Posttranscriptional and posttranslational regulation of actin

Abstract

Actin is one of the most abundant intracellular proteins, essential in every eukaryotic cell type. Actin plays key roles in tissue morphogenesis, cell adhesion, muscle contraction, and developmental reprogramming. Most actin studies have focused on its regulation at the protein level, either directly or through differential interactions with over a hundred intracellular binding partners. However, numerous studies emerging in recent years demonstrate specific types of nucleotide-level regulation that strongly affect non-muscle actins during cell migration and adhesion and are potentially applicable to other members of the actin family. This regulation involves zipcode-mediated actin mRNA targeting to the cell periphery, proposed to mediate local synthesis of actin at the cell leading edge, as well as the recently discovered N-terminal arginylation that specifically targets non-muscle β-actin via a nucleotide-dependent mechanism. Moreover, a study published this year suggests that actin’s essential roles at the organismal level may be entirely nucleotide-dependent. This review summarizes the emerging data on actin’s nucleotide-level regulation.

Introduction

Actin is a highly abundant intracellular protein, present in every eukaryotic cell as a key building block of the actin cytoskeleton. The major functions of actin are based on its ability to polymerize into filaments, which play essential roles in such fundamental processes as cell adhesion, mechanosensing, directional migration, muscle contractility, cell division, and intracellular transport (see, e.g. (Pollard, 1990; Pollard, 2016; Pollard and Cooper, 2009)). In addition to its cytoskeleton functions, actin can also regulate gene transcription, by nuclear import and direct binding to the promoter regions of a number of genes (Falahzadeh et al., 2015; Kristo et al., 2016; Misu et al., 2017; Viita and Vartiainen, 2017; Virtanen and Vartiainen, 2017; Wesolowska and Lenart, 2015). Overall, actin is one of the most essential and abundant eukaryotic proteins mediating a variety of highly diverse functions in vivo.

While lower eukaryotes (e.g., yeast) contain only one actin, multicellular organisms normally contain multiple actin isoforms that play distinct intracellular functions (Chang et al., 1984; Perrin and Ervasti, 2010; Tondeleir et al., 2009). In vertebrates, actin is represented by six homologous isoforms, encoded by six different genes that show distinct tissue-specific expression patterns in muscle and non-muscle cells (Herman, 1993; Rubenstein, 1990). Two non-muscle actins, ubiquitous in every cell type, are nearly identical to each other at the amino acid level (>99%); all the other actins share ≥94% identity (Fig. 1).

Fig. 1. Alignment of the six mammalian actin isoforms at the amino acid level.

Black line indicates stretches of sequence identical between all six isoforms. Residues different between two or more sequences are spelled out in the alignment with letters that are denoted with different colors throughout the sequences (red for the N-termini, black for residues common in at least 4 isoforms, green for residues common in at least two, and blue for residues outside the N-terminus unique for one isoform). Gene symbols for the actin isoforms used in the alignment are listed on the top left, next to the corresponding lines of sequences. Asterisk represents the stop.

In muscle and non-muscle cells, actin interacts with hundreds of different proteins. These proteins can bind to actin monomers and polymers to modulate actin polymerization and disassembly, cross-link actin filaments into the cortical network or tight parallel bundles, mediate different types of actin nucleation, drive actin-dependent movement, or scaffold actin filaments into the force-generating assemblies, such as sarcomeres in the muscle, or stress fibers in non-muscle cells (see, e.g. (Borisy and Svitkina, 2000; Svitkina, 2018a; Svitkina, 2013; Svitkina, 2018b)). Many studies in the field have been devoted to studying actin regulation via actin-binding proteins and the upstream regulatory events that achieve the precise coordination of these proteins’ functions at the global and local levels. Notably, however, relatively little information has been gathered about the processes involved in the direct regulation of actin itself at the post-transcriptional and posttranslational level, and about the functional consequences of this regulation to actin’s diverse intracellular roles.

Actin in vivo is a target for over 16 different posttranslational modifications, starting with the N-terminus, which undergoes co-translational removal of Met and acetylation of the second residue in all actin isoforms, and ending with a sleuth of modifications identified mostly in proteomics studies, which cover nearly every exposed and reactive residue of the actin surface (Terman and Kashina, 2013). Due to high sequence similarity, it is still not known if these modifications uniformly affect all actin isoforms, or if they differentially target only specific actin(s), and whether any of these modifications are segregated throughout the cell. Moreover, almost nothing is known about the functional consequences of actin modifications on different sites. A notable exception on both counts concerns N-terminal arginylation, an addition of Arg mediated by the arginyltransferase ATE1 that specifically targets the N-terminus of β-actin in non-muscle cells but has not been shown to similarly target any other actin isoform (Karakozova et al., 2006; Kashina, 2006; Pavlyk et al., 2018; Saha et al., 2010).

Data emerging in recent years demonstrate that actin regulation does not begin at the amino acid level, but extends to posttranscriptional events, which exert direct functional effects on actin by regulating the availability and translation of its mRNA (Bassell et al., 1994; Condeelis and Singer, 2005; Kislauskis et al., 1994; Kislauskis et al., 1997; Shestakova et al., 2001). This mRNA-level regulation affects not only the overall actin protein levels in each cell type, but also actin dynamics and targeting to different areas and structures within the cell, as well as posttranslational modifications. β -actin’s mRNA can be specifically targeted to the cell periphery, and this targeting is required for normal cell migration, cell-cell, and cell-substrate adhesion (Condeelis and Singer, 2005; Gutierrez et al., 2014; Katz et al., 2012; Shestakova et al., 2001). Actin’s nucleotide coding sequence, rather than its amino acid sequence, controls differential arginylation of non-muscle actins, which occurs specifically on the N-terminus of β- but not γ-actin (Karakozova et al., 2006; Zhang et al., 2010). Moreover, a recent study demonstrated that the divergent functions of non-muscle β- and γ-actin in mouse development and cell migration depend solely on the intact nucleotide sequence of their genes, while the two proteins appear to be interchangeable at the amino acid level (Vedula et al., 2017). This discovery sheds new light on the nucleotide sequence as an essential determinant of actin’s function in vivo.

mRNA-level regulation of actin abundance and targeting

Actin isoforms are differentially expressed in different cell types (Perrin and Ervasti, 2010; Simiczyjew et al., 2017; Tondeleir et al., 2009). These expression patterns are believed to be achieved by regulation of the gene promoters, which define tissue specificity and the levels of each actin isoform’s mRNA in every cell type. This promoter-based regulation is especially critical for muscle actins, which are believed to play their major roles in specific muscle types. The four muscle actins, α–skeletal, α–cardiac, α–smooth muscle, and γ-smooth muscle, are named after the tissue in which each of these actins plays a predominant role and is expressed at levels much higher than the others. α–smooth muscle actin is prevalent in the smooth muscle (also containing high levels of γ-enteric smooth muscle actin), as well as the vasculature. α–cardiac and α–skeletal actin are dominantly expressed in the cardiac and the striated muscle, respectively. However, this tissue specificity is not exclusive. For example, α–smooth muscle actin is also prominently expressed in non-muscle cells (Belyantseva et al., 2009; Bunnell et al., 2011). Two major non-muscle actins, β- and γ- cytoplasmic actin, are ubiquitously expressed in every cell type, even if their ratios to each other and to other actin isoforms change from tissue to tissue (Bunnell and Ervasti, 2011; Perrin and Ervasti, 2010; Tondeleir et al., 2009). Moreover, mRNA levels for different actin isoforms exhibit significant changes during embryogenesis and in pathological processes as cancer ((Bachvarova et al., 1989; Barja et al., 1986; Buckingham, 1985; Gunning et al., 1997; Lloyd et al., 2004), reviewed in (Simiczyjew et al., 2017)). The β-actin gene has been shown to generate two alternatively spliced transcripts: one apparently present mostly in non-muscle tissue (the longer transcript), the other constitutively expressed in every cell type (the shorter transcript) (Ghosh et al., 2008). An alternatively spliced transcript has also been identified for non-muscle γ-actin, proposed to regulate its levels through mRNA decay (Drummond and Friderici, 2013). Thus, the regulation of actin function and the complex balance of different actin isoforms in this regulation starts at the level of transcript generation.

In addition to transcriptional regulation, actin levels are very likely stringently regulated at the translation stage. This regulation has not been investigated, however its existence is evident in studies that compare the relative abundance of actin isoforms’ mRNA and protein in the same cells. For example, in mouse embryonic fibroblasts, β-actin mRNA is ~6-fold higher than γ-actin mRNA, but at the protein levels these two actin isoforms are produced at an approximately 1:1 ratio (Patrinostro et al., 2017), suggesting a large fraction of β-actin transcripts exist in a translationally repressed state. Similar discrepancies between the ratios of actin isoform mRNAs and their protein levels are also seen in other tissues (Erba et al., 1988; Patrinostro et al., 2017). Consequently, it appears likely that the production of proteins from these mRNAs is governed by specialized mechanisms of translational repression and de-repression. These mechanisms remain to be characterized.

In addition to undergoing differentially controlled transcription and translation, β-actin mRNA, unlike that for any other actin isoform, undergoes spatial segregation in cells. This segregation, initially identified by in situ hybridization ((Hill and Gunning, 1993; Sundell and Singer, 1990), and later characterized in detail at the molecular, cellular, and organismal levels (Condeelis and Singer, 2005; Kislauskis et al., 1997; Oleynikov and Singer, 2003; Ross et al., 1997; Shestakova et al., 2001), utilizes zipcode-mediated transport to target a prominent subset (an estimated 10%) of β-actin mRNA to the cell leading edge. This β-actin-specific process has been shown in a body of studies to be essential for normal migration of non-muscle cells, including migratory fibroblasts and, prominently, neuronal growth cones (Bassell et al., 1998; Kislauskis et al., 1997; Shestakova et al., 2001; Zhang et al., 1999). It has been assumed that such mRNA targeting is necessary primarily for localized β-actin translation at the cell periphery. However this conclusion is still debated in the literature, since the estimated high abundance of freely diffusing actin monomer in an average cell type should be able to easily overwhelm any pools of de novo synthesized actin in the cell. It is also not known why this mRNA targeting is specific to only one actin isoform.

Recent data on the importance of nucleotide-level actin regulation (Vedula et al., 2017) for the first time enables us to think about the role of β-actin mRNA targeting from a new angle, combining all the knowledge about unique β-actin regulation at the posttranscriptional and posttranslational level into a new model that links localized β-actin translation to its essential biological functions.

mRNA targeting and actin networks at the cell leading edge.

Directional cell migration and cell-cell adhesion are driven by spatially and temporally regulated actin filament polymerization that mediates the leading edge extension, as well as search and capture of the potential adhesion sites on the extracellular matrix and the surface of other cells (Fig. 2). This protrusive and adhesive activity of the cell leading edge requires coordinated effort of two distinct actin filament networks that exert both pushing and pulling forces (Brevier et al., 2008; Svitkina, 2018b). The pushing force is mediated by polarized elongation of the branched actin filament networks, whose geometry is defined by Rac-coordinated Arp 2/3 complex. The pulling force is exerted via sets of contractile linear actin filaments, polymerized via RhoA-coordinated mDia/EVL and contracting due to myosin II motor activity. At the cell leading edge these two distinct actin filament networks occupy the same cytoplasmic space, and disruption of either network is sufficient to impair cell adhesive function. It has been long known that β-actin mRNA localization, mediated by the zipcode sequence in its 3’UTR that scaffolds it via the zipcode-binding protein ZBP1, is required for the assembly and activity of both types of these actin networks (Condeelis and Singer, 2005), however the exact role of mRNA localization in this process is not understood.

Figure 2. Regulation of actin networks by mRNA targeting, translation, and posttranslational modifications.

Actin mRNA transport, mediated by ZBP1 (orange circle), that is regulated by phosphorylation (yellow circle add-on), facilitates de novo translation of the actin monomers at the cell periphery. During and after translation, actin can be differentially modified and regulated by Rho or Rac signaling, coupled with different actin nucleators (ARP2/3 complex of mDia formin) and myosin contractility, to facilitate the assembly and function of different actin networks.

In addition to β-actin mRNA, ZBP1 mediates the localization of other key transcripts encoding proteins that are required for both the pushing and the pulling activities in the leading edge actin networks, seen in migratory neuronal growth cones, fibroblasts, and epithelial cells (Gu et al., 2009). Thus, ZBP1 is responsible for assembly and targeting of an mRNA operon, composed of a number of key transcripts encoding proteins implicated in cell migration and adhesion, including β-actin, EVL, Arp 2/3 complex components, E-cadherin, and β-catenin. Proper localization and translation of epithelial zipcode transcripts is required for focal adhesion and adherens junction complex maturation (Bassell et al., 1998; Cruz et al., 2015; Gu et al., 2012; Gu et al., 2009; Gutierrez et al., 2014; Katz et al., 2012). Moreover, delocalizing individual epithelial zipcode operon transcripts perturbs directional motility in neurons and fibroblasts and tissue barrier function in epithelial cells demonstrating the importance of localized protein synthesis in controlling cell and tissue function (Bassell et al., 1998; Cruz et al., 2015; Gutierrez et al., 2014). It has been assumed that this ZBP1-mediated targeting is important because of the need to translate these proteins on site, however this assumption is still a topic of debate in the field, and the underlying reasons for mRNA targeting, especially in the case of β-actin, are still not understood.

β-actin mRNA localization and active translation are both required for cell adhesion and motility

A number of studies have shown that both β-actin transcript localization and active protein synthesis are required for directional motility in cell types of mesenchymal morphology exhibiting fibroblast-like movement. Early on, (Shestakova et al., 2001) showed that masking the β-actin mRNA zipcode sequence with antisense oligonucleotides to block ZBP1 binding, a treatment that delocalizes β-actin transcripts from the leading edge of chick embryo fibroblasts, destroys their leading edge morphology and abolishes their directional movement. β-actin transcript localization to the leading edge of 3T3 fibroblast cells is serum-inducible, demonstrating that the localization process in itself is regulated by stimuli (Hill et al., 1994). Recent work has identified the kinesin motor KIF11 as a necessary component of the β-actin mRNP localization complex in primary fibroblasts (Song et al., 2015). Blocking KIF11 motor activity or reducing KIF11 expression delocalizes β-actin mRNA from the leading edge, decreasing persistent movement without affecting total cell migration velocity, confirming the importance of transcript localization to directional movement in fibroblast cells (Song et al., 2015). Imaging studies show that ZBP1-transported β-actin mRNA can be targeted directly to the focal adhesion sites, where it dwells for minutes, providing indirect support for the idea that anchoring of focal adhesions may involve formation of actin filaments from newly synthesized actin (Katz et al., 2012). Of note, cytoplasmic transcripts move more than the focal adhesion-localized transcripts, leading the authors to postulate that these anchored mRNAs are actively translating (Katz et al., 2016).

Further evidence that mRNA localization and lamellar protrusions sites correlate in both space and time is provided by computational analysis of single-molecule FISH images of β-actin mRNA polarization and cell protrusion (Park et al., 2012). This analysis suggests that the half-lives for mRNA polarization and cell protrusion are ~16 minutes and ~4 minutes, respectively, demonstrating that mRNA polarization is a relatively long-lived process (Park et al., 2012). Moreover, net cell migration distance in these experiments was strongly correlated with mRNA polarization, reinforcing the proposed link between directed motility and β-actin mRNA localization (Park et al., 2012).

A number of observations argue that β-actin mRNA transport occurs in the form of packaged mRNA particles that need to be unfolded on site in order to become translation-competent. Correlated imaging of single β-actin mRNA molecules, ZBP1, and ribosomal markers suggest that β-actin mRNA/ZBP1 binding and β-actin/ribosome binding are inversely spatially correlated (Wu et al., 2015). β-actin/ZBP1 binding is highest in the perinuclear cytoplasm and lowest at the leading edge, while β-actin/ribosome binding shows the opposite distribution. These observations indirectly suggest that ZBP1 activity leads to decreased β-actin translation in the perinuclear cytoplasm and its increased translation at the cell periphery, a mechanism ultimately aimed to generate asymmetry in β-actin protein distribution and its enrichment at the cell leading edge (Wu et al., 2015). In fact, staining of actively migrating cells with antibodies to β-actin tends to highlight a bright zone at the very leading edge (see, e.g. (Dugina et al., 2009; Karakozova et al., 2006) for representative examples), supporting the idea that β-actin protein is indeed highly enriched in this zone during cell movement.

Further evidence of translational repression and de-repression of cell leading edge localized β-actin mRNA has been obtained by high-resolution imaging of co-labeled β-actin mRNA and ribosomes that points to the existence of β-actin translation “hot-spots” (Buxbaum et al., 2014; Strohl et al., 2017). Notably, treatment of cells with translation inhibitors increases the number of leading edge-localizing β-actin transcripts, suggesting that mRNA transport and translation cannot occur simultaneously (Katz et al., 2016). Overall, these data argue that mRNA packaging by ZBP1 may constitute at least one of the predicted mechanisms of translation repression that accounts for the high discrepancy between β-actin mRNA and protein levels in the same cells. Indeed, recent work identified a molecular complex containing β-actin mRNA, ZBP1, and Argonaute proteins demonstrating that β-actin transcripts associate with RISC complexes and therefore are in a translationally arrested states within cells (Kourtidis et al., 2017). Furthermore, immunoprecipation experiments with Argonaute antibodies pulled down ZBP1 and β-actin mRNA, confirming that β-actin transcript translation is regulated by RISC complexes (Hock et al., 2007). It remains to be determined whether additional mechanisms of β-actin translational repression also exist in cells.

Localized β-actin monomer synthesis stimulates adherens junction assembly to control epithelial tissue barrier function

Imaging studies tracking de novo synthesized β-actin monomers show that such monomers are enriched at de novo cell-cell contact sites, suggesting that β-actin protein synthesis actively occurs at these sites (Rodriguez et al., 2006; Gutierrez et al., 2014). Recent work demonstrates that this targeting at the epithelial cell-cell contact sites is required to assemble the linear actin cables used to anchor cadherin family transmembrane receptors at adherens junction assembly sites (Cruz et al., 2015; Gutierrez et al., 2014; Rodriguez et al., 2008). Simultaneous imaging of β-actin mRNA transcripts and monomer synthesis sites reveals that only a small subset of the β-actin transcripts localized at the cell-cell contact sites are actively producing protein (Rodriguez et al., 2006). Thus, the cellular distribution of β-actin protein depends on both transcript localization and the local production of de novo synthesized β-actin monomer from these localized transcripts.

Treatment of cells with inhibitors of translation initiation, or delocalizing β-actin monomer synthesis via mRNA zipcode oligonucleotide masking or deletion both perturb epithelial adherens junction assembly in a dose-dependent manner, without changing the total amount of actin in the cell (Gutierrez et al., 2014; Rodriguez et al., 2006). Furthermore, such disruption perturbs epithelial adherens junction assembly and tissue barrier function evidenced by decreased E-cadherin/F-actin fluorescence covariance at cell-cell contacts and increased dextran diffusion (Cruz et al., 2015). This demonstrates that the location of β-actin synthesis is more important for adherens junction assembly than the total cellular concentration of β-actin monomer.

It has been recently demonstrated that cell-cell contact localized β-actin monomer synthesis stimulates linear actin filament polymerization and therefore opposes E-cadherin endocytosis to fine tune epithelial tissue barrier permeability (Cruz et al., 2015). The idea that linear actin filaments, assembled from locally synthesized monomers, function as anchors for E-cadherin to resist removal from the plasma membrane at de novo contact sites is attractive because of the observation that non-trans interacting E-cadherin is preferentially internalized (Izumi et al., 2004). Additionally, the requirement for E-cadherin function to rescue adherens junction assembly defects caused by β-actin monomer synthesis delocalization (Cruz et al., 2015) supports the hypothesis that E-cadherin homodimerization is the event initiating the β-actin monomer localization process (Rodriguez et al., 2008). Importantly, β-actin siRNA knockdown in epithelial cells specifically perturbs adherens junction assembly and tissue barrier permeability (Baranwal et al., 2012). Taken together, these studies demonstrate a critical role for β-actin monomer synthesis localization during adherens junction assembly to control epithelial tissue barrier function.

A new piece to the puzzle: β-actin arginylation is essential for directional cell migration

With all the accumulated data on the importance of actin mRNA targeting in various biological processes, it still appears puzzling that all this complex regulation is apparently inherent to only one of the six actin isoforms. Coincidentally, β-actin is also the most biologically essential member of the actin family. Knockout data in mice show that β-actin knockout produces the most severe phenotype in mice, compared to the knockout of other actin isoforms, leading to early embryonic lethality, which cannot be rescued by compensatory up-regulation of other actins in the same animals (see (Perrin and Ervasti, 2010) for review). This result is especially remarkable, given the high similarities of all actin isoforms at the amino acid level.

Migratory non-muscle cells contain nearly equal levels of β- and γ-non-muscle actin, which are also the two actin isoforms that are the most similar to each other at the amino acid level, with only four conservative substitutions in their N-termini (Simiczyjew et al., 2017)(Fig. 1). Despite this high similarity, numerous studies point to differences in their intracellular functions (see (Perrin and Ervasti, 2010) for review), starting with the fact that only β-actin mRNA is zipcode-targeted (Condeelis and Singer, 2005). Notably, only β-actin undergoes posttranslational arginylation on the N-terminally exposed Asp residue – the only posttranslational modification of actin that has been identified to-date to exclusively target only one actin isoform (Karakozova et al., 2006). This arginylation is essential for directional cell migration and maintenance of the cell leading edge morphology and adhesion (Karakozova et al., 2006), and is rapidly regulated by stimuli (Pavlyk et al., 2018) but the underlying mechanisms are unknown.

Recent work has surprisingly shown that this selective arginylation of β-actin is defined at the level of the coding sequence (Zhang et al., 2010). Due to silent substitutions, β- and γ-actin mRNA differ by approximately 13%. This difference leads to the formation of different secondary structures. γ-actin mRNA forms a stable loop in its 5’ region, which slows down ribosome translocation generating lower translation rates. β-actin mRNA has a more relaxed structure, which causes no specific ribosome stalling and leads to its faster translation. As a result, γ-actin remains stalled on the ribosome as a nascent peptide, where its arginylation leads to its co-translational targeting by Ub-proteasome degradation. In contrast, β-actin translates and folds without pausing, so that its arginylation is not coupled to ubiquitination and degradation. This complex mechanism ensures that only β-actin, but not γ-actin, is arginylated in cells, directly coupling nucleotide sequence to actin’s posttranslationally modified state (Zhang et al., 2010).

While the exact role of β-actin arginylation at the protein level is still unknown, follow-up work suggests that lack of arginylation leads to reduced actin polymerization in the cell (Saha et al., 2010), and that arginylated actin incorporates into an enriched zone at the leading edge of the migrating cells in a stimuli-dependent manner (Pavlyk et al., 2018). Given the fact that this arginylation is likely translation-dependent (Zhang et al., 2010), an attractive hypothesis is that arginylation preferentially targets de novo synthesized actin to facilitate its function on site, and thus its local synthesis is required at the cell leading edge as a pre-requisite for arginylation. Notably, mRNA encoding the enzyme of the arginylation machinery, ATE1, has been recently found to localize to the cell periphery, likely via a putative zipcode-binding sequence (Pavlyk et al., 2018; Wang et al., 2017). Thus, arginylation provides a new piece of the puzzle in understanding the biological significance of zipcode-mediated mRNA targeting, introducing a possibility that at least one of the consequences of such targeting is to generate a functionally distinct pool of actin protein, marked by posttranslational (or co-translational) arginylation. In support, cells lacking arginylation show impairments in both cell-substrate and cell-cell adhesion (Zhang et al., 2012), and reduced directionality and speed of migration (Karakozova et al., 2006; Kurosaka et al., 2010; Saha et al., 2010).

Nucleotide sequence as a global determinant of actin function

All these data make it evident that β-actin is somehow special among the actin isoforms, since it undergoes unique mRNA targeting and unique arginylation, not seen in other actin isoforms. Even more importantly β-actin is also the most essential of all the actin family members, the only one whose knockout in mice leads to early embryonic lethality that cannot be compensated even by up-regulation of the highly similar γ-actin isoform. What are the essential determinants of such β-actin uniqueness among the actin family?

A recent study in mouse models tackled this problem from an unexpected angle (Vedula et al., 2017). In this study, the authors tested the hypothesis that the divergent functions of β- and γ-cytoplasmic actin reside at the level of their nucleotide sequences, rather than their amino acids. To test this, the authors utilized CRISPR/Cas9 gene editing to convert 5 nucleotides at the beginning of the β-actin gene sequence, so that this edited gene, while being nearly intact, would produce a protein identical to γ-actin. Thus, this mouse (termed Actbc-g, for β-coded-γ-actin) had a nearly intact β-actin gene, but completely lacked the β-actin protein. The authors predicted that if β-actin amino acid sequence determines its essential function, this mouse would have a lethal phenotype, similar to that of the β-actin gene knockout. However, surprisingly, this mouse had no phenotype at all, and embryonic fibroblasts derived from this mouse showed no defect in actin cytoskeleton or migration. Thus, this work definitively proves that β-actin’s essential function in organism survival and cell motility is determined entirely at the nucleotide level, and is not specific to β-actin’s amino acid sequence.

In search for the underlying reasons for this effect, the authors turned to the previously published ribosome profiling studies, which compare the ribosome densities on different mRNAs genome-wide. Comparisons of the composite ribosome densities within the actin family showed that β-actin mRNA has over a 1000-fold higher number of ribosomes compared to γ-actin, and orders of magnitude higher numbers compared to any other member of the actin family. While these numbers do not directly reflect translation rate, they suggest that β-actin mRNA molecules tend to exist in ribosome-bound, translation-ready state – a support for the translation burst hypothesis proposed in other studies. It appears likely that this number somehow relates to the fact that β-actin is the most essential member of the actin family that is also uniquely regulated at the posttranscriptional and posttranslational level, by a mechanism that the authors propose to call “the actin code”.

Overall, we propose that peripheral β-actin mRNA localization, coupled with its specifically high ribosome density, provides a mechanism for localized translational bursts of β-actin synthesis, essential for normal cell migration, cell adhesion, and organism’s survival. This mechanism likely supersedes the specific requirements for the amino acid sequence, since γ-actin protein, translated off the β-actin gene, can effectively perform all the same functions in the mouse (Vedula et al., 2017).

Conclusion

It is evident from multiple ongoing studies that intracellular actin regulation starts at the level that far precedes its translation, during transcriptional and posttranscriptional events. This nucleotide-based regulation combines mRNA zipcode-mediated targeting to the cell periphery, nucleotide-encoded translation dynamics, and posttranslational modifications, such as arginylation, that likely occur as soon as the actin’s nascent peptide emerges from the ribosome (Fig. 2). The existing studies, while numerous, have only sketched out this mechanism in application to certain types of cell behavior, such as migration and adhesion. Unraveling further details of this regulation during the functioning of different actin isoforms in different tissues and cell types constitutes an exciting direction of further studies.