Do actin isoforms have unique functionalities at the protein level?

Arising from R. Shah et al. Nature Communications https://doi.org/10.1038/s41467-024-49427-2 (2024)

A longstanding question in the field of cytoskeletal biology is whether the few differences in amino acid sequence between closely related actin isoforms are functionally important in vivo. The recent study by ref. ¹ concludes that “β- and γ-actin perform specialized and non-redundant roles in cytokinesis and cannot substitute for one another.” We feel that the major conclusion of ref. ¹ cannot be true in vivo when considered in the context of two prior studies showing that mice and cells live quite well in the complete absence of β-actin^2,3.

The mammalian actin family consists of highly related muscle and non-muscle isoforms, each encoded by a different gene. Within the family, the ubiquitously expressed, non-muscle β-actin and γ-actin are the most similar, differing by only four biochemically similar amino acids that are all located in the first 10 residues of the proteins. Although these differences seem minor, the sequences of each actin isoform are identical in all birds and mammals, suggesting strong evolutionary pressure maintaining the exact protein sequence of each actin. The reasons for these distinctions have been a longstanding mystery in the field.

One hypothesis related to the conserved amino acid sequences of these actins is that each protein carries out unique functions in cells that cannot be accomplished by the other isoform. An alternative hypothesis is that gene-level differences drive differential transcription patterns and translation dynamics of actin in cells, while the specifics of the protein sequence play a secondary role, and its strict evolutionary conservation largely reflects very unforgiving tolerance for mutations within each gene.

We thought this question was largely settled by two studies published in 2017 and 2018, where mice were engineered to express γ-actin protein from the Actb gene, which normally encodes β-actin^2,3. In these mice, produced by introducing five nucleotide substitutions into the native β-actin gene, both actin genes are nearly intact, but all β-actin protein is fully replaced with γ-actin. Thus, these mice represent a powerful model to address the functionality of the amino acid-level differences between β- and γ-actin, without the influence of nucleotide-level elements. Any phenotypic changes in these mice are a direct result of amino acid changes in the actin isoforms, while the processes in which the actin isoforms are fully interchangeable occur normally.

The results of these studies from our independent groups were very clear and in complete agreement with each other. Mice completely lacking β-actin protein but maintaining normal actin genes were viable, fertile, and did not exhibit any of the major abnormalities previously observed in actin gene knockout models^2,3. Detailed characterization of cells isolated from these mice revealed no differences in cell division or migration^2,3. The only phenotypic difference detected was the eventual degeneration of auditory hair cells^3,4, which are terminally differentiated sensory cells in the inner ear that detect sound. A later study found that long-lived and terminally differentiated retinal cells also degenerated during aging⁵. No changes so far have been reported in younger mice. Thus, it can be concluded that β- and γ-actin are largely interchangeable at the protein level when it comes to an organism’s development and viability. Based on this, it appears likely that mammalian cells require two non-muscle actin genes to control expression level or timing, and only in the most specialized cell types does there seem to be an actual requirement for β-actin protein.

Shah et al. ¹ come to the opposite conclusion, which is that “β- and γ-actin perform specialized and non-redundant roles in cytokinesis and cannot substitute for one another.” Thus, their claim is that β-actin is required for the fundamental process of cell division. Our question is simple: How is it that mice entirely lacking β-actin can live and reproduce^2,3 if their conclusion is correct?

The experiments of ref. ¹ supporting their claim of a specific and non-redundant role for β-actin in cytokinesis are only correlative, however, they do suggest a potentially important advance, which is that DIAPH3 exhibits a preference for β-actin. Unfortunately, the biochemical data presented is confused by the use of γ-actin purified from chicken gizzard, which is actually smooth muscle γ-actin and not the more ubiquitous cytoplasmic γ-actin as suggested by the authors¹. Nevertheless, DIAPH3 mutations in humans are associated with deafness⁶, so it’s possible that β-actin and DIAPH3 have a unique interaction that can explain hearing loss in mice expressing cytoplasmic γ-actin in place of β-actin^3,4. However, the claim made by ref. ¹ that “β- and γ-actin perform specialized and non-redundant roles in cytokinesis and cannot substitute for one another” conflict directly with previous work^2,3 and demand reconciliation.

How can β-actin protein perform a non-redundant irreplaceable role in cytokinesis when cells lacking β-actin protein divide normally in mice? What does it mean when we observe experimental phenomena in a cell, but the phenomena is not apparent when cells are part of an organism? It is possible, of course, that this major discrepancy has a mechanistic explanation—e.g., because of cellular and organismal context that sets apart the human cancer cell line used by ref. ¹ and the mouse models used in the other two studies^2,3. Conceivably, the replacement of β- with γ-actin in early embryogenesis may lead to compensatory changes in the landscape of actin-binding proteins to ensure normal actin cytoskeleton functions at the organismal level, while cancer cells in culture may have developed a different cytokinesis mechanism that relies on the precise differences in actin isoforms’ amino acid sequences. If so, there may well be lessons to learn from understanding those differences in the context of cancer. However, global conclusions about actin isoforms’ unique functionality in such fundamental and essential processes as cytokinesis must be made with full consideration of all the published data^2,3.