MICROSCOPY RESEARCH AND TECHNIQUE 44:293–303 (1999) Dynamic Control of Reversible Cell Adhesion and Actin Cytoskeleton in the Mouth of Beroë SIDNEY L. TAMM* Biology Department, Boston University, Boston, Massachusetts 02115 KEY WORDS dynamic tissue adhesion; actin-based appositions; ctenophores ABSTRACT Cell-cell adhesion in the various types of intercellular junctions of differentiated tissues is relatively stable and permanent. In migrating cells of embryos, or in wound closure, inflammatory responses and tumors of adult tissues, however, bonds between cells are made and broken and made again, i.e., cell-cell adhesions are transient and reversible. These nonjunctional contacts lack the organized structure of intercellular junctions, but may initiate their tissue-specific formation during development. Investigation of dynamic, nonjunctional cell-cell adhesions has been hampered by the asynchronous and heterogeneous distribution of these transient contacts among groups of moving cells. We recently discovered a novel system of reversible cell adhesion in a differentiated tissue that overcomes this difficulty. Here I review our current knowledge of this system, particularly its unique experimental advantages for investigating the mechanisms and control of dynamic cell adhesion. Microsc Res Tech 44:293–303, 1999. r 1999 Wiley-Liss, Inc. THE MOUTH OF BEROË Ctenophores (comb jellies) of the order Beroida are voracious predators of other gelatinous marine zooplankton, especially other kinds of ctenophores. Beroids in surface waters reach lengths of 30 cm, and have a large mouth and a voluminous stomach that occupies most of their miter- or cucumber-shaped body. Beroë actively seeks prey by swimming mouth forward, propelled by the beating of rows of giant ciliary comb plates (Fig. 1) (Tamm, 1982). The mouth remains closed, with the body shape streamlined, until the lips contact prey. Then the mouth opens suddenly and the stomach cavity rapidly expands to suck in the prey (Fig. 2) (Horridge, 1965a; Tamm, 1982). Complete engulfment of prey as big as the Beroë takes only a few seconds (Tamm, 1982; Tamm and Tamm, 1993a), facilitated by activation of tooth-like macrocilia lining the lips or stomodaeum (Horridge, 1965a; Tamm and Tamm, 1993b). The mouth quickly closes and re-seals after ingestion, and the bloated Beroë swims slowly away to digest its meal. MOUTH CLOSURE Beroid species with thin flexible body walls, flattened shape, and wide mouths (B. sp. Gloria, B. forskali, B. mitrata) keep their mouths closed by adhesion between paired strips of epithelial cells (Tamm and Tamm, 1991b, 1993a). In B. sp. Gloria (⫽B. ovata, G.R. Harbison, personal communication), the focus of our work, the epithelial adhesive strips run around the inside of opposing lips; in B. forskali and B. mitrata, the adhesive strips run longitudinally along the midline of the stomodaeum (Tamm and Tamm, 1991b). The two different orientations of the adhesive strips are correlated with different patterns of macrocilia inside the stomodaeum. That the epithelial strips serve as belt-like tissue fasteners can be easily demonstrated in Mg2⫹anaesthetized animals (which lack muscular and neur 1999 WILEY-LISS, INC. ral responses) by pulling apart the adherent lips or stomodaeal walls with forceps (Fig. 3) (Tamm and Tamm, 1991b). Sections through closed mouths shows that the singlelayered, flattened stomodaeal epithelium is markedly thicker in the adhesive strips, due to increased height (20–25 µm) of cells in this region (Fig. 4). The adhesive strips are about 200 µm wide, and consist of more than a hundred thousand cells per lip. The adherent epithelial cells of apposing lips are joined together by numerous close contacts between their surface membranes (Figs. 4–6). At these appositions, the plasma membranes of adjoining cells are highly folded and interdigitated, and run parallel with a uniform separation of approximately 15 nm (Fig. 6; Tamm and Tamm, 1991b). The intercellular space often contains dense flocculent material which sometimes appears periodically disposed (Fig. 6A,C). The apposed plasma membranes are lined by a dense cytoplasmic coat, 15–30 nm thick, which contains numerous 6–8 nm diameter microfilaments (Fig. 6C,D) (Tamm and Tamm, 1991b; 1993a). Rhodamine phalloidin fluorescence confirms that the dense cytoplasmic coat contains F-actin (Fig. 7A) (Tamm and Tamm, 1993a). The regions of close contact alternate with vacuolar intercellular spaces lacking filamentous membrane coats (Figs. 4, 5, 6A). A possible reason for these interruptions of the adhesive contacts will be discussed later. The two adherent epithelial strips thus appear structurally identical, and have a mirror-image symmetry about the plane of contact. Apposing strips appear to be functionally alike as well, and probably make equal Contract grant sponsor: NIH; contract grant number GM 45557. *Correspondence to: Sidney L. Tamm, Biology Department, Boston University, Boston, MA 02215 USA E-mail: email@example.com Received 17 March 1998; revision accepted 20 July 1998 294 S.L. TAMM Fig. 1. Beroë mitrata swimming mouth forward (to the left) with the lips closed into a wide curved slit. The body is flattened in the stomodaeal plane with a blunt front end and tapering rear. Longitudinal rows of ciliary comb plates propel the animal forward. Photograph by Dr. Claude Carré, Station Zoologique, Villefranche-sur-Mer, France. Natural size. Reprinted from Tamm SL and Tamm S. 199b. Reversible epithelial adhesion closes the mouth of Beroë, a carnivorous marine jelly. Biol Bull Mar Biol Lab Woods Hole 181:463–473 with permission from the publisher. Fig. 2. Beroë sp. Gloria (lower) eating a Mnemiopsis (upper). A. The wide lips of Beroë touch the lobes of Mnemiopsis. B. The mouth opens fully and the stomach cavity expands to begin ingestion. C. The Mnemiopsis is half swallowed and is visible inside Beroë’s stomach. D. The entire Mnemiopsis is engulfed and the bloated Beroë closes its mouth to swim away. Bar, 5 mm. Reproduced from video tape from Tamm SL and Tamm S. 199b. Reversible epithelial adhesion closes the mouth of Beroë, a carnivorous marine jelly. Biol Bull Mar Biol Lab Woods Hole 181:463–473 with permission. contributions to adhesion. This symmetric interaction is evident by forcibly pulling apart excised lips from two mouths, cutting each lip in half, and then testing for contra- vs. ipsilateral re-adhesion. After 10–15 minutes of contact, the two halves of one lip stick to each other as firmly as do lip halves from opposite sides of the mouth, showing that any part of the epithelium is similarly adhesive. The epithelial cells within each adhesive strip are linked laterally by adherens-like junctions encircling their apical ends. The junctional membranes are lined by dense cytoplasmic coats of microfilaments (Tamm and Tamm, 1993a; see also Hernandez-Nicaise et al, 1989). cells and/or cell extensions (DeRosier and Tilney, 1984; Elson, 1988; Janmey, 1991). The actin filaments underlying the plasma membranes of the adhesive cells probably serve to mechanically stabilize and strengthen the highly-folded interdigitating surfaces of these cells. Such reinforced interlocking of the cell cortices would provide a stable closure mechanism that fastens the paired epithelial strips together like a jigsaw puzzle, independent of tonic muscular activity. Indeed, Mg2⫹relaxed Beroë swim forward with their mouth tightly closed (Tamm and Tamm, 1991b). The adhesive strips are probably not held together by their interlocking geometry alone, but by specific molecular interactions as well. We recently tested the requirement for extracellular Ca2⫹ in epithelial adhesion to investigate Ca2⫹-dependent cell-cell adhesion molecules such as cadherins, selectins, and some integrins. In preliminary experiments we found that Ca2⫹- MECHANISM OF ADHESION Actin filaments provide mechanical stiffness and support for maintaining the asymmetric shape of many REVERSIBLE CELL ADHESION IN BEROË Fig. 3. Zone of adhesion between excised, Mg-relaxed lip pieces of B. sp. Gloria. The paired lips (left and right) are being pulled apart with forceps, and are joined symmetrically by the adhesive strips (arrowheads) on the stretched stomodaeal walls. The lips edges and macrociliary fields (M) are at the top; the stomodaeum (s) is down. Scale bar, 165 µm. Reproduced from video tape from Tamm SL and Tamm S. 199b. Reversible epithelial adhesion closes the mouth of Beroë, a carnivorous marine jelly. Biol Bull Mar Biol Lab Woods Hole 181:463–473 with permission. free artificial sea water (Ca2⫹-free ASW) and Ca2⫹-Mg2⫹free ASW caused excised adherent lips of B. sp. Gloria to open after 10–15 minutes treatment. Ca2⫹-Mg2⫹-free ASW was more effective in reducing adhesion than Ca2⫹-free ASW, which contains 100 mM Mg2⫹. Control lips in ASW remained tightly closed. When returned to ASW, the Ca2⫹-free ASW-separated lips re-sealed tightly in 20–30 minutes, but the Ca2⫹-Mg2⫹-ASW-opened lips only weakly re-sealed after 1 hour. Thus, epithelial cell adhesion requires extracellular Ca2⫹, which can be partly substituted for by high Mg2⫹ concentration. However, we have not yet checked the ultrastructure of Ca2⫹-free or Ca2⫹-Mg2⫹-free ASW-separated lips to rule out other, nonspecific effects of low divalent cations on the tissues. We also found that treatment of pulled apart lips with the lectin Concanavalin A prevented re-adhesion of epithelial adhesive strips, indicating that proteincarbohydrate binding between lectin domains and oligosaccharide chains of membrane glycoproteins may be involved in lip adhesion. The above results merely suggest that cadherins and/or selectins or even integrins could play a role in lip adhesion. Although homophilic binding by cadherins most readily fits the observed symmetry of cell adhesion between paired epithelial strips, intrinsically heterophilic interactions by selectins and/or integrins could, if populations of partners are mixed on both lips surfaces, render the adhesion functionally homophilic. The ultrastructure of the adhesive cell appositions in Beroë, together with the presence of submembrane actin filaments, more closely resembles cadherin or 295 Fig. 4. Toluidine blue-stained longitudinal thick sections (0.5 µm) through a closed mouth of Beroë sp. Gloria. A. Lip edges are at left, lined with macrocilia (M); stomodaeum (S) is to right. The stomodaeal walls are joined together by paired strips (cut transversely) of thickened epithelia (arrowheads) located inside the lips. B. Higher magnification of thickened, joined stomodaeal epithelial strips. Apposing cell surfaces make numerous close contacts interrupted by vacuolar intercellular spaces. Scale bar: A, 100 µm; B, 17 µm. Reproduced from Tamm SL and Tamm S. 199b. Reversible epithelial adhesion closes the mouth of Beroë, a carnivorous marine jelly. Biol Bull Mar Biol Lab Woods Hole 181:463–473 with permission. integrin-mediated contacts than selectin-based attachments of leucocytes to blood vessels. Selectin-ligand bridges between blood cells and endothelial cells are much greater than 10–20 nm (about 100 nm for P-selectin and PSGL-1 bonds), and a sub-membrane actin cytoskeleton is not obvious (Springer, 1995). Because integrins are more commonly found in cell-matrix adhesions (focal contacts) than in cell-cell adhesions (blood cell-endothelium), cadherins are presently our favorite candidate for possible adhesion molecules in Beroë lips. We plan to do further experiments on the molecular nature of lip adhesion. In particular, we will use pan-cadherin antibodies (Geiger, et al 1990) to look for cadherins in Beroë adhesive strips by immunofluorescence microscopy, and to test whether function-blocking cadherin antibodies and synthetic peptides prevent lip adhesion (Mege, et al 1992). In this regard, some cadherins can certainly function at the high salinity of sea water: two conventional cadherins have been identified in sea urchin embryos, and anti-cadherin antibodies perturb sea urchin development (Ghersi, et al 1993). Similar fluorescence and function-blocking/competing ligand studies will also probe possible involvement of selectins and integrins in lip adhesion. 296 S.L. TAMM Fig. 5. TEM of adherent epithelial strips in a closed mouth of B. sp. Gloria. The zone of adhesion runs diagonally from lower left to upper right (arrows), and consists of numerous close contacts of the cell surfaces alternating with vacular intercellular spaces (s). The plasma membranes at the appositions have a dense cytoplasmic coat. Thin longitudinal muscles (lm), here cut transversely, run next to the mesoglea. ⫻9300. Reproduced from Tamm SL and Tamm S. 199b. Reversible epithelial adhesion closes the mouth of Beroë, a carnivorous marine jelly. Biol Bull Mar Biol Lab Woods Hole 181:463–473 with permission. REVERSIBLE CELL ADHESION IN BEROË 297 MOUTH OPENING Contact of prey (other ctenophores) with any region of the lips of a searching Beroë triggers a local muscular separation of the lips, followed by rapid peeling apart of the adhesive strips and wide opening of the mouth by coordinated muscular activity. The entire response takes 0.2–0.3 second (Tamm and Tamm, 1993a). Not surprisingly, mouth opening involves active deadhesion of the epithelial strips themselves. TEM of mouths induced to open by prey (food-opened mouths) shows a dramatic change in the surfaces of separated adhesive strips (Tamm and Tamm, 1993a). No traces of the specialized appositions remain, nor is there any evidence of disruption or tearing apart of the cell surfaces. Instead, the plasma membranes of the adhesive cells appear uniformly smooth and intact without any sign of the filamentous cytoplasmic coats (Fig. 9). Rhodamine phalloidin staining is also missing in the separated adhesive strips of food-opened mouths (Fig. 7C) (Tamm and Tamm, 1993a). Thus, the interdigitated topography and submembranous actin coat of the cell contacts are rapidly lost— within 0.2–0.3 second—when the mouth opens to engulf prey. However, the actin-coated junctions encircling the apical sides of the epithelial cells remain intact after mouth opening, showing the selective nature of the control of cell adhesion and actin cytoskeleton in these cells (Fig. 9) (Tamm and Tamm, 1993a). In contrast to food-opened mouths, the adhesive cell appositions and submembrane actin coat do not disappear when the mouth is forcibly opened in the absence of prey. Instead, the still-adherent contacts are ripped off the cells, leaving remnants of intact appositions on either lip (Figs. 7B, 8) (Tamm and Tamm, 1993a). As noted above, pulled apart mouths or lips will gradually close and re-adhere within 10–15 minutes, showing the well-developed wound-healing and regenerative powers of ctenophores (cf. Coonfield, 1936). Therefore, disassembly of membrane appositions and actin coats in food-opened adhesive strips cannot be due simply to mechanical forces arising from mouth opening, but must be signaled by the animal itself, before muscular separation of the lips. DE-ADHESION AND MOUTH OPENING The rapid disappearance of the cortical actin cytoskeleton of adhesive cells upon food-induced mouth opening should weaken or collapse the highly-folded surface architecture of these cells. Withdrawal of mechanical support for the jigsaw-like interdigitations of adherent strips should diminish their binding and facilitate their separation when the lips are peeled open by muscular activity. Disassembly of the submembrane actin net- Fig. 6. Membrane appositions of adherent epithelial cells in closed mouths of B. sp. Gloria. The apposed plasma membranes run parallel to each other and are separated by a distance of about 15 nm, even when folded and interdigitated (A). The cytoplasmic side of the membranes is coated with dense material containing microfilaments, which are particularly evident in oblique or tangential views (C, D). The intercellular space is often filled with flocculent material that sometimes appears periodically arranged (A,C). A, ⫻108,400; B, ⫻93,400; C, ⫻86,700; D, ⫻69,000. Reproduced from Tamm SL and Tamm S. 199b. Reversible epithelial adhesion closes the mouth of Beroë, a carnivorous marine jelly. Biol Bull Mar Biol Lab Woods Hole 181:463–473 with permission. 298 S.L. TAMM work should also eliminate the cell binding activity of possible cadherins and integrins. Cadherin and integrinmediated cell adhesion requires linkage of cytoplasmic domains of these proteins, via attachment proteins, to actin filaments, and thus may be regulated by the state of the actin cytoskeleton (Nagafuchi and Takeichi, 1989; Takeichi, 1990; Alberts, et al 1994). Fig. 7. Rhodamine-phalloidin fluorescence of epithelial adhesive strips (as) on the inner surface of lips of B. sp. Gloria. The adhesive strips run diagonally from upper left to lower right in each image. A. From a closed mouth: the surface of the adhesive zone shows uniform, diffuse actin staining. Note fluorescence of belt-like circular muscles and narrow longitudinal muscles in the mesoglea. B. Pulled apart lip: large dark holes in the fluorescent adhesive strip correspond to ripped-off actin-coated appositions on the complementary lip (see Fig. 8). C. Food-opened lip: the uniform actin fluorescence has completely disappeared from the surface of the adhesive zone. Only the pattern of circular and longitudinal muscles remains in the background. Bar, 25 µm. Reproduced from Tamm SL and Tamm S. 199b. Reversible epithelial adhesion closes the mouth of Beroë, a carnivorous marine jelly. Biol Bull Mar Biol Lab Woods Hole 181:463–473 with permission. REGULATION OF TISSUE ADHESION How does prey trigger rapid de-adhesion of the epithelial strips and muscular opening of the mouth? We recently discovered a giant fiber nerve net that may mediate this response (Tamm and Tamm, 1995). TEM reveals a plexus of 6–8 µm diameter neurons (giant by ctenophore standards!) that underlies the epithelium of the adhesive strips (Figs. 10, 11), but not that of the general stomodaeum, and is apparently a regional specialization of the much finer stomodaeal nerve net. Individual giant axons can make synaptic contacts with both longitudinal muscles which run next to the mesoglea, and epithelial adhesive or gland cells (Fig. 10). This is the first example to our knowledge of a neuron contacting more than one type of effector cell. The synapses, identified by their ultrastructure, have the characteristic ‘‘chemical’’ transmission morphology of other presumed nervous elements in ctenophores (Fig. 11) (Hernandez-Nicaise, 1973). Feeding is initiated by contact of prey with mechanoand chemoreceptors on the outer edge of the lips (Tamm and Tamm, 1991b; 1993a). The presumed receptor cells bear bristle-like actin-filled projections, as well as cilia with unusual onion-like basal structures (HernandezNicaise, 1974; Horridge, 1965b; Tamm and Tamm, 1991a). These cells may therefore function as both mechanoreceptors and chemoreceptors, i.e., as double sensory receptors. The presumed receptor cells make synaptic contacts with small neurites of the ectodermal nerve net (Hernandez-Nicaise, 1974; Tamm and Tamm, 1991a). If these neurites are connected to the giant nerve net in the stomodaeal adhesive strips, then sensory receptors on the lips could initiate signals that are rapidly conducted by the giant fiber system to both epithelial cells and longitudinal muscles in the adhesive strips. A multi-effector giant fiber nerve net may thus serve as a final common pathway to rapidly signal de-adhesion of the epithelial strips as well as contractions of longitudinal muscles underlying the adhesive strips, thereby enabling Beroë to open its mouth quickly to engulf prey. How might nervous signals trigger disappearance of the actin-based, interlocking contacts of adhesive cells? Since the giant neurons synapse onto the bases of the epithelial cells, a signal (or signals) must travel 20–25 µm to the apical cell surface and cause disassembly of the appositions within 0.2–0.3 seconds. The remarkable speed of this process requires an electrical response (active or electrotonic membrane depolarization) of the adhesive cells. The subsequent intracellular messengers and pathways that mediate the rapid changes in cell-cell contacts and adhesion are unknown. Nor do we know whether disappearance of the cortical actin cytoskeleton is due to complete depolymerization of actin filaments, or rapid turnover and remodeling into a different pattern elsewhere in the cell. REVERSIBLE CELL ADHESION IN BEROË 299 Fig. 8. TEM of a pulled apart adhesive strip, showing intact appositions (paired arrowheads), vacuolar spaces and cytoplasmic remnants (cr) ripped off the missing strip. ⫻15,100. Inset: Higher magnification of a ripped-off apposition on a pulled-apart strip, showing the dense cytoplasmic coats of the apposed plasma mem- branes (paired arrowheads). ⫻43,300. Reproduced from Tamm SL and Tamm S. 199b. Reversible epithelial adhesion closes the mouth of Beroë, a carnivorous marine jelly. Biol Bull Mar Biol Lab Woods Hole 181:463–473 with permission. Actin filament-severing proteins of the gelsolin and ADF (actin-depolymerizing factor)/cofilin families are known to promote rapid actin filament turnover and reorganization in many cellular responses to a wide variety of stimuli and cell signaling molecules (Moon and Drubin, 1995; Puius, et al 1998; Welch, et al 1997). These actin-binding proteins serve as stimulus-responsive modulators of actin dynamics, and are therefore prime candidates for mediating the rapid loss of actinsupported appositions of Beroë adhesive cells in response to prey. Gelsolin’s actin-severing activity is activated by Ca2⫹ (Yin and Stossel, 1979). Ca2⫹ influx through voltagedependent Ca2⫹ channels, triggered by membrane depolarization, can induce gelsolin-mediated actin depolymerization in neurons (Bernstein and Bamburg, 1985; Furukawa, et al 1997; Neely and Gesemann, 1994). ADF/cofilin is functionally regulated by pH, polyphosphoinositides, and phosphorylation (Aizawa, et al 1997; Moon and Drubin, 1995). In the latter case, stimulusevoked Ca2⫹ influx, via calmodulin-dependent protein phosphatase, can induce rapid dephosphorylation of ADF/cofilin, thereby activating it and causing rapid reorganization of the actin cytoskeleton (Meberg, et al 1998; Moon and Drubin, 1995). Activation of ADF/cofilin is not only temporally regulated by signaling pathways, but spatially regulated as well. Dephosphorylated active ADF/cofilin somehow translocates to appropriate regions of the cell where it promotes actin filament turnover and remodeling (Abe, et al 1996; Nebl, et al 1996; Samstag, et al 1994; Suzuki, et al 1995). In cAMP-induced chemotaxis of Dictyostelium, GFP-cofilin redistributes into new pseudopods in 30–60 seconds (Aizawa, et al 1997). In stimulusactivated neutrophil-like HL-60 cells, cofilin is translocated from the cytosol to the plasma membrane of the leading edge in less than 30 seconds (Suzuki, et al 1995). In Beroë, giant neuron synaptic activity at the bases of adhesive cells might trigger depolarization and opening of voltage-gated Ca2⫹ channels at the apical surfaces, perhaps in the vacuolar intercellular spaces surrounding the appositions. Ca2⫹ influx from the vacuolar fluid (sea water?) could activate gelsolin and/or ADF/cofilin, with or without subsequent translocation, leading to rapid loss of the actin filament coat in the appositions. A major difference, however, between previously described changes in the actin cytoskeleton and that of Beroë adhesive contacts is speed of the response. The disassembly of the submembranous actin coat and cell 300 S.L. TAMM Fig. 9. TEM through the apical end of a food-opened adhesive strip. The appositions have disappeared, and the plasma membrane is smooth and unfolded without a dense cytoplasmic coat. Encircling belt junctions between epithelial cells remain, however (arrowheads). ⫻23,500. Reproduced from Tamm SL and Tamm S. 199b. Reversible epithelial adhesion closes the mouth of Beroë, a carnivorous marine jelly. Biol Bull Mar Biol Lab Woods Hole 181:463–473 with permission. appositions in food-opened epithelial adhesive strips must occur before, or during muscular separation of the lips, i.e., in less than 0.2–0.3 seconds. This is approximately 100-fold faster than the most ‘‘rapid’’ or ‘‘dynamic’’ reorganizations of actin cytoskeletons reported so far. The question is, does Beroë use novel processes to regulate its actin cytoskeleton and cell-cell adhesion so quickly? We hope to find the answer. no physiological evidence yet that it signals deadhesion. As noted above, the interlocking close appositions of the adhesive cells are not continous, but are interrupted by irregular intercellular gaps whose membranes are not lined by dense actin coats. These vacuolar spaces probably do not contribute to adhesion of the epithelial strips. Why are they present? The intercellular spaces may serve to increase the surface area of the paired adhesive strips without unduly increasing their adhesive strength. Wider strips would be more likely to overlap and reestablish contact after the mouth of a bloated Beroë closes over prey. In fact, sections through closed mouths often show a slight offset or mismatch between opposing adherent strips (Fig. 4; Tamm and Tamm, 1991b). If the adherent surfaces were continuous without intervening spaces, the strips themselves would be much narrower, and more likely to miss one another upon closure of the stretched stomach walls after ingestion. REVERSIBILITY AND RE-ADHESION After engulfing prey, Beroë’s mouth closes and the lips re-seal. We have not yet examined the process of re-adhesion of food-opened epithelial strips, so we do not know how the submembranous actin cytoskeleton and interdigitating surface appositions reform. Formation of the elaborately sculptured, interlocking surfaces of apposed adhesive cells may be driven by localized polymerization of membrane-associated actin. It is well known that polymerization and/or remodeling of actin networks can cause changes in cell shape and architecture (Tilney and Inoue, 1982; Welch, et al 1997; Wyman, et al 1990). It will be informative to test agents that inhibit actin polymerization (cytochalasin D, latrunculin) on reformation of actin-based appositions after feeding. Whether the giant neuron system is involved in signaling re-adhesion remains unknown, but we have SIGNIFICANCE We may ask first, not what Beroë’s epithelial strips do for understanding the dynamics of cell adhesion and actin cytoskeleton, but what do they do for Beroë? At the moderate Reynolds number of a cruising Beroë, the streamlined body shape should reduce drag REVERSIBLE CELL ADHESION IN BEROË 301 Fig. 10. Giant neuron lattice (gn) underlying the epithelial adhesive strip. The neuronal branches are continuous and filled with clear vesicles and parallel arrays of microtubules which diverge at intersec- tions to follow the branches. Neuronal branches running perpendicular to longitudinal muscles (lm) interrupt the fibers. Note synapses onto adhesive cells and muscles (arrowheads). Scale bar, 1 µm. and save energy. In thin-walled beroids with stomodaeal adhesive strips, mouth closure does not require muscular or neural activity. Epithelial adhesion thus seems a useful and efficient method for closing the mouth and streamlining the body of an active gelatinous predator that spends most of its time swimming mouth forward in search of prey. In contrast, beroids with a thicker body wall, cucumber-shaped body, and smaller mouth (B. cucumis, B. gracilis) do not have stomodaeal adhesive strips. Neither the lips nor the stomach walls are fastened together in any manner (Tamm and Tamm, 1991b). Evidently, the thicker, firmer body and smaller oral opening provide sufficient resistance to maintain normal body shape during forward swimming, without the need for epithelial adhesion. What Beroë’s mouth can do for us as cell biologists is more relevant here. Reversible tissue adhesion in Beroë shares many structural and functional properties with transient adhesions made between migrating cells in developing embryos of higher animals, wound closure after injury in adult tissues, inflammatory responses, and tumor growth. In all these cases, cell contacts generally do not involve formation of structurally differentiated intercellular junctions, but instead employ nonjunctional appositions, in which adjoining plasma membranes come close together and run parallel, separated by a 10–20 nm space. Submembrane actin filament networks are a common feature in such appositions (Alberts, et al 1994; Heaysman and Pegrum 1973a, 1973b). ‘‘This type of nonjunctional contact may be optimal for cell locomotion—close enough to give traction but not tight enough to immobilize the cell’’ (Alberts, et al 1994, p. 971). Because the apposed cell membranes of Beroë adhesive strips and moving cells in other systems are not structurally bound to one another, ‘‘bonds between the cell surfaces in such appositions would be more readily made and broken and remade again than in junctions’’ (Trinkaus, 1984, p. 171). These nonjunctional cell-cell contacts are thus ‘‘a prime candidate for the kind of adhesion that moving cells require’’ (Trinkaus, 1984, p. 171). Similarly, such appositions would seem well-designed for the reversible type of tissue adhesion required to close and open the mouth of Beroë. Unlike the other systems of transient cell-cell adhesions cited above, however, the epithelial strips of Beroë involve thousands of cells all doing the same thing— adhering or de-adhering—at the same time in response to readily controllable external stimuli. The same re- 302 S.L. TAMM Fig. 11. Multiple synapses (arrowheads) of a giant neuron (gn) onto neighboring adhesive cells (ac) wedged between longitudinal muscles (lm). Note the characteristic ‘‘presynaptic triad’’ composed of a single layer of synaptic vesicles at the cleft, a flattened sac of smooth ER, and one or more closely apposed mitochondria; arrowheads indicate synaptic polarity (see Hernandez-Nicaise, 1973). Scale bar, 1 µm. Reproduced from Tamm SL and Tamm S. 199b. Reversible epithelial adhesion closes the mouth of Beroë, a carnivorous marine jelly. Biol Bull Mar Biol Lab Woods Hole 181:463–473 with permission. gions of the same cells continually and reversibly make and break adhesive contacts. De-adhering epithelial strips show the fastest change in actin-based cell-cell adhesion yet reported. The mouth of Beroë therefore offers unique experimental advantages for investigating the mechanisms and control of dynamic cell adhesions and ultrarapid remodeling of the actin cytoskeleton. Once again, a marine invertebrate provides a model system for analyzing general biological problems. Bernstein BW, Bamburg JR. 1985. Reorganization of actin in depolarized synaptosomes. J Neurosci 5:2565–2569. Coonfield BR. 1936. Regeneration in Mnemiopsis leidyi, Agassiz. Biology Bulletin of the Marine Biology Laboratory at Woods Hole 71:421–428. DeRosier DJ, Tilney LG. 1984. The form and function of actin. A product of its unique design. In: Shay JW, editor. Cell and Muscle Motility. New York: Plenum Press. Vol. 5:139–169. Elson EL. 1988. Cellular mechanics as an indicator of cytoskeletal structure and function. Ann Rev Biophys Biophys Chem 17:397– 430. Furukawa K, Fu W, Li Y, Witke W, Kwiatkowski DJ, Mattson MP. 1997. The actin-severing protein gelsolin modulates calcium channel and NMDA receptor activities and vulnerability to excitotoxicity in hippocampal neurons. J Neurosci 17:8178–8186. Geiger B, Volberg T, Ginsburg D, Bitzur S, Sabanay I, Hynes RO. 1990. Broad spectrum pan-cadherin antibodies reactive with the C-terminal 24 amino acid residues of N-cadherin. J Cell Sci 97:607– 614. Ghersi G, Salamone M, Dolo V, Levi G, Vittorelli ML. 1993. Differential expression and function of cadherin-like proteins in the sea urchin embryo. Mech Develop 41:47–55. Heaysman JEM, Pegrum SM. 1973a. Early contacts between fibroblasts. Exp Cell Res 78:71–78. Heaysman JEM, Pegrum SM. 1973b. Early contacts between normal fibroblasts and mouse sarcoma cells. Exp Cell Res 78:479–481. Hernandez-Nicaise M-L. 1973. The nervous system of ctenophores. III. Ultrastructure of synapses. J Neurocytol 2:249–263. Hernandez-Nicaise M-L. 1974. Ultrastructural evidence for a sensorymotor neuron in Ctenophora. Tissue & Cell 6:43–47. ACKNOWLEDGEMENTS This review is dedicated to G.A. Horridge, a Dionysian and pioneer of structure and function in nervous systems of invertebrates. REFERENCES Abe H, Obinata T, Minamide LS, Bamburg JR. 1996. Xenopus laevis actin-depolymerizing factor/cofilin: a phosphorylation-regulated protein essential for development. J Cell Biol 132:871–885. Aizawa H, Fukui Y, Yahara I. 1997. Live dynamics of Dictyostelium cofilin suggests a role in remodeling actin latticework into bundles. J Cell Sci 110:2333–2344. Alberts B, Bray D, Lewis J, Raff M, Roberts K, Watson JD. 1994. Molecular Biology of the Cell, (3rd edition). New York: Garland Publishing, Inc. REVERSIBLE CELL ADHESION IN BEROË Hernandez-Nicaise M-L, Nicaise G, Reese TS. 1989. Intercellular junctions in ctenophore integument. In: PAV Anderson, editor. Evolution of the First Nervous Systems. New York: Plenum Press, p 21–32. Horridge GA. 1965a. Macrocilia with numerous shafts from the lips of the ctenophore Beroe. Proc R Soc Lond B 162:351–364. Horridge GA. 1965b. Non-motile sensory cilia and neuromuscular junctions in a ctenophore independent effector organ. Proc R Soc Lond B 162:333–350. Janmey PA. 1991. Mechanical properties of cytoskeletal polymers. Curr Opin Cell Biol 2:4–11. Meberg PJ, Ono S, Minamide LS, Takahashi M, Bamburg JR. 1998. Actin depolymerizing factor and cofilin phosphorylation dynamics: response to signals that regulate neurite extension. Cell Mot Cytoskel 39:172–190. Mege RM, Goudou D, Diaz C, Nicolet M, Garcia L, Geraud G, Rieger F. 1992. N-cadherin and N-CAM in myoblast fusion: compared localization and effect of blockade by peptides and antibodies. J Cell Sci 103:897–906. Moon A, Drubin DG. 1995. The ADF/cofilin proteins: stimulusresponsive modulators of actin dynamics. Mol Biol Cell 6:1423– 1431. Nagafuchi A, Takeichi M. 1989. Transmembrane control of cadherinmediated cell adhesion: a 94 kDa protein functionally associated with a specific region of the cytoplasmic domain of E-cadherin. Cell Regul 1:37–44. Nebl G, Meuer SC, Samstag Y. 1996. Dephosphorylation of serine 3 regulates nuclear translocation of cofilin. J Biol Chem 271:26276– 26280. Neely MD, Gesemann M. 1994. Disruption of microfilaments in growth cones following depolarization and calcium influx. J Neurosci 14:7511–7520. Puius YA, Mahoney NM, Almo SC. 1998. The modular structure of actin-regulatory proteins. Curr Opin Cell Biol 10:23–34. Samstag Y, Eckerskorn C, Wesselborg S, Henning S, Wallich R, Meuer SC. 1994. Costimulatory signals for human T-cell activation induce nuclear translocation of pp19/cofilin. Proc Natl Acad Sci USA 91:4494–4498. Springer TA. 1995. Traffic signals on endothelium for lymphocycte 303 recirculation and leucocyte emigration. Annu Rev Physiol 57:827– 872. Suzuki K, Yamaguchi T, Tanaka T, Kawanishi T, Nishimaki MT, Yamamoto K, Tsuji T, Irimura T, Hayakawa T, Takahashi A. 1995. Activation induces dephosphorylation of cofilin and its translocation to plasma membranes in neutrophil-like differentiated HL-60 cells. J Biol Chem 270:19551–19556. Takeichi M. 1990. Cadherins: a molecular family important in selective cell-cell adhesion. Annu Rev Biochem 59:237–252. Tamm SL. 1982. Ctenophores. In: GAB Shelton, Editor. Electrical Conduction and Behaviour in ‘Simple’ Invertebrates. Oxford: Oxford Univ Press, p 266–358. Tamm S, Tamm SL. 1991a. Actin pegs and ultrastructure of presumed sensory receptors of Beroë (Ctenophora). Cell Tiss Res 264:151–159. Tamm SL, Tamm S. 1991b. Reversible epithelial adhesion closes the mouth of Beroë, a carnivorous marine jelly. Biol Bull Mar Biol Lab Woods Hole 181:463–473. Tamm SL, Tamm S. 1993a. Dynamic control of cell-cell adhesion and membrane-associated actin during food-induced mouth opening in Beroë. J Cell Sci 106:355–364. Tamm SL, Tamm S. 1993b. Diversity of macrociliary size, tooth patterns, and distribution in Beroë (Ctenophora). Zoomorphol 113: 79–89. Tamm S, Tamm SL. 1995. A giant nerve net with multi-effector synapses underlying epithelial adhesive strips in the mouth of Beroë (Ctenophora). J Neurocytol 24:711–723. Tilney LG, Inoue S. 1982. Acrosomal reaction of Thyone sperm. II. The kinetics and possible mechanism of acrosomal process elongation. J Cell Biol 93:820–827. Trinkaus JP. 1984. Cells into organs. The forces that shape the embryo, 2nd edition. Englewood Cliffs: Prentice-Hall. Welch MD, Mallavarapu A, Rosenblatt J, Mitchison TJ. 1997. Actin dynamics in vivo. Curr Opin Cell Biol 9:54–61. Wymann MP, Kernen P, Bengtsson T, Anderson T, Baggiolini M, Deranleau DA. 1990. Corresponding oscillations in neutrophil shape and filamentous actin content. J Biol Chem 265:619–622. Yin HL, Stossel TP. 1979. Control of cytoplasmic actin gel-sol transformation by gelsolin, a calcium-dependent regulatory protein. Nature 281:583–586.