Bringing Classical Embryology to C. elegans Gastrulation

Developmental Cell 6

infected wild-type flies, these animals are still able to effectively kill the infecting pathogen. Thus, it is unclear what is actually causing these NO-inhibited animals to die. Perhaps some bacteria survive, in a protected niche, and then go one to cause lethality during pupation. Another possibility is that infection (or the immune response) might cause some irreparable damage, preventing further development. Curiously, if the Toll pathway is blocked (by mutation) and NO production is inhibited, naturally infected larvae are no longer able to kill the infecting bacteria. These results suggest that the Toll and IMD pathways function redundantly to maximize bactericidal activity in the gut. The role of NO in Toll signaling was also examined. NO inhibitors have only a mild inhibitory effect on the induction of Drosomycin, a target of the Toll pathway. On the other hand, NO donors can activate expression of Drosomycin, independent of the IMD pathway. This argues that NO can activate the Toll pathway, but the physiological significance of this activation is not yet clear. The discovery that immune activation in Drosophila requires NO raises a number of interesting questions. How does infection lead to NO production? In the insect immune response, does NO function by the activation of sGC, as it does in other mammalian and Drosophila systems? Or is immune signaling mediated by nitrosylation of other proteins? Lastly, it will be interesting to

Bringing Classical Embryology to C. elegans Gastrulation

In a recent paper, Lee and Goldstein develop an explant assay that recapitulates key aspects of gastrulation in C. elegans and permits classical embryological manipulations. The resulting detailed analysis of cell behavior will ultimately extend to broader issues, such as, whether morphogenesis can be described as the sum of single-cell events or if unique phenomena emerge at the multicellular level. For an understanding of multicellular rearrangements, such as gastrulation, neurulation, and organogenesis, it is important to understand the behaviors of individual cells. The dissection of complex morphogenetic processes has been facilitated by the development of simplified assays in which cell behavior can be directly observed and manipulated (e.g., Keller et al., 1991). A challenge now is to integrate these approaches into traditionally genetic model systems, such as Drosophila and C. elegans. In a recent paper published in Development, Lee and Goldstein extend the use of cultured explant assays in C. elegans to dissect the dynamics of cell movements during gastrulation (Lee and Goldstein, 2003). This approach will make it possible to assess how individual mechanisms, such as cell migration, contractility, and adhesion, work together to drive multicellular reorganization.

learn whether, in addition to its role in signaling, NO also functions as a microbicidal agent in Drosophila macrophages. Given the extensive genetic analysis of NO and immune signaling in Drosophila, we can anticipate that we will soon know more about NO. Neal Silverman Department of Medicine Division of Infectious Disease and Immunology University of Massachusetts Medical School 364 Plantation Road Worcester, Massachusetts 01605 Selected Reading Basset, A., Khush, R.S., Braun, A., Gardan, L., Boccard, F., Hoffmann, J.A., and Lemaitre, B. (2000). Proc. Natl. Acad. Sci. USA 97, 3376–3381. Bogdan, C. (2001). Nat. Immunol. 2, 907–916. Foley, E., and O’Farrell, P.H. (2003). Genes Dev. 17, 115–125. Gibbs, S.M., Becker, A., Hardy, R.W., and Truman, J.W. (2001). J. Neurosci. 21, 7705–7714. Lemaitre, B., Reichhart, J.M., and Hoffmann, J.A. (1997). Proc. Natl. Acad. Sci. USA 94, 14614–14619. Mannick, J.B., and Schonhoff, C.M. (2002). Arch. Biochem. Biophys. 408, 1–6. Silverman, N., and Maniatis, T. (2001). Genes Dev. 15, 2321–2342. Wingrove, J.A., and O’Farrell, P.H. (1999). Cell 98, 105–114.

The onset of C. elegans gastrulation is marked by the ingression of two gut precursor cells into the interior of the embryo, leaving a space at the surface that is filled in by neighboring cells (see Figure, panel A). By successfully culturing embryos after removal of the vitelline membrane, Lee and Goldstein demonstrate that the gut precursors, Ea and Ep, can still rearrange normally with respect to two cells, P4 and MSxx, that come to overly them. Remarkably, this shift in position still occurs in ablated embryos where more than half of the total cells are missing. These results demonstrate that gut cell internalization is not dependent on outside forces from the vitelline membrane or from several other cell types, including cells that move together with P4 and MSxx to close the gap left by Ea and Ep. Interestingly, in cases where the ablated explants adopt an aberrant linear orientation, P4 and MSxx can still reposition relative to Ea and Ep (see Figure, panel B). Therefore, the morphogenetic process of gut cell internalization in C. elegans can be reduced to the study of interactions among only four cells. This simplified explant assay enables the direct manipulation of cells. By removing cells or replacing them in altered orientations, the authors demonstrate that P4 and MSxx can move independently, indicating that they do not simply chemotax toward one another. These experiments also reveal an unexpected behavioral difference between P4 and MSxx: the orientation of P4 can influence its direction of movement with respect to MSxx, but not vice versa. Analysis of cell interactions in explants can thus mechanistically distinguish between

Previews 7

Early Gastrulation Movements in C. elegans (A) Illustration of the ingression of gut precursors, Ea and Ep, and the convergence of neighboring cells, MSxx and P4, in an embryo. (B) Simplified schematic to show the movement of just these four cells within an explant assay, in which more than half of the total cells have been removed.

movements that appear superficially equivalent in the intact embryo. Interestingly, membrane-labeling experiments also reveal that MSxx is not likely to advance using a fibroblast-type migratory mechanism. Instead, the authors propose that an actin/myosin-based apical constriction of Ea and Ep drives their internalization and pulls P4 and MSxx together. In support of this model, membrane labeling demonstrates an apical constriction in Ea and Ep. In addition, cell movements are blocked by pharmacological disruption of the actin cytoskeleton and inhibition of the myosin activator, myosin light chain kinase. Apical constriction is a common early step of cell ingression during gastrulation in many organisms. For example, the bottle cells that form the sea urchin vegetal plate and the Xenopus blastopore lip all apically constrict (Hardin and Keller, 1988; Kimberly and Hardin, 1998). In Drosophila and C. elegans, nonmuscle myosin accumulates at the apical surface of ingressing cells (Young et al., 1991; Nance and Priess, 2002), and activation of myosin contractility is required for cell internalization in C. elegans (Lee and Goldstein, 2003). These results raise the possibility that spatial regulation of myosin-based contractility may be a widely used mechanism for cell internalization. It will be interesting to determine how this spatial regulation is established and controlled. Cell internalization during gastrulation is likely to involve processes in addition to actin/myosin-based apical constriction. Xenopus bottle cells and Drosophila mesoderm precursors undergo considerable cell elongation along the apical/basal axis, and, in Drosophila, the basal surface subsequently expands more than 2-fold (Hardin and Keller, 1988; Leptin and Grunewald, 1990; Sweeton et al., 1991). During neurulation in vertebrates, cell shape changes are accompanied by extensive cell rearrangement and cell division (Schoenwolf and Smith, 1990), suggesting that multiple mechanisms may contribute to cell internalization. In addition to the apical constriction of Ea and Ep in C. elegans, membrane-tracking experiments suggest an active rolling movement of the neighboring MSxx cell, indicating that MSxx is not just being passively dragged. Furthermore, MSxx sometimes moves past the Ea/Ep boundary in explants, suggesting that additional forces can drive cell

translocation. Therefore, even simplified explants can provide enough complexity for an examination of the coordinated action of multiple mechanisms during morphogenesis. During cell rearrangement, where multiple mechanisms are operating and many cells change position, it is important to establish which cells are the active participants as well as which mechanism in an active cell provides the driving force. For example, it is unclear to what extent the apical constriction of Ea/Ep or the rolling of MSxx provides the driving force for cell rearrangement in C. elegans. It should now be possible to address these traditionally difficult questions by generating mosaic explants where individual cells are selectively modified by genetic or pharmacological techniques. It will also be important to determine whether the active influence is mechanical or signaling in nature, clues to which may be found by defining the molecular components involved. To address how the concerted action of individual cells can drive complex multicellular rearrangements, we must first understand the mechanisms that dictate the behavior of single cells. This understanding may ultimately reveal that the most useful way to describe a morphogenetic process may not be at the cellular level, but at the level of cell sheets or coherent groups of cells. The findings of Lee and Goldstein also highlight similarities between cell shape changes of two cells and the internalization of large cell populations during Drosophila and vertebrate gastrulation. Further comparison of cell behaviors in simplified assays to those in intact embryos or to related processes in different species will determine the extent of such similarities and may reveal mechanistic principles. This work offers hope for understanding not only the cellular mechanisms of morphogenesis, but also whether these mechanisms can account for the dynamics of larger cell populations. Rachel E. Dawes-Hoang, Jennifer A. Zallen, and Eric F. Wieschaus Department of Molecular Biology Princeton University Lewis Thomas Lab Washington Road Princeton, New Jersey 08544

Developmental Cell 8

Selected Reading

Nance, J., and Priess, J.R. (2002). Development 129, 387–397.

Hardin, J., and Keller, R. (1988). Development 103, 211–230.

Schoenwolf, G.C., and Smith, J.L. (1990). Development 109, 243–270.

Keller, R., Clark, W.H.J., and Griffin, F. (1991). Gastrulation: Movements, Patterns, and Molecules (New York: Plenum Press).

Sweeton, D., Parks, S., Costa, M., and Wieschaus, E. (1991). Development 112, 775–789.

Kimberly, E.L., and Hardin, J. (1998). Dev. Biol. 204, 235–250.

Young, P.E., Pesacreta, T.C., and Kiehart, D.P. (1991). Development 111, 1–14.

Lee, J.-Y., and Goldstein, B. (2003). Development 130, 307–320. Leptin, M., and Grunewald, B. (1990). Development 110, 73–84.

In Search of Lipid Translocases and Their Biological Functions

In plasma membranes, lipids distribute asymmetrically across the bilayer, a process that requires proteins. Recent work identified novel lipid translocators in yeast, and their activity was functionally correlated to endocytosis, thus boosting investigations on identity, mechanism, and function of lipid translocases.

The asymmetric distribution of lipids in plasma membranes (PM) of eukaryotic cells is well known. Whereas neutral phospholipids, like phosphatidylcholine (PC), sphingomyelin, and (glyco)sphingolipids (GSL), are largely localized in the outer leaflet, the aminophospholipids phosphatidylethanolamine (PE) and phosphatidylserine (PS) reside predominantly in the inner leaflet. Cholesterol is almost equally distributed over both leaflets and, in the outer leaflet, is often found in association with GSL, thereby forming microdomains (“rafts”). Mechanisms underlying the maintenance and/or regulation of this distinct lateral and transbilayer lipid distribution are still poorly understood. Because of thermodynamic constraints, the involvement of integral membrane proteins in translocating lipids across bilayers appears obvious. Potential lipid translocators include an ATP-dependent aminotranslocase (P-type ATPase), a specific inwarddirected pump for aminophospholipids; a lipid nonspecific Ca2⫹-dependent scramblase, which mediates bidirectional translocation; and members of the ATP binding cassette (ABC) transporter family, which mediate outward lipid migration. As these properties were mostly revealed through the use of fluorescent or spin-labeled lipids, direct demonstration of translocation of natural lipids is as yet poorly supported. Furthermore, incomplete protein purification and poor functional reconstitution has frustrated the functional identification of lipid translocases. Blood cells and, more recently, the genetically tractable yeast Saccharomyces cerevisiae are most often employed to study general principles of the mechanism and regulation of phospholipid translocation. In yeast, following their initial insertion in the outer PM leaflet, fluorescent NBD (nitrobenzoxa-diazole) derivatives of PE and PS, as well as PC, are transported across the

bilayer in a protein- and energy-dependent, but endocytosis-independent, manner (Grant et al., 2001). Interestingly, internalization of PC and PE is inhibited in anterograde sec mutants, suggesting that inward translocation requires continuous transport and recycling of the translocase(s) and/or accessory elements to the cell surface. Although it was questioned in early work (Chen et al., 1999), it now becomes apparent (Gomes et al., 2000; Pomorski et al., 2003) that Drsp2, a member of a subfamily of P-type ATPases, is one of the key players. Mutated drs2 generates an aminophospholipid transport-defective phenotype, but a plant homolog from Arabidopsis, ALA1, can complement for the deficiency in PS translocation (Gomes et al., 2000). Although mainly localized to the Golgi, Drs2p is rather dynamic and cycles between PM and late Golgi (Hua et al., 2002; Pomorski et al., 2003). In an elegant contribution, Pomorski et al. (2003) now identify two novel drs2-related P-type ATPases, Dnf1p and Dnf2p, as lipid translocators for PS, PE, and PC in the yeast PM. Overall, a dynamic picture emerges of cycling translocases in the endosomal (Dnf) and Golgi-PM secretory track (Drs2p) (Hua et al., 2002; Pomorski et al., 2003), rationalizing previous observations that secretory mutants display an inhibition of lipid translocation (Grant et al., 2001). Most importantly, Pomorski et al. show that, in dnf1/dnf2 deletion mutants, endogenous PE accumulates at the cell surface and further increases when drs2 is also deleted, suggesting that the ATPases may act in concert as a lipid translocation machinery. It is yet not yet clear where Drs2-mediated PE translocation activity is expressed, i.e., at the Golgi, at the PM, or at both. Neither can it be excluded that, in the drs2 mutant, vesiculation at the Golgi, which is facilitated by Drs2p, might have been impaired (Gall et al., 2002). As a result, PM-directed transport of accessory compounds of the translocation machinery, which may include “activator proteins” not related to ATPases, e.g., Rosp3, could have been impeded. Interestingly, the translocation capacity, but not the apparent affinity for either lipid class (i.e., PE, PC, and PS), of both translocators differed, with Dnf1p, relative to Dnf2p, displaying by far the highest activity. Although the noted lipid preference disqualifies these proteins as specific aminophospholipid translocases (but natural PC as substrate has yet to be determined), sphingolipids, phosphatidic acid, and phosphatidylglycerol are not substrates (Pomorski et al., 2003). Further work on regulation, including the potential homo- or heterooligomerization, will be important to solve issues as to why there should be multiple translocases, displaying similar affinities for