Synergistic assembly, disassembly, and protection of complex forms of bundled F-actin

Many physiological functions, including motility, polarity, cell division, endocytosis, and connectivity, are controlled via rapid assembly and disassembly of the actin cytoskeleton. Actin assembly from its monomeric/globular (G) state into filamentous (F) states provides the supporting structures and dynamics needed for these diverse physiological functions (Blanchoin et al., 2014). Moreover, F-actin’s ability to be assembled into tightly packed parallel bundles further increases its stability (Rajan et al., 2023a)—including in the main core of finger-like cell protrusions (such as filopodia, stereocilia, and microvilli), and stress fibers that assist/maintain cellular adhesion and integrity (Rajan et al., 2023a). A common aspect of F-actin bundles is the presence of multiple classes of actin-bundling proteins—such as fascins, espins, fimbrin/plastins, and villins—in different combinations and molar ratios (Rajan et al., 2023a). For example, although fascin is a well-known major bundling protein, additional bundling proteins such as espins (also called ESPN, DFNB36, and forked) and fimbrin/plastins are also present in fascin-bundled F-actin structures (Rajan et al., 2023a). Therefore, fascin alone may not be optimal for their formation, maintenance, and/or function. Notably, a loss/mutation of any of these bundling proteins dramatically affects cohesive actin bundled structures (Cant and Cooley, 1996; Cant et al., 1994; Krey et al., 2016; Sekerková et al., 2011; Van Audenhove et al., 2016; Zheng et al., 2000). This underscores the importance of both individual and multiple bundling proteins in providing distinct structural and mechanical properties to bundles and tailoring them for specific cellular functions (Rajan et al., 2023a).

Stable actin bundles also need to undergo tightly regulated spatiotemporally targeted disassembly to allow cells to undergo functional remodeling (Blanchoin et al., 2014; Lappalainen et al., 2022). The molecules and mechanisms underlying this regulated disassembly are still poorly understood. Furthermore, to the best of our knowledge, the ability of disassemblers to destabilize bundles composed of multiple bundling proteins has not been examined. Recently, we found that MICAL—a potent disassembler of F-actin (Rajan et al., 2023b)—is also able to robustly disassemble fascin-bundled F-actin (Hung et al., 2010; Rajan et al., 2023c). Likewise, although cofilin, a well-known F-actin disassembler, is a relatively poor disassembler of fascin-bundled F-actin compared with Mical (Rajan et al., 2023c; and also see Breitsprecher et al., 2011; Chikireddy et al., 2024; Elkhatib et al., 2014; Schmoller et al., 2011), it robustly enhances MICAL’s disassembly of fascin-bundled F-actin, and vice versa (Rajan et al., 2023c). In light of these disassembly effects on fascin-bundled F-actin, we wondered how MICAL and cofilin affect F-actin bundled with multiple bundling proteins. For example, both fascin and espin have been well-studied in vitro and in vivo for their ability to form superbundles containing many filaments (Claessens et al., 2008; Krishnan et al., 2021; Tilney and DeRosier, 2005; Tilney et al., 1995; Wulfkuhle et al., 1998). Together, they play key roles in organizing and stabilizing actin structures in filopodia, stereocilia, microvilli, and other subcellular regions in many cell types (Chou et al., 2011; Claessens et al., 2008; Davidson et al., 2019; Dickinson and Thatcher, 1997; Krishnan et al., 2021; Okenve-Ramos and Llimargas, 2014; Tilney et al., 1995; Winkelman et al., 2016). However, the molecular, biochemical, and cellular mechanisms by which complex superbundles are disassembled and remodeled are unknown.

Thus, herein we sought to further define the role of fascin and espin in actin bundles stability, topology, and robustness—and the mechanisms through which they are disassembled. Our results confirm that while fascin and espin colocalize on actin filaments, they do not compete for binding to them and together bundle them robustly. Furthermore, we find that fascin and espin synergize to form wider, longer, and denser actin bundles than expected from the sum of their individual effects. Moreover, despite their increased stability, fascin–espin co-bundled filaments are robustly disassembled by Mical and Mical in combination with cofilin. However, compared with bundles composed of fascin or espin alone, their co-bundled filaments are more resistant to this disassembly. In vivo studies support our biochemical results—including that fascin and espin function together to dampen Mical’s F-actin disassembly—which is critical for controlling cellular remodeling and semaphorin repulsion. Thus, together, our results provide a first understanding of how complex F-actin bundles are dismantled. Our results demonstrate also how the presence of two different bundling proteins confers unique properties to actin filaments to help protect them from disassembly.

Fascin is a 55-kDa protein that uses its four tandem β-trefoil domains (β-T) to form two actin-binding sites (ABS), which bind and bundle two actin filaments (Fig. 1 A). Espins consist of a main C-terminal domain, which is necessary and sufficient for actin binding/bundling (ABD, Fig. 1 A), and shares homology with Drosophila-forked proteins (Bartles, 2000; Rajan et al., 2023a; Sekerková et al., 2006). In mammals, four splice forms (1 to 4) of the single espin gene have been identified (Sekerková et al., 2004). In this study, we used the shortest espin splice form—splice form 4 (∼30 kDa; also called espin-4 or small espin; Fig. 1 A)—because it contains the common espin C-terminal ABD, it has been used in foundational studies of espin (Bartles et al., 1998; Claessens et al., 2008), and it does not contain domains that serve to autoinhibit espin’s function (Liu et al., 2016; Zheng et al., 2014). Furthermore, since we have used the Drosophila model for in vivo studies herein (see below), it is notable that the espin-4 splice variant most closely matches Drosophila espin (forked), and it has the domains sufficient to rescue the defects that occur in vivo when Drosophila forked (espin) is removed (Grieshaber et al., 2001). This allowed us to more closely compare our in vitro and in vivo results. Below, for simplicity, we refer to this espin-4 splice form as “espin.”

First, we compared fascin and espin’s effects on actin bundling by using low-speed sedimentation assays (Figs. 1, B–D; and S1, A and B). The amount of pelleted actin (bundled F-actin) increased with increasing concentrations of either fascin or espin (Fig. 1, B–D; and Fig. S1 B). However, fascin appeared to be a more potent actin bundler than espin, as it induced greater bundles formation at the same molar ratios to actin (Fig. 1 B). We found that the binding density of Drosophila fascin (bound fascin/actin ≈ 0.23; Fig. 1 C) aligned well with the expected fascin:actin ratio in bundles (calculated as 1/13 × 6/2 = 0.23 based on human fascin-1 structural information [Gong et al., 2025]). Drosophila fascin’s binding affinity (K_D) to bundled actin (“bundling affinity”) was also similar (0.57 ± 0.09 µM) (Fig. 1 C) to published work (e.g., similar to zebrafish fascin 2b (K_d ≈ 0.37 µM [Chou et al., 2011]). Espin showed slightly lower actin-bundling affinity (K_D = 0.84 ± 0.07 µM) and an espin:actin-binding density of ∼0.45 (Fig. 1, C and D). Based on high-speed sedimentation assays (Bartles et al., 1998) and small-angle X-ray scattering (SAXS) data (from espin 3A-actin bundles) (Purdy et al., 2007), it has been reported that one espin molecule binds to ∼3–4 actin monomers. Similar to fascin, espin bundled F-actin with moderate cooperativity (Hill cooperativity coefficient [n] = 1.8 ± 0.2 versus 1.8 ± 0.4 for fascin) (Fig. 1, C and D).

To better monitor fascin- and espin-mediated actin bundles formation, light-scattering experiments were performed. Fascin rapidly increased light scattering in a concentration-dependent manner, indicating rapid bundle formation (Fig. 1 E and Fig. S1 C (1) [Rajan et al., 2023c]). Similarly, espin rapidly, and in a concentration-dependent manner, increased light scattering, albeit to a smaller extent than with fascin (Fig. 1 E and Fig. S1 C (2)). Indeed, at the same protein concentrations, fascin and espin exhibited notable light-scattering differences (Fig. 1 E), supporting our low-speed sedimentation results (Fig. 1 B).

To further monitor these bundles, we used transmission electron microscopy (EM). Fascin (at 1:1 molar ratio to actin) organized F-actin into bundles with a well-defined thickness (Fig. 1 F). These bundles had an average width of ∼114 nm and contained ∼18 parallel filaments, with an interfilament distance of ∼8.5 nm (Table 1). These Drosophila fascin–actin bundles parameters are consistent with results obtained using mammalian fascins (Ishikawa et al., 2003; Jansen et al., 2011).

Espin also generated parallel actin bundles (Fig. 1 F). However, the bundles formed either at 1:1 or 2:1 espin:actin molar ratios were thinner than fascin–actin bundles, with F-actin more closely spaced in them (∼ 7 nm; Table 1). The closer spacing may be due to espin’s smaller size than fascin (29 versus 55 kDa). The espin–actin bundle formation at higher espin:actin molar ratios is consistent with previous work, in which 4:1 ratio of espin (espin-4) to actin was used (Claessens et al., 2008). However, wider bundles at a lower ratio of espin to actin (1:7 molar ratio) have also been seen (Bartles et al., 1998). These differences may be due to experimental conditions—specifically, the incubation temperature of actin with espin (37°C in Bartles et al. [1998] versus room temperature [RT] in our experiments). Taken together, these results show robust F-actin bundling by both fascin and espin, but indicate that fascin is more potent in that than espin.

To define better fascin and espin’s F-actin bundling, we examined their localization along F-actin. Previously, another espin isoform (espin-2B) was shown to be associated with fascin domains in actin bundles (Winkelman et al., 2016). Therefore, we used three-color total internal reflection fluorescence microscopy (TIRFM) assays to image simultaneously fascin, espin, and F-actin. Notably, we found that fascin and espin co-localize on actin bundles (Fig. S1 D). In light of this co-localization along F-actin, we tested whether fascin and espin compete for F-actin binding. In sedimentation assays, fascin’s presence reduced only marginally the bound espin in actin bundles (Fig. S1 E). Similarly, only a marginal reduction in fascin binding to bundled F-actin was observed in espin’s presence (Fig. S1 F). These results reveal little or no competition between fascin and espin for their binding to bundled F-actin. Thus, as both these two proteins are associated with common F-actin structures, such as filopodia, microvilli, and stereocilia (Rajan et al., 2023a), we conclude that although fascin and espin colocalize along filaments, they likely bind to different regions in them (as they have different ABS sequences, Fig. 1 A). Based on our data (Fig. 1 C) and earlier studies (Aramaki et al., 2016; Gong et al., 2025; Jansen et al., 2011; Song et al., 2025), under saturating conditions, one fascin molecule binds every 13 actin monomers within bundles and most likely leaves enough space for espin binding between two bound fascins.

Our observations support previous results from multiple labs that fascin and espin are robust F-actin–bundling proteins (Rajan et al., 2023a). Previous results with purified proteins have also revealed that together fascin and espin strongly increase F-actin bundles size (Claessens et al., 2008). We therefore further examined their combined bundling effects. Interestingly, our results—using sedimentation and light-scattering assays—indicated that when fascin and espin were added together, they were more effective at bundling F-actin than the sum of their individual effects—i.e., they showed a synergistic effect in F-actin bundling (Fig. 2, A and B). Thus, we investigated the potential synergistic effect of these two bundlers.

To simplify this task, we chose protein concentrations at which each bundling protein alone yielded limited bundles. Yet, we found that when present together these two proteins strongly increased the amount of pelleted (bundled) actin to ∼80%, i.e., ∼23% over the sum of their individual pelleting effects (14% [fascin] + 43% [espin] = 57%) (Fig. 2 A). We also examined these effects using light-scattering assays, which we found to be more sensitive for monitoring fascin and espin’s synergy. Similarly, we found that together fascin and espin increased light scattering beyond the sum of their individual effects (Fig. 2 B and Fig. S2 A). In each case, at the same total concentrations of fascin or espin, the light-scattering increase was greater when both proteins were present (Fig. S2 B, blue line [only fascin] and gray line [only espin], and compare it with the orange and purple lines in Fig. S2 A (1 and 2), and the fuchsia line in Fig. S2 B), showing their synergistic effects in actin bundling.

Notably, the light-scattering increase was fascin concentration dependent (Fig. S2 C), indicating its greater than espin’s bundling contribution to their synergistic effect. These synergistic effects are also supported by additional assays described below. Thus, fascin and espin synergize to form actin bundles since together their ability to form F-actin bundles is greater than the sum of their individual effects.

To gain more insight into how fascin and espin regulate actin bundles properties, we used TIRFM to compare their individual and combined effects on actin bundles. The images shown in Fig. 2 C are representative of actin bundles after treatment of on-slide polymerized Alexa488-labeled actin with bundling protein(s). First, we measured actin bundles length under these conditions. With fascin (0.05 µM), mainly F-actin was seen, with occasional short bundled actin regions (Fig. 2 C). With espin (0.15 µM), more bundles were seen, with an average length of ∼10 μm (Fig. 2 C). However, with both fascin (0.05 µM) and espin (0.15 µM), the bundles were approximately twofold longer than the combined length of bundles formed with fascin and espin added alone (Fig. 2 D). These results further support fascin and espin’s synergistic effect on F-actin bundling.

Next, we measured the thickness of fascin, espin, and fascin–espin–actin bundles. Using TIRFM, we observed that in the presence of both fascin and espin, versus either fascin or espin alone, there was a trend toward increased bundling and thicker bundles (at least five filaments/bundles) (Fig. S2 D). However, due to fluorescence intensity saturation (after five filaments; Fig. S2 E), only limited filament numbers in the bundles could be measured using TIRFM. Thus, we used EM to measure the bundles’ width.

Consistent with our TIRFM analysis, at low fascin (only) concentrations, mostly single filaments were observed, with just a few bundles of two filaments (Fig. 2, E and F; and Table 1). With espin only, thin bundles (∼30 nm wide) consisting of four to five filaments were observed (Fig. 2, E and F; and Table 1). In the presence of both fascin and espin, the bundles were approximately two to three times wider (∼100 nm) than those of either fascin or espin bundles, with >10 filaments (Fig. 2, E and F; and Table 1). Further increasing fascin and espin levels induced thicker bundles (Fig. S2 F). The interfilament distance in these fascin + espin bundles was ∼7.7 nm, a value intermediate of that observed for fascin or espin-only bundles (Table 1). Our analysis of both TIRFM and EM results are also consistent with there being a correlation between the fascin + espin synergistic increase in length and width (r² = 0.95; Fig. S2 G). Thus, fascin and espin synergize to generate thicker and longer actin bundles compared with the sum of either bundler alone. This synergistic interaction between fascin and espin provides mechanistic insights into previous observations (mentioned above) that the presence of both fascin and espin strongly increases F-actin bundles size (Claessens et al., 2008). This synergistic interaction we have identified also provides mechanistic insights into cellular results revealing that fascin and espin together generate thicker F-actin bundles (Krishnan et al., 2021; Tilney et al., 1995) and longer filopodia (Chou et al., 2011) than either protein on its own.

Our previous results revealed that Mical rapidly and extensively dismantles fascin-bundled F-actin (Hung et al., 2010; Rajan et al., 2023c); hence, we examined its effect on espin-bundled F-actin. Notably, using low-speed sedimentation, we found that Mical markedly reduced the amount of espin-bundled F-actin (Fig. 3 A). These results reveal that Mical robustly disassembles espin-bundled F-actin. Thus, we looked at the effects of Mical on fascin + espin co-bundled F-actin. In particular, the actin bundles in cells contain multiple bundling proteins and undergo tightly regulated spatiotemporal disassembly and turnover to support cellular responses (cytoskeletal remodeling) (Blanchoin et al., 2014; Lappalainen et al., 2022). However, to the best of our knowledge, no proteins have been identified that destabilize bundles composed of multiple bundling proteins. Notably, using low-speed sedimentation, we found that Mical markedly reduced the amount of fascin–espin co-bundled F-actin (Fig. 3 B). We also found using high-speed disassembly assays that Mical not only disassembles these fascin–espin co-bundled filaments to unbundled filaments (as revealed by low-speed sedimentation; Fig. 3, A and B) but also to G-actin (Fig. 3 C). Thus, Mical robustly disassembles complex F-actin bundles—i.e., those simultaneously bundled with more than one type of bundling protein—including down to the G-actin form.

We next compared the impact of Mical on unbundled, fascin-bundled, espin-bundled, and fascin–espin co-bundled F-actin. Consistent with our previous results, we found using pyrene actin assays that fascin-bundled F-actin was more resistant to Mical’s F-actin disassembly than unbundled F-actin (Fig. 4 A [Rajan et al., 2023c]). Notably, we also saw a similar protective effect of espin on F-actin: that espin-bundled F-actin was more resistant to Mical’s disassembly than unbundled F-actin (Fig. 4 A). Furthermore, when both fascin and espin were present in bundles, we found that they were even more resistant to Mical’s disassembly (Fig. 4 A). These results indicate that the joint presence of fascin and espin serves to increase F-actin’s protection against Mical-mediated disassembly.

To better quantify and compare Mical’s effect on these different types of bundled filaments, we reduced its concentration. Using sedimentation assays, we found that these lower Mical concentrations still decreased the amount of pelleted (bundled) actin in all three bundling proteins conditions (Fig. S3 A (1 and 2)). Yet, although we used similar bundling protein concentrations in all conditions (0.4 μM), the percentage decrease in the pelleted (bundled) actin was smaller for bundles containing both fascin and espin than with either fascin or espin alone (Fig. S3 A (2)). We also observed a similar trend in our light-scattering assays (Fig. 4 B). To examine better these effects at a molecular level, we used time-lapsed TIRFM assays. We found that Mical disassembled F-actin bundled with fascin, espin, and fascin + espin (Fig. 4 C and Video 1). However, the joint presence of fascin and espin notably slowed Mical’s F-actin disassembly—these co-bundles persisted longer than those formed in the presence of equal molar concentrations of either bundler alone (Fig. 4 C and Video 1).

Mical has been found to disassemble actin bundles by shortening and thinning them (Rajan et al., 2023c). Moreover, the thinning sites subsequently become bundles severing sites (Fig. 4 C, cyan arrows). Thus, we compared the shortening, thinning, and severing rates of all three bundles types (Fig. 4 C (2–4)). We found that when fascin and espin were present together, Mical’s bundles shortening was slowed (on average) by 2.2-fold and 3.1-fold compared with its effects on similar molar concentrations of fascin- or espin-bundled F-actin (0.48 ± 0.02 µm/min versus 1.06 ± 0.07 µm/min and 1.53 ± 0.24 µm/min, respectively) (Fig. 4 C (2)). So too, for the first 120 s of data acquisition, the thinning of fascin–espin-bundled F-actin by Mical occurred at ∼3.2- and 6.2-fold slower rate as compared with that of F-actin bundled with fascin or espin alone (Fig. 4 C (3)). Moreover, the overall (for 10 min) thinning rate was ∼1.5- and 2.1-fold slower when both bundlers were present than when they were individually present (Fig. S3 B). These observations suggest that initially, fascin–espin co-bundles robustly resist the Mical-mediated thinning, but eventually begin to disassemble. No severing was observed for all three bundles types for the initial 60 s. Severing was observed subsequently in the fascin-only and espin-only bundles. However, for fascin–espin co-bundled F-actin, no severing was observed even after 400 s (Fig. 4 C (4)), despite the observed bundles thinning. We also noted that similar to the greater contribution of fascin to the fascin–espin synergistic bundling effect (Fig. S2 C), both low-speed sedimentation and light-scattering assays revealed that it also provided a greater contribution to fascin–espin co-bundles resistance to their Mical-mediated disassembly (Fig. S3, C and D). Thus, while Mical disassembles actin bundled with both fascin and espin, these proteins synergize to help protect F-actin from disassembly.

To further investigate the mechanisms underlying fascin–espin co-bundled F-actin disassembly, we looked at cofilin’s effects on them. Cofilin disassembles fascin-bundled F-actin, but not as well as Mical (Rajan et al., 2023c; and also see Breitsprecher et al., 2011; Chikireddy et al., 2024; Elkhatib et al., 2014; Schmoller et al., 2011). We therefore examined cofilin’s effects on espin-bundled F-actin. Using light-scattering and TIRFM assays, we found that cofilin disassembled espin-bundled F-actin, having a stronger effect on them than on fascin-bundled F-actin (Fig. 5, A and B; and Video 2). Furthermore, as we saw with Mical, cofilin caused less disassembly of fascin–espin co-bundled filaments than those with equimolar concentrations of fascin or espin (Fig. 5, A and B; and Video 2).

Mical and cofilin synergize to disassemble unbundled and fascin-bundled F-actin (Grintsevich et al., 2016; Grintsevich et al., 2017; Rajan et al., 2023c; Wioland et al., 2021). We therefore wondered whether Mical and cofilin also worked in a similar synergistic way to disassemble complex fascin–espin co-bundled filaments. Notably, pyrene fluorescence, light scattering, and TIRFM assays revealed that compared with the sum of their individual effects, Mical and cofilin synergized to disassemble both espin-bundled and fascin–espin co-bundled F-actin (Fig. 6, Fig. S4, and Video 3). For example, using TIRFM, we found that Mical + cofilin induced rapid and extensive bundles disassembly of fascin, espin, and fascin–espin co-bundled F-actin (Fig. 6 B (1 and 2); and Video 3). Furthermore, comparing the effects of Mical alone (Fig. 4 C), cofilin alone (Fig. 5 B), and Mical + cofilin (Fig. 6 B)—revealed that Mical + cofilin enhanced more strongly the rate and number of bundles shortening, thinning, and severing events (and note the different time scales between Fig. 6 B (1), Fig. 4 C (1), and Fig. 5 B (1)), compared with the sum of Mical and cofilin’s individual effects (Fig. 6 B (3) and Fig. S4 E).

However, pyrene fluorescence, light-scattering, and TIRFM assays revealed also that even though Mical + cofilin strongly disassembled all three actin bundles types, their impact was reduced when both fascin and espin were present on them together (Fig. 6, A and B; Fig. S4, A–D; and Video 3). Notably, in TIRFM assays, fascin and espin alone bundled F-actin was already shorter and thinner than fascin–espin co-bundled F-actin by the time we started collecting TIRFM data, suggesting their excessive severing and disassembly (in Fig. 6 B (1) compare the 60 s for all conditions, Video 3). Additionally, a ∼2.6-fold and 3.1-fold slower bundles thinning of fascin–espin co-bundled F-actin was observed compared with that seen with equimolar concentrations of fascin or espin containing bundles, respectively (Fig. S4 D). Thus, Mical and cofilin synergize to disassemble F-actin bundles composed of two different bundlers, but when present together in these bundles, fascin and espin synergize to slow down this disassembly.

In light of our identification above of molecular mechanisms that robustly disassemble complex F-actin bundles, we explored their biochemical basis. MICALs are oxidoreductase (redox) enzymes that use actin filaments as a substrate, acting on them through a catalytic posttranslational mechanism (Hung et al., 2011; Rajan et al., 2023b). In particular, MICALs bind F-actin, and in the presence of their coenzyme (NADPH), they are enzymatically active and oxidize actin’s methionine (M) residues M44 and M47 (producing Mical-oxidized actin [Mox-actin]), which destabilizes F-actin (Fig. 7 A [Hung et al., 2011; Rajan et al., 2023b]). We found that Mical requires NADPH to disassemble espin and espin–fascin co-bundled F-actin (Fig. 3, A and B), indicating that Mical uses its catalytic mechanism to disassemble also these different types of bundled filaments. We therefore looked at whether espin-bundled and espin–fascin co-bundled F-actin are indeed a Mical substrate. In the absence of a substrate, Mical consumes slowly its NADPH coenzyme (Fig. 7 B [Hung et al., 2011]). Adding espin alone or espin + fascin to Mical did not significantly increase its NADPH consumption rate (Fig. 7 B), showing that they are not Mical substrates. However, similar to unbundled F-actin, espin-bundled and espin–fascin co-bundled F-actin enhanced Mical’s NADPH consumption (Fig. 7 B). Notably, these results add to our previous observations that fascin-bundled F-actin also enhances Mical’s NADPH consumption (Rajan et al., 2023c). These results demonstrate that F-actin bundled with espin, and complex bundles composed of espin + fascin, similar to unbundled F-actin and fascin-bundled F-actin, are Mical substrates.

We therefore further explored the biochemical mechanisms through which Mical’s enzymatic activity disassembles these complex bundles. Notably, no differences were observed in Mical binding to espin-bundled, fascin-bundled, or espin–fascin co-bundled F-actin versus unbundled F-actin (Fig. S5 A). Also, Mical (in NADPH’s absence) did not affect espin or fascin’s F-actin binding (Fig. S5 B). Thus, Mical in the absence of its enzymatic activity does not alter fascin’s binding/unbinding dynamics. However, we found that once Mical oxidizes F-actin, although fascin and espin’s affinity for Mox–F-actin is similar to that for unoxidized F-actin (Fig. S5 C), this Mox–F-actin is poorly bundled by espin and espin + fascin (Fig. 7 C). These results, coupled with previous results that Mox–F-actin is highly unstable and Mox–G-actin only poorly polymerizes (Grintsevich et al., 2016; Grintsevich et al., 2017; Hung et al., 2011), provide a biochemical understanding of how Mical disassembles complex bundled filaments. Namely, Mical, in the presence of its NADPH coenzyme, oxidizes actin within complex bundles. This destabilizes F-actin bundles, leads to a decrease in F-actin that espin and fascin can bundle well, and produces Mox–G-actin that cannot readily polymerize (Fig. 7 D (1)).

Next, we explored how cofilin and Mical synergize in disassembling these complex bundles. Cofilin targets aged, ADP-enriched F-actin segments (Chin et al., 2016; McCullough et al., 2011; Suarez et al., 2011). It binds cooperatively to F-actin, alters its twist (shortens their helical pitch), and rearranges its inter-protomer contacts, resulting in F-actin disassembly (Bobkov et al., 2002; Bobkov et al., 2004; De La Cruz and Sept, 2010; Galkin et al., 2003; Wioland et al., 2017). We found that cofilin concentrations well above those used in our Mical + cofilin disassembly assays (Fig. 5 and Fig. S4) did not affect espin or fascin F-actin binding (Fig. S5 D). Thus, cofilin does not assist Mical’s bundled F-actin disassembly by preventing bundling proteins from interacting with F-actin. Similarly, cofilin did not increase its binding to bundled F-actin at the fascin and espin concentrations used in our Mical + cofilin disassembly assays (Fig. S5 E). These results are consistent with previous results with fascin and cofilin loading (Breitsprecher et al., 2011; Chikireddy et al., 2024) and indicate that fascin and espin do not assist Mical by increasing cofilin loading onto F-actin. In contrast to that, although Mical alone does not affect cofilin’s F-actin binding, Mical’s F-actin oxidation has been found to increase cofilin’s interaction with F-actin (Grintsevich et al., 2016; Grintsevich et al., 2017; Rajan et al., 2023c; Wioland et al., 2021). These results, therefore present a model consistent with Mical’s effects on unbundled and fascin-bundled F-actin. Namely, Mical’s F-actin oxidation alters F-actin’s conformation and stability—including weakening espin and fascin–F-actin bundling (Fig. 7 D (1)), which increases cofilin’s binding and allows it to disassemble Mox-F-actin more rapidly than the unoxidized F-actin (Fig. 7 D (2 and 3)).

We also sought to understand the mechanisms through which these complex co-bundled filaments are more protected from disassembly by Mical. Notably, although espin-bundled and espin–fascin co-bundled F-actin enhanced Mical’s enzymatic activity (Fig. 7 B), they did so to a lesser extent than unbundled F-actin (Fig. 7 B). Yet, we did not observe any difference in Mical’s binding to espin or espin–fascin co-bundled F-actin versus unbundled F-actin (Fig. S5 A). Thus, these results indicate that espin and fascin’s combined presence in bundled F-actin helps to protect F-actin from Mical-induced disassembly by decreasing Mical’s enzymatic action on F-actin. These effects, in combination with a proposed decreased accessibility of Mical to individual filaments within the thicker bundles produced by espin + fascin (compared with espin and fascin alone), provide mechanisms underlying the increased protection of these complex bundles from Mical + cofilin.

Lastly, we turned to in vivo assays to explore further the mechanisms underlying complex bundled F-actin disassembly. F-actin in vivo is often found simultaneously bundled with different bundling proteins (Rajan et al., 2023a). F-actin in Drosophila bristle cells, for example, is well-known to associate with multiple bundling proteins—most notably espin and fascin—and it is a model for studying these complex F-actin bundles in vivo (Fig. 8 A [Tilney and DeRosier, 2005]). Notably, current models of bristle formation indicate espin appears before fascin and initiates F-actin assembly into bundles (Tilney and DeRosier, 2005). Fascin then further bundles/stabilizes espin-bundled filaments and organizes them into maximally cross-linked F-actin (Tilney and DeRosier, 2005). Removal of either espin or fascin (Drosophilaforked or singed–/– mutants, respectively) generates similar types of bristle morphological defects (see below [Tilney and DeRosier, 2005]).

Based on the results of our in vitro experiments, we set up in vivo experiments using this bristle model system to examine Mical’s effects on the disassembly of espin-bundled, fascin-bundled, and espin–fascin co-bundled F-actin. High Mical levels robustly disassemble bundled F-actin in bristle cells, such that there is very little normal structure remaining in them (Hung et al., 2010). Thus, for better quantification of Mical’s effect on different bundled filaments, we reduced its levels (similar to what we have done in vitro [above] to quantify Mical’s effects on different bundled F-actin). Expressing low Mical levels in bristles results in a low-level of bundled F-actin disassembly that induces localized remodeling to generate a branch (Fig. 8 B [Hung et al., 2010]). Removing espin (using a CRISPR-generated espin mutant knockout [–/–] [Baskar et al., 2019]) also results in bundled F-actin disassembly with effects on bristle extension and shape (Fig. 8 C). Notably, this effect of decreasing espin levels strongly resembles the effect of increasing Mical levels—including F-actin disassembly and remodeling that generates a branch (compare Fig. 8, B and C). Thus, increasing Mical levels and decreasing espin levels result in similar observable traits (i.e., they phenocopy one another), supporting that they are part of the same genetic pathway. Moreover, removing espin, in a background in which low levels of Mical are expressed in bristles, significantly increased Mical’s in vivo effects—resulting in increased Mical-mediated bundled F-actin disassembly and cellular remodeling (compare Fig. 8, B–E). Therefore, as we observed with purified proteins, our in vivo results also indicate that espin bundling helps protects F-actin from Mical’s disassembly.

We thus tested whether espin and fascin work together to help protect F-actin from Mical in vivo. Small decreases in espin levels (i.e., espin^+/− heterozygotes) do not produce any observable defects on their own (i.e., these flies look normal/“wild-type”). However, in contrast to that, we found that even small decreases in espin levels in vivo significantly increased Mical’s ability to disassemble bundled F-actin and remodel cells (Fig. 8, F and G). We also observed similar effects when we decreased fascin levels in vivo—including that small decreases in fascin levels, which did not produce any observable defects on their own, significantly increased Mical’s disassembly/remodeling effects in vivo (Fig. 8, F and H). So too, similar to what we observed with purified proteins, changing the levels of fascin alone, compared with changing the levels of espin alone, more strongly altered Mical’s disassembly/remodeling effects (Fig. 8 F). We found also that simultaneously lowering the levels of both espin and fascin, which did not produce any observable defects on their own, increased even more Mical’s effects on bundled F-actin and cellular remodeling (Fig. 8, F–I). Thus, Mical works better to disassemble bundled F-actin in vivo when espin and fascin levels are both decreased. Likewise, fascin and espin help protect F-actin from this disassembly more effectively together than alone. Therefore, as we saw with purified proteins, espin and fascin combine in vivo to more robustly protect F-actin against its Mical-mediated disassembly and remodeling.

Roles of actin bundles, including why they are bundled often simultaneously and seemingly redundantly, by different proteins, is incompletely understood. It is unknown also how these complex, multi-protein–bundled, stable structures are disassembled to allow for their remodeling per functional needs. We find here that two of the most common bundlers, fascin and espin, synergize to form actin bundles—forming wider, denser, and longer bundles than through the sum of their individual effects. We also find that these complex, multi-protein—bundled F-actin structures are effectively destabilized by a synergism of Mical with cofilin (Fig. 9). Our results also reveal that together fascin and espin help protect F-actin from this disassembly more effectively than each one of them alone (Fig. 9). This interplay between multiple bundling proteins and F-actin disassemblers is critical for cellular remodeling in vivo (Fig. 9).

We speculate that the mechanisms underlying espin and fascin’s synergy result from an intersection of their individual properties. Our results support previous observations that fascin and espin (splice forms 3A and 3B) simultaneously bind to F-actin in overlapping patterns (Chanut-Delalande et al., 2006; Chou et al., 2011; Otani et al., 2016; Roy and Perrin, 2018; Vignjevic et al., 2006; Winkelman et al., 2016). Our results also support previous observations that fascin and espin (3A) bundle F-actin in a similarly effective manner (Lieleg et al., 2009), including that they both induce F-actin twisting (∼1° per actin monomer), which may underlie their close coexistence in bundles (Claessens et al., 2008; Gong et al., 2025; Purdy et al., 2007; Shin et al., 2009). Yet, differences also exist between these bundlers—including that espin induces stiffer and more twisted bundles, while fascin produces more flexible bundles (Claessens et al., 2008; Purdy et al., 2007; Shin et al., 2009). Fascin is also more effective at bundling filaments that are pre-aligned (Courson and Rock, 2010; Sherer and Courtemanche, 2022; Tilney et al., 1998; Winkelman et al., 2016; Yang et al., 2013). Thus, we propose that the synergism we observe occurs because espin helps to gather, tether, and rigidly bundle actin filaments and thereby align them to facilitate enhanced fascin-mediated bundling.

Our synergistic results are also supported by several observations: (1) espin is highly expressed prior to fascin in cells that generate espin–fascin bundles (Chanut-Delalande et al., 2006; Guild et al., 2003; Guild et al., 2005; Roy and Perrin, 2018; Tilney et al., 2000; Wulfkuhle et al., 1998), (2) espin and fascin generate together larger/more expanded F-actin bundles in vitro and in vivo (Chou et al., 2011; Claessens et al., 2008; Krishnan et al., 2021; Tilney and DeRosier, 2005; Velez-Ortega and Frolenkov, 2019; Winkelman et al., 2016), and (3) espin and fascin mutants generally phenocopy each other (Dickinson and Thatcher, 1997; Guild et al., 2005; Okenve-Ramos and Llimargas, 2014; Tilney and DeRosier, 2005; Velez-Ortega and Frolenkov, 2019), but cannot substitute for one another (Cant and Cooley, 1996). This indicates that they require each other in vivo.

Our results provide also important insights into disassembly of F-actin structures that are bundled together with multiple bundlers. Bundled F-actin, including that bundled with fascin or espin, is notably resistant to disassembly (Blanchoin et al., 2014; Breitsprecher et al., 2011; Chikireddy et al., 2024; Elkhatib et al., 2014; Huang et al., 2005; Loomis et al., 2003; Michelot et al., 2007; Mizuno et al., 2018; Schmoller et al., 2011; Yang et al., 2022). We previously found that Mical, and Mical in combination with cofilin, rapidly and extensively disassembles fascin-bundled F-actin (Hung et al., 2010; Rajan et al., 2023c). We now find that both Mical and cofilin, and Mical in combination with cofilin, rapidly and extensively disassembles espin-bundled F-actin, as well as complex bundles containing both fascin and espin. These observations provide critical insights into the mechanisms of remodeling complex and highly stabilized cellular actin networks.

It is also noteworthy that espin and fascin have different expression patterns (Rajan et al., 2023a), timing of expression (as mentioned above), and subcellular locations (Krishnan et al., 2021; Tilney and DeRosier, 2005; Velez-Ortega and Frolenkov, 2019; Vignjevic et al., 2006). Thus, their spatiotemporal overlap may promote structural stability in specific types of cells, subcellular regions, and/or at different ages. For example, in some cells, fascin localizes along the length of filopodia, while espin is predominantly in the proximal parts of filopodia (Vignjevic et al., 2006). Thus, because of these regions of espin–fascin overlap, the proximal parts of these filopodia would likely be more protected from Mical–cofilin-driven disassembly than distal regions. Interestingly, semaphorin repulsive guidance cues employ Mical–cofilin to exert their F-actin disassembly effects (Grintsevich et al., 2016; Hung and Terman, 2011; Rajan et al., 2023b), and they have been found to more strongly disassemble distal versus proximal filopodial regions (Fan et al., 1993; Hung and Terman, 2011). Thus, our data suggest specific mechanisms that limit the effects of these critical repulsive guidance cues.

Our in vivo work within the bristle model system supports our in vitro results and indicates that a fine-tuned balance between the assembly and disassembly of actin bundles is essential for proper cell functions. Notably, pathologies including carcinomas and neurological abnormalities are linked to dysregulation of bundlers, including fascin and espin (Rajan et al., 2023a). So too, dysregulation of disassembly proteins, including Mical and cofilin, is linked to the same abnormalities (Rajan et al., 2023b). One notable example of such dysregulation is in the stereocilia of inner hair cells—the mechanosensory structures required to transduce sound into sensory information in vertebrates that share functional and structural similarities with Drosophila bristles (Muller and Littlewood-Evans, 2001). Fascin, espin, cofilin, as well as other proteins important for MICAL’s function, such as myosin Myo15 and the MsrB/SelR reductase that reverses MICAL’s effects (Rajan et al., 2023b), are required for normal hearing (Park and Bird, 2023). Thus, an interplay between these proteins is likely to not only affect Drosophila bristles but also multiple other cell types in mammals, including humans.

In conclusion, our results not only elaborate further the unique individual contributions of F-actin bundlers to bundle structure but also their ability to synergize with each other. Although in some cases bundling proteins can substitute for each other in vivo (Cant et al., 1998; Davidson et al., 2019; Revenu et al., 2012), our observations support results that each protein has its unique contributions to F-actin bundling (Bartles, 2000; DeRosier and Tilney, 2000; Park and Bird, 2023; Rajan et al., 2023a). Indeed, bundling proteins are often co-expressed, and their specific combinations—such as of fascin and α-actinin (Bashirzadeh et al., 2020; Claessens et al., 2008; Schmoller et al., 2008; Tseng et al., 2002; Tseng et al., 2005; Winkelman et al., 2016), fascin and plastin/fimbrin (Van Audenhove et al., 2016), α-actinin and filamin (Esue et al., 2009), and Eps8 and IRSp53 (Disanza et al., 2006)—provide new attributes to F-actin networks. Such combinations (including synergistic interactions) may be a general feature of bundling proteins, allowing them (despite their similar bundling function) to provide specific cells and subcellular regions with different biological properties—such as increased stability. Along these lines, it is also intriguing that larger espin splice forms (espin-1 and espin-2) form thicker actin bundles than espin-4 (the splice form used herein) and have higher F-actin affinity (∼3×) and stronger actin-bundling activity (∼2.5×) due to the presence of additional domains (Bartles et al., 1998; Belyantseva et al., 2025; Chen et al., 1999; Zheng et al., 2014). However, large espin splice forms are autoinhibited, thus limiting their abilities in the absence of activators (Liu et al., 2016; Zheng et al., 2014). Therefore, different splice forms of bundling proteins provide additional means to increase bundles stability.

We also identify specific molecules and mechanisms that allow for the disassembly and remodeling of these multi-protein–bundled actin structures. Thus, our results indicate that a fine-tuned mechanism is likely to exist in all cells to control: (1) the extent of F-actin cross-linking by multiple bundling proteins, which stabilizes bundles and decreases cellular remodeling, and (2) the extent of disassembly proteins activation, which drives decreased bundles stabilization/increased cellular remodeling. It seems likely that specific conditions may alter this balance and lead to disorders/diseases that could selectively affect different tissues. Thus, this balance should be further investigated—including the roles of other bundling proteins and their combinations (including different splice forms) in bundles topology and stability and their interplay with Mical and cofilin disassembly proteins.

Drosophila fascin and Drosophila Mical (Mical^RedoxCH) DNA constructs described by Hung et al. (2010); Rajan et al. (2023c) were used. Human cofilin-1 plasmid was a kind gift from Dr. E.M. De La Cruz (Yale University School of Medicine, New Haven, CT, USA). These particular constructs were used to compare this study with our previous results using these constructs (Rajan et al., 2023c). A protein expression plasmid containing the espin splice variant espin-4 (small espin), which has the domains needed to rescue the in vivo defects that result from a loss of Drosophila espin (forked) (Grieshaber et al., 2001), was kindly provided by Dr. J.R. Bartles (Northwestern University Feinberg School of Medicine, Chicago, IL, USA) (Bartles et al., 1998).

Actin was either purified from rabbit skeletal muscle acetone powder (Pel-Freez) as described by Spudich and Watt (Spudich and Watt, 1971) or obtained from Cytoskeleton, Inc. In brief, lyophilized acetone powder was resuspended in a Pre-G-buffer (5 mM Tris-HCl, pH 8.0, 0.2 mM CaCl₂ supplemented with 0.2 mM ATP, 1 mM DTT, and 0.4 mM PMSF), then cleared by centrifugation for 1 h at 13,000 rpm in a J2-HS rotor, 4°C. The supernatant was filtered through a 0.2-µm syringe filter, and then actin was polymerized by the addition of 2 mM MgCl₂ and 50 mM KCl for 2 h while on slow stirring at 4°C. Next, more solid KCl was added to a final concentration of 0.6 M, and stirring was continued for another 30 min at 4°C. F-actin was pelleted by centrifugation in a Ti-60 rotor at 48,000 rpm at 4°C for 100 min. The pellet was homogenized by a Dounce homogenizer and then dialyzed versus G-buffer (2 mM Tris-HCl, pH 8, 0.2 mM CaCl₂, 0.5 mM DTT, and 0.2 mM ATP) at 4°C for 3 days with buffer changes at 24-h intervals. After 4 days, actin was subjected to centrifugation (48,000 rpm, Ti-60 rotor), 4°C for 100 min, and the supernatant was loaded on a size exclusion Superdex S200 16/60 column (Amersham Biosciences) equilibrated in S200 G-buffer (2 mM Tris-HCl, pH 8, 0.2 mM CaCl₂, 0.5 mM DTT, 0.2 mM ATP, 0.25 mM TCEP, 0.2 mM PMSF, and 0.02% NaN₃). Collected peak fractions were stored at 4°C. Actin was used within 1 mo of its purification.

Mox-actin was prepared and purified as described (Grintsevich et al., 2016; Rajan et al., 2023c). In particular, 20 µM of actin was polymerized (F-actin) by incubating G-actin in 1× polymerization buffer (diluted from a stock 10× PB buffer: 100 mM HEPES, pH 7.4, 500 mM KCl, 10 mM MgCl₂, 10 mM EGTA, supplemented with 2 mM ATP and 10 mM DTT) for 1 h at RT. F-actin was then diluted to 2 μM in 1× PB buffer supplemented with NADPH (0.04 μM) and Mical 0.04 μM (50:1 molar ratio of actin to Mical) and incubated for 1 h at RT. After an hour, the solution was centrifuged at 321,530 × g for 20 min at 4°C, and the supernatant (containing Mox-actin) was mixed for 30 min with 250 μl of HisPur Ni-NTA Resin (Thermo Fisher Scientific) equilibrated with G-buffer to remove the residual Mical enzyme. Then the flow-through was collected and dialyzed overnight against S200 G-buffer. Before using Mox-actin in our experiments, its oxidation was confirmed by a subtilisin digestion assay as described (Grintsevich et al., 2016; Rajan et al., 2023c). In particular, G-actin and Mox-actin were digested with subtilisin (we used 1:1,000 [wt/wt] ratio of subtilisin:actin) in a G-buffer for 15 min. Subtilisin cleaves G-actin between residues 47 and 48 (Schwyter et al., 1989), while Mox-actin remains uncleaved, indicating its modification by Mical at position 47.

Drosophila Fascin was purified as previously described (Hung et al., 2010). In particular, fascin was expressed in ArcticExpress (DE3) cells (Agilent Technologies). Cells were grown at 30°C until OD₆₀₀ = 0.6–0.8, followed by an induction with 0.5 mM IPTG for 24 h at 14°C. The protein was purified by passing the lysate through a HisTrapFF affinity column (Cytiva) followed by its elution using 250 mM imidazole buffer. The elutant was digested overnight at 4°C with SUMO proteinase (in house purified [Hung et al., 2010; Rajan et al., 2023c]) and further purified using a HisTrapFF affinity column (Cytiva) to remove the His-SUMO protease. The fractions containing fascin protein were further passed through a HiTrap Q column (Cytiva) equilibrated with buffer (10 mM Tris-HCl, pH 8.0, 1 M NaCl, 5% glycerol, and 1 mM DTT). Under these conditions, fascin does not bind to a HiTrap Q column and was collected from the flow-through and dialyzed overnight in fascin storage buffer (10 mM Tris-HCl, pH 8.0, 50 mM NaCl, 5% glycerol, and 1 mM DTT).

Drosophila Mical (Mical^RedoxCH) was purified as previously described (Hung et al., 2010). Mical^RedoxCH was expressed in ArcticExpress (DE3) cells (Agilent Technologies). Cells were grown at 30°C until OD₆₀₀ = 0.6–0.8, followed by an induction with 0.2 mM IPTG, for 24 h, at 14°C. The protein was purified by incubating the lysate with HisPur Ni-NTA Resin column (Thermo), followed by elution with 250 mM imidazole buffer, and digestion with thrombin (Cat. No. 69671-3; Millipore). The digested samples were then bound again to the HisTrap HP column (Cytiva) to separate Mical from the Nus tag, and then passed through HiTrap Q FF column (Cytiva) equilibrated with Q buffer (10 mM Tris-HCl pH 8.0, 10 mM NaCl, 5% glycerol, and 2 mM DTT). Mical was eluted then with a linear gradient of NaCl (0–1 M) in 5 column volumes from the HiTrap Q column. The fractions containing the Mical protein were combined and overnight dialyzed versus Mical storage buffer (20 mM Tris-HCl, pH 8.0, 50 mM NaCl, 5% glycerol, and 2 mM DTT) at 4°C.

Human cofilin-1 (referred to as cofilin in this study) were purified as previously described (Grintsevich et al., 2016; Rajan et al., 2023c). In particular, cofilin was expressed in BL21(DE3)pLysS cells (Agilent Technologies). Cells were grown at 37°C until OD₆₀₀ = 0.6–0.8, followed by an induction with 1 mM IPTG for 4 h at 37°C. Clear cell lysate was loaded onto a DE52 (DEAE cellulose, Sigma-Aldrich) column—sequentially connected to a HiTrap SP FF column (GE) at 4°C. Both columns were equilibrated with 10 mM MOPS, pH 7, supplemented with 1 mM DTT and 0.2 mM PMSF. Cofilin was eluted from the SP FF column with a linear gradient of NaCl (0–700 mM in five column volumes). Cofilin containing fractions were combined and further purified using a size exclusion HiLoad 16/60 Superdex 75 column (Amersham Biosciences) equilibrated with 10 mM MOPS, pH 7, 25 mM NaCl, and 1 mM DTT.

Espin-4 was purified by following the instructions developed by J.R. Bartles (who originally identified espins) (Bartles et al., 1998) and as used by others for this espin splice form (e.g., Claessens et al., 2008; Kitajiri et al., 2010; Purdy et al., 2007). Briefly, espin-4 was expressed in Escherichia coli BL21-Codon Plus (DE3)-RIL (Agilent). Cells were grown at 37°C until OD₆₀₀ = 0.6–0.8, followed by an induction with 0.5 mM IPTG, for 8 h, at 25°C. Bacterial pellets were collected by centrifugation and resuspended in an extraction buffer (50 mM Tris-HCl, pH 8.5, 10 mM betamercaptoethanol [βME]) with PMSF and protease inhibitors. The protein was purified by incubating the lysate with HisPur Ni-NTA Resin column (Thermo Fisher Scientific) washed with wash buffer 1 (20 mM Tris-HCl, pH 8.5, 100 mM KCl, 10% glycerol, 20 mM imidazole, and 10 mM βME) and wash buffer 2 (20 mM Tris-HCl, pH 8.5, 1.5 M KCl, 10% glycerol, 20 mM imidazole, and 10 mM βME). The bound recombinant proteins were eluted from the column by a 25–250 mM imidazole gradient. The purified espin was cleaved by 6× His-tagged Tobacco Etch Virus protease (Life Technologies) for 8 h, at 4°C, and the uncleaved protein and protease were removed by passing the solution through a HisPur Ni-NTA Resin column (Thermo Fisher Scientific). Cleaved espin, in the flow-through, was collected and dialyzed overnight into espin storage buffer (10 mM Tris-HCl, pH 7.5, 100 mM KCl, 0.02% NaN₃, 1 mM DTT, and 10% glycerol). The protein was concentrated using a filter concentrator (Amicon filter, Millipore). Similar results with the espin protein purified in this manner were obtained independently in both the Terman and Reisler labs. We also noted, using standard western analysis approaches with a His antibody (1:1,000; #70796; Novagen) (Hung et al., 2010), that espin was susceptible to degradation (Fig. S1 A (2)). Fresh batches of espin protein were employed to overcome this problem.

All purified proteins except actin were stored at −80°C for futher use.

Freshly gel-filtered actin was labeled with Alexa488 succinimidyl ester (SE) (Cat. No. A20000; Life Technologies) as described (Rajan et al., 2023c). In particular, 2 mg of G-actin was polymerized for 1 h, at RT, by adding 2 mM MgCl₂ and 50 mM KCl. The F-actin (in pellet) was collected after centrifugation at 60,000 rpm (TLA110 rotor) for 30 min at 4°C and dialysized against labeling buffer (50 mM PIPES, pH 6.8, 50 mM KCl, 0.2 mM CaCl₂, and 0.2 mM ATP) for 3 h at 4°C. After dialysis, actin was incubated with a threefold molar excess of Alexa488SE dye, overnight at 4°C. The labeling reaction was stopped using 1 mM DTT, and the resulting fluorescently labeled F-actin was pelleted (TLA100 rotor, 90,000 rpm, 20 min, 4°C), depolymerized in G-buffer, and gel-filtered on a S200 10/300 Gl equilibrated with S200 G-buffer.

Fascin and espin were labeled on surface cysteines using Cy3-Monomaleimide and Alexa488-Monomaleimide (Invitrogen), which were resuspended in dimethylformamide (DMF) to a concentration of 1 mM. Purified proteins were first dialyzed overnight versus the labeling buffer (20 mM Tris-HCl, pH 7.9, 100 mM NaCl, 0.2 mM EDTA, 0.01% NaN₃, and 10% glycerol) to remove DTT. A 3 to 5-molar excess of dye was added then to the protein and allowed to label it overnight at 4°C. The labeling reaction was quenched using 1 mM DTT, and the excess dye was removed by three serial dialysis versus 500 ml of the protein storage buffer (labeling buffer + 1 mM DTT) for 2.5 h each.

For sedimentation assays, samples were centrifuged for 20 min either at 10,000 × g (low-speed) or 321,530 × g (high-speed). The supernatants and pellets were separated and analyzed on SDS-PAGE. The gels were stained with Coomassie blue and quantified with ImageJ (RRID:SCR_003070) (Schindelin et al., 2012).

10 µM of Mg-ATP F-actin or Mox-F-actin were prepared by first incubating Ca-ATP-G-actin for 3 min in a polymerization ME exchange buffer (0.05 mM MgCl₂ and 0.2 mM EGTA) to exchange bound Ca²⁺ in G-actin with Mg²⁺. They were then incubated in a 1× polymerization buffer (diluted from a stock 10× PB buffer: 500 mM KCl, 10 mM MgCl₂, 10 mM EGTA, 100 mM HEPES, pH 7.4, supplemented with 2 mM ATP and 10 mM DTT) for 30 min, at 25°C. Actin filaments were then diluted in 1× PB buffer in the absence or presence of bundling proteins and then incubated at RT for 30 min. The formed F-actin or bundles were incubated with Mical (without NADPH) or cofilin (for binding or bundling assays) or Mical/NADPH (for disassembly assays) for 30 min at RT before centrifugation.

To determine the decrease in light scattering, the data were normalized to a maximum intensity of 1, from which the relative decrease in light-scattering intensity was determined over 15 min for Figs. 4 B (2) and Fig. 5 A, and for 1 min for Fig. 6 A (2).

Unlabeled and pyrene-labeled actin were mixed with 10× PB buffer to yield a 10 µM, (10% labeled) actin stock in 1× PB buffer. Before polymerization, Ca²⁺ in G-actin was exchanged with Mg²⁺ by incubating actin with 0.2 mM EGTA and 0.1 mM MgCl₂ for 3 min at RT. The polymerized actin was diluted to 5 µM or 2.5 µM with buffer or buffer with bundling proteins for 1 h. Next, samples of F-actin or bundled-actin were incubated with a buffer containing NADPH, Mical, and/or cofilin. The pyrene fluorescence intensity of actin was monitored in a 96-well fluorescence plate reader (M1000; Tecan Infinite) at 407 nm (excitation at 365 nm). The time delay between mixing of the components and the start of fluorescence data collection was 15–20 s.

Time-lapse TIRFM was performed according to established protocols (Rajan et al., 2023c; Rajan et al., 2024). In short, 12 µl glass flow cells were prepared using GOPTS-mPEG [(3-Glycidyloxypropyl)trimethoxysilane–methoxy(polyethylene glycol)]-coated coverslips. In paricular, the coverslips were tandem-treated first with GOPTS (Cat No. 440167; Sigma-Aldrich) and subsequently with mPEG amine (Cat No. A3085; JenKem Technology USA). Before that, the flow cells were washed with a blocking buffer (1% Pluronic F127 solution [P2443; Sigma-Aldrich] + 0.1 mg/ml casein) and equilibrated with the TIRF buffer (10 mM HEPES, pH 7.4, 50 mM KCl, 2 mM MgCl₂, and 0.2 mM EGTA), supplemented with 50 mM DTT, 0.2 mM ATP, 20 mM glucose, and 1% methylcellulose. The on-slide actin polymerization was initiated by introducing into flow chambers 1.0 µM and 20% Alexa488SE actin sample (50 μl) in 1× TIRF imaging buffer (TIRF buffer supplemented with 0.05 mg/ml casein, 0.25 mg/ml glucose oxidase, and 0.05 mg/ml catalase to minimize photobleaching during the imaging [1× TIRF-IB]). After polymerizing actin for 15 min, bundling protein(s) (at final concentrations indicated in the figure legends) were added into the flow cells. Actin bundles were allowed to form on the slides’ surface for 5 min. 5–10 images of the bundles were taken per each condition (as shown in Fig. 2 C).

For Mical/NADPH and/or cofilin-mediated disassembly assays,the on-slide actin polymerization was initiated by introducing 1.0 µM, 20% Alexa488SE actin sample (50 μl) in 1× TIRF-IB into the flow chambers along with different combinations of bundling proteins (as indicated in Fig. 4 C, Fig. 5 B, and Fig. 6 B). After that, 2CV of Mical/NADPH mixtures and/or cofilin (in 1× TIRF-IB) were added into flow chambers, and the reactions were allowed to proceed for 1–2 min before recording the time-lapse TIRFM movies. The movies were recorded using a 3i Vector TIRF System mounted on a Zeiss Axio Observer 7 Basic Marianas Microscope, with a Definite Focus 2 and α Plan-Apochromat 63×/1.46NA Oil TIRF objective M27 and an Andor iXon3 897 512 × 512 10 MHz EMCCD Camera, using Slidebook 6 software (RRID:SCR_01430; Intelligent Imaging Innovations). The movies were recorded for 5–10 min, with a 2.5 s frame interval at RT.

All TIRFM data were analyzed using Fiji (ImageJ) software (RRID:SCR_002285; NIH) (Schindelin et al., 2012)—the same as described by Rajan et al. (2023c). In particular, before each analysis, background subtraction was done using rolling ball radius algorithm (ball radius 50 pixels). Bundled actin filaments were randomly chosen from the field of view for analysis, but F-actin bundles present at the edges were excluded from the analysis. Unbundled actin filaments—identified based on their fluorescence intensity—were excluded from the analysis. The length of actin filaments was measured manually by using JFilament plugin in Fiji (JFilament 2D) (RRID:SCR_002285). Fluorescence intensities of bundles were analyzed using the line-scan tool of ImageJ software (RRID:SCR_003070). Average intensity values of internal control unbundled filaments (from the same experiment) were corrected for photobleaching and used to calculate the number of filaments per bundle.

To determine the rates of bundles thinning and shortening during disassembly assays, the photobleaching correction of the movies was performed first. Internal control filaments’ fluorescence intensity from the same experiment was monitored over time to obtain a linear photobleaching curve that was utilized for a photobleaching correction. In the presence of disassembly proteins (Mical/NADPH, cofilin, and Mical/NADPH + cofilin), bundles thinning can occur at any point along their length, either at the ends (“peeling mode” as described by Rajan et al. [2023c]) or between them (i.e., thinning by removal of a small filament segment anywhere in the bundle while the remaining bundle’s width stays the same). To avoid bias and to capture both scenarios, we analyzed entire actin bundles, rather than focusing on specific sites. To determine bundles thinning rates (Fig. 4 C, Fig. 5 B, Fig. 6 B, and Fig. S4, C and E), the entire bundles were selected from the observation area with the region of interest (ROI) manager in Fiji, and their fluorescence intensity values were measured over the entire time of data acquisition using the Analyze tool. We calculated the thinning rates of bundles at two different times: the initial thinning rate (over the first 120 s of imaging) and the overall thinning rate (across the entire data acquisition time). We reasoned that each time frame offers distinct insights. The initial rate reflects how rapidly bundles disassembly begins following the addition of disassembly proteins. It is not affected by potential reassociation of severed fragments later in the reaction. In contrast to that, the overall average rate captures the progression of the entire disassembly process, from start to finish. To determine the initial rates of bundles thinning, data for the first 120 s of imaging were fitted by linear regression (Fig. 4 C (3), Fig. 5 B (3), Fig. 6 B (3), and Fig. S4 E). For total rates determination over the entire time of data acquisition (Fig. S3 B and Fig. S4 D), the intensity data were first normalized and then fitted to the exponential decay model with two parameters, using SigmaPlot 14 (RRID:SCR_003210). Small filaments transiently interacting with the surface—before their rapid disappearance—were not included in the analyses.

The shortening rates of bundles were determined over the first 60 s of imaging after addition of the indicated proteins (Fig. 4 C (2), Fig. 5 B (2), Fig. 6 B (3), and Fig. S4, C and E), with their length measured by using the JFilament plugin in Fiji (JFilament 2D) (RRID:SCR_002285). As above, bundles were identified based on their intensity, and their shrinkage was monitored from the point where single filaments extensions at the bundles end (if present) have depolymerized. When observed, paused bundles were excluded from the analysis. The decrease in length due to severing was also not included. Notably, this shrinkage analysis does not account for the thinning mode of disassembly.

Severing events were marked by bundles breaking into two or more distinct fragments. Bundles severing occurs at the sites of drastic thinning in actin bundles due to the action of the added disassembly proteins (in this case Mical/NADPH, cofilin, or Mical/NADPH + cofilin). Severing rates (breaks/micron/minute) in Fig. 4 C (3), Fig. 5 B (3), Fig. 6 B (3), and Fig. S4, C and E were calculated by measuring the initial length of bundles at t = 60 s (first frame of the video, i.e., the necessary delay between the addition of disassembly proteins and the start of video acquisition), and then manually identifying and counting the severing events frame by frame, starting from the frame t = 62.5 s in the time-lapse images after the flow-in of the indicated proteins and for the next 60 s. The measurements were performed only on 5 µm or longer segments of bundles because shorter fragments (<5 µm) tended to diffuse away—so as in our previous publication (Rajan et al., 2023c), they were thus excluded from the analysis. In the absence of disassembly proteins, bundles length and thickness remain constant during the 60 s in which severing was monitored.

To calculate the percentage of remaining bundles as shown in Fig. 6B (2) (comparing the effects of Mical/NADPH + cofilin on the three different types of bundles [fascin alone, espin alone, or fascin + espin]), we sought to find a quantitative measure that best highlighted the observed differences that Mical/NADPH + cofilin had on different bundle types. For these experiments, we randomly selected 15–20 bundles per bundle type and monitored their length for 150 s after the start of imaging (at intervals of 30 s) using the ROI Manager in Fiji. Bundles or bundles fragments that were ≥7 µm were considered, while the rest were omitted from the final count of the bundles present in the field of view. In particular, since severing of bundles leads to an increase in the total number of fragments, we focused solely on the longer fragments (≥7 µm) to simplify the quantification process. Bundles present on the edge of the field of view were excluded from our analysis. The number of bundles at t = 60 s (first frame of the video) was normalized to 100% (individually for each video, n = 3 experiments/conditions). The percentage of bundles remaining was quantified using equation: Fraction bundled (%) = (number of bundles at specific time/total number of bundles at t = 60 s) × 100. Time-dependent distribution of the percentage of remaining actin bundles was fitted by linear regression to obtain the rates of bundles disappearance.

For three-color TIRFM analysis (Fig. S1 D), elongation rates of individual actin bundles were determined by monitoring over many time points a ROI of individual bundles in the actin-phalloidin 647 channel. The actin ROIs were then applied to the fascin (555 nm) and espin (488 nm) channels. A kymograph of all three channels was created from the ROIs using Multi Kymograph - ImageJ plugin (RRID:SCR_003070).

10 μM G-actin was polymerized for 1 h using the PB buffer. The polymerized actin was incubated with bundling proteins for 1 h at RT to form bundles. These samples were diluted then to 1 µM and applied to glow-discharged, carbon-coated EM grids. After 60 s of adsorption, the grids were blotted dry and treated with 2% uranyl acetate for 45 s at RT. The grids were examined in a FEI Technai T12 electron microscope operated at 120 kV. Actin bundles images were taken at a 26,000× magnification, defocus from −3 to −8 with a Gatan Ultrascan1000 4 MP CCD (2 by 2 k) camera, and Gatan digital micrograph software (RRID:SCR_014492) was used for image acquisition. Actin bundles were selected in a randomized manner across multiple grids squares that were identically fixed and stained, with no bias toward a specific morphology and size. The collected images were analyzed using the Gatan digital micrograph software (RRID:SCR_014492) and ImageJ software (RRID:SCR_003070). Even though we imaged more than five bundles per condition (when using high molar ratio of bundling proteins to actin), not all bundles were suitable for quantitative analysis due to overlapping filaments or disassembled filaments on the top of the bundles. Therefore, the only bundles regions in bundles where individual filaments could be seen to make reliable measurements were considered.

We also used our EM and TIRF assays to look for a correlation between the increase in bundle length and width. In particular, for exact correlation determination of these parameters, the length and width of the same actin bundles is required. However, as mentioned, we found that due to fluorescence intensity saturation (after five actin filaments; Fig. S2 E), the thickness of bundles could not be measured using TIRFM assays (which were used to determine the length of the bundles). Furthermore, EM analysis does not allow us to measure over time the length of the filaments/bundles. Thus, to see if our results at least point to the idea that these two parameters are correlated, we plotted our results using TIRFM assays to determine the length of the bundles and EM assays to determine the width of the bundles and presented our results in Fig. S2 G.

To determine whether espin alone or espin-bundled F-actin activates Mical’s catalytic activity, we looked at Mical’s utilization/consumption of its co-enzyme. Unbundled actin filaments activate the catalytic activity of Mical (Hung et al., 2011). Fascin-bundled actin filaments also activate Mical’s enzyme activity, but at a lower rate and extent than unbundled filaments (Rajan et al., 2023c). Therefore, we wondered if actin filaments bundled with espin or fascin + espin could also activate Mical—or if espin’s or fascin + espin bundling of F-actin might prevent/block Mical’s active site binding to or activation by F-actin. NADPH consumption (conversion of NADPH to NADP⁺) by Mical was measured as described by Grintsevich et al. (2016); Hung et al. (2011); Rajan et al. (2023c), with a decrease in the reduced form of NADPH determined from the decreased light absorption at 340 nm. In particular, espin, fascin + espin, or F-actin preincubated with espin or espin + fascin in 1× PB buffer was added to a polymerization buffer containing NADPH and supplemented with Mical. The change in NADPH consumption was calculated by subtracting its consumption in the absence of Mical from that in its presence.

Fly stocks, including Mical, were the same as previously reported (Hung et al., 2010; Rajan et al., 2023c), except for stocks of espin, fascin, and espin and fascin double mutants, which were obtained from the Bloomington Drosophila Stock Center, Bloomington, Indiana, USA, or were a kind gift of U. Abdu (Ben-Gurion University of the Negev, Be'er Sheva, Israel). The B11-GAL4 bristle driver was employed for bristle expression, as previously done (Hung et al., 2010; Rajan et al., 2023c). Multiple espin (forked [f]) and fascin (singed [sn]) alleles and their combinations were examined including f^36a (which is a well-characterized null allele; RRID:BDSC_43 [Petersen et al., 1994]) and sn³ (a well-characterized hypomorphic allele; RRID:BDSC_113 [Cant et al., 1994; Paterson and O'Hare, 1991]).

Fascin and espin are associated with common F-actin structures, such as filopodia, microvilli, stereocilia, and Drosophila bristles (Bartles, 2000; Bartles et al., 1998; Chou et al., 2011; DeRosier and Tilney, 2000; Grieshaber and Petersen, 1999; Grieshaber et al., 2001; Perrin et al., 2013; Petersen et al., 1994; Rajan et al., 2023a; Shin et al., 2010; Shin et al., 2013; Tilney et al., 2004). Drosophila bristles, in particular, have become a model for studying these complex F-actin bundles in vivo (Tilney et al., 1995; Tilney et al., 1998; Tilney et al., 2000; Tilney and DeRosier, 2005; Wulfkuhle et al., 1998). We therefore used Drosophila bristles as a model for studying these complex F-actin bundles in vivo and the effects of Mical on them. Analysis of bristle cell remodeling, genetic interactions, and their imaging, drawings, quantification, and statistics were done as described (Hung et al., 2010; Rajan et al., 2023c). In particular, flies were crossed at 25°C and newly emerging adult offspring were collected, genotyped, and analyzed under a dissecting microscope (Leica Stereo Zoom S8 APO). Bristles were then examined for defects in morphology. The number of branches on each scutellar bristle was counted, and the results were presented as the mean number of branches per bristle (± SEM). Data were plotted and statistically analyzed using GraphPad Prism software (RRID:SCR_002798). Adult bristle imaging and drawings were done with the aid of a Zeiss Discovery M2 Bio stereomicroscope using a Zeiss 1.5× PlanApo S objective at RT (22°C) with a motorized focus and zoom, a Zeiss Axiocam HRc camera, three-dimensional reconstruction software (Zeiss Axiovision [RRID:SCR_002677] and Extended Focus [a kind gift from B. Lee (via Zeiss, Oberkochen, Germany)]), and Microsoft Office Powerpoint software (RRID:SCR_023631).

Imaging of Drosophila pupae for bundled F-actin was also done as described (Hung et al., 2010; Rajan et al., 2023c). In particular, white prepupae from each genotype were identified with the aid of Tb balancers and were marked to collect. After 44 h of prepupal development, the pupae were collected onto double-sided tape placed in a Petri dish. The pupae were then carefully removed from their pupal cases and positioned dorsal side down on glass-bottom Petri dishes (Part No. P35GC-1.5-14-C; MatTek). Imaging of notal bristles was performed at RT (22°C) using a Zeiss LSM880 confocal microscope with Zen 2.3 software (RRID:SCR_013672; Zeiss) equipped with a Zeiss 20 × Plan-Apochromat objective (0.8 numerical aperture) to visualize actin labeled with GFP. Then, representative images of bristles (in JPEG format) were analyzed using ImageJ software (RRID:SCR_003070). Fluorescent regions were isolated by adjusting the color threshold. The ROI in each bristle was manually selected using the polygon selection tool. Measurement parameters were configured to record area, integrated density, and mean gray value. Fluorescence intensity within each ROI was quantified using the “Measure” function, and the integrated density values were recorded for each genotype. This allowed us to determine the percentage of disrupted bundles for a given genotype as compared with the total bundle area. This also allowed us to determine the area in µm² of bundle disassembly for each analyzed genotype. Data were plotted and statistically analyzed using GraphPad Prism software (RRID:SCR_002798).

The following genotypes were used in Fig. 8: Fig. 8 A: normal (UAS:GFP-actin5C/+, B11-GAL4/+). Fig. 8 B: Bristle Mical⁺⁺⁺ (low) left image (UAS:GFP-actin5C/+, B11-GAL4/+, UAS:Mical/+). Bristle Mical⁺⁺⁺ (low) right image (B11-GAL4/+, UAS:Mical/+). Fig. 8 C: espin (–/– mutant) left image (espin^CRISPRKO/espin^CRISPRKO, UAS:GFP-actin5C/+, B11-GAL4/+). espin (–/– mutant) right image (espin^CRISPRKO/espin^CRISPRKO).Fig. 8 D: Bristle Mical⁺⁺⁺+ espin^–/– top image (espin^CRISPRKO/espin^CRISPRKO, UAS:GFP-actin5C/+, B11-GAL4/+, UAS:Mical/+). Bristle Mical⁺⁺⁺+ espin^–/– Bottom Image (espin^CRISPRKO/espin^CRISPRKO, B11-GAL4/+, UAS:Mical/+). Fig. 8 E: Bristle Mical⁺⁺⁺ (UAS:GFP-actin5C/+, B11-GAL4/+, UAS:Mical/+), espin^–/– (espin^CRISPRKO/espin^CRISPRKO, UAS:GFP-actin5C/+, B11-GAL4/+), Mical⁺⁺⁺+ espin^–/– (espin^CRISPRKO/espin^CRISPRKO, UAS:GFP-actin5C/+, B11-GAL4/+, UAS:Mical/+). Fig. 8 F, top graph: Bristle Mical⁺⁺⁺ (UAS:GFP-actin5C/+, B11-GAL4/+, UAS:Mical/+), bristle Mical⁺⁺⁺+ decreased espin (+/−) (f^36a/+, UAS:GFP-actin5C/+, B11-GAL4/+, UAS:Mical/+), Bristle Mical⁺⁺⁺+ decreased fascin (+/−) (sn³/+, UAS:GFP-actin5C/+, B11-GAL4/+, UAS:Mical/+). Bristle Mical⁺⁺⁺ + decreased espin (+/−) and fascin (+/−) (f^36a, sn³/+, B11-GAL4/+, UAS:Mical/+). Fig. 8 F, bottom graph: Bristle Mical⁺⁺⁺ (B11-GAL4/+, UAS:Mical/+), Bristle Mical⁺⁺⁺+ decreased espin (+/−) (f^36a/+, B11-GAL4/+, UAS:Mical/+), Bristle Mical⁺⁺⁺+ decreased fascin (+/−) (sn³/+, B11-GAL4/+, UAS:Mical/+). Bristle Mical⁺⁺⁺ + decreased espin (+/−) and fascin (+/−) (f^36a, sn³/+, B11-GAL4/+, UAS:Mical/+). Fig. 8 G: Bristle Mical⁺⁺⁺+ decreased espin (+/−) left image (f^36a/+, UAS:GFP-actin5C/+, B11-GAL4/+, UAS:Mical/+), Bristle Mical⁺⁺⁺+ decreased espin (+/−) right image (f^36a/+, B11-GAL4/+, UAS:Mical/+). Fig. 8 H: Bristle Mical⁺⁺⁺+ decreased fascin (+/−) left image (sn³/+, UAS:GFP-actin5C/+, B11-GAL4/+, UAS:Mical/+). Bristle Mical⁺⁺⁺+ decreased fascin (+/−) right image (sn³/+, B11-GAL4/+, UAS:Mical/+). Fig. 8 I: Bristle Mical⁺⁺⁺ + decreased espin (+/−) and fascin (+/−) top image (f^36a/+, sn³/+, UAS:GFP-actin5C/+, B11-GAL4/+, UAS:Mical/+). Bristle Mical⁺⁺⁺ + decreased espin (+/−) and fascin (+/−) bottom image (f^36a/+, sn³/+, B11-GAL4/+, UAS:Mical/+).

For each representative protein purification, image, gel, immunoblot, graph, or in vivo experiment, the experiments were repeated at least two separate independent times, and there were no limitations in repeatability. No statistical method was used to predetermine the sample size, which was based on what is published in the field. Differences between experimental and control animal conditions were highly notable, with relatively little variability, and so the sample size was larger than needed to ensure adequate power analysis to detect an effect. Animal studies complied with the Institutional Animal Care and Use Committee guidelines. Both males and females of Drosophila pupae and adults were used. Animal studies were based on pre-established criteria to compare against age-matched animals. Animal experiments were not randomized. Animals of the correct genotype were determined, and those collected of that genotype were included as data. Likewise, for biochemical experiments, samples were grouped together based on experimental conditions and collected data points for those experiments are presented. For all genetic experiments, the genotype needed to be determined based on different genetic/chromosome markers, so blinding was not employed. Likewise, for the biochemical experiments, different proteins and reagents for the particular data set needed to be added and then analyzed using specific technical approaches and expertise, and so blinding was not employed. For both genetic and biochemical experiments, differences between the control and experimental conditions were highly notable and reproducible in both biological and technical replicates. Images and gels show representative examples of experimental results, including that brightness, contrast, sharpness, background, and/or color balance were adjusted uniformly across the whole image using Adobe Photoshop software (RRID:SCR_014199) or Microsoft Office Powerpoint software (RRID:SCR_023631). Graphs show mean ± SEM or the percentage of animals with that defect. For each graph, the value of n and what n represents are stated in the figure legend. For statistical analysis, the statistical test used and the P value for each comparison are stated in the figure legend. Statistical methods used were based on what is standard in the field, and no statistical tests were used to determine whether the data met assumptions of the statistical approach. All statistical analyses were performed using GraphPad Prism Version #10.5 (RRID:SCR_002798). A P value of P > 0.05 is not considered statistically significant. Single asterisk (*) indicates P < 0.05, double asterisks (**) indicate P < 0.01, triple asterisks (***) indicate P < 0.001, and quadruple asterisks (****) indicate P < 0.0001. To the best of our knowledge, the statistical tests are justified as appropriate.

Fig. S1 shows further analysis of fascin and espin interactions with F-actin. Fig. S2 shows further analysis of the effect of different combinations of bundling proteins on actin bundles formation and topology. Fig. S3 shows further analysis of the Mical’s disassembly effects on different combinations of bundling proteins in actin bundles. Fig. S4 shows further analysis of Mical and cofilin synergy in disassembling fascin, espin, and fascin–espin co-bundled F-actin—and fascin and espin’s combined protection against this disassembly. Fig. S5 shows further analysis of the biochemical mechanism underlying Mical and cofilin’s effects on bundled forms of F-actin. Video 1 shows Mical/NADPH-mediated disassembly of fascin-bundled, espin-bundled, and fascin + espin co-bundled actin filaments. Video 2 shows cofilin-mediated disassembly of fascin-bundled, espin-bundled, and fascin + espin co-bundled actin filaments. Video 3 shows Mical/NADPH and cofilin-mediated disassembly of fascin-bundled, espin-bundled, and fascin + espin co-bundled actin filaments.

The data are available in the published article and its online supplemental material and are also available from the corresponding authors upon request.

We thank I. Martini and the University of California, Los Angeles (UCLA)’s Molecular Instrumentation Center for fluorescence spectrophotometer usage; M. Quinlan (UCLA) for TIRFM usage; and U. Abdu (Ben-Gurion University of the Negev, Be’er Sheva, Israel), J.R. Bartles (Northwestern University Feinberg School of Medicine, Chicago, IL, USA), E.M. De La Cruz (Yale University School of Medicine, New Haven, CT, USA), B. Lee (via Zeiss, Oberkochen, Germany), and the Bloomington Stock Center for reagents.

This study was supported by National Institutes of Health (NIH) grants NS073968 and MH085923 (J.R. Terman) and GM077190 (E. Reisler).

Author contributions: Sudeepa Rajan: conceptualization, data curation, formal analysis, investigation, methodology, project administration, supervision, validation, visualization, and writing—original draft, review, and editing. Jimok Yoon: conceptualization, data curation, formal analysis, investigation, methodology, project administration, resources, supervision, validation, visualization, and writing—original draft, review, and editing. Heng Wu: Conceptualization, data curation, formal analysis, investigation, methodology, project administration, resources, validation, visualization, and writing—original draft, review, and editing. Raju Baskar: conceptualization, data curation, formal analysis, investigation, methodology, project administration, resources, validation, visualization, and writing—original draft, review, and editing. Roman Aguirre: formal analysis and investigation. Jonathan R. Terman: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, supervision, validation, visualization, and writing—original draft, review, and editing. Emil Reisler: conceptualization, data curation, formal analysis, funding acquisition, methodology, project administration, supervision, validation, visualization, and writing—original draft, review, and editing.