Please inhibit responsibly: Natural and synthetic actin toxins as useful tools in cell biology

Abstract

The actin cytoskeleton drives many critical cell functions, including motility, division, and vesicular trafficking. To fulfill these functions, actin networks are dynamic and tightly regulated by dozens of proteins that cause actin to assemble and disassemble at the proper time and place. Given the importance of actin to a cell’s biology, it is not surprising that some organisms produce toxins that target actin dynamics to incapacitate prey, win turf wars, or as a defense against predation. For decades, cell biologists have leveraged these toxins and synthesized new ones to cause defects in the structure and function of the actin cytoskeleton. Here, we provide an overview of commonly used actin inhibitors and their origins, as well as best practices for their use in biological studies.

INTRODUCTION

The actin cytoskeleton is a network of actin filaments and actin-binding proteins that support and remodel membranes to control many cellular processes. Precisely regulating the assembly of actin monomers into polymers and higher-order structures, as well as controlling their disassembly, allows actin networks to be responsive and dynamic (Velle et al., 2024). Small molecules that bind either actin monomers or polymers to prevent these dynamics, as well as those that bind to actin-binding proteins and indirectly alter actin dynamics or function, can therefore cause actin-dependent phenotypes. In this Perspective, we provide a brief overview of such “actin inhibitors” that have potential applications in organisms across the eukaryotic tree. We discuss their mechanisms of action, origins, and some cautions in using them in cell biological experiments, especially in nontraditional model systems.

INHIBITORS OF ACTIN MONOMERS AND POLYMERS

Some actin inhibitors bind directly to actin monomers (Figure 1). For example, latrunculins (LatA and LatB) bind to the ATP/ADP-binding cleft of actin monomers at 1:1 stoichiometry (Morton et al., 2000), which is thought to sequester actin from polymerization. Because many actin networks undergo fast turnover, preventing newly disassembled actin monomers from repolymerizing likely results in the observed rapid loss of actin structures within the cell (Spector et al., 1983; Ayscough et al., 1997). Their rapid effects, potency, membrane permeability, and reversibility have made latrunculins some of the most widely used “go-to” actin inhibitors in cell biological research. Like latrunculins, swinholide A is an inhibitor that results in the net loss of actin polymer by binding to actin monomers, although at 1:2 stoichiometry (Bubb et al., 1995), suggesting that this drug binds and sequesters actin dimers. In addition to inhibiting new actin polymerization, both swinholide A and LatA can disassemble existing polymers by severing (Bubb et al., 1995; Fujiwara et al., 2018), and LatA can also promote disassembly at filament ends (Fujiwara et al., 2018).

FIGURE 1:

Actin inhibitors target many aspects of actin dynamics and/or organization. Commonly used actin drugs are shown as a cartoon with inhibitors that bind actin shown at their approximate binding sites, and inhibitors of actin-binding proteins shown using inhibitory arrows.

Many other classes of inhibitors bind to different surfaces of actin filaments (Figure 1). Cytochalasins (such as cytochalasins B and D), for example, cap the rapidly growing “plus” (also called “barbed”) end of an actin filament, thereby blocking the incorporation of additional monomers and causing a net loss of actin filaments (MacLean-Fletcher and Pollard, 1980; Cooper, 1987). Due to their efficacy and early discovery in 1960s, cytochalasins B and D were used in foundational studies elucidating the role of actin filaments in various cellular processes, such as cytokinesis, morphogenesis, and muscle contraction in animals, and chloroplast movement in green algae (Carter, 1967; Schroeder, 1970; Wessells et al., 1971; Wagner et al., 1972). Other inhibitors, like jasplakinolide and phalloidin, bind along the length of actin filaments, stabilizing them by interacting with three subunits (Cooper, 1987; Bubb et al., 1994; Pospich et al., 2020). Jasplakinolide is membrane permeable, and when used in vivo, it hyperstabilizes preexisting actin filaments and induces polymerization (Bubb et al., 1994). Fluorescently conjugated derivatives of jasplakinolide, like SiR-actin, have also been useful in low concentrations for live-imaging studies (Lukinavicius et al., 2014) (also see Melak et al., 2017 for additional live imaging probes). In contrast, phalloidin is membrane impermeable, limiting its uses for live cells except for some special cases (e.g., microinjection; [Hamaguchi and Mabuchi, 1982; Schmit and Lambert, 1990; Wehland et al., 1977]). For experiments using fixed cells, however, fluorescently conjugated phalloidin is used widely and considered the gold standard, to which other actin-visualization methods are often compared (Wulf et al., 1979; Melak et al., 2017).

INHIBITORS OF ACTIN-BINDING PROTEINS

In contrast with small molecules that bind actin directly to influence its dynamics, other compounds target actin-binding proteins (Figure 1). Actin rarely assembles spontaneously in cells, and instead is highly regulated by layers of upstream proteins and signals (Velle et al., 2024). Among these regulators are actin nucleators—a class of proteins and complexes that initiate the assembly of a new filament by stabilizing an actin dimer or trimer. Inhibitors of actin nucleators and other actin-binding proteins can therefore be used to impair specific types of actin networks and activities.

Branched actin networks are assembled by the Arp2/3 complex—an actin nucleator that, when activated, can bind to the side of a preexisting filament and initiate the assembly of a new filament to form a branch (Mullins et al., 1998). These networks are ideal for generating expansive forces that push on membranes. Arp2/3-derived actin networks drive myriad cellular processes, including crawling motility, vesicular trafficking, endocytosis, and phagocytosis (Velle et al., 2024), making the Arp2/3 complex a useful target for studying these phenotypes. The small-molecule CK-666 directly binds the Arp2/3 complex at the interface of Arp2 and Arp3 subunits, and prevents it from nucleating new filaments (Baggett et al., 2012; Hetrick et al., 2013; Nolen et al., 2009). Similarly, the small-molecule CK-869 locks the Arp2/3 complex in an inactive conformation by binding to Arp3 (Hetrick et al., 2013; Nolen et al., 2009). These compounds have been useful for studies of diverse systems, including plant cells (Xu et al., 2024), single-celled eukaryotes (Velle and Fritz-Laylin, 2020; Prostak et al., 2021; Bigge et al., 2023a), animal cells (e.g., lamellipodia in sea urchin cells [Henson et al., 2015] and mammalian cells [Abu Taha et al., 2014; Haynes et al., 2015], as well as spindle positioning and cytokinesis in mouse oocytes [Sun et al., 2011; Yi et al., 2011]), and for investigating host–pathogen interactions (Dhanda et al., 2018; Miller et al., 2018; Pfanzelter et al., 2018; Velle and Campellone, 2018; Malych et al., 2025). Looking upstream of the Arp2/3 complex, wiskostatin is a compound that binds the Arp2/3 complex activator N-WASP, locking it in its autoinhibited conformation (Peterson et al., 2004). Although useful for impairing only the N-WASP pathway of Arp2/3 activation (and not the ~10 other activators), wiskostatin has well-documented off-target effects such as reducing cellular ATP levels (Guerriero and Weisz, 2007).

Contractile actin networks often drive phenotypes like migration, cytokinesis, and membrane blebbing (Velle et al., 2024). These networks typically involve actin filaments assembled by formin family proteins and require the motor protein myosin II. Formins can both nucleate and elongate actin filaments, and these activities can be perturbed by SMIFH2, a broad inhibitor of formin family proteins thought to bind the FH2 domain (Rizvi et al., 2009). Additionally, SMIFH2 has been shown to inhibit multiple myosins (Nishimura et al., 2021), which makes this drug potentially useful for broadly assessing the roles of actin and myosin networks. However, due to this lack of specificity (likely due to high electrophilicity [Baell, 2010]) and additional known off-target effects (Innocenti, 2023), caution should be taken in interpreting any result with SMIFH2.

Myosin II can also be specifically inhibited by the small-molecule blebbistatin (Straight et al., 2003). Blebbistatin works by binding myosin II when its motor head is complexed with ADP-Pi and slowing the release of phosphate (Kovacs et al., 2004). Combining blebbistatin with fluorescence imaging can be problematic, as it is inactivated by blue light (Kolega, 2004; Sakamoto et al., 2005), making it incompatible with GFP-tagged proteins and some fluorescent dyes. Although this could be leveraged as a way to inactivate blebbistatin in specific cells or subcellular regions, caution should be taken as it becomes toxic in cells following blue light inactivation (Kolega, 2004; Sakamoto et al., 2005; Mikulich et al., 2012). Due to these problems and its low solubility in aqueous solutions, several derivatives of blebbistatin like para-aminoblebbistatin have become popular alternatives (Kepiro et al., 2012; Kepiro et al., 2014; Varkuti et al., 2016). Blebbistatin and its derivatives have been used for studying myosin II-dependent processes in budding and fission yeasts (Mishra et al., 2013; Malla et al., 2022), chytrid fungi (Robinson et al., 2022), Dictyostelium amoebae (Shu et al., 2005), and mammalian cells (Guha et al., 2005; Murthy and Wadsworth, 2005). Although these examples represent a wide array of diverse organisms, blebbistatin is notably ineffective against Drosophila melanogaster’s myosin II in cells and in vitro (Straight et al., 2003; Heissler et al., 2015), highlighting the importance of carefully testing even widely used drugs when introducing them into a new system.

ORIGINS AND ECOLOGICAL ROLES OF ACTIN INHIBITORS

Although CK-666, CK-869, SMIFH2, and blebbistatin are all synthetic compounds, many inhibitors of actin networks are derived from nature (Figure 2). Actin toxins are produced by a diverse array of organisms, including marine sponges (latrunculins [Kashman et al., 1980] and swinholide A [Carmely and Kashman, 1985]), fungi (cytochalasins [Aldridge et al., 1967] and phalloidin [Wieland, 1968]), dinoflagellates (pectenotoxin [Yasumoto et al., 1984] and goniodomin [Sharma et al., 1968]), and bacteria (occidiofungin [Lu et al., 2009] and ACD toxins [Sheahan et al., 2004]). Some of these toxins may prevent predation; pectenotoxin, for example, is produced by dinoflagellates and has been suggested as a defense against copepod grazing (Pourdanandeh et al., 2025). Some others may be used in turf wars for resources between species; cytochalasins and other fungicides are produced at the site of contact when two fungi are cocultured (Knowles et al., 2019). Toxins may even function as offensive weapons, as some cytochalasins confer the ability of fungal pathogens to infect plants (Saiwai et al., 1983; Tsurushima et al., 2005). Compared with our understanding of how these compounds work in some model cells in the laboratory and in experiments in vitro, little is known about how they are synthesized, their natural targets, or how the targeted cells respond to them. Here, we detail one of the most well-studied actin inhibitors, latrunculin.

FIGURE 2:

Frequently used inhibitors can impair actin dynamics in many but not all species. Inhibitors are listed in a table with their source, binding site, and target. For actin-binding inhibitors, a few key residues for binding corresponding to the full rabbit muscle actin sequence (XP_002722940.2) are listed, although this list is not exhaustive (see also: Supplemental Figures S5–S6 in [Velle and Fritz-Laylin, 2020], Figure 1 in [Wirshing et al., 2025] and [Faulstich et al., 1993; Belmont et al., 1999; Morton et al., 2000; Nair et al., 2008; Pospich et al., 2017; Kumari et al., 2020]). Model organisms and their phylogenetic relationships are also shown, with circles, indicating if an inhibitor causes an actin-related phenotype (black circle), if it has been tested but does not cause a phenotype (white circle), if the target proteins are absent (slash), or if there is insufficient information to determine if an inhibitor is effective against actin networks in that organism (gray circle). The effects of phalloidin are based on positive actin staining using fixed and/or permeabilized cells, except for a few microinjection experiments and experiments using red algae. Cytochalasin D has only been shown to be effective in permeabilized yeast. Asterisks: LatB is effective against one actin isoform (encoded by IDA5 gene) but not another (NAP1) in Chlamydomonas; Chlamydomonas lacks Myosin II, but blebbistatin has effects that may be consistent with inhibiting other myosins. Figure References: Dictyostelium: (Gerisch et al., 2004; Shu et al., 2005; Mondal et al., 2008; Yumura et al., 2014; Dong et al., 2016; Ecke et al., 2020); Mammals (mouse, rat, rabbit, or human): (Carter, 1967; Spector et al., 1983; Bubb et al., 2000; Straight et al., 2003; Nolen et al., 2009; Rizvi et al., 2009); Budding yeast (Saccharomyces cerevisiae): (Adams and Pringle, 1984; Li et al., 1995; Ayscough et al., 1997; Nolen et al., 2009; Mendes Pinto et al., 2012; Mishra et al., 2013; Toshima et al., 2016); Chytrid fungus (Batrachochytrium dendrobatidis): (Fritz-Laylin et al., 2017; Prostak et al., 2021; Robinson et al., 2022); Arabidopsis: (Sawitzky et al., 1999; Collings et al., 2006; Cao et al., 2016; Xu et al., 2024); Green Algae (Chlamydomonas or Micrasterias): (Hoffman and Goodenough, 1980; Dentler and Adams, 1992; Holzinger and Meindl, 1997; Avasthi et al., 2014; Onishi et al., 2016; Craig et al., 2019); Red Algae (Porphyra, Bostrychia, Cyanidium, or Cyanidioschyzon): (Suzuki et al., 1995; Wilson et al., 2002; Ackland et al., 2007; Maschmann et al., 2020); Tetrahymena: (Hirono et al., 1989; Zackroff and Hufnagel, 2002; Hosein et al., 2005); Diatoms (various species): (Poulsen et al., 1999); Reticulomyxa: (Koonce et al., 1986); Giardia: (Paredez et al., 2011); Naegleria: (Velle and Fritz-Laylin, 2020).

Latrunculins are isolated from the Red Sea sponge Negombata magnifica (formerly Latrunculia magnifica) (Nèeman et al., 1975; Spector et al., 1983). Although LatA can be synthesized in vitro (Furstner et al., 2007), the biosynthesis pathway in the sea sponge is still unknown and may involve the sponge’s symbionts. Once produced, sponges may use the toxin for self-defense against predators such as carnivorous fish (Nèeman et al., 1975; Proksch, 1994). Alternatively, these toxins may be utilized to “paralyze” the prey microbes that sponges filter-feed upon. Regardless of what latrunculins are used for, a question is raised: How are sponges not intoxicated by their own latrunculins? One potential answer is by diversifying their actin sequences. Although genomic information is not available for N. magnifica, many sponges encode multiple actin genes. For example, a well-studied sponge, Halichondria panicea, has one conventional (accession XP_064390159.1, 97% identical to human α-actin) and one highly divergent actin (accession XP_064405707.1, 67% identical to human α-actin). Because the latter has a substitution in one of the residues involved in direct hydrogen bonding between human α-actin and LatA (tyrosine 71, which is positioned inside the ATP/ADP-binding pocket and substituted with asparagine in XP_064405707 [Morton et al., 2000]), it is possible that this actin does not bind to LatA and therefore confers resistance to the cells expressing this toxin.

This strategy of diversifying actin sequences to achieve resistance is not limited to toxin producers. Nudibranch mollusks in the genus Chromodoris feed on sponges and collect LatA, which in turn is used for self-defense against predators (Cheney et al., 2016). Although the precise mechanism of how they avoid self-toxication by LatA is unclear, the divergent actins in their genomes have substitutions in residues involved in LatA-actin binding (Hertzer et al., 2023). In another reported example, the green alga Chlamydomonas reinhardtii has two actins, one conventional (IDA5) and one divergent (NAP1) (Kato-Minoura et al., 1997). Interestingly, NAP1 is not expressed in vegetative cells under normal growth conditions, but it is transcriptionally upregulated upon exposure to latrunculins and depolymerization of filaments made of IDA5 (Onishi et al., 2016; Onishi et al., 2018). NAP1 has a substitution at the position equivalent to tyrosine 71 in human α-actin (substituted with histidine), and indeed, NAP1 forms filaments resistant to the toxins. Thus, isolation of nap1 mutants (which are sensitive to latrunculins) was crucial for studying the roles of actin filaments in processes such as cell division, endocytosis, ciliary assembly, and chloroplast division (Onishi et al., 2016; Onishi et al., 2020; Bigge et al., 2023b; Clark-Cotton et al., 2025). The exact reason why Chlamydomonas (and related algae in the Volvocine group; Kato-Minoura et al., 2003) has evolved this apparent defense mechanism against latrunculin toxins is unclear, but divergent actins are found throughout the eukaryotic tree. Some of these actins are reported to be resistant to latrunculins and/or have mutations similar to those reported to inhibit latrunculin-binding (Ayscough et al., 1997; Belmont and Drubin, 1998; Belmont et al., 1999; Morton et al., 2000; Fujita et al., 2003; Procaccio et al., 2006; Schwarzerova et al., 2010; Riviere et al., 2012; Johnston et al., 2013; Onishi et al., 2016; Filipuzzi et al., 2017; Hertzer et al., 2023). Examples include plant Arabidopsis thaliana (ACT9, AT2G42090.1), ciliate Paramecium tetraurelia (e.g., XP_001459509.1; [Sehring et al., 2007b]), intestinal parasite Giardia lamblia (Paredez et al., 2011), tunicate Ciona intestinalis (XP_002128924.1), amoeba Dictyostelium discoideum (e.g., XP_636187), Pipistrellus bats (CAK6444077.1, XP_036265340.1), Fusarium fungi (e.g., UPK93335), and Symbiodinium dinoflagellates (e.g., CAE7723233.1, CAE7265032.1, CAE7619015.1). These divergent sequences suggest that actins can evolve at some decent rate while maintaining their essential functions, and some of the evolution may occur as a result of molecular cat-and-mouse between actin and actin-targeting toxins. Thus, when using actin inhibitors in experiments, it is important to consider the possibility that the target organism or cell type may exhibit some unexpected resistance or sensitivity to these drugs. Comparisons of in vivo and in vitro results should be done with some caution, especially when actin proteins from different genomes are used in them. Conversely, when we consider the evolution of actin proteins, nature has already done some genetics experiments and provides rich resources that have not been fully utilized (Velle et al., 2024).

USE AND MISUSE OF ACTIN INHIBITORS

Small-molecule inhibitors have been a staple in actin cytoskeletal research for decades, and continue to be a valuable tool for probing actin-based phenotypes. Many of these inhibitors work on a wide array of model systems and have well-documented activities (Figure 2). Inhibitors also have some advantages over genetic tools: effects typically occur on a rapid timescale—seconds to minutes— and many inhibitors are reversible and can be washed out. Furthermore, because the actin cytoskeleton is vital to so many cellular processes, knocking out actin itself is often not a viable option. Moreover, the formin inhibitor SMIFH2 has an advantage over genetic tools, in that it inhibits many different formins (as well as myosin II) simultaneously, so it can be used to quickly assess the contributions of formin-derived contractile networks to a particular phenotype (this would, of course, be best paired with further experiments, including genetic manipulations of individual genes).

Despite the many advantages of small-molecule inhibitors, certain precautions should be taken, particularly when introducing them to a new model system. Here are some tips:

1. Check the binding site: If a particular drug has not been used in the literature for your model system, it is typically worth checking if the residues at the drug’s binding site are conserved (Figure 2) (Sehring et al., 2007a; Paredez et al., 2011; Onishi et al., 2016; Velle and Fritz-Laylin, 2020; Wirshing et al., 2025). For example, many ascomycete fungi have a mutation (valine at position 75) in their actin genes at a phalloidin-binding site, and fluorescent phalloidins fail to stain actin filaments in these species. Amazingly, a single “reversion” mutation at this position to isoleucine is sufficient to allow actin visualization in one of the species, Aureobasidium pullulans (Wirshing et al., 2025). However, just because a binding site is conserved, or even if a drug binds in vitro, this does not guarantee the compound will be effective in living cells.

2. Test a wide array of concentrations: Using the literature as a guide, test a range of concentrations. Effective in vivo drug doses for different organisms and cell types can vary by orders of magnitude. For example, to completely depolymerize actin filaments, 100 μM LatA is needed for budding yeast (Ayscough et al., 1997), while 1 μM is sufficient for mammalian fibroblasts (Spector et al., 1989). Organisms with cell walls, low membrane permeability, and/or strong efflux pumps may require a higher dose of an inhibitor. Alternatively, mutations in efflux pumps (Asadi et al., 2017; Ayscough, 2000), or a pairing with an efflux pump inhibitor such as verapamil (Lukinavicius et al., 2014) can reduce the required concentration. It is best practice to use the lowest concentration that still results in a robust phenotype. A drug that only has an effect at concentrations orders of magnitude above what is needed for a similar organism can be a red flag. Be aware, however, that the efficacy of inhibitors may be variable from one batch/source to another. As an example, LatB (which is produced by a single supplier) efficacy has not been consistent across several lot numbers (e.g., Onishi et al., 2018), which might have contributed to the notion that LatB is less effective/stable than LatA.

3. Use controls: Always use a vehicle control and a control with nothing added. These negative controls can be especially important for drugs dissolved in DMSO—a common solvent that can affect a cell’s biology (Fukui and Katsumaru, 1980; Holthaus et al., 2018; Verheijen et al., 2019; Prostak et al., 2021). Inactive controls, including CK-689 for CK-666 (Nolen et al., 2009) and the inactive (+)-blebbistatin enantiomer (Straight et al., 2003), may also be useful. Positive controls are helpful if they are available for your system; for example, if knocking down Arp2/3 complex phenocopies CK-666 treatment, and if there are no additive effects when the two are combined, the drug is likely hitting its target (Rotty et al., 2017; Xu et al., 2024).

4. Beware of off-target effects: Some inhibitors, including SMIFH2, wiskostatin, and cytochalasins, have well-documented off-target effects (Zigmond and Hirsch, 1972; Jarett and Smith, 1979; Guerriero and Weisz, 2007; Cossar et al., 2023; Innocenti, 2023). If your system is genetically tractable, check inhibitors of actin-binding proteins by knocking out the target of the inhibitor before treatment to reveal other phenotypes caused by the inhibitor (e.g., Shu et al., 2005). If latrunculin is effective in your system, treating with latrunculin first and then adding the inhibitor of interest can show whether a phenotype is actin linked—if the phenotype persists in the absence of actin filaments, the inhibitor is likely off-target. Finally, actin inhibitors can compete with actin-binding proteins, upsetting the normal balance between actin and its regulators (e.g., Klenchin et al., 2003; Wang et al., 2019), and potentially causing emergent phenotypes that are not directly related to the inhibitor’s target.

5. Be mindful of different isoforms: An inhibitor may not inhibit every version of its target, especially if multiple genes encode multiple versions. In addition to the well-documented latrunculin-resistant actins (see above), in mammalian cells there are eight different possible Arp2/3 complexes, which are differentially inhibited by CK-666 and CK-869 (Cao et al., 2024).

FUTURE OUTLOOK

Although inhibitors of actin and actin-binding proteins have been used for decades, their relevance is only growing. Testing well-described compounds in new and emerging model systems is important for understanding the fundamentals of their cytoskeletal biology, as well as evolution. For example, recent studies have shown that some Asgard archaeans possess components of a eukaryotic-like cytoskeleton (reviewed in Charles-Orszag et al., 2024). Having inhibitors that can impair these archaeal actin networks could be critical for defining the contributions of actin to their biology and tracing the evolution of actin phenotypes. Additionally, investigating the roles of actin toxins as biological weapons can tell us more about ecology and the context in which some actins may have evolved. Finally, screens of natural and synthetic products have led to the discovery of additional compounds, which could be useful for treating some eukaryotic pathogens like fungal infections (Ravichandran et al., 2019), toxoplasmosis (Kelsen et al., 2023), or malaria (Moussaoui et al., 2023). New screens or modifications of existing compounds could also be useful for identifying new ways to visualize actin in live cells; although SiR actin (based on Jasplakinolide) is a valuable tool for mammalian cells, it has not been successfully used in most other eukaryotic cell types. By continuing to screen both new compounds and new species, actin inhibitors are certain to remain powerful tools for studying cell biology, evolution, ecology, and pathogenesis.