteins related to pTα are found only in mammals, suggesting that αβTCR develop-ment in other vertebrates does not rely on pre-TCR signalling6. So what bonus was conferred by the emergence of a pTα chain? It has been argued7 that, if rearrangements of the TCRα and the TCRβ genes were to happen concomitantly, and if their outcome were to be checked in a sin-gle step, developing T cells would not meet the stringent specificity requirements that charac-terize TCRαβ selection in mammals. Therefore,
G L a c i o L o G Y
Ice-sheet advance in AntarcticaReliable forecasts of sea-level rise depend on accurately modelling the dynamics of polar ice sheets. A numerical framework that better reflects ice-sheet basal drag adds greater realism to such models.
F a B i e N G i L L e T- c H a U L e T & G a ë L D U r a N D
Sea levels are rising, with about one-third of the increase being attributable to the release of ice from coastal glaciers in
Greenland and Antarctica. A decade ago came news that these glaciers were accelerating, and they have continued to do so. But our poor understanding of the processes that are driving ice discharge casts doubt on the reliability of sea-level forecasts for the coming centuries1. In a paper in Geophysical Research Letters, Mor-lighem et al.2 provide a leap forward in the modelling of ice-sheet dynamics that promises to improve matters considerably.
Confident predictions of the evolution of a system obviously require a proper descrip-tion of its current state. For polar ice sheets, the advent of remote-sensing techniques has been a great help, and it was with such data that their significant mass loss was diagnosed. A particular example of rapid change is that two-thirds of the mass loss from the West Antarctic Ice Sheet can be attributed to the Pine Island Bay sector, where ice release has doubled over the past decade3. From a broader perspective, we know the surface elevation and surface velocities of a fair portion of both the Green-land and Antarctic ice sheets, and are starting to have a reasonable view of the changes that
have occurred over the past ten years. But other crucial factors operating at the base of the ice sheets, such as the drag exerted by the underlying bedrock on the ice, are much less well known.
Although description of the ice sheets’ current state is reaching maturity, forecasts require a numerical model that adequately describes the thermo-mechanical processes involved. Given the continental scale of the ice masses concerned, and the limited com-puting resources, most ice-sheet models use approximations of the ice-flow equations4. These approximations are fully justified when applied to the part of an ice sheet that is grounded (resting on bedrock), or the float-ing extensions (the ice shelves). But they fail to properly describe the flow at the transition zone in between, called the grounding line5 (Fig. 1). We are therefore faced with deficient models for the very location at which most changes are now occurring. Moreover, ice-sheet models were mainly designed to study the long-term variations of ice masses, and so were focused on palaeoclimatic simulations and reconstruction of the area covered by ice in the past. So modelling of the configuration of present-day ice sheets arose more as a by-product — a final step — of such simulations, rather than as an objective in itself. These
it seems likely that the pre-TCR maximizes the generation of a large collection of T-cell clones, each expressing a TCR endowed with a unique antigenic specificity. This allows mammals to successfully cope with any antigen.
Pang and co-workers’ structure1 is of sig-nificance not just because it elucidates the unique mechanism by which assembly of the pre-TCR is directly coupled to its signalling role. The ligand-independent dimerization that the authors describe for pre-TCR might occur more generally during the development of adaptive immunity. Developing B cells also use a pre-B-cell receptor that, although built up differently to pre-TCR8, monitors successful rearrangement at the level of the gene encoding the immunoglobulin heavy chains. Whether, akin to the pre-TCR, the pre-B-cell receptor also functions through ligand-independent dimerization remains to be seen. ■
Bernard Malissen and Hervé Luche are at the Centre d’Immunologie de Marseille Luminy, INSERM, CNRS, Université de la Méditerranée, Parc Scientifique de Marseille Luminy, Case 906, 13288 Marseille Cédex 9, France. e-mail: [email protected]
1. pang, S. S. et al. Nature 467, 844–848 (2010).2. Saint-ruf, c. et al. Science 266, 1208–1212
(1994).3. Yamasaki, S. et al. Nature Immunol. 7, 67–75
(2005).4. wilson, a., Maréchal, c. & MacDonald, h. r.
J. Immunol. 166, 51–57 (2001).5. Bentley, G. a., Boulot, G., Karjalainen, K. & Mariuzza,
r. a. Science 267, 1984–1987 (1995).6. Kreslavsky, t., Gleimer, M. & von Boehmer, h. Curr.
Opin. Immunol. 22, 185–192 (2010).7. Malissen, B., ardouin, L., Lin, S. Y., Gillet, a. &
Malissen, M. Adv. Immunol. 72, 103–148 (1999).
8. Bankovich, a. J. et al. Science 316, 291–294 (2007).
7 9 4 | N A T U R E | V O L 4 6 7 | 1 4 O c T O b E R 2 0 1 0
NEWS & VIEWSRESEaRch
© 20 Macmillan Publishers Limited. All rights reserved10
50 YEARS AGOIn a recent address given in the University of Nottingham, Sir William Slater, secretary of the Agricultural Research Council, made a strong plea that universities should lead rather than follow contemporary thought, and in particular that they should be sufficiently independent financially to be able to allocate sufficient funds for research on lines of work entirely of their own choosing. Sir William was discussing the relative values of such research and of sponsored research, undertaken with the assistance of outside bodies ... The problem of course arises out of the extent to which the universities are now dependent on public funds, and how far, even with the University Grants Committee, it is possible for their independence from outside pressure to be fully maintained.From Nature 15 October 1960.
100 YEARS AGOThe second congress on radiology and electricity was held at Brussels on September 13–15 ... [The] president called upon Madame Curie to give an account of the recent experiments made in Paris to isolate metallic radium. It will be remembered that this metal has hitherto not been separated from its salts, although a radium amalgam was obtained some years ago by Coehn. The beautiful experiments described by Madame Curie, resulting in the isolation of metallic radium, must be regarded as a triumph in chemical manipulation when it is remembered that ... the operations had to be carried out with minute quantities of material in such a way as to avoid loss of the precious substance during the process. These experiments should remove all possible doubt that radium is, in fact, an element belonging to the same group of metals as barium.From Nature 13 October 1910.
models, then, only partly represent current ice-sheet geometry and flow.
Help is now at hand through the appli-cation of inverse methods. Such methods involve extracting model parameters from available data, and have shown their worth in improving model forecasts in meteorology and oceanography. Although they were introduced in glaciology almost 20 years ago6, they have remained largely unexploited. But new initia-tives have recently emerged7,8, among which the work of Morlighem et al.2 offers the most accomplished attempt to tackle the weaknesses in existing ice-sheet models.
First, using the finite-element method, Mor-lighem et al. solved the full set of flow equations (the Stokes equations) and two approxima-tions of those equations commonly used in ice-sheet models. Second, they generalized a particular inverse method, namely the con-trol method9, to plausibly infer the basal drag, exerted by the bedrock on the over lying glacier, from measured surface velocities. Finally, they applied their methodology over the Pine Island drainage basin, successfully reducing the gap between surface observations and model results. The comparison of the basal-drag patterns, obtained by the inverse method for each flow approximation, confirms the need for careful flow resolution in the vicinity of the grounding line.
Challenges remain, however, in terms of both method and computation. Robust simu-lations will require computation of grounding-line motion, which is at the heart of a 40-year debate. Recent progress has been made in this respect, from both theoretical10 and numeri-cal perspectives11, showing that a significant grid refinement (to the subkilometre scale)
is required in this transition zone between grounded and floating ice. Such a refinement, together with the extension of Morlighem and colleagues’ methodology from a drainage basin to a whole ice sheet, will require massive par-allel computation facilities. Nonetheless, we have reached the stage at which more robust numerical tools will soon allow model simula-tions to be run forward in time, and proposal of prognostic simulations of the behaviour of polar glaciers. ■
Fabien Gillet-Chaulet and Gaël Durand are in the Laboratoire de Glaciologie et Géophysique de l’Environnement (LGGE), CNRS/UJF-Grenoble I, 38400 St Martin d’Hères, France. e-mails: [email protected]; [email protected]
1. Solomon, S. D. et al. (eds) Climate Change 2007: The Physical Science Basis (cambridge Univ. press, 2007).
2. Morlighem, M. et al. Geophys. Res. Lett. 37, doi:10.1029/2010GL043853 (2010).
3. rignot, e. et al. Nature Geosci. 1, 106–110 (2008).
4. ritz, c., rommelaere, v. & Dumas, c. J. Geophys. Res. 106, doi:10.1029/2001JD900232 (2001).
5. vieli, a. & payne, a. J. J. Geophys. Res. doi:10.1029/2004JF000202 (2005).
6. Macayeal, D. r. J. Glaciol. 39, 91–98 (1993).7. heimbach, p. & Bugnion, v. Ann. Glaciol. 50, 67–80
(2009).8. arthern, r. J. & Gudmundsson, G. h. J. Glaciol. 56,
527–533 (2010).9. vieli, a. & payne, a. J. Ann. Glaciol. 36, 197–204
(2003).10. Schoof, c. J. Geophys. Res.
doi:10.1029/2006JF000664 (2007).11. Durand, G., Gagliardini, o., de Fleurian, B.,
zwinger, t. & Le Meur, e. J. Geophys. Res. doi:10.1029/2008JF001170 (2009).
12. www.nasa.gov/topics/earth/features/antarctica-outline.html
EastAntarctica
Pine Island glacier
WestAntarctica
na
Sa
Figure 1 | The Antarctic grounding line. The grounding line is the transition point of a glacier from flow on bedrock to floating ice shelf, and is a crucial location for estimating ice loss to the ocean. Morlighem and colleagues’ numerical treatment2, as applied to the Pine Island glacier, will improve modelling of ice flow in this zone. Identifying the location of the grounding line around Antarctica is another task, and the red line on this map, generated from satellite data, shows just such a depiction. The map was presented earlier this year by a team led by researchers at NASA’s Goddard Space Flight Center (see ref. 12).
1 4 O c T O b E R 2 0 1 0 | V O L 4 6 7 | N A T U R E | 7 9 5
NEWS & VIEWS RESEaRch
© 20 Macmillan Publishers Limited. All rights reserved10
