Actin nucleators in the nucleus: an emerging theme

Summary

Actin is an integral component of the cytoskeleton, forming a plethora of macromolecular structures that mediate various cellular functions. The formation of such structures relies on the ability of actin monomers to associate into polymers, and this process is regulated by actin nucleation factors. These factors use monomeric actin pools at specific cellular locations, thereby permitting rapid actin filament formation when required. It has now been established that actin is also present in the nucleus, where it is implicated in chromatin remodelling and the regulation of eukaryotic gene transcription. Notably, the presence of typical actin filaments in the nucleus has not been demonstrated directly. However, studies in recent years have provided evidence for the nuclear localisation of actin nucleation factors that promote cytoplasmic actin polymerisation. Their localisation to the nucleus suggests that these proteins mediate collaboration between the cytoskeleton and the nucleus, which might be dependent on their ability to promote actin polymerisation. The nature of this cooperation remains enigmatic and it will be important to elucidate the physiological relevance of the link between cytoskeletal actin networks and nuclear events. This Commentary explores the current evidence for the nuclear roles of actin nucleation factors. Furthermore, the implication of actin-associated proteins in relaying exogenous signals to the nucleus, particularly in response to cellular stress, will be considered.

Introduction

Actin microfilaments are key elements of the cytoskeleton. They are composed of actin subunits, known as globular actin (G-actin), that assemble to form filamentous actin (F-actin). These filaments are polarised, with a fast-growing (barbed or plus) end and a slow-growing (pointed or minus) end (Kasai et al., 1962; Huxley, 1963; Oosawa, 1977; Wang, 1985). Actin dynamics are influenced by a vast number of regulatory proteins in the cell, which bind to and modify G- and F-actin, resulting in the formation and disassembly of a variety of macromolecular structures, such as lamellipodia (Chhabra and Higgs, 2007). An important stage of actin regulation is the initial step of actin filament formation, known as nucleation, in which actin monomers first combine. During the nucleation process in vitro, actin monomers assemble to form dimeric and then trimeric complexes, which are known as nucleation seeds. This process is the rate-limiting step of actin polymerisation in vitro, where the actin concentration is limiting, whereas the subsequent addition of further actin monomers to these stable nucleation seeds is energetically favourable and results in rapid filament elongation (Fig. 1A). In cells, spontaneous actin polymerisation is inhibited by proteins that sequester actin monomers, such as profilin and thymosin β4 (Safer et al., 1990; Schlüter et al., 1997). Consequently, cells use factors that overcome this inhibition. These nucleation factors contain actin-binding sites and either mimic or stabilise spontaneously formed actin nucleation seeds, which enable the fast generation of actin networks in response to specific signals. Additionally, the localisation of such nucleation factors allows cells to direct filament formation to particular sites.

Fig. 1.

Actin filament formation involves three classes of actin nucleation factors. (A) Actin filaments assemble from actin monomers. The initial dimerisation and trimerisation steps are energetically unfavourable. Nucleation factors overcome this, allowing rapid polymerisation to occur. (B) The Arp2/3 complex, once activated by an NPF nucleates a new filament from the side of an existing filament, causing filament branching at a 70° angle (left). A formin FH2-domain dimer associates with the barbed end of an actin filament and the FH1 domains recruit and deliver profilin-bound actin to the barbed end. The formin moves processively with the barbed end as the actin filament elongates (middle). Tandem monomer-binding nucleators bring together actin monomers through tandem G-actin-binding motifs to form an actin nucleus (right). The diagram depicts cordon bleu, which nucleates a filament by assembling and stabilising a trimeric actin nucleus and might remain associated with the pointed end of the actin filament as the filament elongates. A, acidic region; C, central connector region; L, linker region; W, WH2 domain.

Actin nucleation factors fall into three major classes: (1) the Arp2/3 (actin-related protein 2/3) complex together with the nucleation-promoting factors (NPFs), (2) the formins and (3) the tandem-monomer-binding nucleators. The Arp2/3 complex was the first physiologically relevant actin nucleation factor to be identified. It is composed of the actin-related proteins ARP2 and ARP3 and five additional subunits. In vitro, it promotes actin nucleation after binding to a pre-existing ‘mother’ filament in the presence of NPFs to generate branched actin filaments. The class I NPFs are the major regulators of Arp2/3-mediated nucleation and are defined by the presence of a WCA domain, which consists of a Wiskott–Aldrich syndrome protein (WASP) homology 2 (WH2) domain (W), a central, or connector, region (C) and an acidic region (A). The WCA region is the minimal sequence element that is required for potent activation of Arp2/3-mediated actin nucleation (Fig. 1B) (Machesky et al., 1994; Machesky et al., 1999).

The second major class of actin nucleators are the formin proteins, which were shown to participate in actin nucleation independently of the Arp2/3 complex and NPFs and are able to generate unbranched actin filaments (Fig. 1B) (Pruyne et al., 2002; Sagot et al., 2002). The identification of formins was followed by the discovery of a third class of actin nucleators, denoted tandem-monomer-binding nucleators. This group comprises proteins such as SPIRE and cordon bleu (COBL). Through their tandem G-actin binding motifs (typically WH2 domains), these proteins bring together actin monomers to form an actin nucleation seed (Fig. 1B) (Quinlan et al., 2005; Ahuja et al., 2007). Recently, two additions have been made to this latter group. The first, junction-mediating and regulatory protein (JMY), was shown to be an independent tandem-monomer-binding nucleation factor in addition to acting as a WCA-domain-containing activator of the Arp2/3 complex (and thus also a class I NPF) (Coutts et al., 2009; Zuchero et al., 2009). The second, adenomatous polyposis coli (APC), lacks WH2 domains and its minimum actin-nucleating domain has been shown to form a dimer in vitro (Okada et al., 2010). It therefore recruits actin monomers through an unknown mechanism (Okada et al., 2010).

All three groups of actin nucleators function in the cytoplasm, where they promote actin filament formation (Pruyne et al., 2002; Quinlan et al., 2005; Goley and Welch, 2006; Ahuja et al., 2007; Coutts et al., 2009; Zuchero et al., 2009; Okada et al., 2010). However, decades of research have established the presence of actin and actin-associated proteins, such as myosin I, in the nucleus (Scheer et al., 1984; Gonsior et al., 1999; Bettinger et al., 2004; Philimonenko et al., 2004; de Lanerolle et al., 2005). Furthermore, an increasing number of roles for nuclear actin have been elucidated (Olave et al., 2002; Shumaker et al., 2003; Fomproix and Percipalle, 2004; Hofmann, 2009). Nuclear actin functions in transcription, nuclear export, intranuclear transport and chromatin remodelling (Olave et al., 2002; Shumaker et al., 2003; Fomproix and Percipalle, 2004; Hofmann, 2009). Indeed, actin is a component of all three types of RNA polymerase (classes I, II and III), which emphasises the involvement of actin in gene regulation (Hofmann et al., 2004; Hu et al., 2004; Philimonenko et al., 2004; Yoo et al., 2007).

Despite the knowledge that actin functions in the nucleus, the nuclear roles of proteins involved in cytoskeletal actin filament formation so far remain unclear, and the regulation of cytoskeletal and nuclear actin are considered separate events. Perhaps this is, in part, a result of a lack of studies providing the direct detection of nuclear F-actin in cells. For this reason, the presence of filamentous actin in the nucleus is still a matter of debate. Notably, the presence of dynamic nuclear polymeric actin was demonstrated by McDonald and colleagues using fluorescence recovery after photobleaching (FRAP) in conjunction with actin mutants that cannot form polymers and compounds that alter actin polymerisation (McDonald et al., 2006). This observation and other studies suggest that actin polymers in the nucleus and the cytoplasm are structurally different. For example, immunofluorescence studies using antibodies against actin demonstrate the presence of nuclear actin aggregates that cannot be detected by the F-actin-binding toxin phalloidin (Gonsior et al., 1999; Schoenenberger et al., 2005). Perhaps F-actin structures in the nucleus are altered by actin-binding proteins, or are kept very short, which would hamper their detection by classical methods. Furthermore, cytoplasmic polymeric actin is typically associated with myosin II. Yet there is little evidence for the nuclear presence of myosin II (Obungu et al., 2003), which might also indicate an absence or the formation of a structurally different form of F-actin in the nucleus. However, F-actin, which is capable of phalloidin binding, was recently detected by live-cell imaging in nuclei transplanted into Xenopus oocytes (Miyamoto et al., 2011). Improved imaging and novel experimental systems might therefore clarify the situation in the future.

It is becoming increasingly evident that proteins that directly promote actin nucleation, and thus filament formation are found in the nucleus (Gieni and Hendzel, 2009). Is this nuclear localisation physiologically relevant? If so, do these nucleation factors function in the polymerisation of nuclear actin or do they have distinct nuclear roles? It is possible that they provide a temporal link between cytoskeletal actin dynamics and transcription, thereby relaying exogenous signals to the nucleus. As we learn more about the actin nucleation machinery, a trend is beginning to emerge for the multi-localisation of cytoskeletal factors to both the cytoplasm and the nucleus. It is therefore important to understand what regulates the compartmentalisation of actin nucleation factors, their nuclear function and the cellular impact of their dual roles.

This Commentary will summarise the current knowledge of the repertoire of actin nucleation factors that are known to shuttle from the cytoplasm to the nucleus. Moreover, we will explore the possible reasons why such shuttle systems have developed.

The class I NPFs

There are four typical class I NPFs. They do not possess intrinsic actin nucleation activity, but only nucleate actin in association with the Arp2/3 complex. The WCA region of NPFs interacts with and thereby activates the Arp2/3 complex. Specifically, contact between the ARP2 subunit and the C region and between the ARP3 subunit and the A region promote a substantial conformational change within the Arp2/3 complex, which brings ARP2 and ARP3 into close proximity (Goley et al., 2004; Rodal et al., 2005; Egile et al., 2005; Boczkowska et al., 2008). The WH2 domain then delivers an actin monomer to this site, forming a trimer of ARP2, ARP3 and actin, which acts as the nucleation seed for the new actin filament (Boczkowska et al., 2008). The class I NPF family members are WASP, SCAR (suppressor of cAMP receptor, also known as WAVE), WASH (for WAS protein family homologue) and WHAMM (for WAS protein homologue associated with actin, Golgi membranes and microtubules). Of these, both WASP and WASH have been implicated in nuclear events.

WASP

WASP, and its closest homologue neural WASP (N-WASP), are the best characterised members of the class I NPF proteins and were the first class I NPFs to be identified through their role in cytoskeletal reorganisation (Miki et al., 1996; Symons et al., 1996; Machesky and Insall, 1998). Mammalian WASP is expressed specifically in haematopoietic cells (Derry et al., 1994) and its mutation results in deficiencies in cell migration, phagocytosis and T-cell signalling, which lead to Wiskott–Aldrich syndrome (Bosticardo et al., 2009). N-WASP is ubiquitously expressed and is implicated in a range of processes, including neuritogenesis, endocytosis, the maintenance of cell–cell contacts, transcription and cell motility (Innocenti et al., 2005; Otani et al., 2006; Wu et al., 2006; Yarar et al., 2007; You and Lin-Chao, 2010).

Much is known about the influence of WASP on cellular functions through the generation of cytoplasmic F-actin structures. However, N-WASP contains both a nuclear localisation (NLS) and nuclear export signal (NES) and has been shown to be present in the nucleus (Suetsugu and Takenawa, 2003). Studies have indicated that the nuclear localisation of N-WASP is regulated by its activation and phosphorylation states (Suetsugu and Takenawa, 2003; Wu et al., 2004). Specifically, there is evidence that the active, open conformation of N-WASP localises more readily to the nucleus than inactive N-WASP and that phosphorylation of N-WASP by Src family kinases or by focal adhesion kinase (FAK) at the conserved Tyr256 residue (which is analogous to Tyr291 in WASP) results in its nuclear export (Vetterkind et al., 2002; Suetsugu and Takenawa, 2003; Wu et al., 2004). Notably, one study has demonstrated that phosphorylation at Tyr256 abolishes nuclear localisation of N-WASP and enhances N-WASP-dependent cell motility (Wu et al., 2004). This highlights the impact that subcellular localisation can have on N-WASP function. However, it should be noted that the in vivo role of tyrosine phosphorylation remains controversial. There is evidence that phosphorylation mutants of N-WASP act in a manner that is both consistent and inconsistent with the in vitro observations that demonstrate the role of tyrosine phosphorylation in enhancing Arp2/3 activation (Dovas and Cox, 2010). Despite this, it has been postulated that nuclear N-WASP induces the polymerisation of nuclear actin to regulate RNA-polymerase-II-dependent transcription in association with the Arp2/3 complex (Wu et al., 2006). More recently, nuclear WASP has been shown to associate with histone-modifying proteins and has been implicated in the regulation of histone modifications and chromatin structure (Taylor et al., 2010). Notably, the depletion of N-WASP does have an effect on transcription in vivo (Wu et al., 2006). However, Wu and co-workers used nuclear extracts in an in vitro actin polymerisation assay supplemented with recombinant N-WASP and actin to infer a role for nuclear actin polymerisation in this process (Wu et al., 2006) and so far, no other studies have directly shown NPF-mediated nuclear actin polymerisation. Therefore, more direct biochemical data will be necessary to demonstrate whether nuclear actin polymerisation occurs in vivo, and moreover, whether N-WASP can promote it. Nonetheless, these studies imply that a characterised modulator of actin microfilaments can influence gene expression and it will be interesting to clarify the role of WASP in transcription.

Of note, the NCK adaptor protein 1 (NCK1), which is an activator of N-WASP, translocates from the cytoplasm to the nucleus in association with suppressor of cytokine signalling 7 (SOCS7) and G-actin in response to DNA damage (Fig. 2A) (Kremer et al., 2007). Moreover, nuclear NCK is essential for the activation of p53 in response to UV-induced DNA damage (Kremer et al., 2007). The localisation of NCK to the nucleus under damage conditions subsequently has an impact on the actin cytoskeleton, by reducing the number of stress fibres and causing loss of cell polarity (Kremer et al., 2007). This highlights a role for NCK in coupling the DNA damage response to actin dynamics, and suggests that the localisation of WASP under damage conditions should be examined. Structural changes in the cytoskeleton during DNA damage might provide a general mechanism for relaying signals to the nucleus through actin-associated proteins and we will return to this idea later in this Commentary.

Fig. 2.

Actin regulators link the state of cytoskeletal actin to nuclear events. (A) In response to DNA damage, NCK translocates from the cytoplasm, where it activates N-WASP, to the nucleus, in association with SOCS and G-actin. Nuclear NCK is essential for p53 activation in response to UV-induced DNA damage. N-WASP can also translocate to the nucleus and affect gene expression; however, it is unknown whether this occurs following cellular stress. (B) In non-stressed cells, cytoplasmic JMY can influence actin nucleation through its WH2 domains and thereby enhance cell motility. Following DNA damage, the levels of monomeric actin are reduced, which exposes an NLS in the WCA region of JMY and enables its nuclear import. In the nucleus, JMY functions as a transcriptional co-activator, and its relocalisation reduces cell motility. The ability of JMY to nucleate actin filaments might have a role in the nucleus, influencing transcription. Phosphorylated Strap protein surrounds P300. (C) APC forms a homodimer and nucleates actin filaments in the cytoplasm. It shuttles to the nucleus, possibly in response to increases in cell density, where it sequesters β-catenin. This prevents β-catenin- and LEF1-mediated transcription, which would otherwise drive cell proliferation and can result in transformation.

WASH

WASH has been purified from cells as part of a multiprotein complex (Derivery et al., 2009; Gomez and Billadeau, 2009). It has an essential functional role in endosome sorting by facilitating tubule fission through activating the Arp2/3 complex, thereby determining endosome morphology and trafficking (Derivery et al., 2009; Gomez and Billadeau, 2009; Duleh and Welch, 2010). Moreover, studies from Drosophila melanogaster suggest that WASH works in concert with the formin Cappuccino (CAPU) and the tandem-monomer-binding nucleator SPIRE to control actin and microtubule dynamics (Liu et al., 2009).

Interestingly, vertebrate WASH proteins have conserved predicted nuclear localisation and export signals as well as putative SUMOylation sites, which suggests a possible role for these proteins in the nucleus (Linardopoulou et al., 2007). Additionally, Drosophila WASH has been identified as a component of a nuclear complex, which contains various transcription factors and chromatin modifiers that modulate the transcription of genes involved in DNA replication and proliferation (Hochheimer et al., 2002). Notably, WASH was found to co-precipitate with the protein ISWI (for imitation SWI) (Hochheimer et al., 2002), a highly conserved member of the SWI2–SNF2 family of ATPases, which acts as the catalytic subunit of three chromatin-remodelling complexes in D. melanogaster (Cairns, 1998). Further studies are required to establish the nuclear localisation of WASH and to provide more direct evidence for these potential interactions and their impact on gene expression. However, these observations hint at a role for WASH, as for WASP, in chromatin structure reorganisation, and might indicate that NPFs in general are able to affect epigenetic control.

Formins

Formins bind to barbed actin filament ends to accelerate polymer elongation (Pruyne et al., 2002; Zigmond et al., 2003; Kovar and Pollard, 2004; Romero et al., 2004; Kovar et al., 2006). The cellular role of formins has been attributed to various processes, including cell migration, adherens junction assembly, endosome movement and cytokinesis (Chesarone et al., 2010). There are 15 mammalian formin genes, and formins can be further separated into seven different subclasses: diaphanous (DIA), formin-related proteins in leukocytes (FRLs), dishevelled-associated activators of morphogenesis (DAAMs), formin-homology-2-domain-containing proteins (FHODs), formins, delphilin and inverted formins (INFs) (Higgs and Peterson, 2005). FHOD1 and mDia2 have both been reported to shuttle into the nucleus.

FHOD1

FHOD1 colocalises with F-actin structures at the cell periphery, induces cell elongation and enhances cell migration (Koka et al., 2003). Moreover, constitutively active FHOD1 induces the formation of F-actin stress fibres and coordinates F-actin and microtubule networks (Gasteier et al., 2003; Gasteier et al., 2005). Recently, FHOD1 has been shown to associate and synergise with Rho-associated, coiled-coil-containing protein kinase (ROCK) to promote plasma membrane blebbing (Hannemann et al., 2008). Notably, caspase-3-mediated cleavage of FHOD1 generates a C-terminal product that is almost entirely nuclear (Ménard et al., 2006). Overexpression of this cleavage product inhibits RNA polymerase I transcription, as indicated by it blocking 5-bromouridine 5′-triphosphate (BrUTP) incorporation (Ménard et al., 2006), which implicates a nuclear FHOD1 cleavage product in the regulation of transcription. Further studies will be required to establish the signalling pathway that leads to FHOD1 cleavage and the connection between FHOD1 and transcription, but it is tempting to speculate that FHOD1 might modulate transcription in response to cell stresses that trigger apoptosis through caspase-3 activation.

mDia

The formin mDia is expressed as three isoforms, mDia1, mDia2 and mDia3. It is an effector of Rho GTPases and is involved in actin nucleation and elongation and has wide-ranging roles, including contractile ring formation, endocytosis and the formation of lamellipodia and filopodia (Gasman et al., 2003; Pellegrin and Mellor, 2005; Hotulainen and Lappalainen, 2006; Carramusa et al., 2007; Yang et al., 2007; Sarmiento et al., 2008; Watanabe et al., 2008; Ryu et al., 2009). mDia proteins influence cell division and migration and have been implicated in metastasis (Gupton et al., 2007; Yang et al., 2007; Watanabe et al., 2008; Narumiya et al., 2009). Studies also hint at a potential nuclear role for mDia. Inhibition of the nuclear export factor exportin XPO1 (also known as CRM1) results in nuclear accumulation of mDia2 (Miki et al., 2009). mDia2 contains at least one NES that is recognised by the nuclear export receptor CRM1 and a bipartite basic NLS, which facilitate continuous shuttling between the nucleus and the cytoplasm. Although mDia1 and mDia3 have been reported not to accumulate in the nucleus (Miki et al., 2009), previous work has suggested their presence in the nucleus (Stüven et al., 2003; Yasuda et al., 2004). Specifically, mDia1 has been found to co-precipitate with CRM1, and mDia3 has been shown to interact with heterochromatin protein 1 (also known as CBX5) (Stüven et al., 2003; Yasuda et al., 2004), which again implies a function for actin nucleators in chromatin remodelling.

mDia also has a role in the regulation of nuclear serum response factor (SRF) activity by influencing actin dynamics in the cytoplasm (Fig. 3) (Copeland and Treisman, 2002). SRF is a transcription factor that binds to the serum response element (SRE) in the promoter region of target genes that are involved in the regulation of cellular processes such as cell cycle progression, apoptosis, cell growth and differentiation. SRF activity is modulated by SRF transcription co-factors called myocardin-related transcription factors (MRTFs). In addition to binding SRF, MRTFs also contain G-actin-binding sites (Miralles et al., 2003; Miano et al., 2007). The interaction between MRTFs and G-actin favours the cytoplasmic localisation of MRTFs and thus the suppression of SRF-dependent transcription (Miralles et al., 2003). Rho GTPase signalling to mDia1 has been shown to enhance actin polymerisation, thereby decreasing the G-actin levels and subsequently leading to MRTF relocalisation to the nucleus and SRF activation (Copeland and Treisman, 2002). This allows a feedback mechanism that enables the transcriptional machinery to sense which cytoskeletal components require biosynthesis, for example, to maintain cell motility. The MRTF–SRF circuit has been extensively reviewed recently (see Olson and Nordheim, 2010) and is likely to become a paradigm for studying how the dynamics of the actin cytoskeleton are integrated with nuclear events and gene expression. Perhaps actin nucleators that translocate to the nucleus mediate similar circuits through their interactions with G-actin, enabling information from the cytoplasm to be relayed to the nucleus. Interestingly, research is now emerging that suggests the nuclear accumulation of actin nucleators in response to cellular stress. It is possible that stress signalling pathways are acting through these proteins to alter cytoskeletal events and reprogramme transcription simultaneously, facilitating a fast cellular response. The possible involvement of cytoskeletal actin nucleation regulators in the DNA damage response will be explored in more detail below.

Fig. 3.

The MRTF–SRF circuit. MRTFs are sequestered in the cytoplasm by monomeric G-actin. Rho GTPase signalling enhances actin polymerisation and thereby decreases G-actin levels, which releases MRTF. Following its release from actin, MRTF translocates to the nucleus and interacts with SRF. SRF binds DNA (at the CArG box) to activate the transcription of genes encoding actin and other cytoskeletal components.

Tandem-monomer-binding nucleators

JMY

Recent studies have revealed the unique ability of JMY to act both as a class I NPF that activates the Arp2/3 complex and as an Arp2/3-independent tandem-monomer-binding nucleation factor (Zuchero et al., 2009). Thus, JMY is the first activator of the Arp2/3 complex that can also nucleate actin filaments independently of Arp2/3. Under conditions without cellular stress, this actin-regulatory function of JMY has been implicated in the formation of lamellipodia and/or filopodia at the leading edge of the cell, affecting cell migration and invasion (Coutts et al., 2009; Zuchero et al., 2009). In addition, the depletion of JMY leads to increased E-cadherin expression, which antagonises migration, presumably by promoting cell–cell adhesion (Coutts et al., 2009). Following DNA damage, for example, as a result of exposure to ultraviolet (UV) radiation, JMY translocates to the nucleus where it augments p53-dependent gene expression and is no longer able to influence cell migration (Fig. 2B) (Coutts et al., 2007; Coutts et al., 2009). A very recent study has elucidated a mechanism for the observed nuclear import of JMY, which is governed by the cytoskeletal environment (Zuchero et al., 2012). It was shown that the WH2-domain-containing region in JMY overlaps with an atypical, bipartite NLS, and that the binding of actin to this region blocks the binding of importins and thus the nuclear import of JMY. Reduced monomeric actin levels are thought to occur as a result of DNA-damage-induced actin polymerisation and this coincides with the relocation of JMY to the nucleus following enhanced importin binding (Zuchero et al., 2012). Nuclear JMY results in the appearance of actin-containing structures, suggesting that the region containing the JMY WH2 domain is active in the nucleus (Coutts et al., 2009). Moreover, the actin polymerisation inhibitor latrunculin A substantially reduces the observed increase in p53 activity that occurs as a result of the co-activator role of nuclear JMY (Coutts et al., 2009). Of note, latrunculin A does not deplete JMY from the nuclei of DNA-damaged cells (Zuchero et al., 2012). This implies that JMY might influence transcription through actin polymerisation. Crucially, the nature of the actin-containing structures needs to be clarified and direct biochemical evidence for JMY-induced nuclear actin polymerisation should be sought. Oligomerisation of actin by JMY might directly enhance RNA-polymerase-II-dependent transcription or activate chromatin remodelling complexes and future studies should address these intriguing possibilities.

APC

The most recently identified tandem-monomer-binding nucleation factor is APC, which is a tumour suppressor protein that has been linked to colorectal cancer. APC forms a homodimer to recruit four actin monomers, thereby leading to the nucleation and generation of unbranched actin filaments (Fig. 2C) (Okada et al., 2010). APC is thought to synergise with mDia1 to promote actin filament assembly in vivo, which might promote cell protrusions and cell–cell contacts, and thus influence migration (Oshima et al., 1997; Harris and Nelson, 2010; Okada et al., 2010). Interestingly, APC also has a defined nuclear role (Neufeld, 2009). APC shuttles between the nucleus and the cytoplasm, and this process is dependent on CRM1 (Neufeld et al., 2000a; Zhang et al., 2000; Galea et al., 2001; Brocardo et al., 2005). Nuclear APC has been shown to affect the Wnt signalling pathway by promoting both the sequestration and nuclear export of β-catenin (Henderson, 2000; Neufeld et al., 2000b; Rosin-Arbesfeld et al., 2000). This prevents stable β-catenin accumulation in the nucleus, where it would otherwise form a complex with the lymphoid enhancer-binding factor 1 (LEF1, also known as TCF) family of transcription factors and activate transcription (Fig. 2C). TCF-mediated transcription ultimately leads to cell proliferation and transformation (Hurlstone and Clevers, 2002; Nelson and Nusse, 2004), thus nuclear APC counteracts these oncogenic processes. Phosphorylation of APC, which possibly changes according to the proliferative state of the cell, influences its subcellular localisation (Zhang et al., 2000; Zhang et al., 2001; Rosin-Arbesfeld et al., 2003; Caro-Gonzalez et al., 2012). It remains to be seen whether DNA damage influences the subcellular localisation of APC, but its bipartite function in nuclear events and organisation of the actin cytoskeleton seems to be regulated by extracellular signals. Thus there is growing evidence for an emerging facet of cytoskeleton-nuclear communication that bridges actin microfilament dynamics and the cellular stress response.

Nuclear ARP proteins

Although it has been suggested that the Arp2/3 complex works in concert with N-WASP in the nucleus to promote RNA polymerase II-dependent transcription (Yoo et al., 2007), there is limited evidence for a nuclear role for the Arp2/3 complex. However, the nuclear localisation of other ARP proteins has been described, initially in budding yeast and, more recently, in vertebrates (reviewed by Oma and Harata, 2011). Specifically ARP4, ARP5, ARP6 and ARP8 localise to the nucleus in vertebrates (Harata et al., 1999; Kato et al., 2001; Ohfuchi et al., 2006; Aoyama et al., 2008; Kitayama et al., 2009) and ARP4 and ARP5 have been shown to shuttle between the nucleus and the cytoplasm in human cell lines (Lee et al., 2003; Kitayama et al., 2009).

The nuclear ARP proteins, along with actin, are components of large chromatin-modifying complexes, including nucleosome remodelling and certain histone acetyltransferase (HAT) complexes, and are involved in multiple nuclear functions including transcription, DNA repair and replication, and chromosome segregation (reviewed in Oma and Harata, 2011). For example, ARP4, ARP5 and ARP8, together with actin, form subunits of the mammalian INO80 chromatin-remodelling complex that functions in transcriptional regulation, DNA-damage-induced double strand break repair and the nucleotide excision repair (NER) pathway (Jin et al., 2005; Jiang et al., 2010). ARP8 is required for the recruitment of mammalian INO80 to sites of DNA damage (Kashiwaba et al., 2010), and ARP5 has been shown to have a specific role in the INO80-mediated NER pathway in response to UV lesions (Jiang et al., 2010). Moreover, ARP5 enhances the accumulation of phosphorylated histone H2AX, which might be required for the recruitment of repair factors (Kitayama et al., 2009). It will be interesting to determine whether DNA-damaging agents affect the localisation and nuclear accumulation of these ARP proteins.

Importantly, it is currently unknown whether any of the nuclear ARP proteins can promote actin filament formation. A recent study has shown that ARP4 and ARP8 act synergistically to inhibit actin polymerisation, and that ARP4 could have a role in maintaining the pool of monomeric actin in the nucleus (Fenn et al., 2011). Perhaps nuclear ARPs interact with nucleators, such as JMY, in response to DNA damage to affect a transcriptional stress response. Indeed, nuclear ARP proteins have also been shown to have a role in regulating the activity of transcription factors. For example, ARP4 can influence p53-dependent transcription (Lee et al., 2005; Wang et al., 2007). It will be important to investigate the involvement of nuclear ARPs in DNA damage circuits, to examine whether they can interact with known actin nucleators and to reveal whether they can promote actin polymerisation

Conclusions and future directions

It is well known that cytoskeletal events that are driven by actin filament formation require altered gene expression, and there is a growing repertoire of relay systems that link cytoskeletal actin to nuclear events. This is illustrated by the actin–MRTF–SRF feedback loop, in which signals are relayed to the transcriptional machinery when cytoskeletal components require biosynthesis. In addition, LIM-domain-containing proteins that are involved in focal adhesion formation, in association with actin fibres, can act as nuclear transcription co-factors (Kadrmas and Beckerle, 2004). Moreover, the loss of cell–cell contacts and attendant actin reorganisation enhances RhoA dependent activation of NF-κB transcription (Cowell et al., 2009).

It is also apparent that facilitators of microfilament nucleation physically shuttle into the nucleus to influence nuclear events. This process might have evolved to allow the direct integration of seemingly disparate events (Fig. 4). For example, the translocation of actin nucleators such as JMY, WASP and APC to the nucleus will concomitantly cease the promotion of actin dynamics that facilitate migration at the same time as modulating gene expression. This might be particularly important in transmitting exogenous signals within the cell. Indeed, this Commentary has highlighted the potential of some actin regulators to directly communicate UV-induced DNA damage to the cell periphery. Moreover, the roles of nuclear ARP proteins are being elucidated, and it is increasingly evident that they are involved in the repair of UV-induced DNA lesions and damage-induced double-strand breaks. Perhaps they work alongside actin nucleators to mediate a response to exogenous damaging agents.

Fig. 4.

Integration of events mediated by cytoplasmic–nuclear shuttling of actin modulators. (A) A general scheme demonstrating how regulators of actin nucleation might directly relay external signals to the nucleus to modulate gene expression. (B) Nuclear import of JMY, as a specific example of how regulators of actin nucleation relay external signals to the nucleus to modulate gene expression. Following DNA damage, actin polymerises, reducing the concentration of actin monomers. This, in turn, enables importins to bind to the JMY NLS and import JMY into the nucleus. Once in the nucleus, JMY ceases to influence cytoskeletal actin polymerisation and cannot drive cell motility. JMY also enhances p53-dependent transcription, which might depend on the polymerisation state of actin.

We have drawn attention to growing evidence that actin nucleation factors localised in the nucleus can influence gene expression. This might occur through the reorganisation of heterochromatin to influence RNA-polymerase-II-dependent transcription or through direct interactions with RNA polymerases. However, the involvement of actin in this process needs to be clarified. Notably, the presence of typical actin microfilaments in the nucleus has not been directly demonstrated and therefore it remains unknown whether polymerisation of actin can occur as a result of actin nucleators localised in the nucleus. Further research is needed to identify the structure of nuclear actin and to fully understand how actin, its nucleation factors and the transcriptional machinery interact in the nucleus. The presence of nuclear Arp2/3 requires verification and it is possible that nuclear class I NPFs are not involved in Arp2/3 activation. Perhaps they function independently or work in concert with nuclear ARP proteins. Overall, understanding the roles of cytoskeletal proteins that influence transcription will be important for our comprehension of cellular biology and, given the range of cellular processes that they influence, will in all probability impinge on the pathology of a variety of diseases.