protein synthesis,and the inhibition ofDNAreplication following stress-induced releaseofthe protein nucleolin8.There has been a remarkable convergenceofrecent evidence — including the Rubbi andMilner paper1— suggesting that nucleoli areimportant in monitoring cellular stress. Thehealth ofthe nucleolus is an excellent surrogatefor the health of the cell,and conditions thatlead to nucleolar disruption are unlikely to besafe for continued cell proliferation. Thenotion that intact nucleoli are necessary tohold the p53 response in check provides anattractive model in which a default pathway ofp53 induction and inhibition of cell growth isovercome only by the maintenance of nucleo-lar well-being. These ideas reinforce the grow-ing realization that the nucleolus — longregarded as a mere factory for assembling ribo-somal subunits — is a vital command unit inmonitoring and responding to stress.■Henning F. Horn and Karen H. Vousden are at theBeatson Institute for Cancer Research, SwitchbackRoad, Glasgow G61 1BD, UK.e-mail: [email protected]. Rubbi,C. P. & Milner, J. EMBO J. 22, 6068–6077 (2003).2. Leonardo,A. D., Linke, S. P., Clarkin, K. & Wahl, G. M. GenesDev.8, 2540–2551 (1994).3. Siegel, J., Fritsche, M.,Mai, S.,Brandner,G. & Hess,R. D.Oncogene 11, 1363–1370 (1995).4. Lu,X. & Lane, D. P. Cell 75, 765–778 (1993).5. Sherr, C.J & Weber, J. D.Curr. Opin. Genet. Dev. 10,94–99 (2000).6 Colombo, E.,Marine, J.-C., Danovi, D., Falini, B.& Pelicci,P. G.Nature Cell Biol. 4, 529–533 (2002).7. Tsai, R.Y. & McKay, R.D.Genes Dev. 16,2991–3003 (2002).8. Daniely, Y., Dimitrova,D. D. & Borowiec, A.Mol. Cell. Biol.22,6014–6022 (2002).9. Blander, G.et al. J. Biol. Chem. 274, 29463–29469 (1999).10.Lohrum,M. A.E., Ludwig, R.L., Kubbutat, M. H.G.,Hanlon, M. & Vousden, K. H. Cancer Cell 3, 577–587 (2003).11.Zhang,Y. et al. Mol. Cell. Biol. 23, 8902–8912 (2003).12.Mazumder,B. et al. Cell 115, 187–198 (2003).As judged by external appearances,the left and right sides of vertebratebodies are (more or less) identical.There are,however, consistent left–right dif-ferences in the structure and placement ofthe internal organs. The heart,for instance,usually forms on the left, the liver on theright. In recent years, researchers haveuncovered several different molecularevents that are involved in establishing thisleft–right asymmetry as embryos develop1.But the picture that has emerged from thesenews and viewsstudies contains significant gaps. The paperby Raya et al.2on page 121 of this issue goessome way towards completing this picture,revealing an explicit link between an early,temporary asymmetry and later, stable patterns of asymmetric gene expression.The events that lead to the initial breakingof left–right symmetry in vertebrateembryos are not fully understood, but theyare believed to provide only weak transientbiases3.So additional mechanisms must existto amplify these biases,converting them intostable and heritable asymmetric patterns ofgene expression1. The earliest detected fea-ture of left–right asymmetry that is commonto all vertebrates studied is the expression ofthe secreted growth-factor protein Nodal onthe left side ofthe ‘node’.This region,locatedon the midline ofthe embryo, acts as an orga-nizing centre during development.In mice,Nodal expression has been shown to dependon a second signalling pathway, centred onthe cell-surface-located receptor Notch4,5.But how the Notch pathway becomes acti-vated to a sufficient degree to trigger Nodalexpression only on the left side of the noderemains an open question.Raya et al.2use a combination of model-ling and experimentation to address thisproblem in chick embryos. Having deter-mined the patterns of expression of variouskey genes around the node,the authors capi-talize on this information to construct amathematical model of the network of mol-ecular interactions underlying Notch activa-tion and Nodal expression. As Nodalenhances its own production,it can act as anon–off switch: only a transient increase inactivity of the Notch pathway is required toinduce stable Nodal expression. Raya et al.find that the simplest way to achieve this intheir model is to enhance the affinity ofNotch for its activating partners (ligands) —the Delta-like 1 (Dll1) and Serrate 1 (Srr1)proteins.So the model suggests that a tran-sient lateral bias in this affinity should beenough to convert the initially symmetricpattern of gene expression into one that ismanifestly asymmetric.The authors carry out a range of experi-ments that show that this is indeed the case.In doing so, they uncover a chain of eventsthat lead from a left–right asymmetry in theelectrochemical potential across the mem-branes of cells around the node, to the left-specific expression of Nodal.The first step inthis cascade is a previously described left-sided reduction in the activity of a mem-brane-spanning ion pump (the H&/K&-ATPase);this reduction results in membranedepolarization6.Raya et al. find that thisdepolarization leads to a transient increasein the extracellular concentration of Ca2&ions on the left ofthe node. And this in turn isnecessary for left-sided Nodal expression —suggesting that it could be Ca2&that modu-lates the affinity of Notch for its ligands.InDevelopmental biologyAsymmetric fixationNick MonkComputer simulations and laboratory experiments have shed light on howan asymmetric pattern of gene expression is fixed in vertebrate embryos— an early step towards asymmetric development of the internal organs.Figure 1Fixing asymmetry in vertebrates.According to convention,embryos are viewedfrom the ‘front’— so the left-hand side of theembryo appears on the right of this diagram.An early manifestation of asymmetry in chickembryos is the expression of the Nodal gene on the left of the ‘node’(oval).Raya et al.2putforward a model for how this occurs.It wasknown from studies in mice that Nodalexpression depends on the Notch pathway,which is in turn activated by Dll1 and Srr1.a,At stage 5 of development (19–22 hours afterfertilization),Dll1 expression extends furthertowards the head (the anterior) on the left thanon the right. This is the earliest indication thatNotch activity is higher on the left (as Dll1is atarget of Notch activity). b,During stage 6(23–25 hours after fertilization), the Dll1andSrr1 expression domains are symmetrical. But,as the fifth pulse of expression of the Lfng genesweeps up the embryo,it moves further to