Register for Mechanobiology 2016!

Twinfilin, a molecular mailman for actin monomers
Sandra Palmgren, Maria Vartiainen, Pekka Lappalainen


Twinfilin is a ubiquitous actin-monomer-binding protein that is composed of two ADF-homology domains. It forms a 1:1 complex with ADP-actin-monomers, inhibits nucleotide exchange on actin monomers and prevents assembly of the monomer into filaments. The two ADF-H domains in twinfilin probably have 3D structures similar to those of the ADF/cofilin proteins and overlapping actin-binding sites. Twinfilin also interacts with PtdIns(4,5)P2, which inhibits its actin-monomer-sequestering activity in vitro. Mutations in the twinfilin gene result in defects in the bipolar budding pattern in S. cerevisiae and in a rough eye phenotype and aberrant bristle morphology in Drosophila melanogaster. These phenotypes are caused by the uncontrolled polymerization of actin filaments in the absence of twinfilin. Studies on budding yeast suggest that twinfilin contributes to actin filament turnover by localizing actin monomers, in their `inactive' ADP-form, to the sites of rapid filament assembly. This is mediated through direct interactions between twinfilin and capping protein. Therefore, twinfilin might serve as a link between rapid actin filament depolymerization and assembly in cells.


The actin cytoskeleton participates in a wide variety of cellular functions, including endocytosis, secretion, motility, cell division and intracellular signal transduction. In cells, actin exists as both filamentous (F-actin) and monomeric (G-actin) forms, and these dynamic pools of actin are regulated by a large number of actin-binding proteins. Most actin-binding proteins interact only with filaments and can promote nucleation, crosslinking and capping of actin filaments. Other actin-binding proteins interact with both actin monomers and filaments; some bind only the monomers ( Sheterline et al., 1998).

Actin-monomer-binding proteins play an important role in cells by controlling the size and localization of the cytoplasmic actin-monomer pool and by regulating the incorporation of actin monomers into filaments. Three classes of small actin-monomer-binding protein exist in organisms as diverse as yeasts and mammals: ADF/cofilin, profilin and twinfilin ( Bamburg et al., 1999; Pollard et al., 2000; Lappalainen et al., 1998). In addition to these three ubiquitous proteins, a fourth class of small actin-monomer-binding protein, β-thymosins, is present in vertebrates.β -thymosins are small ATP-actin-monomer-binding proteins whose role is to sequester a large pool of unpolymerized actin in cells ( Pollard et al., 2000).β -thymosins are expressed typically in highly motile cells such as platelets and differentiating neurons ( Border et al., 1993).

ADF/cofilins are small (15-19 kDa) proteins that bind both actin monomers and filaments. They promote rapid actin dynamics in cells by increasing the depolymerization rate at the minus end (pointed end) of actin filaments ( Carlier et al., 1997; Rosenblatt et al., 1997; Lappalainen and Drubin, 1997). ADF/cofilins may also have a weak actin filament-severing activity, which would account for the formation of new filament ends ( Chan et al., 2000). ADF/cofilins interact with ADP-acin-monomers and filaments with higher affinities than ATP-actins and inhibit the spontaneous nucleotide exchange on actin monomers ( Bamburg et al., 1999). These small actin-binding proteins appear to be essential for viability in all eukaryotes: null mutations in Saccharomyces cerevisiae, Caenorhabditis elegans and Drosophila melanogaster ADF/cofilin genes are lethal ( Moon et al., 1993; McKim et al., 1994; Gunsalus et al., 1995).

Profilins are small (13-16 kDa) proteins that bind only actin monomers. Profilins promote assembly at the plus end (barbed end) of the filament, but they also suppress spontaneous filament nucleation. Therefore, profilins function as actin-monomer-sequestering proteins in the absence of free filament plus ends ( Pantaloni and Carlier, 1993; Vinson et al., 1998). By contrast with ADF/cofilins, profilins bind ATP-actin-monomers with higher affinity than ADP-actin (Goldschmidt-Cleremont et al., 1991; Pantaloni and Carlier, 1993; Vinson et al., 1998). Furthermore, profilins, enhance the nucleotide exchange on actin monomers and, at least in the yeasts S. cerevisiae and S. pombe, this activity is essential in vivo ( Wolven et al., 2000; Lu and Pollard, 2001).

The third evolutionarily conserved class of small actin-monomer-binding protein, twinfilins, was identified more recently. The biological roles and actin interactions of twinfilin are less well characterized than those of ADF/cofilins and profilins. Here we review recent advances in our knowledge about this ubiquitous actin-monomer-binding protein and discuss its possible role in actin filament turnover.

Structure and biochemical properties of twinfilin

Twinfilin was originally identified in Saccharomyces cerevisiae through its sequence homology to ADF/cofilin proteins ( Lappalainen et al., 1998; Goode et al., 1998). Twinfilin homologues have since been identified in other eukaryotes, including the mouse and Drosophila ( Table 1), which suggests that the protein is present in eukaryotes ranging from yeast to mammals ( Vartiainen et al., 2000, Wahlström et al., 2001); however, twinfilin homologues have not been found in any plants.

View this table:
Table 1.

Biochemical and cell biological activities of mouse, yeast, and Drosophila twinfilins

Twinfilin is a 37-40 kDa protein composed of two ADF/cofilin-like (ADF-H) domains connected by a short linker region and followed by a C-terminal tail of ∼20 residues ( Fig. 1A). The two ADF-H domains are ∼20% homologous to ADF/cofilin and to each other ( Lappalainen et al., 1998). Interestingly, the individual twinfilin ADF-H domains are more similar across species than are the N- and C-terminal ADF-H domains within a species ( Fig. 1B). This suggests that the ADF-H domain duplicated once before the divergence of the fungal and animal lineages. The positions of secondary structural elements are relatively well conserved between ADF/cofilin and twinfilin; both ADF-H domains probably therefore have tertiary structures similar to that of ADF/cofilin proteins. Furthermore, insertions in the twinfilin ADF-H domains are located in the predicted loop regions ( Lappalainen et al., 1998).

Fig. 1.

The domain structure of twinfilin. (A) Twinfilin is composed of two domains homologous to the ADF/cofilin proteins (ADF-H domain), separated by a short,∼ 30 amino acid, linker and followed by a C-terminal tail region. The ADF-H domains are approximately 20% identical to each other, and they both appear to be involved in interactions with actin monomers. (B) A phylogenetic tree of twinfilin ADF-H domains from different species as produced by Clustal-X software. The N-terminal and C-terminal ADF-H domains of twinfilins from different species form two independent branches in the tree, demonstrating that each one of the ADF-H domains between different species are more homologous than the N- and C-terminal ADF-H domains within any particular twinfilin. This suggests that the two ADF-H domains of twinfilin were already present before the divergence of fungal and animal lineages. Databases and accession numbers for twinfilin sequences are: S. pombe twinfilin (EMBL, AL034490); S. cerevisiae twinfilin (EMBL, Z72865); C. elegans twinfilin (EMBL, U46668); D. melanogaster twinfilin (EMBL, AE003703); H. sapiens twinfilin-1 (PIR, A55922); H. sapiens twinfilin-2 (EMBL, Y17169); M. musculus twinfilin-1 (GenBank, U82324); and M. musculus twinfilin-2 (EMBL, Y17808).

All twinfilins characterized so far are biochemically similar. The human homologue of twinfilin was originally identified as a tyrosine kinase ( Beeler et al., 1994), but recent studies have demonstrated that neither mammalian ( Vartiainen et al., 2000; Rohwer et al., 1999) nor fungal ( Goode et al., 1998) twinfilins have kinase activity. Twinfilins also lack sequence homology to known protein kinases. Instead, all twinfilins characterized so far are actin-monomer-binding proteins ( Goode et al., 1998; Vartiainen et al., 2000; Wahlström et al., 2001). They appear to form a 1:1 complex with actin monomers: in actin filament sedimentation assays, yeast twinfilin shifts actin monomers to the supernatant fraction in a ∼1:1 molar ratio ( Goode et al., 1998). In addition, the migration of the mouse twinfilin—actinmonomer complex in a sucrose gradient is consistent with a 1:1 molar ratio ( Vartiainen et al., 2000). The ability of twinfilin to prevent actin filament assembly in sedimentation assays and in kinetic pyrene actin assembly assays shows that twinfilin is an efficient actin-monomer-sequestering protein ( Goode et al., 1998; Vartiainen et al., 2000; Wahlström et al., 2001). Furthermore, yeast twinfilin inhibits the spontaneous nucleotide exchange on actin monomers in a manner similar to that of ADF/cofilins ( Goode et al., 1998).

The residues important for actin-monomer binding in ADF/cofilins are well conserved in both twinfilin ADF-H domains ( Lappalainen et al., 1998), and recent mutagenesis studies on yeast twinfilin demonstrated that these corresponding residues are important for actin-monomer binding ( Palmgren et al., 2001). This suggests that twinfilins and ADF/cofilins interact with actin monomers at partly or completely overlapping interfaces. Both ADF/cofilin ( Maciver and Weeds, 1994; Carlier et al., 1997) and twinfilin ( Palmgren et al., 2001) prefer to interact with ADP-actin-monomers; in contrast, the actin-monomer-binding proteins profilin and thymosin-β4 bind ATP-actin-monomers with higher affinities ( Carlier et al., 1993; Pantaloni and Carlier, 1993; Vinson et al., 1998). Therefore, ADF/cofilin and twinfilin may interact with newly depolymerized, assembly-incompetent ADP-actin-monomers, whereas profilin and thymosin-β4 regulate the dynamics of the assembly-competent ATP-actin-monomer pool.

Despite these similarities between twinfilin and ADF/cofilin, the proteins differ in other aspects. In contrast to ADF/cofilins, twinfilins do not detectably bind actin filaments and do not promote the depolymerization of actin filaments ( Goode et al., 1998; Vartiainen et al., 2000; Wahlström et al., 2001). Such a lack of actin filament binding is further supported by the fact that the residues essential for binding of ADF/cofilin to actin filaments ( Lappalainen et al., 1997; Ono et al., 2001) are not conserved in twinfilins ( Lappalainen et al., 1998).

Localization and biological function of twinfilin

The cellular role of twinfilin has been explored in three different experimental systems: Saccharomyces cerevisiae, Drosophila melanogaster and cultured murine cell lines. In the multicellular organisms, twinfilin is expressed ubiquitously in tissues and during development ( Vartiainen et al., 2000; Wahlström et al., 2001), which indicates that it functions in basic cell biological processes common to all cell types.

The subcellular localization of twinfilin strongly suggests its involvement in actin filament dynamics. In budding yeast, twinfilin is diffusely cytoplasmic but is also concentrated at cortical actin patches, which are sites of rapid actin filament turnover ( Palmgren et al., 2001). In cultured mammalian cells, twinfilin has a strong punctate cytoplasmic staining pattern, and a significant proportion is also localized to cellular processes rich in actin monomers and filaments ( Vartiainen et al., 2000). In developing Drosophila bristles, twinfilin is diffusely cytoplasmic and concentrated at the ends of actin bundles ( Wahlström et al., 2001). Further support for the involvement of twinfilin in actin filament dynamics comes from the observation that overexpressing twinfilin in yeast and mammalian cells results in the formation of abnormal actin structures: in yeast the actin cytoskeleton is depolarized and actin bars accumulate in the cytoplasm ( Goode et al., 1998), and in mammalian cells there are fewer stress fibers and worm-like actin filament structures appear ( Vartiainen et al., 2000).

Deletion of the twinfilin gene in budding yeast does not produce a clear phenotype, but the twfΔ mutants appear to have somewhat abnormal cortical actin patches and defects in the bipolar bud-site selection pattern ( Goode et al., 1998; Fig. 2B). The role of twinfilin during the development of a multicellular organism has been investigated in a Drosophila strain carrying a P-element insertion in the first intron of the twinfilin gene. These flies show a dramatic decrease in twinfilin mRNA and protein. The mutants are smaller and less active than wild-type flies. Furthermore, they have morphological defects such as a rough eye phenotype, but the most obvious external phenotype is seen in bristles ( Wahlström et al., 2001). Drosophila bristles are sensory organs formed from a single cell as a long extension that is supported by actin filament bundles. Each bundle is composed of repeated units attached end to end. As the bristle elongates, actin filaments are formed at the tip of the bristle and then progressively packed together to form new units that have maximally crosslinked filaments ( Tilney et al., 2000). In the twinfilin mutants, the bristles are shorter than those of wild-type flies and are often bent ( Fig. 2D). The mutant bristles also have a rough surface morphology: the ridges and grooves of the bristles are highly irregular. Phalloidin staining of the developing bristles demonstrated that the actin bundles in the mutant bristles are often twisted and misoriented and that the twinfilin mutant bristles contain many abnormal F-actin-containing spots or tiny actin bundles. These abnormal actin structures explain the defective mutant bristle morphology observed in the adult ( Wahlström et al., 2001).

Fig. 2.

Effects of mutations in yeast and Drosophila twinfilin. The most obvious phenotype of twinfilin deletion (Δtwf) yeast strains is the random bipolar budding pattern. Wild-type diploid yeast cells (A) show a bipolar budding pattern, whereas the bud-scars, as visualized by calcofluor staining, show random distribution in Δtwftwf yeast cells (B). A strong hypomorphic mutation in the Drosphila twinfilin gene results in multiple morphological defects, including aberrant bristle morphology. In wild-type flies (C) the macrochaetae (arrows) are straight and show regular longitudinal groove and ridge structures. In twinfilin mutant flies (D) the macrochaetae (arrows) are often bent and show irregular surface morphology. Bars, 5 μm (A,B); 25 μm (C,D).

The role of twinfilin in actin filament turnover

Twinfilin is an abundant protein present at a 1:10 twinfilin:actin ratio in budding yeast ( Palmgren et al., 2001). Its abundance and high affinity for ADP-actin-monomers suggests that twinfilin can sequester a large proportion of actin monomers in the cell at any given time. Furthermore, the preference of twinfilin for ADP-actin-monomers and its inhibitory effect on nucleotide exchange suggest that it significantly affects the size and nucleotide status of the actin-monomer pool in yeast cells. This is especially notable considering that the actin-monomer pool in S. cerevisiae is much smaller than that in the cells of motile organisms ( Karpova et al., 1995).

All biochemical data indicate that twinfilin is an actin-monomer-binding protein, but it localizes to the cortical actin filament structures in yeast and mammalian cells, and to the ends of the actin filament bundles in developing Drosophila bristles ( Palmgren et al., 2001; Vartiainen et al., 2000; Wahlström et al., 2001). In yeast, the localization of twinfilin to cortical actin patches is dependent on the presence of intact actin filaments, because twinfilin is diffusely cytoplasmic in cells treated with the actin-filament-depolymerizing agent latrunculin-A ( Palmgren et al., 2001). These data suggest that twinfilin also interacts with other components of the cortical actin cytoskeleton.

We screened several yeast strains with mutations in known components of the cortical actin cytoskeleton and discovered that twinfilin does not localize to the cortical actin patches in strains carrying deletions in CAP1 and CAP2, which encode the two subunits of the heterodimeric actin filament plus end capping protein ( Palmgren et al., 2001). Deleting either subunit of the capping protein leads to a loss of the other subunit ( Amatruda et al., 1992). Twinfilin co-immunoprecipitates with capping protein from yeast extracts, and the purified proteins directly interact as shown by native PAGE ( Palmgren et al., 2001). The twinfilin-capping protein interaction appears to be conserved during evolution, because mouse twinfilin and α1β2 capping protein also interact in vitro ( Palmgren et al., 2001). Furthermore, in developing Drosophila bristles, twinfilin localizes to the plus ends of the actin bundles, where capping protein is reported to be ( Wahlström et al., 2001; Hopmann et al., 1996). Correct localization of twinfilin to the cortical actin cytoskeleton also appears to require an interaction with an actin monomer, because a mutant twinfilin that is unable to interact with actin monomers in vitro is diffusely cytoplasmic when expressed in yeast ( Palmgren et al., 2001).

It is possible that the biological role of twinfilin is to recycle actin monomers, in their `inactive' ADP-form, to the sites of actin filament assembly in cells ( Fig. 3). In such a model, newly depolymerized ADP-actin-monomers could be delivered to twinfilin by ADF/cofilin, as suggested by the ability of twinfilin to form a stable complex with ADP-actin-monomers, and the overlapping actin-binding interfaces of twinfilin and ADF/cofilin. Twinfilin might keep the delivered actin monomers in their `inactive' ADP form, because yeast twinfilin inhibits the spontaneous nucleotide exchange on actin monomers ( Goode et al., 1998). Therefore, twinfilin would prevent actin filament assembly at undesired locations in the cells. The twinfilin—ADP-actin-monomer complexes could then be concentrated to the cellular sites of rapid actin filament assembly through interactions between twinfilin and capping protein. After dissociating from twinfilin, the actin monomer would undergo nucleotide exchange and assemble into the plus end of the filament either spontaneously or by a profilin-catalyzed mechanism ( Fig. 3). Further support for a role for twinfilin as an actin-monomer-recycling/localizing protein is provided by the genetic interactions between twinfilin and cofilin ( Goode et al., 1998; Wahlström et al., 2001) and between twinfilin and profilin ( Wolven et al., 2000). In yeast, the twinfilin-null mutant shows synthetic lethality with the cofilin mutant cof1-22: an allele that exhibits diminished actin filament depolymerization activity ( Lappalainen and Drubin, 1997) and shows synthetic lethality with the profilin mutant pfy1-4, which has defects in actin nucleotide exchange activity ( Wolven et al., 2000). In combination with possible defects in the actin filament localization activity of Δtwf1 cells, either one of these mutations would be expected to result in a significant decrease in the number of assembly competent (ATP-actin) monomers at the sites of rapid actin filament assembly.

Fig. 3.

A model for the role of twinfilin in actin filament turnover. ADF/cofilin feeds new monomers to the cytoplasmic pool by depolymerizing actin filaments at their minus-ends. Because twinfilin binds ADP-actin-monomers with a high affinity and through an overlapping interface with ADF/cofilin, it may sequester actin monomers from ADF/cofilin. Twinfilin inhibits the spontaneous nucleotide exchange on actin monomers and may localize actin monomers, in their inactive ADP-form, to the sites of rapid actin filament assembly. The localization of twinfilin to the sites of rapid actin filament assembly is mediated through interactions with capping protein. After localization, the ADP-actin-monomer is released from twinfilin and subsequent nucleotide exchange and assembly into the plus-end of the filament may be catalyzed by profilin. The activities of cofilin, profilin and twinfilin are downregulated by phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2).

The activities of many actin-binding proteins are regulated by phosphorylation or by binding to phosphatidylinositols. In mammalian cells, the activity of twinfilin is regulated by phosphorylation (M.V. and P.L., unpublished), whereas phosphorylated twinfilin has not been detected in yeast (S.P. and P.L., unpublished). However, yeast twinfilin interacts with phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P2) in vitro, and this interaction downregulates the actin-monomer-sequestering activity of twinfilin ( Palmgren et al., 2001). The interaction with PtdIns(4,5)P2 might keep twinfilin from sequestering actin monomers at those cell regions where the nucleation and assembly of actin filaments is rapid. In general, the proteins that promote the assembly of actin filaments (e.g. WASP) are upregulated by PtdIns(4,5)P2 ( Rohatgi et al., 2000; Higgs and Pollard, 2000), whereas the activity of the proteins that promote the disassembly of actin filaments (e.g. ADF/cofilin) is downregulated ( Yonezawa et al., 1990). PtdIns(4,5)P2 also promotes the assembly of actin filaments under in vivo conditions, as demonstrated by studies of actin polymerization by PtdIns(4,5)P2-containing vesicles in cell extracts ( Ma et al., 1998) and living cells ( Rozelle et al., 2000; Tall et al., 2000). It remains to be shown whether the interaction between twinfilin and PtdIns(4,5)P2 takes place in vivo and if this interaction has any physiological relevance.

Conclusions and perspectives

Twinfilin is an abundant ADP-actin-monomer-binding protein concentrated at the cortical actin cytoskeleton and at the ends of actin filament bundles ( Vartiainen et al., 2000; Palmgren et al., 2001; Wahlström et al., 2001). Localization of twinfilin to the cortical actin cytoskeleton requires a direct interaction with capping protein. Studies on yeast twinfilin suggest that this ubiquitous actin-monomer-sequestering protein contributes to cytoskeletal dynamics by recycling/localizing ADP-actin-monomers to the cellular sites of rapid actin filament assembly ( Palmgren et al., 2001). Therefore, twinfilin might function as a link between the ADF/cofilin-induced rapid actin filament depolymerization and the profilin-induced actin filament assembly ( Fig. 3). Deleting twinfilin in budding yeast leads to enlargement of the cortical actin patches and to defects in bipolar bud-site selection pattern. Furthermore, a yeast twinfilin-null mutation is synthetically lethal in combination with certain cofilin and profilin mutants, which suggests that twinfilin, together with cofilin and profilin, is intimately involved in rapid actin filament turnover ( Goode et al., 1998; Wolven et al., 2000). In multicellular organisms, the activity of twinfilin is required for several developmental processes. A mutation in Drosophila twinfilin results in small adult size, a rough eye phenotype and aberrant bristle morphology. The defects in bristle morphogenesis arise from the uncontrolled nucleation/polymerization of actin filaments in developing bristles in the absence of twinfilin, which further supports a role for twinfilin in regulating the dynamics and localization of the cytoplasmic actin-monomer pool ( Wahlström et al., 2001).

In the future, it remains to be shown whether twinfilin's role is indeed to localize ADP-actin-monomers in cells, or whether it has a more complex role in actin dynamics. Furthermore, it will be important to elucidate the molecular mechanism of the twinfilin—actin-monomer interaction. It is still unclear why twinfilin, which forms a 1:1 complex with actin monomers, has two ADF-H domains. At least in theory, both ADF-H domains of twinfilin should be able to interact with one actin monomer. Further studies must also elucidate how actin monomers are released from twinfilin and whether the interaction with capping protein affects the actin-monomer-binding activity of twinfilin. Finally, it will be interesting to examine how various signal transduction pathways regulate the activity of twinfilin in different cell types and during various cell biological processes.


We thank Tapio Heino and Gudrun Wahlström for critical reading of the manuscript and help with Fig. 2, and Pauli Ojala for advice on phylogenetic analysis of twinfilins. This work was supported by grants from Academy of Finland and Biocentrum Helsinki.


View Abstract