NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Madame Curie Bioscience Database [Internet]. Austin (TX): Landes Bioscience; 2000-.

Cover of Madame Curie Bioscience Database

Madame Curie Bioscience Database [Internet].

Show details

Origins and Evolution of the Actin Cytoskeleton

* and .

* Corresponding Author: Center for Biochemistry and Center for Molecular Medicine Cologne, Medical Faculty, University of Cologne, Joseph-Stelzmann-Str. 52, 50931 Köln, Germany. Email:


The presence of a complex cytoskeletal system is a hallmark feature of eukaryotic cells, distinguishing them from their prokaryotic (bacterial or archaeal) “cousins”. No extant prokaryote studied so far possesses obvious homologues of major cytoskeletal proteins shared universally among eukaryotes, such as e.g., actin or tubulin. However, several proteins exhibiting limited sequence similarity with certain cytoskeletal components, as well as the ability to form filaments, have been found.1-3 These include, among others, relatives of actin and actin-associated proteins that will be discussed in detail below, the FtsZ family of bacterial and archaeal tubulin-related proteins participating in cell division4 and an intermediate filament-like protein (crescentin) from Caulobacter.5

Evidence from large-scale genome sequencing indicates that eukaryotic nuclear genomes arose as a result of a “merger” of at least two ancestors.6-8 There are even two subclasses of genomic DNA with different compositional characteristics; one of them, resembling recent Bacteria, has contributed e.g., the molecular apparatus of the energy metabolism, while the other, close to Archaea, brought, among other genes, components of the replication and proteosynthetic machinery.9 The cytoskeleton has been also suggested as a candidate for archaeal heritage,10 however, its origins remain mysterious. It has even been proposed that genes for cytoskeletal proteins come from a third ancestor of eukaryotes, a hypothetical “chronocyte”, long extinct and surviving only through its eukaryotic descendants.11 In any case, a cytoskeleton-like apparatus must have been present at least at the point of acquisition of endosymbionts that later gave rise to mitochondria.

Here we focus on a subset of cytoskeletal proteins, namely actin and a core of associated proteins participating in the control of actin dynamics, especially filament nucleation (reviewed in ref. 12). Identification of components shared by evolutionarily distant eukaryotic lineages (such as plants, yeast, Metazoa, slime molds and other selected protists), and in a few cases also the discovery of related proteins in prokaryotes, may provide the first step towards reconstructing the composition, and possibly also functional characteristics, of the initial set of “actin-associated modules” of the common ancestor of eukaryotes (Table 1).

Table 1. Evolutionary distribution of ancient domains related to the eukaryotic microfilament system and their prokaryotic relatives.

Table 1

Evolutionary distribution of ancient domains related to the eukaryotic microfilament system and their prokaryotic relatives.

We shall discuss selected parts of the actin-associated apparatus separately. First, we focus on the actin monomer itself, together with a class of evolutionarily conserved monomer-binding proteins that modulate the balance of monomeric and filamentous actin. Next, we will examine complexes that serve as “primers” nucleating new actin filaments. We shall then move to an assortment of actin-binding proteins that either regulate filament dynamics or mediate association of other cellular structures with actin filaments, including also actin-dependent motors. Finally, we shall summarize the potential evolutionarily conserved aspects of the regulatory mechanisms controlling the structure and function of the actin network.

The Actin Cytoskeleton in the Cellular Context

While much of actin's fame derives from studies of metazoan muscle actin/myosin complex, nonmuscle actin participates in a range of essential processes of eukaryotic cell morphogenesis, in motility of (nonmuscle) metazoan and amoeboid cells and in intracellular transport. Actin fibers contribute to intracellular movement either by providing direct locomotive force through filament assembly, or serving as “tracks” along which structures travel, driven by molecular motors such as myosins. The polymerization-based mechanisms are believed to be evolutionarily older than those involving molecular motors,13 thus justifying our focus on the actin nucleation machinery.

Actin filament assembly is believed to participate in the “exploratory behavior” of soft-bodied cells (typical metazoan cells, such as fibroblasts or neurons, or amoeboid cells), i.e., to the formation of filopodia and lamellipodia, as well as membrane ruffles.14,15 Actin “comets” can also propel organelles and intracellular parasites across the cytoplasm, utilizing filament assembly forces.16,17 Even in wall-encased cells of plants or fungi filament assembly contributes to cell shape development, as documented for yeast buds,18 plant trichomes,19 tip-growing root hairs20,21 and pollen tubes.22 However, the resulting networks of actin filaments are believed to serve mainly as tracks for motor-driven delivery of exocytotic vesicles to the expanding regions of the cell surface (e.g., see ref. 23). Perhaps with the exception of trichomes, even these cases can be considered examples of “exploratory behavior”, as nonmotile cells indeed can explore the environment only by expanding (growing) into it.

Actin is also indispensable for essential processes of the cell cycle. Cytokinesis usually involves exocytosis, depending on vesicle delivery along actin tracks, and at least in metazoan and amoeboid cells also constriction of a subcortical actomyosin ring.24,25 Very recently, actin has been also implicated in chromosome congression during oocyte mitosis.26 Thus, actin is central to at least two basic functions of life, namely “tactile” interaction with the environment and cell multiplication.

Actin and Monomer-Binding Proteins

Actin exists in the cytoplasm of eukaryotic cells either as soluble monomers (G-actin) or as filaments (F-actin). The in vivo balance between F- and G-actin, as well as filament turnover, is to a large extent controlled by proteins that regulate the availability and nucleotide-bound state of monomers. This, in turn, affects the rate of filament polymerization (predominantly at plus or barbed ends) and depolymerization (preferentially at minus or pointed ends), resulting in net growth, shrinkage or treadmilling. While some modulators of actin dynamics are restricted only to certain eukaryotic lineages, such as e.g., the actin-sequestering proteins β-thymosin of Metazoa (see discussion of WH2 domains below) or toxofilin of Toxoplasma, others are shared by most eukaryotes and therefore obviously ancient, in particular profilin and proteins of the actin depolymerizing factor (ADF)/cofilin family, which together act as major determinants of cytoplasmic actin dynamics.27,28


Actin is a remarkably conserved protein, with overall identity about 85% between the most divergent family members. However, most species possess multiple actin isoforms exhibiting both structural and functional differences. The number of actin genes per genome ranges from 1 (in yeasts) to almost 100 in some plants (reviewed in refs. 12,29). The closest eukaryotic relatives of actin comprise the family of actin-related proteins (ARPs),30 some of which will be discussed below. The evolutionary separation between actin sensu stricto and the ARPs apparently occurred already in the ancestral eukaryote. The actin/ARP gene family can thus be viewed as a single unit if we are considering the early steps of eukaryotic evolution.

Distant but undisputable prokaryotic homologues of actin have been found. The MreB protein forms filaments both inside cells of rod-like bacteria and in vitro.1,31,32 Both actin and MreB are considered members of an ancient superfamily including also the bacterial cell division protein FtsA, the ATPase domain of Hsp70 and even the enzyme hexokinase. The split between the MreB and actin lineages and the rest of the family apparently occurred very early in evolution,1,2 presumably before the establishment of molecular mechanisms guaranteeing a relatively low mutation rate, and possibly already in the hypothetical era of the RNA-based life.11 Most prokaryotic members of the MreB family are bacterial, with just a handful of homologues identified in Archaea (ref. 10 and our database searches). Although this could reflect limited availability of data from Archaea, the possibility that actin's ancestors originated in Bacteria and/or arrived into the proposed archaeal ancestor by horizontal gene transfer also cannot be excluded.


Profilin, an abundant small protein that may aid “charging” of G-actin by ATP, promoting thereby filament assembly, is ubiquitous in eukaryotes, with the exception of greatly reduced parasites such as Giardia. Like actin, it exists in multiple isoforms in many organisms.33,34 However, it is somewhat less conserved: profilins within a single organism (such as Dictyostelium), may share as little as 55% identity.35 This also complicates searches for possible prokaryotic relatives.

Profilin is related to members of an ancient family including the Roadblock/LC7-related proteins, as well as prokaryotic (both bacterial and archeal) MglB proteins.36,37 This family contains a number of proteins implicated in ATPase or GTPase regulation, including dynein light chains (i.e., a subunit of a tubulin-dependent motor complex). MglB is implicated in gliding motility of microbial cells. Curiously, in some bacteria it resides in the MglAB operon that includes also MglA, a small GTPase,36 suggesting a very ancient relationship between G-proteins and ancestors of the actin module.

ADF/Cofilin and the Gelsolin Repeat

Members of the actin depolymerization factor (ADF)/cofilin family preferentially bind ADP-associated G- and/or F-actin and increase the G-actin level via filament severing and depolymerization. The ADF/cofilin family, defined by the presence of the structurally conserved ADF homology domain, is ubiquitous or at least widespread in eukaryotes,38 suggesting that this domain and the associated severing and depolymerizing activity was present in the common ancestor of eukaryotes. However, members of the ADF family are relatively poorly conserved, and no candidate prokaryotic relatives have been found so far. Bikonts possess only proteins with a single ADF domain, usually without extensions, but the domain appears duplicated or with characteristic extensions in unikonts (Fig. 1).

Figure 1. Reconstruction of the evolutionary history of proteins of the ADF/cofilin and gelsolin repeat families.

Figure 1

Reconstruction of the evolutionary history of proteins of the ADF/cofilin and gelsolin repeat families. The diagram is based on protein repertoires of selected representatives of each lineage: A. thaliana, P. falciparum, D. discoideum, S. cerevisiae (plus (more...)

Interestingly, the ADF domain shares structural similarity with the gelsolin repeat (including the actin binding site), despite lack of sequence similarity.39,40 The ADF and the gelsolin repeat might thus be descendants of an ancient actin-severing protein (see Fig. 1 for a detailed description). The gelsolin repeat is found as a tandem of three or more copies in diverse proteins. The prototype of this family, gelsolin, contains six repeats and acts as a calcium-regulated protein that caps the barbed ends of actin filaments, promotes nucleation and severs existing filaments.41 Typical gelsolin repeat proteins are present in bikonts and in plants (Table 1), indicating that a gelsolin-like protein must have been present in the common ancestor of eukaryotes. Their absence in other lineages is thus apparently due to secondary loss. Moreover, proteins of the Sec23 family, conserved in all eukaryotes, contain a single diverged gelsolin repeat at their C-terminus, supporting the ancient status of the ADF/gelsolin domain. Sec23 is a component of COPII coated vesicles involved in recruiting and formation of prebudding complexes whose interaction with actin has not been investigated so far.42

Actin Nucleation Complexes

Actin can form filaments in vitro even in the absence of other proteins. However, elongation of preexisting filaments at extendable ends (i.e., usually free barbed ends) is more efficient and requires lower actin concentration than establishment of new filaments out of G-actin. Such ends can arise either by fragmentation of existing filaments (mediated e.g., by proteins of the ADF/cofilin family), by removal of proteins capping a preexisting end, or by de novo nucleation aided by a specific protein complex. Nucleation of new filaments in vivo may present a unique regulatory node where multiple signaling pathways converge, resulting in precise control of the actin network structure. So far, two independent nucleation mechanisms have been studied in detail – Arp2/3 mediated nucleation and nucleation mediated by formins (for a review see refs. 12, 43).

The Arp2/3 Complex

The Arp2/3-dependent nucleation complex consists of seven subunits (Arp2, Arp3, and ARPC1 to ARPC5). Arp2 and Arp3 are members of the ARP family mentioned above. All subunits are well-conserved throughout major eukaryotic lineages, although some losses apparently occurred, in particular in Parabasalia and Diplomonadida44,45 (Table 1). ARPC1 is a member of the WD40-like protein family that has also bacterial and archaeal members but no prokaryotic relatives have been found for the remaining four subunits.

Curiously, at least in some organisms, loss of certain subunits is compatible with survival. In Arabidopsis, homozygous mutants lacking single genes Arp2, Arp3, ARPC5 or one out of two isoforms of ARPC2, are viable and fertile, although they exhibit a distinct mutant phenotype of distorted trichomes and malformed epidermal cells.19,46 For the ARPs themselves this may reflect a partial complementation of the mutant defect by other members of the family or even actin; however, the nonessentiality of ARPC5, together with its absence in several eukaryotic lineages suggests that this subunit may be a later addition in the Arp2/3 nucleation complex. Alternatively, most of the tasks requiring actin nucleation may have been taken over by formins in plants (see below). However, deletion of genes encoding Arp2, Arp3, ARPC1 and ARPC5 is lethal in the budding yeast, while loss of ARPC 2, 3 and 4 results in growth defects of varying severity47 (see also; perhaps the ARPC5 subunit might have become indispensable in the specific context of the yeast cell, since budding heavily depends on establishment of Arp2/3-dependent actin structures.18


Formins are defined by the presence of the approximately 400 residues long, predominantly α-helical FH2 domain, capable to form a ring-shaped flexible dimer that caps the barbed end and allows processive elongation of the actin filament. The FH2 domain can be found in most eukaryotes (Table 1) and is usually preceded by the proline-rich FH1 domain that interacts with profilin-actin and funnels actin monomers to the nucleation site. The FH1-FH2 combination probably constitutes the minimal core fully functional in terms of actin nucleation and elongation activity; diverse formins differ in their comparative capping vs. nucleating activities, as well as in their requirements for cofactors such as profilin.48-51 A plant formin has been recently found to possess an unique ability to bundle actin filaments.52

The FH1-FH2 core is usually accompanied by additional domains. Diverse formin classes differ mostly in their N-terminal regions, which generally have regulatory and targeting roles.53-55 A common architecture characterized by the presence of an N-terminal GTPase binding motif (GBD/FH3) and a C-terminal autoinhibition domain can be found among formins of Amoebozoa, Fungi and Metazoa.54 This domain combination, allowing regulation of formin activity by activated Rho GTPases, and establishing thereby a direct link between the nucleation machinery and regulatory signaling pathways, apparently arose only within the unikont lineage,56 since it can be found neither in the formins of Apicomplexa, Kinetoplastida and Ciliophora, nor in those of plants, which acquired either N-terminal membrane anchors or a variant Pten-like domain, possibly also allowing association to membranes.53

Other Actin-Binding Proteins

In addition to the predominantly or exclusively G-actin binding proteins, there is a growing list of proteins that bind to actin filaments and exert actions as diverse as capping, severing, crosslinking, attachment to other cellular structures and force transmission. A number of domains with bona fide actin-binding properties that are shared by many proteins throughout various lineages has been identified; duplications and domain shuffling apparently produced much of the present diversity of ABPs. We shall briefly discuss several representatives: the heterodimeric capping protein, the calponin homology (CH) domain and the small VHP and WH2 domains.

The Heterodimeric Capping Protein

Capping proteins bind tightly to the barbed end of actin filaments and prevent the addition or loss of actin subunits. They are composed of two subunits, an α subunit of 32-36 kDa and α β subunit of 28-32 kDa. Interestingly, the α and β subunits of chicken skeletal muscle capping protein have a strikingly similar folding, despite lacking sequence similarity, so that the entire molecule has a pseudo 2-fold rotational symmetry.57 We suggest that the capping protein was initially homodimeric, but a gene duplication followed by substantial divergence resulted in a heterodimeric protein. The heterodimer must have brought a significant selective advantage, because it has undergone relatively little change since the initial diversification of the two subunits. This event must have taken place very early, because the heterodimeric capping protein can be found in all lineages studied, except in the greatly simplified parasitic diplomonads (Table 1).

The Calponin Homology Domain

The CH domain is a module of about 100 residues with a globular α-helical fold, present in a large family of proteins that can be subdivided into subfamilies according to their domain composition.58 Not all CH domains bind to actin. A typical CH-containing actin-binding motif consists of a tandem pair of CH1 and CH2 domains, although a sole CH1 domain also can bind to actin. CH1-CH2 proteins are numerous in unikonts, but there are a few examples in bikonts (Table 1). Variant domains such as CH3, CHe and CHc apparently do not bind actin: the CH3 domain is found in many signaling and a few cytoskeleton proteins of many, predominantly multicellular, lineages; CHe binds microtubules and is ubiquitous in eukaryotes; 59 the CHc domain is found in choline/carnitine-O-acyltransferases, enzymes involved in fatty acid metabolism and transport found in Metazoa, Fungi and Kinetoplastida.60

The distribution of CH domains suggests that they were present already in the ancestral eukaryote. We postulate that the proto-CH domain did not bind actin and diversified already very early. An ancient duplication gave rise to the tandem CH1-CH2. The actin binding properties could have appeared either before or after this duplication, and the tandem arrangement was clearly advantageous in terms of interaction with F-actin because few proteins have either CH1 or CH2 only; these could have originated by subsequent loss of one of the domains. A further duplication of the CH1-CH2 tandem produced the actin-bundling protein fimbrin, documented in lineages as diverse as unikonts, plants (where they constitute the only CH1-CH2 proteins) and the chromalveolate Tetrahymena thermophila.61 In unikonts the CH1-CH2 family expanded and diversified considerably by domain shuffling, leading to acquisition of tail regions composed of spectrin repeats, filamin repeats and other domains.58

The VHP and WH2 Domains

Actin-binding modules such as the villin headpiece (VHP) and the WASP homology domain 2 (WH2) are well documented mainly in multicellular eukaryotes (Table 1). However, their small size (about 35 residues) hampers reliable database searches, therefore their presence in other eukaryotes cannot be ruled out. Also the alleged bacterial VHP or WH2 domains, as well as viral proteins with WH2 domains (listed in the InterPro resource62) have to be interpreted very cautiously in the absence of functional data.

The VHP appears at the extreme C-terminus of diverse proteins, alone or in combination with other domains, particularly the gelsolin repeat. The VHP binds to F-actin (although there are exceptions) and confers actin-bundling properties to villin and related proteins.63 The WH2 domain binds preferentially ATP-associated G-actin. It can be found alone as a single domain (β-thymosins) or as a tandem of two (actobindin), three (ciboulot) or four (spire) copies. Isolated WH2 domains such as in β-thymosins sequester G-actin and maintain it in a nonpolymerizable form. In contrast, two or three WH2 domains in tandem as in actobindin or ciboulot promote elongation of barbed filament ends similar to profilin,64 while the four domains of spire together promote nucleation of new filaments, independent of “classical” nucleation complexes.65 The WH2 domain may also associate with other domains, as in WASP, verprolin-related proteins and cyclase-associated protein.66

A similarity between the actin-binding regions of the VHP and the WH2 domain has been proposed.39 Both domains might have evolved from a short domain or loop that diverged before the unikont/bikont split to accommodate the different properties, F-actin binding vs. G-actin binding, of the VHP and the WH2 domain, respectively. Nevertheless, it cannot be ruled out that the apparent sequence similarity results from evolutionary convergence.

Other Domains and Proteins

The number of proteins able to interact with actin is very large, and we cannot make a comprehensive account on the evolutionary history of all of them. We will not discuss domains whose presence in a protein does not automatically correlate with actin-binding properties, although they are shared by numerous actin-binding proteins (such as the WD repeat, the kelch repeat and the LIM domain), proteins apparently specific for a limited number of lineages, or those lacking reliably recognizable domains. Future structural studies may reveal relationships among proteins that have passed unnoticed because of apparent absence of sequence similarity. This already happened e.g., in case of the β-trefoil fold, initially described in fascin but later discovered in the Dictyostelium membrane-associated protein hisactophilin on the basis of structural data. It has been proposed that the β-trefoil fold arose by duplication of an ancestral gene encoding a homotrimeric single-repeat protein.67 Similarly, the I/LWEQ domain, named after the conserved initial residues in each of four repeated blocks, might have originated by duplications of an ancestral single-repeat protein. This domain is characteristic of two classes of proteins involved in actin organization, namely the focal adhesion protein talin (Amoebozoa, Metazoa) and the polarisome protein Sla2p/HIP-1 (Amoebozoa, Fungi and Metazoa), and apparently originated within the unikont lineage, prior to the branching between Amoebozoa and Opisthokonta.68

Effectors and Regulators of the Actin Cytoskeleton

The actin-binding proteins discussed in the last section present only a limited selection of molecules that serve as “interfaces” ensuring the integration of the actin network into the cellular web of interactions. Thorough analysis of all proteins that could be considered “effectors” of the actin cytoskeleton, as well as of its regulatory inputs, would exceed the scope of this review. Instead, we shall introduce selected examples that can provide additional insights into the early stages of cytoskeleton evolution. First, we shall focus on the major actin-dependent motor protein, myosin, which is responsible for the movement of a variety of cargoes along the actin network; later we shall discuss the evolutionarily conserved aspects of regulatory pathways controlling the structure and dynamics of the actin cytoskeleton.

Myosin, the Prototype Motor

In terms of sequence, myosins can be recognized by the presence of an ancient, well-conserved domain, the myosin head, an ATPase capable of converting the chemical energy from ATP hydrolysis into mechanical movement along an actin filament. The myosin head has been so far found in nearly all eukaryotes studied (Table 1), indicating that, similar to actin itself, it apparently arose no later than in the common eukaryotic ancestor. Indeed, it has been proposed that the myosin motor domain might have originated from a common ancestor with the microtubule-dependent motor kinesin, as they share a similar 3D structure of the core.69 Although no readily identifiable homologue of the myosin head can be found in prokaryotes, both myosin and kinesin motor domains are related to proteins of the P-loop NTPase superfamily. This superfamily includes both ATPases and GTPases and has also prokaryotic members, suggesting the possible evolutionary root of both motor domains.70,71

Myosins have blossomed into an abundant and diverse protein family during eukaryote evolution. At least 18 myosin classes have been established on the basis of both myosin head sequence and overall domain composition;72,73 however, a recent detailed analysis of 23 genomes covering the whole eukaryotic kingdom distinguishes already 37 myosin classes, with representatives of up to 13 classes found in a single species.74

Evolutionary events documented in myosin evolution include mutations, domain shuffling, domain fusions, partial deletions, duplications, and losses, which makes evolutionary studies extremely difficult. Nevertheless, a thorough phylogenetic analysis, based on domain structure rather than on sequence analysis only, not only provided supportive evidence for the unikont/bikont model of early eukaryote evolution, but also allowed identification of three supposed ancestral myosin families.74 One of them corresponds to myosin I, previously suggested to be one of the oldest myosin classes whose members may have originally functioned as generalists, while more recently evolved families, limited to particular lineages, may have been optimized for specialized functions.72

Regulatory Inputs Controlling the Actin Cytoskeleton

While a number of cellular components, including the constituents of the actin cytoskeleton discussed above, is well conserved throughout evolution, many others are not. In particular, this is often the case of components of the regulatory circuits controlling the function and mutual interactions of conserved “core” molecular modules. Such variable regulatory connections —or “protocols” sensu Csete and Doyle75—between well-conserved molecular assemblies may provide means for generating the great diversity of form and function from a relatively small set of molecular building blocks, as observed in present cells (see also ref. 12). Pathways controlling the structure and function of the actin cytoskeleton provide a good example, since they may seem to be almost entirely lineage-specific on the first glance. We shall now focus on the control of the actin nucleation machinery to illustrate this point; however, a similar argument could be constructed also for other actin-related regulatory pathways.

The Arp2/3 complex, which alone is inactive, can be activated by a variety of cofactors.76 These include e.g., fungal myosin I and Abp1p, metazoan-specific cortactin, the multidomain protein CARMIL, found in Amoebozoa and Metazoa, and several proteins of broader distribution, such as e.g., coronin (see ref. 12). Prominent among Arp2/3 regulators is the large family of conserved WH2 domain-containing WAVE (WASP family verprolin homology) proteins that form a core of a large multiprotein regulatory complex. This complex has been long considered specific to Metazoa, Fungi and Amoebozoa; however, homologues have been recently found also in plants,19 indicating that the WAVE-associated complex is ancient. It remains unclear whether an analogous system can be found also in the remaining deep-branching eukaryotic lineages; sequence searches for homologues are hampered by low sequence complexity of this extremely proline-rich protein. Also formins are embedded in complex signal networks.12,77

Despite the evolutionary plasticity of molecular mechanisms controlling the actin nucleation, some common motifs emerge if we follow the regulatory pathways backward from their actin targets to the upstream inputs. Remarkably, many regulators of actin nucleation, including WAVE proteins and at least some formins, are themselves controlled by Rho GTPases, a class of regulatory proteins with multiple outputs, initially described as major regulators of actin remodeling but later found to participate also in microtubule dynamics, endocytosis, vesicle trafficking, gene transcription, the response to oxidative stress, cytokinesis, cell cycle progression and apoptosis.78-80 Small GTPases of the Rho family are present in all eukaryotes, although the “classical” subfamilies of Rho proper, Rac and Cdc42 are probably specific to Metazoa and Fungi.

Detailed discussion of Rho GTPases and their cofactors would exceed the scope of this review; a thorough evolutionary analysis of small GTPases has been published by others81 (see also chapter by Balch). However, we should at least mention the fact that although the pathways whereby Rho GTPases control their downstream effectors vary greatly across eukaryotic lineages (e.g., see ref. 82), the common motif of actin control by a small GTPase appears to be almost invariant, despite of the varying molecular implementations (see ref. 12 for more detailed discussion). Remarkably, roots of this arrangement could be traced to very early stages of evolution, since MglB, the prokaryotic relative of profilin, often resides within the same operon with a small GTPase (see above and ref. 36).


We hope that careful evaluation of the above outlined data may allow an attempt at reconstructing the microfilament system of the ancestral eukaryote. We can assume that actin-binding domains shared by most eukaryotes (in particular those found on both sides of the unikont-bikont divide74) were present in the common ancestor of the eukaryotes. This means that any protein found both in a representative of the Fungi/Metazoa/Amoebozoa group (the unikonts) and in any of the remaining eukaryotic lineages examined (plants or nonamoeboid unicellular eukaryotes) is likely to be of ancestral origin. Quite many such proteins can be found (Table 1); however, interpretation of their phyletic distribution is not always straightforward. Only actin itself is indeed found in all lineages. However, if we disregard the greatly reduced and incompletely characterized genome of the diplomonad Giardia, and allow for occasional gene loss and/or divergence beyond recognition in rapidly evolving lineages (in particular the unicellular ones), we have to realize that the common eukaryotic ancestor must have had a fairly elaborate cytoskeletal apparatus. This could have been expected, since the ancestor must have been already able to internalize the prokaryotes that later became endosymbionts and gave rise to mitochondria; therefore it must have been capable of some form of phagocytosis.

The ancestral eukaryote thus already possessed not only actin and myosin, but all relevant basic activities required for remodeling of the microfilament system. A profilin-like protein sequestered actin monomers and promoted nucleotide exchange, rendering the monomers ready for polymerization. Nucleation was achieved both by the Arp2/3 complex and by the dimeric FH2 domains. A dimeric capping protein capped the fast growing end of the filaments. This protein was initially homodimeric, but a gene duplication followed by substantial divergence resulted in the current heterodimeric protein. An ADF/gelsolin-related protein was responsible for severing and depolymerization of the filaments. Also proteins responsible for other activities, in particular bundling, crosslinking or membrane association of filaments, were probably already present; however, it has to be said that the evolutionary fate of these activities is more difficult to reconstruct, because in general they cannot be attributed to a single or very few well defined domains. These activities apparently evolved independently several times, frequently by lineage-specific combination and “fine tuning” of preexisting ancient domains (such as the CH domain, the VHP, the WH2 domain), or utilizing novel domains evolved in a single lineage (such as the I/LWEQ domain, which is restricted to the unikonts). Another universal, and most likely ancestral, feature of the actin cytoskeleton is its regulation by means of small GTPases.

Existence of prokaryotic relatives of several components of the actin cytoskeleton may provide clues towards reconstruction of its origin. The simplest predecessor of the above-described machinery might have consisted of a “core” of just a few proteins. The polymerization dynamics of MreB-related “protoactin” filaments could have been controlled by an MglB-like “protoprofilin”, which might have already had some regulatory connection with a small GTPase. Nucleation of novel filaments might have first occurred spontaneously, perhaps on broken filament ends, later a specific conformation of two protoactin monomers supported by a WD40-like protein provided a more efficient nucleation core. Gene duplication supplied material for evolution of “nucleation-optimized” protoactin variants, nowadays known as Arp2 and Arp3, the specific WD40-like protein evolved into ARPC1, and the nucleation complex later acquired additional subunits. All the other components, including the alternative nucleation complex based on dimeric FH2 domains, could have appeared later (although still prior to the acquisition of mitochondria) either de novo or by recruitment of preexisting domains (possible in case of the VHP and WH2 domains), perhaps with exception of a motor protein that could have interacted already with the protoactin filaments. Such a scenario can in principle be only speculative; however, we hope that at least some of the issues can be resolved by future work in the field of molecular phylogenetics, as well as by functional characterization of prokaryotic relatives of the recent cytoskeletal proteins.


The authors would like to thank Marek Eliás and Angelika Noegel for helpful discussion. Work in authors' laboratories has been supported by GACR grant 204/05/0268 (FC) and by grants of the Deutsche Forschungsgemeinschaft, the Center for Molecular Medicine Cologne and the Köln Fortune Program of the Medical Faculty of the University of Cologne (FR).


Doolittle RF, York AL. Bacterial actins? An evolutionary perspective. BioEssays. 2002;24:293–296. [PubMed: 11948613]
Egelman EH. Actin's prokaryotic homologs. Curr Opin Struct Biol. 2003;13:244–248. [PubMed: 12727519]
Margolin W. Bacterial shape: Concave coiled coils curve Caulobacter. Curr Biol. 2004;14:R242–R244. [PubMed: 15043836]
Bramhill D. Bacterial cell division. Annu Rev Cell Dev Biol. 1997;13:395–424. [PubMed: 9442879]
Ausmees N, Kuhn JR, Jacobs-Wagner C. The bacterial cytoskeleton: An intermediate filament-like function in cell shape. Cell. 2003;115:705–713. [PubMed: 14675535]
Gupta RS, Golding GB. The origin of the eukaryotic cell. Trends Biochem Sci. 1996;21:166–171. [PubMed: 8871398]
Horiike T, Hamada K, Shinozawa T. Origin of eukaryotic cell nuclei by symbiosis of Archaea in Bacteria supported by the newly clarified origin of functional genes. Genes Genet Syst. 2002;77:369–376. [PubMed: 12441648]
Rivera MC, Lake JA. The ring of life provides evidence for a genome fusion origin of eukaryotes. Nature. 2004;431:152–155. [PubMed: 15356622]
Rivera MC, Jain R, Moore JE. et al. Genomic evidence for two functionally distinct gene classes. Proc Natl Acad Sci USA. 1998;95:6239–6244. [PMC free article: PMC27643] [PubMed: 9600949]
Li JY, Wu CF. Perspectives on the origin of microfilaments, microtubules, the relevant chaperonin system and cytoskeletal motors - a commentary on the spirochaete origin of flagella. Cell Res. 2003;13:219–227. [PubMed: 12974612]
Hartman H, Fedorov A. The origin of the eukaryotic cell: A genomic investigation. Proc Natl Acad Sci USA. 2002;99:16128–16133. [PMC free article: PMC122206] [PubMed: 11805300]
Cvrcková F, Bavlnka B, Rivero F. Evolutionarily conserved modules in actin nucleation: Lessons from Dictyostelium discoideum and plants. Protoplasma. 2004;224:15–31. [PubMed: 15726806]
Mitchison TJ. Evolution of a dynamic cytoskeleton. Philos Trans R Soc Lond B Biol Sci. 1995;349:299–304. [PubMed: 8577841]
Small JV, Stradal T, Vignal E. et al. The lamellipodium: Where motility begins. Trends Cell Biol. 2002;12:112–120. [PubMed: 11859023]
Bader MF, Doussau F, Chasserot-Golaz S. et al. Coupling actin and membrane dynamics during calcium-regulated exocytosis: A role for Rho and ARF GTPases. Biochim Biophys Acta. 2004;1742:37–49. [PubMed: 15590054]
Goldberg MB. Actin-based motility of intracellular microbial pathogens. Microbiol Mol Biol Rev. 2001;65:595–626. [PMC free article: PMC99042] [PubMed: 11729265]
Pollard TD, Borisy GG. Cellular motility driven by assembly and disassembly of actin filaments. Cell. 2003;112:453–465. [PubMed: 12600310]
Pruyne D, Legesse-Miller A, Gao L. et al. Mechanisms of polarized growth and organelle segregation in yeast. Annu Rev Cell Dev Biol. 2004;20:559–591. [PubMed: 15473852]
Szymanski DB. Breaking the WAVE complex: The point of Arabidopsis trichomes. Curr Opin Plant Biol. 2005;8:103–112. [PubMed: 15653407]
Baluska F, Salaj J, Mathur J. et al. Root hair formation: F-actin-dependent tip growth is initiated by local assembly of profilin-supported F-actin meshworks accumulated within expansin-enriched bulges. Dev Biol. 2000;227:618–632. [PubMed: 11071779]
Ringli C, Baumberger N, Diet A. et al. ACTIN2 is essential for bulge site selection and tip growth during root hair development of Arabidopsis. Plant Physiol. 2002;129:1464–1472. [PMC free article: PMC166735] [PubMed: 12177460]
Vidali L, McKenna ST, Hepler PK. Actin polymerization is essential for pollen tube growth. Mol Biol Cell. 2001;12:2534–2545. [PMC free article: PMC58611] [PubMed: 11514633]
Mathur J, Hulskamp M. Microtubules and microfilaments in cell morphogenesis in higher plants. Curr Biol. 2002;12:R669–R676. [PubMed: 12361589]
Field C, Li R, Oegema K. Cytokinesis in eukaryotes: A mechanistic comparison. Curr Opin Cell Biol. 1999;11:68–90. [PubMed: 10047527]
Hales KG, Bi E, Wu JQ. et al. Cytokinesis: An emerging unified theory for eukaryotes? Curr Opin Cell Biol. 1999;11:717–725. [PubMed: 10600712]
Lénárt P, Bacher CP, Daigle N. et al. A contractile nuclear actin network drives chromosome congression in oocytes. Nature. 2005;436:812–818. [PubMed: 16015286]
Bamburg JR, Drubin DG. Actin depolymerizing factor (ADF)/cofilin. In: Kreis T, Vale R, eds. Guidebook to the Cytoskeletal and Motor Proteins. Oxford: Oxford University Press. 1999:19–23.
Paavilainen VO, Bertling E, Falck S. et al. Regulation of cytoskeletal dynamics by actin-monomerbinding proteins. Trends Cell Biol. 2004;14:386–394. [PubMed: 15246432]
McCurdy DW, Kovar DR, Staiger CJ. Actin and actin-binding proteins in higher plants. Protoplasma. 2001;215:89–104. [PubMed: 11732068]
Frankel S. Arps, divergent members. In: Kreis T, Vale R, eds. Guidebook to the Cytoskeletal and Motor Proteins. Oxford: Oxford University Press. 1999:49–52.
van den Ent F, Amos LA, Löwe J. Prokaryotic origin of the actin cytoskeleton. Nature. 2001;413:39–44. [PubMed: 11544518]
Jones LJ, Carballido-Lopez R, Errington J. Control of cell shape in bacteria: Helical, actin-like filaments in Bacillus subtilis. Cell. 2001;104:913–922. [PubMed: 11290328]
Pollard TD. Profilins. In: Kreis T, Vale R, eds. Guidebook to the Cytoskeletal and Motor Proteins. Oxford: Oxford University Press. 1999:117–120.
dos Remedios GC, Chhabra D, Kekic M. et al. Actin binding proteins: Regulation of cytoskeletal microfilaments. Physiol Rev. 2003;83:433–473. [PubMed: 12663865]
Rivero F, Eichinger L. The microfilament system of Dictysotelium discoideum. In: Loomis WF, Kuspa A, eds. Dictyostelium Genomics. Norfolk: Horizon Bioscience. 2005:125–171.
Koonin EV, Aravind L. Dynein light chains of the Roadblock/LC7 group belong to an ancient protein superfamily implicated in NTPase regulation. Curr Biol. 2000;10:R774–R776. [PubMed: 11084347]
Kurzbauer R, Teis D, de Araujo ME. et al. Crystal structure of the p14/MP1 scaffolding complex: How a twin couple attaches mitogen-activated protein kinase signaling to late endosomes. Proc Natl Acad Sci USA. 2004;101:10984–10989. [PMC free article: PMC503730] [PubMed: 15263099]
Maciver SK, Hussey P. The ADF/cofilin family: Actin-remodeling proteins. Genome Biol. 2002;3:R3007. [PMC free article: PMC139363] [PubMed: 12049672]
Van Troys M, Vanderkerckhove J, Ampe C. Structural modules in actin-binding proteins: Towards a new clasification. Biochim Biophys Acta. 1999;1448:323–348. [PubMed: 9990286]
Hatanaka H, Ogura K, Moriyama K. et al. Tertiary structure of destrin and structural similartity between two actin-regulating protein families. Cell. 1996;85:1047–1055. [PubMed: 8674111]
Kwiatkowski DJ. Functions of gelsolin: Motility, signaling, apoptosis, cancer. Curr Opin Cell Biol. 1999;11:103–108. [PubMed: 10047530]
Bi X, Corpina RA, Goldberg J. Structure of the Sec23/24-Sar1 prebudding complex of the COPII vesicle coat. Nature. 2002;419:271–277. [PubMed: 12239560]
Higgs HN, Pollard TD. Regulation of actin filament network formation through Arp2/3 complex: Activation by a diverse array of proteins. Annu Rev Biochem. 2001;70:649–676. [PubMed: 11395419]
Machesky LM, Gould KL. The Arp2/3 complex: A multifunctional actin organizer. Curr Opin Cell Biol. 1999;11:117–121. [PubMed: 10047519]
Beltzner CC, Pollard TD. Identification of functionally important residues of Arp2/3 complex by analysis of homology models from diverse species. J Mol Biol. 2004;336:551–565. [PubMed: 14757065]
Mathur J. The ARP2/3 complex: Giving plant cells a leading edge. BioEssays. 2005;27:377–387. [PubMed: 15770684]
Giaever G, Chu AM, Ni L. et al. Functional profiling of the Saccharomyces cerevisiae genome. Nature. 2002;418:387–391. [PubMed: 12140549]
Evangelista M, Pruyne D, Amberg DC. et al. Formins direct Arp2/3-independent actin filament assembly to polarize cell growth in yeast. Nature Cell Biol. 2002;4:32–41. [PubMed: 11740490]
Li F, Higgs HN. The mouse formin mDia1 is a potent actin nucleation factor regulated by autoinhibition. Curr Biol. 2003;13:1335–1340. [PubMed: 12906795]
Kovar DR, Kuhn JR, Tichy AL. et al. The fission yeast cytokinesis formin Cdc12p is a barbed end actin filament capping protein gated by profilin. J Cell Biol. 2003;161:885–887. [PMC free article: PMC2172974] [PubMed: 12796476]
Otomo T, Tomchick DR, Otomo C. et al. Structural basis of actin filament nucleation and processive capping by a formin homology 2 domain. Nature. 2005;433:488–494. [PubMed: 15635372]
Michelot A, Guerin C, Huang S. et al. The Formin Homology 1 domain modulates the actin nucleation and bundling activity of Arabidopsis FORMIN1. Plant Cell. 2005;17:2296–2313. [PMC free article: PMC1182490] [PubMed: 15994911]
Cvrcková F, Novotny M, Pícková D. et al. Formin homology 2 domains occur in multiple contexts in angiosperms. BMC Genomics. 2004;5:44. [PMC free article: PMC509240] [PubMed: 15256004]
Rivero F, Muramoto T, Meyer A-K. et al. A comparative sequence analysis reveals a common GBD/FH3-FH1-FH2-DAD architecture in formins from Dictyostelium, fungi and metazoa. BMC Genomics. 2005;6:28. [PMC free article: PMC555941] [PubMed: 15740615]
Higgs HN, Peterson KJ. Phylogenetic analysis of the formin homology 2 domain. Mol Biol Cell. 2005;16:1–13. [PMC free article: PMC539145] [PubMed: 15509653]
Simpson AG, Roger AJ. The real ‘kingdoms’ of eukaryotes. Curr Biol. 2004;14:R693–R696. [PubMed: 15341755]
Wear MA, Cooper JA. Capping protein: New insights into mechanism and regulation. Trends Biochem Sci. 2004;29:418–428. [PubMed: 15362226]
Korenbaum E, Rivero F. Calponin homology domains at a glance. J Cell Sci. 2002;115:3543–3545. [PubMed: 12186940]
Tirnauer JS, Bierer BE. EB1 proteins regulate microtubule dynamics, cell polarity, and chromosome stability. J Cell Biol. 2000;149:761–766. [PMC free article: PMC2174556] [PubMed: 10811817]
van FR, Huijkman NC, Boomsma C. et al. Genomics of the human carnitine acyltransferase genes. Mol Genet Metab. 2000;71:139–153. [PubMed: 11001805]
Watanabe A, Yonemura I, Gonda K. et al. Cloning and sequencing of the gene for a Tetrahymena fimbrin-like protein. J Biochem. 2000;127:85–94. [PubMed: 10731670]
Mulder NJ, Apweiler R, Attwood TK. et al. The InterPro Database, 2003 brings increased coverage and new features. Nucleic Acids Res. 2003;31:315–318. [PMC free article: PMC165493] [PubMed: 12520011]
Vardar D, Chishti AH, Frank BS. et al. Villin-type headpiece domains show a wide range of F-actin-binding affinities. Cell Motil Cytoskeleton. 2002;52:9–21. [PubMed: 11977079]
Hertzog M, van Heijenoort C, Didry D. et al. The beta-thymosin/WH2 domain: Structural basis for the switch from inhibition to promotion of actin assembly. Cell. 2004;117:611–623. [PubMed: 15163409]
Quinlan ME, Heuser JE, Kerkhoff E. et al. Drosophila Spire is an actin nucleation factor. Nature. 2005;433:382–388. [PubMed: 15674283]
Paunola E, Mattila PK, Lappalainen P. WH2 domain: A small versatile adapter for actin monomers. FEBS Lett. 2002;513:92–97. [PubMed: 11911886]
Ponting CP, Russell JB. Identification of distant homologues of fibroblast growth factors suggests a common ancestor for all beta-trefoil proteins. J Mol Biol. 2000;302:1041–1047. [PubMed: 11183773]
McCann RO, Craig SW. The I/LWEQ module: A conserved sequence that signifies F-actin binding in functionally diverse proteins from yeast to mammals. Proc Natl Acad Sci USA. 1997;94:5679–5684. [PMC free article: PMC20838] [PubMed: 9159132]
Kull FJ, Sablin EP, Lau R. et al. Crystal structure of the kinesin motor domain reveals a structural similarity to myosin. Nature. 1996;380:550–555. [PMC free article: PMC2851642] [PubMed: 8606779]
Kull FJ, Vale RD, Fletterick RJ. The case for a common ancestor: Kinesin and myosin motor proteins and G proteins. J Musc Res Cell Motil. 1998;19:877–886. [PubMed: 10047987]
Leipe DD, Wolf YI, Koonin EV. et al. Classification and evolution of P-loop GTPases and related ATPases. J Mol Biol. 2002;317:41–72. [PubMed: 11916378]
Thompson RF, Langford GM. Myosin superfamily evolutionary history. Anat Rec. 2002;268:276–289. [PubMed: 12382324]
Hodge T, Cope MJ. A myosin family tree. J Cell Sci. 2000;113:3353–3354. [PubMed: 10984423]
Richards TA, Cavalier-Smith T. Myosin domain evolution and the primary divergence of eukaryotes. Nature. 2005;436:1113–1118. [PubMed: 16121172]
Csete ME, Doyle JC. Reverse engineering of biological complexity. Science. 2002;295:1664–1668. [PubMed: 11872830]
Weaver AM, Young ME, Lee WL. et al. Integration of signals to the Arp2/3 complex. Curr Opin Cell Biol. 2003;15:23–30. [PubMed: 12517700]
Deeks MJ, Hussey P, Davies B. Formins: Intermediates in signal transduction cascades that affect cytoskeletal reorganization. Trends Plant Sci. 2002;7:492–498. [PubMed: 12417149]
Johnson DI. Cdc42: An essential Rho-type GTPase controlling eukaryotic cell polarity. Microbiol Mol Biol Rev. 1999;63:54–105. [PMC free article: PMC98957] [PubMed: 10066831]
Etienne-Manneville S, Hall A. Rho GTPases in cell biology. Nature. 2002;420:629–635. [PubMed: 12478284]
Sorokina EM, Chernoff J. Rho-GTPases: New members, new pathways. J Cell Biochem. 2005;94(2):225–231. [PubMed: 15543593]
Jekely G. Small GTPases and the evolution of the eukaryotic cell. BioEssays. 2003;25(11):1129–1138. [PubMed: 14579253]
Cotteret S, Chernoff J. The evolutionary history of effectors downstream of Cdc42 and Rac. Genome Biol. 2002;3:R0002. [PMC free article: PMC139012] [PubMed: 11864373]
Copyright © 2000-2013, Landes Bioscience.
Bookshelf ID: NBK5970
PubReader format: click here to try


  • PubReader
  • Print View
  • Cite this Page

Related information

  • PMC
    PubMed Central citations
  • PubMed
    Links to pubmed

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...