Redox and actin, a fascinating story

Abstract

Actin is an extraordinarily complex protein whose functions are essential to cell motility, division, contraction, signaling, transport, tissular structures, DNA repair, and many more cellular activities critical to life for both animals and plants. It is one of the most abundant and conserved proteins and it exists in either a soluble, globular (monomeric, G-actin) or an insoluble, self-assembled (polymerized or filamentous actin, F-actin) conformation as a key component of the cytoskeleton. In the early 1990's little, if anything, was known about the impact of reactive oxygen species (ROS) on the biology of actin except that ROS could disrupt the actin cytoskeleton. Instructively, G-actin is susceptible to alteration by ROS, and thus, purification of G-actin is typically performed in the presence of strong antioxidants (like dithiothreitol) to limit its oxidative degradation. In contrast, F-actin is a more stable conformation and thus actin can be kept relatively intact in purified preparations as filaments at low temperature for extended periods of time. Both G- and F-actin interact with a myriad of intracellular proteins and at least with a couple of extracellular proteins, and these interactions are essential to the many actin functions. This review will show how, over the past 30 years, our understanding of the role of ROS for actin biology has evolved from noxious denaturizing agents to remarkable regulators of the actin cytoskeleton in cells and consequent cellular functions.

Keywords: ADF/Cofilin; Actin; Atherosclerosis; Bubbles; Cancer; Cell motility; Cytoskeleton; EGF-Receptor; Hydrogen peroxide; Inflammation; Kaposi's sarcoma; Lamellipodium; MICAL; Membranes; NADPH-Oxidase; NOX-1; NOX-2; PDGF-Receptor; Phosphatases; Phosphatidylinositol 4,5 bisphosphate; Phospholipase C-y1; Profilin; Rac1; Ras; Reactive oxygen species; Regeneration; Superoxide; Tissue repair.

Fig. 1

Adapted from Varland et al. (Fig. 1) [10]. A. Structures of selected actin post translation modifications (PTMs): methylation, acetylation, arginylation, phosphorylation, methionine (met) oxidation, cysteine (cys) oxidation, S-nitrosylation, and S-glutathionylation (SG represents glutathione). B. Interaction sites for specific actin binding proteins (ABPs) are represented by the green (Gelsolin G1 & G3 interaction sites), burgundy (Cofilin), pink (DNase I), and blue (Profilin) boxes. Key actin residues that are modified by PTMs are indicated by red text (M44 and M47 by Met-oxidation; C374 by Cys-oxidation, including nitrosylation and glutathionylation); this demonstrates how residues in the aforementioned ABP binding sites are the subject of PTMs that have significant effects on actin's ability to interact with ABPs.

Fig. 2

Roles of profilin and ADF/cofilin in actin treadmilling. Adapted from Pinto-Costa & Sousa (Fig. 2) [25]. (a) Structure of human profilin1 and its ligand binding domains: for actin (blue), PLP (yellow), and for PIP2 (red), adapted from Jockusch, Murk, and Rothkegel [26]. (b) In vitro, with purified protein, in the presence of salts, ATP, and antioxidants, actin can treadmill. This is the consequence of the difference in actin critical concentration at both ends. At the barbed end, the critical concentration of actin (0.1 μM) is ten times lower than at the pointed end (1 μM). Hence, at actin concentrations between 0.1 and 1.0 μM, monomers can add at the barbed end and detach at the pointed end. In addition, treadmilling of actin functions as an ATP triphosphatase, for each mole of actin incorporated in filaments, 1 mole of phosphate is being released, first attached to filaments then detached from filaments. (c) In vivo, G-actin concentration is well above the critical concentration of actin at both ends and thus filaments should elongate at both ends, but treadmilling can still occur, synergistically powered by profilin and ADF/cofilin proteins. At barbed ends, profilin (circles) enhances polymerization. Apart from its actin monomer sequestering activity which inhibits spontaneous nucleation of actin [box 3] to ensure selective actin polymerization driven by Wasp, Scar, Arp2 and Arp3 [27], profilin accelerates enzymatically ADP/ATP nucleotide exchange on G-actin [box 1] including actin monomers that have been moderately oxidized (gold oval dots), producing a pool of polymerizable (oxidized or not) actin subunits bound to ATP (dark gray G-actin). The profilin-G-actin (bound to ATP) complex can readily add to fast-growing-barbed ends, and thus profilin can shuttle G-actin to an F-actin barbed end then detaches from it [box 2]. ADF/cofilin loads cooperatively onto microfilament sides enriched of adenosine diphosphate (ADP)-actin (light gray G-actin) leading to F-actin severing and disassembly at pointed ends, hence the actin treadmilling observed in cells.

Fig. 3

Actin treadmilling and the lamellipodia. (a) This first graph was copied from a video of Julie Theriot showing a fish keratinocyte in motion. Upon receiving extracellular signals like growth factors, many responsive cells polarize and develop a flattened expansion at one pole (lamellipodia), and a bulky pole at the opposite end containing the nucleus and other cellular organelles [31]. Such local flattening of the cell is mediated by criss-crossing actin filaments with anchors in both the ventral (attached to the substratum via focal adhesions) and dorsal plasma membrane, likely supported by myosin for contractility [3,30,31]. (b) Electron micrograph of a keratocyte provided by Tanya Svitkina and previously displayed by Pollard and Borisy [3], showing branched actin filaments at the very edge of the lamellipodia and longer actin filaments deeper within the lamellipodia. The diagram below shows the locations of key proteins [3]. The curves show regions of polymerization (red) and depolymerization (blue) of F-actin within the lamellipodia's advancing edge. (c) Such internal compression of the lamellipodia results in increased pressure at the advancing edge of the lamellipodia (considering Pierre Simon Laplace law and the cell plasma membrane as a complex bubble), PC/r where P is the pressure inside a bubble, C in cells is a constant, and r is the radius of the bubble [35]. Hence, the smaller the bubble the greater the pressure inside the bubble. This continuously advancing actin compression machinery results in pressurizing the plasma membrane forward while stabilizing each advance through forward treadmilling and branching of actin filaments. Equilibrium of pressure across the cell is maintained by the cytoskeleton (possibly involving the wrinkles) [31] and the myosin mediated contraction at the rear end of the cell which detaches from the substrate as the cell advances. When two balloons of identical initial size when deflated, are inflated with one less than the other, upon connecting the balloons with a straw, the transfer of air is from the smaller balloon to the larger balloon, in a way that illustrates the bubble experiments of Laplace. (d) Molecular activities at the edge of the lamellipodia that support the compression machinery (figure adapted from the article of Thomas Pollard and Gary Borisy) [3]. Reorganization of the actin cytoskeleton involves the metabolism of membrane phosphatidylinositols, followed by activation of Rac1 through interaction with guanine nucleotide exchange factor Tiam-1 which is bound to phosphatidylinositol 4,5 bisphosphate and accelerates the binding of Rac1 to GTP. In turn, GTP-Rac1 activates NADPH oxidase (NOX1) with production of superoxide and ROS including H₂O₂ (green gradient in Fig. 3d), which contributes to the activation of WASP/Scar proteins. WASP/Scar bring together the Arp2/3 complex that nucleates new filaments and to form new branches of actin filaments at the edge of the lamellipodia. Such branching of filaments are necessary to generate the compression machinery of the lamellipodia at the edge. Capping proteins can limit the length of filaments, and filaments age by hydrolysis of ATP bound to each actin subunit of filaments (light blue to dark blue) followed by dissociation of the phosphate to complete the ATPase cycle of actin (see Fig. 3). ADF/cofilin promotes phosphate dissociation, severs ADP-actin filaments and promotes dissociation of ADP-actin from filament pointed ends. MICAL oxidation of actin filaments can accelerate markedly the depolymerization of filaments produced by ADP/cofilin. Profilin catalyzes the exchange of ADP for ATP (turning the subunits light blue again), thus returning actin subunits to the pool of ATP-actin bound to profilin, ready to elongate barbed ends as they become available. For MICAL oxidized G-actin, an additional step of reduction catalyzed by methionine sulfoxide reductase B1 (MsrB1) is required for G-actin to join the pool of ATP-actin bound to profilin and ready for polymerization.

Fig. 3

Actin treadmilling and the lamellipodia. (a) This first graph was copied from a video of Julie Theriot showing a fish keratinocyte in motion. Upon receiving extracellular signals like growth factors, many responsive cells polarize and develop a flattened expansion at one pole (lamellipodia), and a bulky pole at the opposite end containing the nucleus and other cellular organelles [31]. Such local flattening of the cell is mediated by criss-crossing actin filaments with anchors in both the ventral (attached to the substratum via focal adhesions) and dorsal plasma membrane, likely supported by myosin for contractility [3,30,31]. (b) Electron micrograph of a keratocyte provided by Tanya Svitkina and previously displayed by Pollard and Borisy [3], showing branched actin filaments at the very edge of the lamellipodia and longer actin filaments deeper within the lamellipodia. The diagram below shows the locations of key proteins [3]. The curves show regions of polymerization (red) and depolymerization (blue) of F-actin within the lamellipodia's advancing edge. (c) Such internal compression of the lamellipodia results in increased pressure at the advancing edge of the lamellipodia (considering Pierre Simon Laplace law and the cell plasma membrane as a complex bubble), PC/r where P is the pressure inside a bubble, C in cells is a constant, and r is the radius of the bubble [35]. Hence, the smaller the bubble the greater the pressure inside the bubble. This continuously advancing actin compression machinery results in pressurizing the plasma membrane forward while stabilizing each advance through forward treadmilling and branching of actin filaments. Equilibrium of pressure across the cell is maintained by the cytoskeleton (possibly involving the wrinkles) [31] and the myosin mediated contraction at the rear end of the cell which detaches from the substrate as the cell advances. When two balloons of identical initial size when deflated, are inflated with one less than the other, upon connecting the balloons with a straw, the transfer of air is from the smaller balloon to the larger balloon, in a way that illustrates the bubble experiments of Laplace. (d) Molecular activities at the edge of the lamellipodia that support the compression machinery (figure adapted from the article of Thomas Pollard and Gary Borisy) [3]. Reorganization of the actin cytoskeleton involves the metabolism of membrane phosphatidylinositols, followed by activation of Rac1 through interaction with guanine nucleotide exchange factor Tiam-1 which is bound to phosphatidylinositol 4,5 bisphosphate and accelerates the binding of Rac1 to GTP. In turn, GTP-Rac1 activates NADPH oxidase (NOX1) with production of superoxide and ROS including H₂O₂ (green gradient in Fig. 3d), which contributes to the activation of WASP/Scar proteins. WASP/Scar bring together the Arp2/3 complex that nucleates new filaments and to form new branches of actin filaments at the edge of the lamellipodia. Such branching of filaments are necessary to generate the compression machinery of the lamellipodia at the edge. Capping proteins can limit the length of filaments, and filaments age by hydrolysis of ATP bound to each actin subunit of filaments (light blue to dark blue) followed by dissociation of the phosphate to complete the ATPase cycle of actin (see Fig. 3). ADF/cofilin promotes phosphate dissociation, severs ADP-actin filaments and promotes dissociation of ADP-actin from filament pointed ends. MICAL oxidation of actin filaments can accelerate markedly the depolymerization of filaments produced by ADP/cofilin. Profilin catalyzes the exchange of ADP for ATP (turning the subunits light blue again), thus returning actin subunits to the pool of ATP-actin bound to profilin, ready to elongate barbed ends as they become available. For MICAL oxidized G-actin, an additional step of reduction catalyzed by methionine sulfoxide reductase B1 (MsrB1) is required for G-actin to join the pool of ATP-actin bound to profilin and ready for polymerization.

Fig. 4

Nucleotide triphosphatase cycle of actin, and oxidation. G-actin is activated upon switching from the ADP-bound to the ATP-bound state. Actin binds ATP with a dissociation constant in the nanomolar range, well below the cellular ATP concentration. Cellular ATP concentration is always saturating, and thus physiological change in ATP concentrations doesn't regulate ATP binding to G-actin. Moreover, because ATP is always in large excess over ADP in cells, whenever ATP or ADP dissociates from G-actin it is nearly always replaced by an ATP. Spontaneous hydrolysis of G-actin bound ATP is usually slow, but ATP hydrolysis is 7000-folds faster when actin is polymerized (F-actin) than when it is monomeric (G-actin). ATP hydrolysis is the irreversible step in the actin cycle, forcing actin to pass through an ADP-bound intermediate state. ADP actin is more susceptible to oxidation, and oxidized ADP-G-actin is no longer polymerizable, whereas oxidized ATP-actin retains its ability to polymerize [24]. Hence, the impact of profilin on the replacement of ADP for ATP for oxidized G-actin is critical to physiological actin responses. The green and orange shades contain the conformations of actin that are less or more susceptible to oxidation, respectively. The blue oval shade indicates the conformations impacted by profilin.

Fig. 5

Dependence of KS tumorigenesis on Rac 1 and ROS. Adapted from Ma et al. [46] KS-like tumors on the tail (a and b) and nose (c) of mice expressing a constitutively active Rac1^V12 (RacCA^+/+). H&E staining of early (d) and late stage (e) neoplastic tissue of RacCA^+/+ mice showing pathologic features identical to that of human KS. The necessary presence of RacCA^+/+ for the development of KS is shown in (f). The dependence on ROS for KS generation in RacCA^+/+ mice is demonstrated by the impact of N-acetyl-l-cysteine (NAC) on tumor-free survival of homozygous male RacCA^+/+ mice (g).

Fig. 6

Increasing ROS production and tumor progression: the colorectal cancer example. Adapted from Fearon and Vogelstein [72]. As we have shown, oncogenes like HRas^V12 can induce ROS production which can activate signaling pathways that promote cell proliferation. Further elevation of ROS levels can induce DNA mutations, leading to genomic instability—a hallmark of cancer cells as they support the development of carcinoma and metastasis.

Fig. 7

ROS, cell motility, and wound healing. Adapted from Moldovan et al. [42] Confluent monolayers of endothelial cells (EC) were scratched with a glass tip to create a wound. The generation of intracellular ROS was monitored by preloading the cells with CM-DCF-DA that fluoresces upon exposure to oxidants, especially H₂O₂. Cells before scratching or cells distant from the scratched area displayed very low CM-DCF-DA fluorescence (a–c). Differential interference contrast (DIC) image and fluorescence image of the same field were overlaid (b). EC flanking the wound (wound margin or WM) were consistently and significantly more fluorescent than more distant cells (intact monolayer or INT) when observed at 1 and 5 h (c). Bar in lower right corner = 100 μm. Furthermore, migration of EC into the “empty” wound space was shown to be dependent on the presence of ROS (see reference 42 for videos of cell motility with and without ROS). Actin polymerization was accelerated in EC with increased production of ROS, specifically H₂O₂. Thus, when ROS production was inhibited by DPI or MnTMPyP, migration of EC was suppressed markedly and polymerization of actin measured by incorporation of fluorescent actin (A488A) into filaments was inhibited (d), possibly due to a limiting number of barbed ends for monomers to be added (e) [13,42]. Since WASP requires PIP2 for activation and because ROS increase PIP2 (increased synthesis, decreased activity of phosphatases) antioxidants might reduce available barbed ends by preventing WASP activation. Representation showing four zebrafish: Before (top) and after (next three) partial caudal fin amputation, right after amputation (second from top), then later after amputation, either in conditions that block H₂O₂ (cut and no ROS, third from top), or healing without inhibition of H₂O₂ (cut with ROS, fourth from top). Regeneration of the Zebrafish's caudal fin has been shown to be dependent on release of H₂O₂ along the wound margin to block phosphatases. When this initial burst of H₂O₂ is inhibited post wounding (cut and no ROS), regeneration of the caudal fin does not occur (f).

Fig. 8

Inflammation, a ROS-dependent process, can drive tissue repair or destruction. Adapted from Goldschmidt-Clermont et al. (Fig. 2) [95]. Atherosclerosis, an inflammatory process that leads to the obstruction of arterial conduits, remains the major cause of death and morbidity. Arterial injuries due to smoking, high blood pressure, diabetes mellitus, or elevated blood lipids, and other environmental stressors, happen daily. There is a mechanistic link between arterial repair and inflammation. When a competent bone marrow is present, the inflammatory reaction following arterial injuries is self-limited and constructive by triggering the circulation of progenitor cells and the production and release of exosomes produced by these cells which promote tissue repair and immune modulation (arterial homeostasis). Indeed, vascular progenitor cells that are capable of arterial repair can egress the bone marrow and participate in repair processes that are highly dependent of Rac1 and Rac2, and thus, on the production of ROS [101]. The self-limited nature of this productive inflammatory reaction is due to the fact that arterial wall repair leads to the cessation of the signal triggering the inflammatory response. But as we become older, age becomes the dominant risk for atherosclerosis [96]. It is highly likely that loss of competent progenitor cells in the bone marrow capable of repair with aging contributes significantly to the aging risk [100]. Hence, in the presence of incompetent bone marrow, there is a lack of arterial repair which perpetuates, and potentially, exacerbates atherosclerotic inflammation. This yields a destructive positive feedback loop, with progressive senescence and dysfunction of the arterial wall, arterial wall remodeling with atherosclerotic plaques whose rupture leads to myocardial infarction, stroke, and peripheral arterial occlusion (lack of arterial homeostasis).