Nuclear actin dynamics and functions at a glance

Nuclear architecture refers to the non-random assembly of genomes, nuclear bodies and the nucleoskeleton, and is associated with the regulation of gene expression (Cremer and Cremer, 2001; Schneider and Grosschedl, 2007). The nucleoskeleton is composed of various proteins, mainly lamins and lamin-associated proteins, which form a mesh-like network beneath the inner nuclear membrane at the outer boundary of the chromatin mass, providing structural support and regulating nuclear shape (Adam, 2017; Zhong et al., 2010). Additionally, other structural elements, such as nuclear pore complexes (NPCs), spectrins and chromatin-associated proteins, contribute to the overall organization and function of the nucleoskeleton (Adam, 2017; Miyamoto and Harata, 2021). However, the contribution of other nucleoskeletal proteins, like actin, to genome organization and maintenance is less well known.

Both actin and actin-binding proteins constantly shuttle between the cytoplasm and the nucleus via NPCs that cover the nuclear envelope (NE) (Grosse and Vartiainen, 2013). With a size of 43 kDa and diameter of ∼5 nm, the globular actin (G-actin) monomer is right at the limit of free diffusion through the NPCs (Keminer and Peters, 1999; Wente and Rout, 2010). However, actin can also be imported into the nucleus in complex with cofilin proteins, facilitated by importin 9 (IPO9) (Dopie et al., 2012). Notably, a study performed in Drosophila has revealed that even in Ipo9-null flies, nuclear actin levels remain high, indicating that actin employs additional mechanism(s) and import factors to enter the nucleus (Borkúti et al., 2022). Meanwhile, export of actin is achieved in complex with profilin proteins and is mediated by exportin 6 (XPO6), which tightly controls nuclear actin levels (Baarlink et al., 2017; Bohnsack et al., 2006; Stüven et al., 2003). Nevertheless, nuclear actin levels might vary between cell types, with one ‘pool’ of nuclear actin appearing to be more readily exported, and another pool of G-actin bound to nuclear complexes, which is more stably associated and thus less export competent (Skarp et al., 2013).

Many cellular functions of actin are dependent on its ability to dynamically assemble and disassemble, which is controlled by a plethora of actin regulators such as capping proteins, actin depolymerizing factors and nucleators. Formins comprise the largest family of actin nucleators, which promote the assembly of linear actin filaments exerting an array of cellular functions (Courtemanche, 2018; Faix and Grosse, 2006). Unlike formins, the actin-related proteins-2/3 (Arp2/3) complex polymerizes branched actin filaments downstream of nucleation-promoting factors (NPFs), such as class I NPFs WASP (also known as WAS), N-WASP (WASL) and SCAR/WAVE proteins (Goley and Welch, 2006).

The presence and position of the centromere on each chromosome is stably maintained to ensure accurate chromosome segregation after each round of cell division (Liu and Mao, 2017). In this context, the diaphanous formin mDia2 (also known as DIAPH3) has been described to dynamically polymerize short nuclear actin filaments to constrain or fence centromere movement during loading of the defining histone H3 variant centromere protein A (CENP-A) (Liu and Mao, 2016; Liu et al., 2018). Activation of mDia2 occurs downstream of small GTPase signaling, which in turn is dependent on the GTPase-activating protein MgcRacGAP (also known as RACGAP1). Consistent with this, members of the small Rho GTPase family have been shown to localize to the nucleus (Guilluy et al., 2011; Sandrock et al., 2010). Following mitosis during the early G1 phase, another nuclear F-actin network aids in nuclear volume expansion and chromatin decompaction (Baarlink et al., 2017). This network consists of both short and long actin filaments and requires actin-bundling activity by the non-muscle α-actinin 4 (ACTN4), which is crucial for ensuring proper post-mitotic nuclear organization (Krippner et al., 2020). These actin filaments are relatively long lasting (over about an hour); however, F-actin can also rapidly assemble and disassemble within minutes in the nuclear compartment downstream of extracellular signal induction. For instance, it has been demonstrated that physiological ligands of G-protein-coupled receptors (GPCRs), such as thrombin, can induce Ca²⁺ transients that trigger the assembly of a dynamic nuclear F-actin network through activation of the formin inverted formin 2 (INF2) (see poster). This leads to a reduction in heterochromatic structures and actin-dependent changes in chromatin dynamics (Wang et al., 2019).

The idea of actin-dependent global chromatin reorganization has also been suggested in other studies using β-actin-knockout mouse embryonic fibroblasts (MEFs) (Mahmood et al., 2021, 2023; Xie et al., 2018a). Although none of these studies have emphasized a role for F-actin, they also do not exclude the potential impact of dynamic actin assembly and/or disassembly in this context. Notably, chromatin dynamics and 3D organization are also influenced by DNA damage; following double-strand breaks (DSBs), chromatin shifts from transcriptionally inactive B compartments to active (open) A compartments, facilitating DSB clustering and the generation of repair hubs (Zagelbaum et al., 2023). It has been recently shown that these DSB-induced shifts are attenuated upon pharmacological inhibition of the Arp2/3 complex, the nucleator of branched actin filaments. This indicates an involvement of actin-based movement in this process of widespread chromosome translocations (Zagelbaum et al., 2023). These findings concur with a multitude of functions that have been described for nuclear actin in the DNA damage and replication stress response, which are discussed below.

DNA damage is defined as a chemical or structural perturbation that can block genome replication and transcription. When unrepaired, these lesions can lead to mutations or more widespread genomic aberrations that threaten the viability of a cell (Jackson and Bartek, 2009). DNA DSBs are one of the most toxic DNA lesions and occur as a result of exogenous or endogenous triggers, including ultraviolet (UV) light, chemical agents, metabolic reactions and/or replication stress (Mehta and Haber, 2014). There are two major pathways for DSB repair, non-homologous end-joining (NHEJ) and homologous recombination (HR), which both include a complex sequence of events that allow for the sensing of the DSB, recruitment of DNA repair proteins and ultimately the repair of the break itself (Andrin et al., 2012; Lieber, 2008; San Filippo et al., 2008). Interestingly, recent evidence has linked nuclear F-actin of various shapes to different aspects of the DNA damage response (DDR), as recently reviewed by Wollscheid and Ulrich (2023).

Over a decade ago, Andrin and colleagues (2012) described a particular role for nuclear polymerized actin in DNA DSB repair. In vitro studies have revealed direct binding of F-actin to DNA repair proteins including the Ku70–Ku80 heterodimer (Ku70 and Ku80 are also known as XRCC6 and XRCC5, respectively), which acts as a mediator in the NHEJ repair pathway (Andrin et al., 2012). Moreover, interfering with actin polymerization via pharmacological intervention or overexpression of a nuclear-targeted non-polymerizable mutant (actin^R62D) alters the retention of Ku80 at DNA damage sites and ultimately inhibits DNA DSB repair in vivo (Andrin et al., 2012). Similarly, knockdown of the actin nucleators formin-2 (FMN2) or simultaneous knockdown of Spire-1 and Spire-2 prevented DNA damage-induced nuclear actin assembly and its dependent clearance of DSBs (Belin et al., 2015). Once DSBs are detected for homology-directed repair, they are translocated outside of the heterochromatic DNA domain to prevent aberrant recombination (Chiolo et al., 2011; Ryu et al., 2015). New light was shed on the underlying mechanism when two groups independently published their findings in 2018 (Caridi et al., 2018; Schrank et al., 2018). In Drosophila cells, nuclear actin polymerizes at ionizing radiation (IR)-induced heterochromatic DSB repair sites in an Arp2/3-dependent manner. These filaments grow towards the nuclear periphery, allowing for myosin-based directed motions of the lesions for efficient clearance in repair foci outside of the heterochromatic region (Caridi et al., 2018; Wollscheid and Ulrich, 2023). Whether DNA DSBs can cluster in mammalian cells has been a subject of debate. However, experimental evidence from the Legube group shows that DSBs cluster in a FMN2-dependent manner, but only when induced within transcriptionally active genes (Aymard et al., 2017), and that they establish 3D contacts to form a new dedicated chromatin compartment, supporting their efficient repair (Arnould et al., 2023). Inhibiting the Arp2/3 complex via the small-molecule CK666 or preventing nuclear actin polymerization by overexpression of a nuclear-localized actin mutant (actin^R62D-NLS) also affects DSB movement into discrete subnuclear clusters, reducing DSB repair efficiency in human cells (Schrank et al., 2018). These actin-driven movements are particularly important for homology-directed repair, as Arp2/3 enrichment is not detected at DSBs repaired by NHEJ. Besides its role in intranuclear movement of DNA DSBs, F-actin has recently been described to also function as a scaffold for the recruitment and proper localization of promyelocytic leukemia (PML) nuclear bodies (NBs) to DNA lesion sites for an efficient DDR (Cobb et al., 2022). Here, the formin mDia2 has been identified as the responsible actin nucleator. However, the exact mechanistic details behind mDia2 activation following DNA damage require further investigation.

Replication stress (RS) is defined as any hindrance of the DNA replication process and usually leads to slowing or stalling of replication fork progression (Zeman and Cimprich, 2014). Mild replication interference typically leads to single-stranded DNA (ssDNA) accumulation followed by fork reversal – that is, the conversion of replication forks into four-way junctions – which assists cells in restarting DNA synthesis once the source of replication stress is resolved. Prolonged fork stalling requires protection of stalled and/or reversed forks from nucleolytic degradation and can lead to fork breakage, both of which require homologous recombination factors (Berti et al., 2020). Nuclear actin was first implicated in regulating unperturbed DNA replication by the Fisher group (Parisis et al., 2017). They showed that overexpression of a diaphanous autoregulatory domain (DAD) construct, which activates nuclear mDia2 by releasing its autoinhibition (Baarlink et al., 2013), and polymerization-favoring actin mutants (Posern et al., 2002) both inhibit S-phase progression. This indicates a role for tightly controlled actin turnover, rather than actin filament stabilization (Parisis et al., 2017), akin to the role of transient nuclear actin assembly facilitating nuclear growth during early G1 phase (Baarlink et al., 2017). Multiple recent reports have since revealed a crucial role for the actin nucleoskeleton in the RS response upon mild or prolonged replication interference (see poster) (Han et al., 2022; Lamm et al., 2020, 2021; Nieminuszczy et al., 2023; Palumbieri et al., 2023; Shi et al., 2023). For example, transient Arp2/3-mediated nuclear F-actin structures have been recently implicated in the immediate response to mild replication interference, where they limit the engagement of PrimPol – a DNA polymerase and primase that mediates the restart of DNA synthesis at DNA lesions – and promote fork reversal, thereby protecting the integrity of replicating DNA (Palumbieri et al., 2023). Moreover, when replication forks reach a complete stall, for example, because of nucleotide depletion or potent DNA polymerase inhibition, NPFs such as WASP directly associate with replication forks, helping to deliver ssDNA-binding proteins, such as RPA and RAD51, thereby limiting nascent strand degradation and fork collapse (Han et al., 2022; Nieminuszczy et al., 2023). Strikingly, not only actin itself, but also the actin-based molecular motor protein myosin VI, has been shown to contribute to the stabilization of stalled or reversed replication forks (Shi et al., 2023). When fork stalling persists for a long time and is likely associated with fork breakage, an increase in F-actin-positive S-phase nuclei is observed, along with a distinct morphology of the observed actin structures, which are nucleated by the Arp2/3 complex downstream of WASP activation (Lamm et al., 2020). These thick, long and persistent filaments were proposed to provide mechanical forces to counteract RS-induced nuclear deformations and to promote the mobility of stalled or broken replication forks towards the nuclear periphery.

RS is a major cause of genomic instability and mutations, and is therefore closely linked to cancer development and metastasis (da Costa et al., 2023). Given that nuclear actin dynamics has been demonstrated to play pivotal roles in transcriptional regulation and the DNA damage response, it is very likely that its dysregulation participates in tumor onset and progression. Interestingly, a recent study has established a link between epithelial-to-mesenchymal transition (EMT)-associated resistance to chemotherapy and nuclear actin (Debaugnies et al., 2023). Using mouse models, the group showed that the small GTPase RHOJ, which interacted with nuclear actin regulators, enhanced the DNA damage response, enabling tumor cells to rapidly repair DNA lesions induced by chemotherapy. When interfering with actin polymerization using the F-actin inhibitor latrunculin B, tumor cells were more sensitive to chemotherapy-induced cell death. However, further studies are needed to unravel the possible mechanism. Notably, nuclear actin polymerization has also been observed in vivo in xenograft tumors after chemotherapeutic intervention using RS-inducing drugs (Lamm et al., 2020). A recent study has demonstrated that the F-actin-bundling protein fascin is actively transported into the nucleus where it regulates the DDR (Lawson et al., 2022). However, enhanced nuclear accumulation and sustained nuclear F-actin bundling reduces invasion and induces apoptosis of cancer cells. These findings highlight the importance of tightly controlled nuclear actin dynamics, not only in physiological conditions but also in the context of oncogenesis, and provide insights into potential routes for selective cancer treatments.

A role for cytoskeletal proteins in the direct control of gene transcription is undisputed in the scientific community, at least since the discovery of actin-dependent regulation of myocardin-related transcription factor-A (MRTF-A) (Miralles and Visa, 2006; Miralles et al., 2003; Vartiainen et al., 2007). In unstimulated cells, MRTF-A, a transcriptional co-activator of serum response factor (SRF), is bound to actin monomers that mask its intrinsic nuclear localization signal (NLS), thus retaining the protein in the cytoplasm (Vartiainen et al., 2007). Once cells are stimulated to polymerize actin into filaments, G-actin is released from MRTF-A, allowing MRTF-A to translocate to the nucleus, where it activates MRTF-A–SRF target genes (Pawlowski et al., 2010; Ulferts et al., 2020). Interestingly, not only cytosolic actin assembly but also mDia1- or mDia2-mediated nuclear actin assembly, is sufficient for MRTF-A–SRF gene regulation (Baarlink et al., 2013; Plessner et al., 2015). Actin itself has been further proposed to associate with the phosphorylated C-terminal domain of RNA Pol II In complex with heterogeneous nuclear ribonucleoprotein (hnRNP) U, where it is involved in transcription initiation as well as elongation processes (Hofmann et al., 2004; Kukalev et al., 2005; Obrdlik et al., 2008). It has also been reported that RNA polymerase I transcriptional activity not only depends on G-actin binding, but also on actin polymerization (Ye et al., 2008). Further investigation is necessary to provide more mechanistic insight into the role of actin in RNA polymerase functions.

Since these discoveries, other examples of actin dynamics regulating gene expression have also begun to emerge (see poster). For example, dynamic nuclear actin assembly is also involved in the regulation of cytokine gene expression upon T-cell activation (Tsopoulidis et al., 2019); engagement of the T-cell receptor (TCR) causes a rapid and transient burst of actin polymerization in the nucleus of CD4⁺ cells. This generation of a dynamic filament network is induced by fast elevation of nuclear Ca²⁺ concentrations, which is similar to what has been described downstream of GPCR-mediated nuclear F-actin assembly (Wang et al., 2019). Directly interfering with actin polymerization with latrunculin B treatment or by inhibition of the actin nucleator Arp2/3 both lead to impaired production of effector cytokines that drive T-cell proliferation and antibody production. How exactly these Ca²⁺ transients regulate nuclear actin polymerization remains elusive, but might involve calmodulin binding to N-WASP to activate nuclear Arp2/3 (Tsopoulidis et al., 2019). In a follow-up study, the authors propose that isoforms of the Arp2/3 subunits, ARPC5 and ARPC5L, coordinate distinct signaling events in CD4⁺ T-cells, with ARPC5L mediating cytokine expression upon TCR engagement and ARPC5-dependent F-actin formation in response to aphidicolin (Sadhu et al., 2023).

In recent years, new insights have emerged that help our understanding of how nuclear actin dynamics mechanistically impact gene transcription (Hu et al., 2008; Knerr et al., 2023; Wei et al., 2020). A key role for nuclear actin in regulating gene expression during neurogenesis and osteogenesis has been reported (Gjorgjieva et al., 2020; Xie et al., 2018b); however, a specific role for the dynamic assembly of nuclear actin had not been shown so far. It is now well known that active RNA Pol II organizes into membraneless subcompartments, termed transcriptional condensates (or droplets) and several mechanisms describing their assembly and maintenance have been proposed, including phase separation (Cho et al., 2018; Rippe, 2021; Rippe and Papantonis, 2022). Phase separation refers to the dynamic and reversible process where molecules demix into distinct liquid-like compartments or condensates through multivalent interactions involving low-affinity binding modules and intrinsically disordered regions (Banani et al., 2017; Mehta and Zhang, 2022). Using super-resolution imaging techniques, one study found that RNA Pol II clustering could be induced in human cells upon serum stimulation or interferon-γ treatment or Ca²⁺ signaling in a nuclear actin-dependent manner (Wei et al., 2020; Ulferts and Grosse, 2024 preprint). Here, de novo F-actin polymerization was induced by the N-WASP–Arp2/3 axis. Interestingly, not only actin, but also the nuclear motor protein myosin VI and its interaction with filaments or short actin polymers, is required for RNA Pol II organization (Hari-Gupta et al., 2022), adding another layer of complexity to actin-driven nuclear organization (see poster).

Nuclear actin also specifically regulates ligand-induced androgen receptor (AR) activity, a transcription factor belonging to the family of steroid hormone receptors (Hu et al., 2008; Knerr et al., 2023). This connection became apparent when functional mutations were identified in the formin and actin nucleator DAAM2 in patients with androgen insensitivity syndrome (Knerr et al., 2023). The authors were able to visualize dynamic actin assembly at AR clusters (see poster) and proposed a model in which nuclear actin and the AR directly interact, suggesting DAAM2 thereby drives clustering of the AR for the formation of transcriptionally active condensate-like droplets via the assembly of highly dynamic actin polymers.

Together, these examples highlight that dynamic nuclear actin polymerization regulates gene transcription in response to different signaling cues by modulating the formation of distinct transcriptional factories. It will be an exciting task for future research to elucidate whether and how F-actin-dependent formation of transcriptional condensates represents a more general mechanism for gene regulation.

Functions of nuclear actin assembly have been investigated using female germ cells in which abundant nuclear actin is observed. Notably, a prominent nuclear F-actin structure can be observed in Xenopus laevis oocytes at meiotic prophase I (Bohnsack et al., 2006). The formation of nuclear F-actin in the oocyte nucleus is attributed to the lack of XPO6 expression, which specifically exports nuclear actin to the cytoplasm. Xenopus oocytes have thus been used to investigate nuclear F-actin functions, revealing its roles in transcriptional reprograming (Miyamoto et al., 2011), maintenance of nuclear structure (Feric and Brangwynne, 2013) and cooperation with F-actin-interacting proteins (Miyamoto et al., 2013). A recent study has also found nuclear F-actin in mouse oocytes (Scheffler et al., 2022). Unlike Xenopus oocytes, mouse oocyte nuclei do not show stable F-actin formation, but dynamic F-actin assembly can be observed for subsequent meiotic progression (see poster). Thus, the formation of nuclear F-actin in the oocyte nucleus can be seen across species and its functions in meiosis needs to be further revealed.

In mice, upon the completion of meiotic maturation, oocytes are fertilized by sperm, forming zygotes that contain nuclei, referred to as pronuclei, which display nuclear F-actin structures (see poster) (Okuno et al., 2020). This zygotic nuclear F-actin is required for embryonic development because the forced depolymerization of nuclear F-actin results in impaired development to term. Furthermore, zygotic nuclear F-actin is important for DNA damage repair at this stage (Okuno et al., 2020). Interestingly, nuclear F-actin was also found in early embryos of other species such as Xenopus and zebrafish (Oda et al., 2017, 2023). In all species where embryonic nuclear F-actin is observed, its abundance disappears by the time major embryonic transcription starts. Further studies are needed to understand how accumulated nuclear F-actin in early embryos is related to the regulation of gene expression. Nevertheless, the accumulation of nuclear F-actin and its disassembly by the time of major embryonic gene activation seems to be a common feature of very early embryos immediately after the initiation of development. An intriguing question is whether the proper formation of nuclear F-actin in zygotes is related to the developmental ability of embryos. The abnormal formation of zygotic nuclear F-actin has been reported in two different cases. Shindo et al. (2021) have shown that interconnected nuclear F-actin is not observed in many cloned mouse embryos that fail to develop normally. Meanwhile, Shi et al. (2022) assessed the safety of decabromodiphenyl ethane (DBDPE), a novel flame retardant, on female reproduction and found that oocytes derived from mice exposed to DBDPE exhibited abnormal formation of zygotic nuclear F-actin after fertilization. Importantly, such embryos showed reduced development to pups, and even resulted in cognitive impairment when grown. In conclusion, nuclear F-actin formation at the zygote stage is necessary for normal development, and it is important to know how nuclear F-actin dynamics regulates the health of the embryos.

The study of nuclear actin dynamics has rapidly grown and has become more widely integrated into how we perceive nuclear organization and architecture. Although many open questions and unsolved problems remain, nuclear F-actin has emerged as an integral part of the nucleoskeleton. In particular, its potential key roles in DNA repair and replication stress mechanisms open new perspectives, as well as challenges, to tackle the precise molecular functions actin might have in these highly complex processes, while simultaneously opening new avenues for potential future therapeutic concepts; for example, how nuclear actin dynamics might assist in resistance to cytostatic drugs and whether this can be exploited for anticancer therapy. Consistent with this, we can expect further progress in our understanding of how actin dynamics shape spatial control of transcription in response to specific physiological signals. Another important question is whether the actin nucleoskeleton participates or exerts canonical actin filament functions, such as regulating nuclear shape, stiffness or plasticity. More sophisticated imaging tools and techniques are required to tackle these questions and to aid in visualizing the actin nucleoskeleton in multicellular contexts such as organoids, tissues and in vivo.

We thank our laboratory members for their helpful discussions. We are most grateful to the three anonymous reviewers for their in-depth and insightful remarks and suggestions. We apologize to all colleagues that could not be cited due to space restrictions. We could not include the diverse array of functions of nuclear actin dynamics involving viruses and other pathogens due to space restrictions.

A high-resolution version of the poster and individual poster panels are available for downloading at https://journals.biologists.com/jcs/article-lookup/doi/10.1242/jcs.261630#supplementary-data.