Dispatch R705

Morphogenesis: Unravelling the cell biology of hole closureAntonio Jacinto and Paul Martin

Recent studies show that the underlying amnioserosaworks actively to help bring together the two epithelialsheets and close the embryonic hole during dorsalclosure in fruitfly development.

Address: Department of Anatomy and Developmental Biology,University College London, Gower Street, London WC1E 6BT, UK.E-mail: paul.martin@ucl.ac.uk

Current Biology 2001, 11:R705–R707

0960-9822/01/$ – see front matter © 2001 Elsevier Science Ltd. All rights reserved.

It is not a good thing to have a hole in your epithelium. Inthe adult, such a defect is probably the result of someinjury and a tissue repair cascade will rapidly kick-in toheal this wound [1]. In the embryo, these holes can be theconsequence of a series of morphogenetic movements — aclassic example of this is the gaping epithelial hole on thedorsal aspect of a Drosophila embryo at the tail end ofembryogenesis. One of the last acts of fly embryogenesis isto seal up this hole, drawing lateral epithelium from thetwo sides of the embryo up and over the exposed extraem-bryonic amnioserosa to form a neat and subsequentlyinvisible midline seam where the two segmented epithe-lial edges meet one another [2]. Two new studies [3,4] —one published recently in Current Biology [4] — haveshown that the amnioserosa is far from being just a passivesubstratum during this hole closure process: it appears toactively direct epithelial advancement, while also playing atugging role to draw the epithelium forward.

Much effort has been put into the genetic dissection of thefascinating morphogenetic process of dorsal closure [5].Upwards of 30 mutant fly embryos fail in some way indorsal closure, leading to various shapes and sizes of holein cuticle preparations of the resulting larvae — generallythey have names that illustrate the defect graphically, suchas kayak, basket, canoe and coracle. These results haverevealed the importance of several signalling cascades indirecting dorsal closure, as well as the likely structural andmotor components of the process. For example, severalsmall GTPase molecular switches appear to play an essen-tial role, as mutants or transgenic flies expressing domi-nant-negative forms of Rho, Rac or Cdc42 all fail toproperly close the hole [6–10], while zipper mutants, whichare null for non-muscle myosin, are so named becausethey struggle to zipper the epithelial faces together [11].

Until recently, it has been assumed that the key tissueplayer in dorsal closure is the epithelium, which has a cableof actin in its leading edge [11] and dynamic filopodial

processes extending from front row cells which are criticalfor zippering up the epithelium at anterior and posteriorends of the hole [12]. The amnioserosa was somewhatignored, and simply considered a passive substratum duringthe whole process. But, like all biology, the deeper you digthe more complexity you discover. Analogies with woundhealing, which is clearly driven partially by re-epithelialisa-tion and partially by wound-bed connective tissue contrac-tion, should have suggested that the amnioserosa might beplaying a more active role, and indeed recent studies [3,4]have now shown that it almost certainly is.

Video analysis of the amnioserosa during the dorsal closureperiod has revealed how individual cells within this sheetshrink their apical surface with a contraction bias in thedorsoventral axis. In fact, some of the cells contract somuch that they appear to drop down out of the focal planeof the amnioserosal sheet (Figure 1). Transmission electronmicroscopy studies have confirmed how amnioserosa cellsconstrict their apices massively, particularly towards theend of dorsal closure [13], but of course these descriptivedata do not formally rule out that the amnioserosa is pas-sively being squeezed by the advancing epithelial sheets.

The best evidence that the amnioserosa is exerting sometension that might aid in drawing the adjacent epithelialsheets over it comes from the elegant laser ablationexperiments carried out in Dan Kiehart’s lab [3]. In thisapproach, small holes are created by laser ablation either inthe amnioserosa or in the adjacent lateral epithelium,locally releasing any potential tension within those tissues.A hole in the amnioserosa leads to gaping of the adjacentepithelium, which is very suggestive that the amnioserosais an active player in the closure process. Moreover, a holejust back from the leading edge of the lateral epitheliumresults in rapid advancement, rather than retraction, of thatregion of leading edge, ruling out the possibility ofpushing forces from the epithelium. These observationsnow make sense of earlier studies showing that mutantembryos in which the amnioserosa failed to form properlyor died early were defective in germ-band retraction anddorsal closure [14].

In the enlightened knowledge that dorsal closure is in factthe result of two tissue movements acting in concert, itbecomes critical to figure out how these two tissuescoordinate with one another. It seems that the epitheliumadvances forwards over a contracting substrate, just like aperson walking slowly up a moving escalator. How are thetwo tissues physically linked? And what signals passbetween the tissues as they ratchet forward? Clearly any

R706 Current Biology Vol 11 No 17

tethering links between the advancing epithelium and thecontracting amnioserosa must be dynamic. These adhe-sions may be at the tips of filopodia and lamellipodial pro-trusions from the front row epithelial cells, but as theseprotrusions appear to touch down transiently on theamnioserosa ahead of them [12], it seems more likely thatfirmer links exist further back between the body of thefront-row cells and the underlying amnioserosal cells.Immunostaining and green fluorescent protein (GFP)

fusion protein techniques for visualising cell–cell adhesioncomponents (such as α-catenin), and cell–matrix adhesionreceptors (the integrins), will reveal whether one or both ofthese classes of adhesion molecules are present at the levelof the actin cable of front-row cells. Either adhesion strat-egy might allow the epithelium to stay linked to, but alsomove forward over, a moving amnioserosal substratum.

Much is already known about the signalling cascadesoccurring within the front few rows of the advancing epithe-lia, and how these signals might direct the shape changesand motility machinery of the epithelial cells [5]. But in arecent paper from Lipshitz and colleagues [4] we are nowtold information about key signalling events that occurwithin the amnioserosa also. It is well established that acti-vation of the Jun N-terminal kinase (JNK) cascade withinleading epithelial cells plays some role in driving dorsalclosure. But it seems that the same cascade, or moreimportantly, shutting down of the same cascade, withinamnioserosal cells is also critical.

The new data [4] show that hindsight, which encodes a zincfinger transcription factor, is responsible for repressing JNKactivity in the amnioserosa, resulting in switching off of atleast two known JNK effectors, the transforming growthfactor β (TGFβ)-related diffusible factor Decapentaplegic(Dpp), and a phosphatase, Puckered (Puc). In hindsightmutants, where JNK signalling is not repressed, dorsalclosure halts. The authors speculate that this failure indorsal closure is because the amnioserosa is now emittingprecocious ‘stop’ signals to adjacent epithelial cells, andindeed they show evidence for disruption of the normalactin processes in leading epithelial cells in these mutants.But it could also be that the downregulation of JNK inamnioserosal cells is a necessary step in order for theamnioserosa itself to contract.

As with all new observations, these data raise several furtherquestions. Perhaps first off, it would be nice to know moreabout the contractility machinery of the amnioserosa andmore precisely how it interacts with the leading edgeepithelial cells. This will entail some ultrastuctural studiesand immunostaining for cytoskeleton and adherens junc-tion components, coupled with live analysis of transgenicembryos expressing, for example, GFP–actin or GFP–α-catenin. We also need to figure out precisely how theamnioserosa coordinates contraction of cells within itself —is there a diffusible signal that leads to synchronous con-traction, analogous to the folded gastrulation gene product[15], which coordinates the apical contractions of epithelialcells in the ventral furrow region during fly gastrulation?

Dorsal closure in Drosophila has now become one ofthe paradigms for studying morphogenetic movementsduring development. As such, it is a prime candidate for

Figure 1

(a,b) Two stills taken 20 minutes apart from a video of a Drosophilaembryo undergoing dorsal closure that expresses GFP–α-cateninfusion protein. The exposed amnioserosa cells are constricting as thelateral epithelial sheets are drawn over them by zippering from theanterior and posterior ends. Note the amnioserosa cells highlighted inred that are clearly contracting their apical surfaces. One of the markedcells shrinks dramatically as it leaves the plane of the amnioserosalsheet. (c) A schematic transverse section through the region of anembryo as indicated with dotted line in (b). The leading edge epithelialcells extend filopodia but are also adherent to the underlyingamnioserosa cells (pale green). As the epithelia advances forward theamnioserosa cells contract at their apices. (Adapted from [13].)

microarray studies which will reveal a fuller glossary ofthe genes involved in directing epithelial advancementand amnioserosal contraction. Soon we will know far moreof the genetic players involved in dorsal closure, and it willbe important that our cell biology studies keep pace withthe genomics. And maybe one day, when we understandhow a fly embryo closes a hole during development, wewill come up with fundamental and conserved mecha-nisms which will help us in the search for better therapiesto enhance the repair of skin wound holes in adult humansin the clinic!

AcknowledgementsThe authors thank Alfonso Martinez-Arias for his constant advice and support.

References1. Jacinto A, Martinez-Arias A, Martin P: Mechanisms of epithelial

fusion and repair. Nat Cell Biol 2001, 3:E117-123.2. Martinez-Arias A: Development and patterning of the larval

epidermis of Drosophila. In The Development of Drosophilamelanogaster. Edited by Bate AM-AaM. Cold Spring Harbor: ColdSpring Harbor Laboratory Press; 1993: 517-607.

3. Kiehart DP, Galbraith CG, Edwards KA, Rickoll WL, Montague RA:Multiple forces contribute to cell sheet morphogenesis for dorsalclosure in Drosophila. J Cell Biol 2000, 149:471-490.

4. Reed BH, Wilk R, Lipshitz HD: Down-regulation of Jun kinasesignaling in the amnioserosa is essential for dorsal closure of theDrosophila embryo. Curr Biol 2001, 24 July issue.

5. Stronach BE, Perrimon N: Stress signaling in Drosophila. Oncogene1999, 18:6172-6182.

6. Harden N, Loh HY, Chia W, Lim L: A dominant inhibitory version ofthe small GTP-binding protein Rac disrupts cytoskeletalstructures and inhibits developmental cell shape changes inDrosophila. Development 1995, 121:903-914.

7. Ricos MG, Harden N, Sem KP, Lim L, Chia W: Dcdc42 acts inTGF-beta signaling during Drosophila morphogenesis: distinctroles for the Drac1/JNK and Dcdc42/TGF-beta cascades incytoskeletal regulation. J Cell Sci 1999, 112:1225-1235.

8. Harden N, Ricos M, Ong YM, Chia W, Lim L: Participation of smallGTPases in dorsal closure of the Drosophila embryo: distinctroles for Rho subfamily proteins in epithelial morphogenesis.J Cell Sci 1999; 112:273-284.

9. Magie CR, Meyer MR, Gorsuch MS, Parkhurst SM: Mutations in theRho1 small GTPase disrupt morphogenesis and segmentationduring early Drosophila development. Development 1999,126:5353-5364.

10. Genova JL, Jong S, Camp JT, Fehon RG: Functional analysis ofCdc42 in actin filament assembly, epithelial morphogenesis, andcell signaling during Drosophila development. Dev Biol 2000,221:181-194.

11. Young PE, Richman AM, Ketchum AS, Kiehart DP: Morphogenesis inDrosophila requires nonmuscle myosin heavy chain function.Genes Dev 1993, 7:29-41.

12. Jacinto A, Wood W, Balayo T, Turmaine M, Martinez-Arias A, Martin P:Dynamic actin-based epithelial adhesion and cell matching duringDrosophila dorsal closure. Curr Biol 2000, 10:1420-1426.

13. Rugendorff A, Younossi-Hartenstein A, Hartenstein V: Embryonicorigin and differentiation of the Drosophila heart. Roux Arch DevBiol 1994, 203:266-280.

14. Lamka ML, Lipshitz HD: Role of the amnioserosa in germ bandretraction of the Drosophila melanogaster embryo. Dev Biol 1999,214:102-112.

15. Costa M, Wilson ET, Wieschaus E: A putative cell signal encodedby the folded gastrulation gene coordinates cell shape changesduring Drosophila gastrulation. Cell 1994, 76:1075-1089.

Dispatch R707