Fluorescent F-actin Probes in Living Cells

Fluorescent actin and fluorescent actin-binding domain peptides           

Fluorescently-labeled actin (e.g., GFP-actin, eGFP-actin) is one of the original reagents used for studying dynamic remodeling of the actin cytoskeleton in living cells across a range of species^{10,11,30,31,32}. It is often employed for fluorescence recovery after photobleaching (FRAP) microscopy¹⁰. Although still utilized by actin researchers, it has several drawbacks. First, the size of GFP (~28 kDa) can impair polymerization^10,33, meaning that GFP-actin localization may not tell the whole story in regards to where actin re-organization is happening and to what extent, as GFP-actin is known to differentially label F-actin structures^10,31. Second, some actin-binding proteins (e.g., nucleators within the formin family) might prevent incorporation of GFP-actin into actin seeds or growing polymers due to steric hindrance^10,14,15. Third, as both G-actin and F-actin are labelled, there is a relatively high fluorescent background signal from non-filamentous actin³⁴, and finally, expression of eGFP-actin can affect cell behavior^10,35,36.

Another method for live cell imaging of actin is yeast- or human-derived actin binding domains fused to GFP, eGFP, or m-Cherry fluorophores^10,11,19. The three most commonly used of these genetically-encoded F-actin probes are Lifeact, utrophin, and F-tractin. The best characterized is Lifeact, a 17 amino acid peptide from yeast Abp140^16,17. Lifeact has been used in a variety of mammalian and non-mammalian cells for live cell imaging of the actin cytoskeleton and is considered one of the gold standards for F-actin visualization^10,31. Indeed, Lifeact transgenic mice and zebrafish have been created^10,17,37,38. However, Lifeact has several negative characteristics, including the possibility of affecting actin dynamics (so-called Lifeact-induced artifacts) and inhibiting the binding of actin-associated proteins such as cofilin^39-41. Furthermore, Lifeact-GFP has a strong affinity for G-actin which results in a high background fluorescence signal ¹⁶. Indeed, athough it binds to F-actin with a Kd of 2.2 ± 0.3 μM, its binding affinity for G-actin is 10-fold higher than for F-actin¹⁶. Additionally, Lifeact is not a comprehensive actin marker as it does not bind all actin-containing structures^10,39. Another downside is that Lifeact is introduced into the cell through transfection rather than simply adding it into the medium as is done for the SiR/SiR700/SPY probes¹⁰. In addition, at high expression levels, Lifeact can influence actin dynamics. For instance, Lifeact induces assembly of nuclear actin¹⁸. Similar to other F-actin live cell imaging reagents, titrations are necessary to select an optimal concentration that selectively and specifically labels F-actin while not affecting actin dynamics. A recent study⁴¹ illustrates the difficulty in finding a concentration of Lifeact-TagGFP2 that does not create artifacts such as reduced F-actin dynamics that affect stress fibers and the actin cytoskeleton itself, likely due to disruption of cofilin binding. These subcellular F-actin changes translated into altered cell motility. Notably, at high expression levels, Lifeact-TagGFP2 also affected the network organization of microtubules and intermediate filaments⁴¹. In the continuing search for live cell F-actin probes, Patel et al.¹⁹ screened four different Lifeact-GFP probe constructs for their ability to affect axonal and dendritic processes in primary hippocampal mouse neurons. All four Lifeact constructs affected neuron morphology in one or more of the following parameters: total axon length, length of dendrites (tree, branch, and/or shaft), and dendritic complexity¹⁹.

UtrCH is another actin binding domain live cell imaging reagent and is based on the tandem calponin homology domains (CH1 and CH2) of utrophin^10,42. UtrCH consists of the first 261 amino acid residues of human utrophin, an actin binding protein⁴³. The CH domains bind to actin with a Kd of ~18 μM⁴⁴. Utrophin-based probes have been used successfully across a wide range of cell types and species^10,19. Similar to Lifeact, at high concentrations, utrophin-based probes can exert deleterious effects on actin cytoskeletal dynamics in both cell nuclei and cytoplasm^18,45. A recent study in neurons further demonstrates that care must be taken in the use of utrophin-based probes as one (Utr261) increased primary axon length, dendritic branching, and dendritic complexity, while another (Utr230) did not bind F-actin in neurites of primary hippocampal mouse neurons¹⁹.

A third actin-binding domain peptide is F-tractin, a 43 amino acid peptide derived from the rat actin-binding inositol 1,4,5-triphosphate 3-kinase A which binds F-actin with a Kd of ~10 μM^10,46,47. Fewer studies have utilized this actin live cell imaging probe¹⁰. Due to its larger size (in comparison to other probes), it is possible that F-tractin would sterically hinder binding of actin-binding proteins that regulate and/or facilitate polymerization¹⁰. In primary hippocampal neurons, F-tractin-EGFP modified dendritic complexity¹⁹. Interestingly, in at least one cell type (i.e., Drosophila nurse cells), F-tractin does not alter actin dynamics in comparison to Lifeact and UtrCH^10,45. Thus, for some cell lines, F-tractin might be preferred and worth evaluating versus other F-actin probes¹⁰.

Actin-directed nanobodies

A new approach to monitoring actin dynamics in living cells is use of single-domain antibodies, so-called nanobodies. This type of antibody binds to a single antigen and does not require preservation of F-actin’s quaternary structure. An example is Actin-Chromobody which is fused to a GFP-variant for fluorescence detection. As with Lifeact, a downside to cells expressing nanobodies is a high background signal associated with binding to G-actin¹⁰. However, in contrast to Lifeact and some of the other F-actin probes, Actin-Chromobody seems to avoid disruption of actin dynamics even at high expression levels¹⁰. However, this technology remains very novel and needs a great deal more study and testing¹⁰.

Affimer proteins for F-actin

Actin “affimers” are synthetic, actin-binding proteins isolated from a phage library screen^20,48,49. Four recently identified actin affimers (designated 2, 6, 14, and 24) were characterized in in vitro and in vivo experiments²⁰. As eGFP-fusion proteins, affimers 6, 14, and 24 labeled F-actin in live cells (1:1 binding stoichiometry, Kds of 0.31 ± 0.17, 0.30 ± 0.05 and 0.38 ± 0.14 μM, respectively; the fourth affimer-GFP fusion protein did not bind F-actin strongly)²⁰. Based on FRAP experiments, eGFP-Affimer 6 behaved most similarly to Lifeact and F-tractin²⁰. Affimer 6 labeled actin structures throughout the cell; however, it failed to colocalize with mCherry-actin in dynamic ruffles, suggesting preferential binding to stable actin filaments²⁰. Affimers 14 and 24 were more likely to either associate with bundled F-actin or were more likely to induce F-actin bundling. Affimer 24 displayed increased signal in structures that resembled focal adhesions. Based on the totality of FRAP results, eGFP-Affimers may preferentially bind to a subset of actin filaments in the cell, and affimers 14 and 24 may be altering actin organization²⁰. To delve deeper into possible affimer-mediated dynamic reorganization of the actin network, live cell, time-lapse imaging for eGFP-affimers 6, 14, and 24 was performed, revealing that all three affimers bound to actin filaments and bundles and could be used to monitor actin dynamics, but to varying degrees and subcellular locations²⁰. For instance, eGFP-affimer 6 labeled both distinct populations of stable actin filaments and those able to depolymerize²⁰. Conversely, eGFP-affimer 14 labeled  short F-actin bundles but is not suitable for depolymerization studies²⁰. eGFP-affimer 24 binding revealed dynamic F-actin changes associated in a lamellar area involved in cell adhesion. In cells transfected with the eGFP-affimers, filopodia were short and poorly labelled²⁰. In the time-lapse live cell imaging conditions employed, actin dynamics did not appear to be affected by the three affimers²⁰. However, similar to actin-directed nanobodies, affimer studies are in their infancy, meaning that a great deal more characterization needs to be done to identify any differences in how the affimers may differentially bind F-actin based on distinct structures, cell types, and/or experimental conditions, as well as confirming that actin dynamics are not altered.