Portal de Periódicos da CAPES

Putting on the Brakes

2000; Cell Press; Volume: 101; Issue: 7 Linguagem: Inglês

10.1016/s0092-8674(00)80879-x

1097-4172

Laura M. Machesky,

Hippo pathway signaling and YAP/TAZ

Cell migration is essential throughout the life of multicellular organisms, but is especially important during development. Two studies in this issue of Cell shed light on the molecular pathways coordinating cell motility and development. 2Bear J.E. Loureiro J.J. Libova I. Fässler R. Wehland J. Gertler F.B Cell. 2000; 101 (this issue,): 717-728Abstract Full Text Full Text PDF PubMed Scopus (365) Google Scholar and 1Bashaw G.J. Kidd T. Murray D. Pawson T. Goodman C.S Cell. 2000; 101 (this issue,): 703-715Abstract Full Text Full Text PDF PubMed Scopus (358) Google Scholar have challenged current ideas regarding the role of the Ena/VASP family of proteins in cell motility. Integration of signals when cells move is a very complex process. For example, the mechanisms by which neurons navigate to their correct targets via attractive and repulsive cues involve many receptors and require multiple decisions. Ena/VASP proteins have been implicated in axon guidance as well as fibroblast migration, platelet activation, and numerous other systems. Ena/VASP proteins appear to be multifunctional and bind to several targets, complicating attempts to pinpoint their exact role in cell motility. The large number of genetic, biochemical, and cell biological studies showing connections between Ena/VASP proteins and cell motility preclude a simple linear pathway from receptor to actin assembly or translocation. However, the two exciting studies featured in Cell this month point to a number of new considerations and force us to revise our models for Ena/VASP function in cells. They also make potentially important distinctions between cell translocation and cell motility. Although these two terms have been used interchangeably by some authors, for this review I will define motility as any activity that includes actin-based protrusion or shape change of the cell and translocation as directed migration resulting in displacement of the entire cell. The Ena/VASP family of proteins has been studied in many systems and appears to have a universal role in control of cell motility and actin dynamics. Ena (Enabled) was originally discovered as a genetic suppressor of mutations in Drosophila Abl (Abelson) tyrosine kinase. Mammals contain a family of Ena-related proteins, including VASP (vasodilator-stimulated phosphoprotein), EVL (Ena-VASP like), Mena (Mammalian Enabled), and Mena+, Mena++, and Mena+++ (three Mena splice variants; 4Gertler F.B. Niebuhr K. Reinhard M. Wehland J. Soriano P Cell. 1996; 87: 227-239Abstract Full Text Full Text PDF PubMed Scopus (541) Google Scholar). Together, Drosophila Ena and its mammalian homologs are commonly referred to as the Ena/VASP family. Members of this family share common sequence motifs (Figure 1), including an EVH1 N-terminal domain that binds to proteins containing a D/E FPPPPXD motif and targets family members to focal adhesions (15Niebuhr K. Ebel F. Frank R. Reinhard M. Domann E. Carl U.D. Walter U. Gertler F.B. Wehland J. Chakraborty T EMBO J. 1997; 16: 5433-5444Crossref PubMed Scopus (326) Google Scholar). The middle portions of the Ena/VASP family are more divergent, with Ena being glutamine rich and the mammalian homologs having cAMP and cGMP-dependent protein kinase sites. Nearer the C terminus on all Ena/VASP family proteins is a proline-rich sequence, which binds to the actin cytoskeletal protein profilin, SH3 domain containing proteins and WW domain containing proteins. The far C terminus of family members contains an EVH2 domain, which binds to F-actin in vitro and may also be important for multimerization (Figure 1 and 10Lanier L.M. Gertler F.B Curr. Opin. Neurobiol. 2000; 10: 80-87Crossref PubMed Scopus (175) Google Scholar). Ena/VASP proteins were first implicated directly in actin assembly in studies on the intracellular pathogenic bacterium Listeria monocytogenes. Listeria has been used extensively as a simplified model system for eukaryotic cell motility. It requires only one bacterial protein, ActA, to recruit host cytoskeletal proteins and assemble an actin tail to propel itself through the cytoplasm. In infected cells, Ena/VASP proteins bind directly to ActA and are recruited to the surface of the bacterium during actin tail assembly (15Niebuhr K. Ebel F. Frank R. Reinhard M. Domann E. Carl U.D. Walter U. Gertler F.B. Wehland J. Chakraborty T EMBO J. 1997; 16: 5433-5444Crossref PubMed Scopus (326) Google Scholar). Deletion of the region of ActA that binds to Ena/VASP proteins causes reduced speed and a reduced percentage of moving bacteria. This year, the minimal set of essential proteins from the host that are required for actin-based bacterial translocation has been identified (13Loisel T.P. Boujemaa R. Pantaloni D. Carlier M.F Nature. 1999; 401: 613-616Crossref PubMed Scopus (778) Google Scholar). Only three components in addition to actin are required: the actin nucleating Arp2/3 complex (for review see 14Machesky L.M. Insall R.H J. Cell Biol. 1999; 146: 267-272Crossref PubMed Scopus (212) Google Scholar), capping protein, and the actin depolymerizing protein ADF/cofilin (13Loisel T.P. Boujemaa R. Pantaloni D. Carlier M.F Nature. 1999; 401: 613-616Crossref PubMed Scopus (778) Google Scholar). In this reconstituted system, VASP enhanced translocation speed, but was not essential, in agreement with data from live infected cells. The mechanism of action of Ena/VASP proteins in Listeria translocation has been somewhat controversial. One widely held view is that the interaction of profilin-actin complexes with the proline-rich regions of Ena/VASP serves to shuttle actin monomers to the surface of the bacterium where new filaments are nucleated. This fits well with studies showing clearly that profilin is recruited to the Listeria surface and colocalizes with Ena/VASP. However, in human platelet extracts, a mutant profilin which has been reported not to bind to proline-rich sequences can still enhance the translocation of Listeria in a similar fashion to wild type. Furthermore, VASP still enhances translocation of Listeria in profilin-depleted extracts (11Laurent V. Loisel T.P. Harbeck B. Wehman A. Groebe L. Jockusch B.M. Wehland J. Gertler F.B. Carlier M.F J. Cell Biol. 1999; 144: 1245-1258Crossref PubMed Scopus (282) Google Scholar). Thus, at least in cell extracts, VASP and profilin can act independently of one another. It would be interesting to test whether Ena/VASP mutants that are unable to bind to profilin can still enhance Listeria translocation. Laurent et al. speculated that the ability of VASP to bind to actin filaments and thus connect the actin tail to the bacterium suggests that in uninfected cells, Ena/VASP proteins might connect newly assembling actin networks to the membrane/cortical areas. Clearly this is an area that needs more future study. In addition to an important role in Listeria translocation, some evidence suggests that Ena/VASP proteins could have a positive role in eukaryotic cell motility. In fibroblasts, targeting Mena+, the neural enriched variant of Mena, to the plasma membrane induced actin-rich protrusions. VASP is also concentrated at the leading edges of live cells in direct proportion to the speed of protrusion (17Rottner K. Behrendt B. Small J.V. Wehland J Nat. Cell Biol. 1999; 1: 321-322Crossref PubMed Scopus (257) Google Scholar) suggesting a role in promoting actin assembly and thus protrusion. However, this positive correlation between VASP and lamellipodial protrusion could also be consistent with a restrictive role for VASP. It would perhaps be fruitful to expand on the immunoelectron microscopy of Rottner et al. to determine the kinds of actin filament structures associated with Ena/VASP (e.g., actin bundles or Arp2/3 complex-containing actin branches; see 18Svitkina T.M. Borisy G.G J. Cell Biol. 1999; 145: 1009-1026Crossref PubMed Scopus (864) Google Scholar, 3Blanchoin L. Amann K.J. Higgs H.N. Marchand J.B. Kaiser D.A. Pollard T.D Nature. 2000; 404: 1007-1011Crossref PubMed Scopus (420) Google Scholar). In addition to a possible role for Ena/VASP proteins in actin assembly, there is also some evidence that they could be involved in cell–cell or cell–matrix adhesion. Ena/VASP proteins localize to focal adhesions, in addition to the cytoplasm and leading edges of cells. In keratinocytes, VASP and Mena are found in cell–cell adhesions and fragments of VASP thought to act as dominant negatives inhibit the junction assembly (19Vasioukhin V. Bauer C. Yin M. Fuchs E Cell. 2000; 100: 209-219Abstract Full Text Full Text PDF PubMed Scopus (913) Google Scholar). Perhaps the VASP and Mena are promoting actin assembly in junctions, as the authors observed filopodial protrusion between neighboring cells during junction formation. Previous studies using VASP null mice showed reduced cAMP and cGMP inhibition of collagen-induced platelet aggregation and increased activation of the αIIbβ3 integrin. This suggests a negative regulatory role for VASP in adhesion in platelets. Clearly there is some connection between Ena/VASP and adhesion, but it may depend on the cell type and context. 2Bear J.E. Loureiro J.J. Libova I. Fässler R. Wehland J. Gertler F.B Cell. 2000; 101 (this issue,): 717-728Abstract Full Text Full Text PDF PubMed Scopus (365) Google Scholar used molecular genetic techniques to determine the consequences of Ena/VASP depletion and overexpression on cell translocation. Overexpression of Ena/VASP slowed cells down to less than half of their wild-type speed. This was unexpected, given the role of Ena/VASP in facilitating Listeria actin tail assembly and translocation. The authors also used a targeting sequence from Listeria ActA to sequester all of the Ena/VASP on the mitochondrial surface, depleting Ena/VASP from the leading edges and focal adhesions. These cells moved faster than the wild-type cells. Reexpression of GFP-Mena slowed the translocation of mena−/− vasp−/− double null mouse cells (also apparently lacking EVL) providing elegant genetic support for the depletion and overexpression studies. To understand whether Ena/VASP at the leading edge of cells, in the cytoplasm, or in focal contacts was functioning in translocation, a competing FPPPP protein fragment was expressed in cells. This peptide removed Ena/VASP from focal adhesions but not from the leading edge or the cytoplasm. It had no detectable effect on cell translocation or protrusion. It is therefore likely that the Ena/VASP at the leading edge of cells is responsible for restricting translocation. This also agrees with data showing that targetting Ena/VASP to the plasma membrane with an FPPPP peptide containing a CAAX box slows cells and reduces membrane protrusion. The exact mechanism of reduced protrusion and slowed translocation is still unclear—it could be due to altered adhesion at the leading edge or effects on actin assembly/organization. Not only do Ena/VASP proteins restrict cell translocation in culture, but in both Drosophila and mouse, they appear to mediate some aspects of axon guidance. Axon guidance involves the neuronal growth cone using both repulsive and attractive cues to follow pathways to the correct target. 1Bashaw G.J. Kidd T. Murray D. Pawson T. Goodman C.S Cell. 2000; 101 (this issue,): 703-715Abstract Full Text Full Text PDF PubMed Scopus (358) Google Scholar now show that Ena interacts with the repulsive neuronal guidance receptor Robo and may at least partly mediate Robo signaling. Loss of Robo resulted in axons aberrantly crossing over the midline, indicating a loss of repulsion (7Kidd T. Brose K. Mitchell K.J. Fetter R.D. Tessier-Lavigne M. Goodman C.S. Tear G Cell. 1998; 92: 205-215Abstract Full Text Full Text PDF PubMed Scopus (700) Google Scholar). When Ena is mutated, the phenotype is much more complex, with general guidance defects, including some midline crossing defects, suggesting that binding to Robo may only be one part of Enas role. However, Ena mutants that are heterozygous for Robo lose midline repulsion and show severe defects in CNS axon guidance. Furthermore, reduction of Ena by 50% in slit,robo/+ transheterozygotes (lacking one copy of slit, one copy of robo and additionally lacking one copy of ena) also caused a large increase in inappropriate midline crossovers. Taken together, these results strongly suggest that Ena and Robo function together in the same signaling pathway. Evidence for a direct interaction between Ena and Robo is provided by in vitro and in vivo experiments. GST-fusion proteins of Ena or Robo can pull down in vitro–translated Robo or Ena, respectively. In embryonic extracts, antibodies to Robo coprecipitate Ena and myc-tagged Robo can precipitate Ena as well. Thus Robo and Ena physically interact. In the model in Figure 2, Ena is drawn as an effector downstream of Robo signaling via a direct interaction with Robo, which seems the simplest explanation for the elegant mix of genetic and biochemical studies of Bashaw et al. Robo is a member of a family of repulsive axon guidance receptors that respond to the slit proteins (Figure 2; 7Kidd T. Brose K. Mitchell K.J. Fetter R.D. Tessier-Lavigne M. Goodman C.S. Tear G Cell. 1998; 92: 205-215Abstract Full Text Full Text PDF PubMed Scopus (700) Google Scholar, 6Kidd T. Bland K.S. Goodman C.S Cell. 1999; 96: 785-794Abstract Full Text Full Text PDF PubMed Scopus (765) Google Scholar). Robo appears to be ubiquitous in higher organisms, with homologs in C. elegans, human, and rat (7Kidd T. Brose K. Mitchell K.J. Fetter R.D. Tessier-Lavigne M. Goodman C.S. Tear G Cell. 1998; 92: 205-215Abstract Full Text Full Text PDF PubMed Scopus (700) Google Scholar). The extracellular domains of Robo are well conserved, with five Ig motifs and three fibronectin III motifs (Figure 2; 7Kidd T. Brose K. Mitchell K.J. Fetter R.D. Tessier-Lavigne M. Goodman C.S. Tear G Cell. 1998; 92: 205-215Abstract Full Text Full Text PDF PubMed Scopus (700) Google Scholar) indicating that Robo could also have some connection with cell adhesion, although there is no experimental evidence to support this. The large cytoplasmic domain is less well-conserved among species except for three short proline-rich motifs (Figure 2; 7Kidd T. Brose K. Mitchell K.J. Fetter R.D. Tessier-Lavigne M. Goodman C.S. Tear G Cell. 1998; 92: 205-215Abstract Full Text Full Text PDF PubMed Scopus (700) Google Scholar). The Ena binding site is a consensus LPPPP motif (Figure 2) similar to Ena/VASP binding motifs in other proteins such as ActA, vinculin, and zyxin. Mutation of this motif in Robo weakens its repulsive output, providing further evidence that Ena acts as a Robo effector. The name Enabled comes from the original observation that it is in a pathway that opposes the Abelson tyrosine kinase. Mutants of the gene encoding Ena can suppress Abl mutants. The current study by Bashaw et al. shows that Abl antagonizes Robo signaling, in agreement with the concept that Ena and Abl function in opposition (Figure 2). Abl can also bind directly to Robo's cytoplasmic domain in GST-fusion protein pulldown experiments. A constitutively active human c-Abl variant induces tyrosine phosphorylation of hRobo1 (human Robo1) (Figure 2). Tyrosine phosphorylation of this residue likely suppresses Robo signaling in vivo, as a mutant Robo that cannot be tyrosine phosphorylated is hyperactive. While Figure 2 presents an attractive model, Bashaw et al. stress in their paper that in vivo levels of Robo tyrosine phosphorylation must still be established, as well as the effects of mutating abl or overexpressing Abl on tyrosine phosphorylation of endogenous Drosophila Robo. The role of Abl in cell motility is well established, but the molecular mechanisms may be multiple and are still unclear. Drosophila Abl can directly phosphorylate Ena, but this function does not seem to be conserved in mammalian Ena/VASP proteins, with the possible exception of Mena+ (4Gertler F.B. Niebuhr K. Reinhard M. Wehland J. Soriano P Cell. 1996; 87: 227-239Abstract Full Text Full Text PDF PubMed Scopus (541) Google Scholar). Human Abl can also directly bind to actin and bundle actin filaments, however, the in vivo significance or the relevance to Drosophila of this function has not been tested. However, the kinase activity of Abl appears to be required for its function in axon guidance. Abl is genetically linked to both cadherin/catenin adhesion and the Rho/Rac GEF Trio in Drosophila (10Lanier L.M. Gertler F.B Curr. Opin. Neurobiol. 2000; 10: 80-87Crossref PubMed Scopus (175) Google Scholar). This could strengthen the connection between Ena/VASP and Abl and adhesion, in agreement with another study showing that human c-Abl kinase activity is dependent on integrin-mediated cell adhesion (12Lewis J.M. Baskaran R. Taagepera S. Schwartz M.A. Wang J.Y Proc. Natl. Acad. Sci. USA. 1996; 93: 15174-15179Crossref PubMed Scopus (266) Google Scholar). In mammals, Abl has been reported to be activated (2- to 3-fold) downstream of activation of the PDGF and EGF receptors, both potent regulators of cell motility (16Plattner R. Kadlec L. DeMali K.A. Kazlauskas A. Pendergast A.M Genes Dev. 1999; 13: 2400-2411Crossref PubMed Scopus (357) Google Scholar). Although recent years have yielded many insights into the molecular basis for cell motility, this complex process still holds many secrets. Recent models have suggested that a major active signal to nucleate new networks of actin filaments comes from a pathway known as the WASP-Arp2/3 pathway (reviewed in 14Machesky L.M. Insall R.H J. Cell Biol. 1999; 146: 267-272Crossref PubMed Scopus (212) Google Scholar). The Ena/VASP proteins facilitate the formation of new networks through enhancement of filament elongation and filament capture (11Laurent V. Loisel T.P. Harbeck B. Wehman A. Groebe L. Jockusch B.M. Wehland J. Gertler F.B. Carlier M.F J. Cell Biol. 1999; 144: 1245-1258Crossref PubMed Scopus (282) Google Scholar). Now we must incorporate the idea that Ena/VASP proteins may actually slow eukaryotic cell translocation although they speed up Listeria translocation (1Bashaw G.J. Kidd T. Murray D. Pawson T. Goodman C.S Cell. 2000; 101 (this issue,): 703-715Abstract Full Text Full Text PDF PubMed Scopus (358) Google Scholar, 2Bear J.E. Loureiro J.J. Libova I. Fässler R. Wehland J. Gertler F.B Cell. 2000; 101 (this issue,): 717-728Abstract Full Text Full Text PDF PubMed Scopus (365) Google Scholar). How can we reconcile these apparently contradictory findings? First, it is probably much too simplistic to assume that increased actin polymerization correlates directly with increased translocation. For example, Dictyostelium cells show a maximum of assembled actin in response to a pulse of extracellular cAMP at a point when the cells are completely rounded up prior to polarization and increased translocation. This cringe response is not a migratory state, but the cells have up to 3-fold more F-actin than resting cells (5Hall A.L. Schlein A. Condeelis J J. Cell. Biochem. 1988; 37: 285-299Crossref PubMed Scopus (107) Google Scholar). Furthermore, many of the signaling molecules generally thought to trigger actin polymerization, such as the small GTPases Rac1 and RhoA, can trigger neurite retraction or even growth cone collapse. It is unclear whether a rise in F-actin accompanies a pause, retraction or even collapse, but this should be testable. Small local amounts of the actin polymerization blocking drug cytochalasin D can actually increase neurite extension, perhaps by allowing the microtubule cytoskeleton to penetrate the cell cortex and advance the leading edge (10Lanier L.M. Gertler F.B Curr. Opin. Neurobiol. 2000; 10: 80-87Crossref PubMed Scopus (175) Google Scholar). Clearly, the role of F-actin in cell translocation is complex, so simply equating F-actin amount or even turnover with rate of translocation is oversimplistic. Listeria cells appear to have a simple built-in polarity—so any increase in actin polymerization would be expected to enhance the translocation speed—but eukaryotic motile cells exhibit much more complex types of movement. Perhaps the shape and polarity of a eukaryotic cell may be a better indicator of its speed than the amount of filamentous actin or the rate of actin filament turnover. Rapidly moving fibroblasts or Dictyostelium cells most often show a very polarized phenotype—the cell is elongated with a clear uropod at the rear and a small but consistent lamellipodium at the front. Stationary cells or cells that are changing direction often show multiple pseudopodia and lamellipodia in all directions. This could be analogous to the pause and turn state of an axon during pathfinding. If the axon needs to make a decision, it may use Ena, whether downstream of the repulsive receptor Robo or other attractive/repulsive receptors, to enhance the production of actin-based filopodia and lamellipodia while it pauses and explores the environment. Furthermore, neurons could require actin assembly for both protrusion and retraction. Contractility, as induced by Rho GTPases, requires large actin-myosin networks, which are probably assembled in response to cues such as repulsion (8Kozma R. Sarner S. Ahmed S. Lim L Mol. Cell Biol. 1997; 17: 1201-1211Crossref PubMed Scopus (527) Google Scholar). Figure 3 shows a cartoon sketch of some different shapes of motile cells in what could be termed pause and explore versus rapid polarized translocation. It is easy to imagine that the actin polymerization requirements for pause and explore might be quite high and that a cell might concentrate its efforts into only a small leading edge area when it translocates in a single direction. If this model holds any truth, we might expect to see both repulsive and attractive receptors coupling to Ena and other parameters, such as the location of activation, to be variable. Another plausible way to think of Ena/VASP proteins as translocation restrictors is that the kinds of effects that Ena/VASP have on actin filaments may be to promote a stable network that enhances adhesion and/or pauses the leading edge. Both cell–cell and cell–substratum adhesion have the effect of slowing down cell translocation and both are likely to require actin polymerization. The effects of Ena/VASP proteins at the leading edges of cells might be to promote adhesion or alternatively to modify the structure of actin networks nucleated by the Arp2/3 complex. 9Lanier L.M. Gates M.A. Witke W. Menzies A.S. Wehman A.M. Macklis J.D. Kwiatkowski D. Soriano P. Gertler F.B Neuron. 1999; 22: 313-325Abstract Full Text Full Text PDF PubMed Scopus (337) Google Scholar found Mena at the tips of growth cone filopodia, so perhaps Ena/VASP reshape actin into parallel bundles, allowing structures such as filopodia to protrude from branched lamellipodial networks. These filopodia could then form the basis for cell–cell junctions in keratinocytes or cell–substratum exploration in neurites or fibroblasts. It would be interesting to use fluorescence microscopy to examine directly the effects of Ena/VASP proteins on dendritic networks produced by the Arp2/3 complex in vitro (3Blanchoin L. Amann K.J. Higgs H.N. Marchand J.B. Kaiser D.A. Pollard T.D Nature. 2000; 404: 1007-1011Crossref PubMed Scopus (420) Google Scholar). Clearly, many questions arise as to the reconciliation of the role of Ena/VASP proteins in Listeria translocation, in vitro and in eukaryotic cells. Some of the questions will probably be answered by resolution of the differences between genetic systems, where expression levels and cellular context within a whole organism are directing the results and biochemical systems that are oversimplified at times but allow direct access to molecular mechanisms. Other questions will be of great biological interest, such as the real function of filopodia and lamellipodia in an advancing neurite, the different contexts in which actin assembly is triggered, and the fine-tuning of the cytoskeleton by numerous signaling and actin binding proteins that results in all of the complex behaviors of motile cells.