Shape is an important property of all cells, from unicellular bacteria to the individual cells of complex metazoans. Even for unicellular organisms, the establishment, maintenance and changing of cell morphology is a dynamic process in time and space, and is the result of both intrinsic genetic programmes, such as those that govern division, and of extrinsic factors, such as chemotaxis or osmotic shock. However, the need to regulate cell shape is most acute in animal cells which, having done away with their cell wall, are both more vulnerable to physical and chemical changes in the environment, and freer to alter their form in a matter of minutes if required, as when engulfing food or migrating (Kunda et al., 2008; Pollard and Borisy, 2003). Local changes in the shape of fungal and animal cells are driven largely by alterations in the organisation of the cortical actin cytoskeleton. In fungi, actin filaments regulate cell form through their ability to guide local growth and cell-wall deposition (Moseley and Goode, 2006). In animal cells, actin meshworks and structures underlie the plasma membrane and, together with myosin motors, generate the required pushing or pulling forces (Chhabra and Higgs, 2007; Pollard, 2007).
In both fungi and metazoans, the architecture and dynamics of the actin cytoskeleton are regulated by a large set of actin-binding proteins, many of which are subject to precisely controlled changes in activity and position in time and space (dos Remedios et al., 2003; Goode et al., 2000; Winder and Ayscough, 2005). As a result of its central role in these fundamental processes, dysregulation of the actin cytoskeleton is associated with various human diseases (Fanzo et al., 2006; Scott and Olson, 2007; Symons et al., 1996; Yamazaki et al., 2005). A more complete understanding of how the actin cytoskeleton is regulated, and how it controls cell form, should therefore illuminate our understanding of normal development and homeostasis, and of the aetiology of various pathological states, such as cancer.
Despite years of study, there are still many aspects of actin cytoskeletal regulation that remain unclear. This is partly due to the fact that local actin cytoskeletal dynamics in a cell at any one time are the net result of the actions of a large ensemble of proteins. The situation is also complicated by the fact that overlapping sets of regulatory proteins sculpt several types of macroscopic actin structures that perform different functions. Nevertheless, a large number of genomes have now been sequenced, and many groups have carried out systemic loss-of-function studies on regulatory proteins in both yeast (Giaever et al., 2002; Winzeler et al., 1999) and animal cells (Boutros et al., 2006; Downward, 2004; Kamath et al., 2003; Kiger et al., 2003; Kittler et al., 2007; Simpson et al., 2008). Datasets such as these provide two new windows into the regulation of the actin cytoskeletal network.
Although the regulation of yeast cell shape is vastly simpler than that of metazoan cells, many of the fundamental processes are homologous. Understanding the extent to which actin regulatory proteins and pathways are conserved throughout evolution will shed light on which actin regulators probably evolved together, and might suggest long-standing functional relationships between particular actin regulators. In addition, by surveying proteins that contain actin-binding domains, we can see how actin regulation has been refined during evolution to accomplish new tasks. In parallel, the recent surge in large-scale genomic screening technology in cell culture provides us with increasing information about the range of phenotypes that result from changes in actin organisation, and allows the actin regulators to be organised into pathways or architectural categories.
In this poster article, we present an overview of the similarities and differences in actin regulation of eukaryotic cell morphogenesis between the budding yeast Saccharomyces cerevisiae and metazoans, and of how various cell-shape changes and structures are produced. Although it is beyond the scope of this simplified overview, extending the comparison to plants and Dictyostelium discoideum has also proved useful (Cvrckova et al., 2004) and, historically, studying how various parasites and viruses can subvert the host-cell actin cytoskeleton has also been illuminating (Baum et al., 2006). We focus here on actin-based structures, but it is worth remembering that cell architecture is also controlled by many additional factors, including myosin and other motor proteins, microtubules, and extrinsic extracellular and mechanical forces (Ingber, 2006; Lecuit and Lenne, 2007; Quintin et al., 2008; Schott et al., 2002). We also concentrate our scrutiny fairly proximal to actin nucleation because its upstream signalling is very complex and cell-type dependent (Moseley and Goode, 2006; Ridley, 2006).
Actin filaments: building the blocks
In the poster panel ‘Actin polymerisation and depolymerisation’, we have depicted a highly simplified view of the fundamental process of actin polymerisation and depolymerisation, along with the immediate fates of the resulting actin double-helical filament [for more detailed discussion, please see Moseley and Goode (Moseley and Goode, 2006) and Pollard (Pollard, 2007)]. Briefly, G-actin—ADP subunits, of which the self-assembly into filamentous actin (F-actin) is poorly favoured, are charged by polymerisation factors such as profilin, forming G-actin—ATP, and added to the growing end of the filament (the barbed end or plus end). As the filament ages, actin-associated ATP at the plus end decays to ADP+Pi and then to ADP. Near the pointed end (or minus end), effectors such as ADF or cofilin cause the severing and depolymerisation of actin, releasing G-actin—ADP back into the uncharged pool, ready for another round of polymerisation. G-actin can also be sequestered by the conserved cyclase-associated protein (CAP) or, in metazoans only, β-thymosins. This core process, which results in net growth or net shortening of actin filaments (or, in principle, in their steady-state ‘treadmilling’, although this is probably of minor importance), is the same in both yeast and metazoans. Metazoans are slightly more complex of course; for example, whereas yeast has one gene encoding cofilin and one encoding actin, flies have two and six genes, respectively (Moseley and Goode, 2006). In many cases, the different isoforms have developed specialised functions, as is the case for flight-muscle actin in Drosophila (Jacinto and Baum, 2003).
Crucially, F-actin assembly is regulated and effected in association with the plasma membrane, often exactly when and where it is needed. Once F-actin is formed, it can be altered to suit the needs of the cell by being variously capped, severed to expose a fresh barbed end, subjected to further linear or branched growth, stabilised (e.g. by tropomyosin), crosslinked, bundled into higher-order structures or recycled into new structures. All of these filaments can be transported from the plasma membrane to other cellular locations (Moseley and Goode, 2006).
Different signals can lead to several F-actin fates through the activation of distinct WH2-domain-containing actin-nucleation or actin-assembly factors (see poster). For example, the activity of small Rho GTPases such as Rac1 and cdc42 at the membrane can stabilise the nucleation-promoting complex of the SCAR/WAVE and WASP families, which is represented by Wasp/Las17 in yeast and by more diverse isoforms in metazoans. These in turn activate the Arp2/3 assembly-factor complex, leading to the elaboration of actin branches off the main filament at a 70-degree angle, repeated iterations of which can result in elaborate meshworks. By contrast, other signals can induce assembly factors of the formin family to elaborate longer linear structures (Heasman and Ridley, 2008; Ibarra et al., 2005; Pollard, 2007). In addition to Arp2/3 and various formins, less-well-characterised assembly factors that catalyse growth of linear actin filaments are found in metazoans only, including Spire and Cordon-bleu, or in pathogens that subvert the actin pathway, such as the VopF and VopL proteins of Vibrio or TARP of Chlamydia (Baum and Kunda, 2005; Renault et al., 2008). Presumably, the need to build more elaborate structures in higher organisms has forged the evolution of diverse assembly proteins.
In summary, although there are many differences between fungi and metazoans with respect to these core processes of actin organisation, actin itself is conserved, along with two conserved assembly mechanisms, in the form of Arp2/3- and formin-dependent assembly. There are also similarities in how actin subunits are balanced between pools of F- and G-actin (through the action of profilin and cofilin). Even upstream, where mechanisms diverge more extensively, WH2-domain-containing nucleation promotion factors are employed in both cases, and small Rho GTPases are employed to help regulate these events in both yeast and metazoans (Cvrckova et al., 2004).
Playing with the actin tool kit
Once actin filaments are produced, how do metazoan and fungal cells use them to create structures or to undergo shape changes? On our poster, we summarise how the diverse actin-dependent cell-architectural changes are achieved in yeast and metazoans. A yeast cell is shown at the bottom of the poster, and a metazoan cell at the top; common protein categories are shown in the same colour for both (disregarding multiple homologues).
The similarities
Yeast and metazoan cells have several common cell-shape needs that are mediated in part by actin-cytoskeletal building blocks. Both must have mechanisms for scission during cell division (Field et al., 1999; Moseley and Goode, 2006) (see poster). Both must also take up extracellular materials. Yeast achieves this requirement solely using endocytosis, which is completely dependent on actin-mediated action for the uptake step. By contrast, besides receptor-mediated endocytosis, metazoans have additional ways of taking up materials, including phagocytosis (engulfment of larger items) and macropinocytosis (engulfment of smaller items), both of which rely on actin and some of its conserved regulators (Girao et al., 2008; Moseley and Goode, 2006). Finally, both yeast and metazoans produce actin cables composed of actin filaments bundled with various actin-binding proteins, which traverse the cell. In yeast, these cables form underneath a structure known as the polarisome, to help to define the long axis of the mother with respect to its budding daughter, perpendicular to the future site of scission. Actin cables in yeast also act as scaffolds to facilitate the delivery of materials, such as particular mRNAs (Moseley and Goode, 2006), a role fulfilled primarily by microtubules in metazoan cells. In certain metazoan cells in culture, long actin cables known as stress fibres form at sites of mechanical stress or focal adhesions (Lock et al., 2008; Naumanen et al., 2008). Finally, both yeast and metazoans elaborate a bundled ring composed of actin and myosin II during the process of division (Field et al., 1999; Moseley and Goode, 2006).
The differences
Cells that contain rigid walls and vacuoles, such as yeast, must balance their growth against turgor pressure; actin is present at the growth zones and defines where cell shape changes will take place (Harold, 2002). By contrast, for structural purposes, metazoan cells must establish an actin cortex, which is a loosely organised meshwork lying beneath the plasma membrane. Also, complex multicellular organisms have a greater need for structural diversity. They must form junctions with neighbouring cells (Hartsock and Nelson, 2008; Niessen and Gottardi, 2008), both in resting tissues and during the sometimes extreme conditions of developmental morphogenesis (Jacinto and Baum, 2003; Quintin et al., 2008). They must sometimes produce elaborate, specialised structures such as neurons (Kawauchi and Hoshino, 2008) or sensory bristles (Tilney and DeRosier, 2005). Moreover, they must occasionally migrate, a process accompanied by various protrusions [for example, ruffles, filopodia, microvilli, blebs and lamellipodia (Charras, 2008; Mattila and Lappalainen, 2008; Pollard and Borisy, 2003; Ridley, 2006)] as the underlying actin structures help to generate the forces that are required to enable the cell to advance or retreat. In many cases, the precise nature of upstream signalling, and even the assembly factor that catalyses actin-filament formation for some of these structures, is still unclear. Also, the jury is still out on the relationship between many protrusions seen in metazoan cells in culture and their function in vivo (Chhabra and Higgs, 2007).
Perspectives
Metazoan genomes contain a number of largely uncharacterised genes containing sequences that are predicted to encode actin-binding domains or domains in common with nucleation-promoting factors or those of their upstream regulation machinery. Although some of these are likely to be redundant, we expect that many new players remain to be characterised (Goldstein and Gunawardena, 2000). At the same time, we must not overestimate the diversity by assuming that all predicted isoforms of a protein will actually have the function ascribed to their founding member(s). For example, there are at least 15 genes in mammals that encode proteins with predicted formin homology domains, but not all have been characterised biochemically for actin-assembly function, and it remains possible that some have been co-opted to play roles in unrelated pathways (Higgs and Peterson, 2005). Nevertheless, by combining unbiased forward genetic screens with investigation of proteins whose sequences are suggestive of a role in actin or cell-shape regulation, and by merging in vitro biochemistry with cell-biological analysis of live cells in time and space, the big picture of how the actin cytoskeleton is regulated should continue to be fleshed out.
