Developmental Cell
Previews of extracellular ferritin is unclear. Release of ferritin is not the result of cell death and may result from an active secretory process but the evidence for secretion is not yet compelling. One reason the role of extracellular ferritin is enigmatic is that while intracellular ferritin is iron rich, serum ferritin is relatively iron poor. High levels of serum ferritin are found during inflammation or in patients with iron-overload disease. Over the years, ferritin receptors have been reported on a wide variety of cell types. Some of the ferritin receptor candidates have not held up in the era in which molecular biology has raised the standard of identification to necessary and sufficient. One exception is Tim-2, a cell surface protein on murine lymphocytes that shows a preference for H-ferritin (Chen et al., 2005). The role of Tim-2 as a putative ferritin receptor fits with studies that show that ferritin secretion results from inflammation and that ferritin may affect immune function (Morikawa et al., 1995; Recalcati et al., 2008). The identification of Scara5 as a ferritin receptor also meets the criteria of necessary and sufficient. Of interest is that Scara5 seems to prefer L-chain enriched ferritin rather than the H-chain ferritins
preferred by Tim-2. Both H and L chain ferritin monomers contribute to the ferritin nanocage. H-ferritin monomers are required as they contain the ferroxidase activity essential for iron incorporation into the ferritin nanocage. The inclusion of L-chains into the ferritin nanocage permits a higher iron content and is thought to be involved in mineralization (Theil, 2004). The amount of H chain and L chain within ferritin can vary dramatically. Ferritin nanocages in tissues that play an important role in iron storage are enriched in L chains. Different cell types synthesize ferritin with different subunit compositions and it might be that the cell surface receptor for ferritin reflects the function of ferritin as a marker for inflammation or as an iron source for organ development. For those who work in the iron field, the results of Li et al. (2009) are satisfying. Iron is such an important element that one expected to find multiple transport systems. Even the simplest eukaryote, S. cerevisiae, has at least four. The results of Li et al. (2009) lead to interesting questions. For instance, what is the basis of the cell type specificity for Tf- dependent and Tfindependent iron transport systems, and
are there high levels of serum ferritin in developing embryos? And finally, what is the mechanism of ferritin secretion? REFERENCES Bernstein, S.E. (1987). J. Lab. Clin. Med. 110, 690–705. Chen, T.T., Li, L., Chung, D.H., Allen, C.D., Torti, S.V., Torti, F.M., Cyster, J.G., Chen, C.Y., Brodsky, F.M., Niemi, E.C., et al. (2005). J. Exp. Med. 202, 955–965. Kobayashi, K. (1978). Cell Tissue Res. 195, 381–394. Levy, J.E., Jin, O., Fujiwara, Y., Kuo, F., and Andrews, N.C. (1999). Nat. Genet. 21, 396–399. Li, J.Y., Paragas, N., Ned, R.M., Qui, A., Viltard, M., Leete, T., Drexler, I.R., Chen, X., Sanna-Cherchi, S., Mohammed, F., et al. (2009). Dev. Cell 16, this issue, 35–46. Morikawa, K., Oseko, F., and Morikawa, S. (1995). Leuk. Lymphoma 18, 429–433. Ponka, P., Beaumont, C., and Richardson, D.R. (1998). Semin. Hematol. 35, 35–54. Recalcati, S., Invernizzi, P., Arosio, P., and Cairo, G. (2008). J. Autoimmun. 30, 84–89. Theil, E.C. (2004). Annu. Rev. Nutr. 24, 327–343. Wheby, M.S., and Umpierre, G. (1964). N. Engl. J. Med. 271, 1391–1395.
Chaos Begets Order: Asynchronous Cell Contractions Drive Epithelial Morphogenesis Ewa Paluch1,2 and Carl-Philipp Heisenberg1,* 1Max-Planck-Institute
of Molecular Cell Biology and Genetics, 01307 Dresden, Germany Institute of Molecular and Cell Biology, 02-109 Warsaw, Poland *Correspondence:
[email protected] DOI 10.1016/j.devcel.2008.12.011 2International
Apical cell contraction triggers tissue folding and invagination in epithelia. During Drosophila gastrulation, ventral furrow formation was thought to be driven by smooth, purse-string-like constriction of an actomyosin belt underlying adherens junctions. Now Martin et al. report in Nature that ventral furrow formation is triggered by asynchronous pulsed contractions of the apical acto-myosin cortex in individual cells. Epithelial folding and invagination are observed in numerous developmental processes including neural tube formation and mesoderm invagination (Ettensohn, 1985). Apical cell contraction is commonly thought to trigger these processes by
changing the morphology of epithelial cells from cuboidal to bottle-shaped, which causes the tissue to bend and fold. In Drosophila, ventral furrow formation has been for many years a model system for the study of the molecular and
4 Developmental Cell 16, January 20, 2009 ª2009 Elsevier Inc.
cellular mechanisms underlying epithelial folding in development (Leptin, 1999). Synchronous and continuous apical constriction of epithelial cells, driven by a purse-string-like contraction of a cortical actin-myosin ring underlying apical
Developmental Cell
Previews
Figure 1. Schematic Representation of Pulsed Asynchronous Apical Cell Contractions during Drosophila Ventral Furrow Formation (A) View of the apical side of ventral furrow cells during furrow formation. (B) Rearrangements of the apical actin (red rods) and myosin (green dumbbells) cortex in a representative cell (marked in [A]) during contraction. Adapted from Martin et al., 2008.
adherens junctions, has previously been proposed to drive epithelial folding (Ettensohn, 1985). Now, Martin et al. (2008) show in Drosophila ventral furrow formation that although the tissue globally contracts at a steady rate, individual cells contract neither synchronously nor continuously. Instead, each cell displays a pulsing behavior, alternating between phases of rapid contraction and pauses where the cell apical surface area remains stable (Figure 1). Strikingly, these contraction pulses are asynchronous between cells within the epithelium. Moreover, by precisely quantifying cell surface variations and the dynamics of myosin at the apical cortex, Martin et al. demonstrate that the constriction phases correlate with coalescence of an apical myosin network rather than a gradual accumulation of myosin at adherens junctions.
This suggests that apical myosin contractions drive the pulses. These observations prompt a number of questions, such as, ‘‘How are the pulsed contractions regulated on a molecular level?’’ and ‘‘What is the mechanical basis for pulsed contractions driving epithelial folding?’’ To address the question of molecular regulation of the contractions, Martin et al. examine the roles of two transcription factors, twist and snail. Twist and snail are key regulators of ventral furrow formation in Drosophila. They have previously been shown to be required for apical contraction of ventral furrow cells, although their specific function in this process is not yet fully understood (Leptin, 1999). Martin et al. have now been able to dissect the function of twist and snail in ventral furrow formation by driving snail expression in a twist loss-of-function background, an
approach previously introduced to circumvent the requirement of twist for snail expression (Seher et al., 2007). By doing this, Martin et al. found that snail is specifically required for contraction of ventral furrow cells, while twist is required for stabilizing the cortex in its contracted state. How twist and snail function in these different processes is still unclear. For snail, a transcriptional repressor, no direct targets have yet been identified during ventral furrow formation. However, snail homologs in other developmental processes have been implicated in junctional disassembly underlying epithelialto-mesenchymal (EMT) transformation (Barrallo-Gimeno and Nieto, 2005). EMT itself has been associated with myosindriven cell cortex contraction, suggesting that snail-dependent contraction of ventral furrow cells might be part of its EMT regulatory function. For twist, several transcriptional targets, including snail, have been identified during ventral furrow formation (Leptin, 1999). Two of these targets, the G protein-coupled receptor folded gastrulation (fog) and the transmembrane protein T48, are known to be involved in twist-induced cell shape changes (Kolsch et al., 2007; Leptin, 1999). Fog and T48, downstream of twist, are thought to facilitate apical contraction of ventral furrow cells by translocating RhoGEF2 and, as a consequence, myosin to the apical membrane. Apical myosin, in turn, has been suggested to move subapical adherens junctions toward the apical side, thereby transforming subapical into apical junctions. It is therefore conceivable that twist stabilizes the cell apical surface by promoting the assembly of apical junctions. The observations of Martin et al. also raise a number of questions related to the mechanics of epithelial folding. For instance, why are apical contractions pulsed and what stabilizes the apical cell surface during pauses between contractions? Observations of myosin dynamics suggest that myosin bursts and coalescence of myosin spots drive pulsed contractions. However, it is unclear what determines the maximum cortical contraction achievable during a given pulse. One possibility is that the cell contracts as much as possible until all apical myosins have coalesced into a spot, and further contraction is only
Developmental Cell 16, January 20, 2009 ª2009 Elsevier Inc. 5
Developmental Cell
Previews possible once myosin motors have redistributed over the apical surface. Another possibility is that a cell contracts as much as interactions with its neighboring cells allow. Contraction of a cell within an epithelium will inevitably lead to stretching of its neighbors and thus the extent by which the cell contracts might be limited by the elastic resistance of its neighboring cells. Studying the spatial and temporal correlation between contraction phases of neighboring cells might help to resolve this issue by providing insights into the correlation between contraction of one cell and stretching of its neighbors. Importantly, cell surface area must be stabilized between contractions for pulses to result in a net decrease of tissue apical surface. Martin et al. propose that tension in the remaining acto-myosin cortex opposes stretching by neighboring cells during pauses between contractions. This implies that the cortical network displays a high elastic modulus that primarily depends on the level and nature of crosslinkers (Bausch and Kroy, 2006). It will thus be interesting to know whether twist is involved in the regulation of actin cortex crosslinking. Together with the observation of twist-dependent apical
junction assembly, it also raises the question of whether actin crosslinking, cortical stiffening, and apical junction formation are interrelated processes. Mapping the tension distribution during ventral furrow formation by, for example, ablating single cells or the cortex within individual cells and correlating tension with the localization of junctional and cytoskeletal components would be helpful in resolving how pulsed apical cell contraction and stabilization are achieved. Continuous tissue deformation driven by asynchronous shape changes of individual cells is likely to be a common feature in development. For example, convergent extension in Xenopus embryos proceeds at a constant rate, although it is driven by apparently uncoordinated cell movements (Keller et al., 2008). A similar mechanism at a different scale is also observed during muscle contraction, where sarcomere shortening is achieved by asynchronous steps of individual motors (Duke, 1999). Future studies will have to determine which developmental processes result from asynchronous behaviors and where (and how) synchrony is achieved (see also Duke, 1999 for a discussion of synchrony in sarcomere contraction). Theoretical
modeling of cells interacting in a tissue (Farhadifar et al., 2007) may help reveal whether asynchronous pulsed contractions of individual cells provide a more efficient and/or robust tissue contraction mode than continuous synchronized contraction of all the cells together. REFERENCES Barrallo-Gimeno, A., and Nieto, M.A. (2005). Development 132, 3151–3161. Bausch, A.R., and Kroy, K. (2006). Nat. Phys. 2, 231–238. Duke, T.A. (1999). Proc. Natl. Acad. Sci. USA 96, 2770–2775. Ettensohn, C.A. (1985). Q. Rev. Biol. 60, 289–307. Farhadifar, R., Roper, J.C., Aigouy, B., Eaton, S., and Julicher, F. (2007). Curr. Biol. 17, 2095–2104. Keller, R., Shook, D., and Skoglund, P. (2008). Phys. Biol. 5, 15007. Kolsch, V., Seher, T., Fernandez-Ballester, G.J., Serrano, L., and Leptin, M. (2007). Science 315, 384–386. Leptin, M. (1999). EMBO J. 18, 3187–3192. Martin, A.C., Kaschube, M., and Wieschaus, E.F. (2008). Nature, in press. Published online November 23, 2008. 10.1038/nature07522. Seher, T.C., Narasimha, M., Vogelsang, E., and Leptin, M. (2007). Mech. Dev. 124, 167–179.
Relaying the Checkpoint Signal from Kinetochore to APC/C Rene´ H. Medema1,* 1Department of Medical Oncology and Cancer Genomics Center, University Medical Center Utrecht, Str.2.118, Universiteitsweg 100, 3584 CG Utrecht, the Netherlands *Correspondence:
[email protected] DOI 10.1016/j.devcel.2008.12.008
The mitotic checkpoint delays chromosome segregation until the last chromosome has correctly attached to the spindle. Exactly how this unattached chromosome can generate a checkpoint signal and inhibit the anaphase promoting complex/cyclosome (APC/C) is unknown. Two Developmental Cell papers in this issue by Kulukian et al. and Malureanu et al. now provide insight into how checkpoint components Mad2 and BubR1 relay the checkpoint signal from kinetochores to APC/C.
A normal eukaryotic cell will not segregate its duplicated sister chromatids until the very last one of them properly attaches
to the mitotic spindle (Rieder et al., 1994), because unattached kinetochores generate a checkpoint signal that
6 Developmental Cell 16, January 20, 2009 ª2009 Elsevier Inc.
prevents anaphase onset (Rieder et al., 1995). These studies suggested that unattached kinetochores serve as a platform