Nature Structural Biology
8, 735 - 736 (2001)
doi:10.1038/nsb0901-735
Actin allostery again?Edward H. EgelmanEdward H. Egelman is in the Department of Biochemistry and Molecular Genetics, Box 800733, University of Virginia, Charlottesville, Virginia 22908-0733, USA.
egelman@virginia.edu The first crystal structure of an actin molecule not associated with another protein has been solved. Remarkably, the most variable part of actin's structure is also the most conserved part of its sequence.Actin is one of the most abundant and highly conserved eukaryotic proteins. It is involved in muscle contraction, cell motility, control of cell shape and numerous other aspects of cellular function. In many cells it accounts for >10% of the total cellular protein. Actin is also highly dynamic and undergoes repeated cycles of assembly into filaments and disassembly from these filaments. Cells couple the polymerization and depolymerization of actin filaments to actin's hydrolysis of ATP. Actin can be modified by mutations, chemical crosslinks and proteolysis so that the binding of myosin is not changed and the activation of myosin's ATPase activity is not reduced, but motility is inhibited1,
2,
3,
4. This suggests that the internal dynamics of the actin subunit, the actin filament or both are essential for productive force generation in the actomyosin interaction, and that force production is not due solely to a conformational change in myosin. The first crystal structure5 of an actin molecule that is not in complex with any other protein, reported recently in Science, provides support for previous models of the internal dynamics of the actin subunit within the filament and also highlights potentially large allosteric interactions within the actin subunit. Understanding the significance of this new structure requires an appreciation of some emerging insights into why actin may be so conserved.
In higher eukaryotes actin is found in different isoforms that have a tissue-specificity (for example,  -muscle actin and  -cytoplasmic actin), and most sequence divergence is tissue-specific rather than species-specific. For a particular isoform, the extent of sequence conservation over evolution is truly remarkable. During the course of evolutionary divergence between chickens and humans, not a single amino acid has been substituted in the skeletal muscle isoform. Yeast cytoplasmic actin has 90% identity with bovine cytoplasmic -actin, and most of the amino acid changes between the two are conservative. A comparison between the crystal structures of yeast cytoplasmic actin (PDB code 1YAG) and bovine cytoplasmic -actin6 (PDB code 2BTF) shows that the C atoms superimpose with <0.9 Å root mean square (r.m.s.) deviation. It is also possible that most of these differences may be due to different proteins bound to actin (for example, profilin bound to bovine actin and gelsolin segment 1 bound to yeast actin), different crystal forms and different bound ligands rather than intrinsic differences between the protein from yeast and cows.
The traditional explanation has been that this remarkable conservation may have been maintained by the fact that actin has specific interactions with  100 other proteins and that these interactions constrain actin against mutation. But many of these actin-binding proteins are far less conserved. For example, the actin depolymerizing factor (ADF) proteins from plant and human cells share only 30% identity7, yet they bind to nearly identical actin filaments in plants and humans. The biologically active form of actin is a helical filament (F-actin), and existing crystal structures of monomeric actin (G-actin) need to be interpreted in the context of this filament. Fig. 1 shows an electron microscopic reconstruction of F-actin, with a crystal structure of G-actin8 fit into the filament. When the residues that are not identical between two different isoforms are highlighted, the majority fall on the surface of the structure, precisely where actin-binding proteins would be interacting. Thus, it seems that the conservation of the remaining 90% of the residues may not be easily explained by the interaction with actin binding proteins. Rather, the filament structure of F-actin, particularly the existence of multiple states within this filament, may have a larger a role in the conservation of actin sequence than the direct interaction with other proteins. Two examples of such multiple states involve the variable twist of the filament9,
10,
11 and the variable tilt of subunits within the filament11. Both of these modes most likely require that multiple discrete states of the subunit−subunit interface must exist.
 | | |
The crystal structure of uncomplexed G-actin5 has now been solved in an ADP state. When compared with previous ATP−actin structures, there is a large shift in subdomain 2 — most strikingly, in the 'DNase I binding loop' (so named since this forms the main contact between actin and the nuclease DNase I within subdomain 2 (Fig. 2). DNase I is a secreted enzyme that is involved in the degradation of DNA during both digestion and apoptosis, and its activity is inhibited when bound to G-actin. The biological basis for this tight binding between the two proteins is still not well understood. The DNase I binding loop within actin was disordered and not visualized in one previous complex12 and was folded as a -strand in two other complexes6,
13, but in the new ADP−actin structure5, it is an -helix. Such conformational variability might be expected to be associated with large sequence variability, but subdomain 2 is actually the most conserved part of the actin molecule. While there is one conservative amino acid substitution (Val to Ile) between bovine and yeast cytoplasmic actins in this subdomain, there are no substitutions in subdomain 2 between the bovine cytoplasmic and human muscle isoforms.
| | |
The different folds that are seen in subdomain 2 are reminiscent of the type of molecular switching that has been observed in viral capsids14. In capsids, such switching is necessary for chemically identical subunits to exist within different environments in the capsid and make different contacts with adjacent subunits. It has previously been suggested that subdomain 2 of actin might serve as a similar molecular switch in actin, since electron microscopic analysis showed that subdomain 2 was the most variable part of the structure and could be found in different discrete states15,
16,
17,
18,
19. A large rotation of subdomain 2 was also seen in a previous crystallographic study8, and a rotation or disordering of subdomain 2 was invoked to explain how filaments of actin were destabilized following ATP hydrolysis and phosphate release15,
18. This change in subdomain 2 would alter or weaken the contacts between two actin subunits along the same long-pitch helical strand in the filament (Fig. 1). The possible dynamic role of this subdomain in muscle contraction is implicated by the observations that both proteolytic cleavage of the 'DNase I binding loop'3 and crosslinking of this loop to the C-terminus of an adjacent subunit4 inhibit force generation in the actomyosin interaction.
Although there is no doubt that the new structure reinforces previous speculation about conformational switching in subdomain 2, is it necessarily true that this structural change is due only to the transition from an ATP to an ADP state? A remarkable property of actin is that large allosteric interactions exist within the subunit. Specifically, modifications to the C-terminus of actin have been shown to be coupled to conformational changes in subdomain 2 and vice versa. These allosteric interactions exist in both monomeric G-actin20,
21 and in filamentous F-actin22,
23. Because the G-actin crystallized by Otterbein et al.5 was maintained in a monomeric state by a modification to the C-terminus (the covalent addition of the fluorescent dye tetramethylrhodamine maleimide), the structural switching observed may be the first high resolution visualization of this allosteric interaction between the C-terminus and subdomain 2. This possibility makes their result no less interesting. Rather, it may, in fact, provide more basis for understanding the remarkable conservation of actin over evolution by elucidating how such long-range allosteric interactions occur and how they have been maintained over evolution.
Actin, like histones, has been considered by many to be a ubiquitous and boring protein. But just as our picture of histones has dramatically changed over the past five years, with the knowledge that covalent modifications of histones play a large role in regulation of chromatin structure and gene activation24, I expect that the dynamic properties of actin will be appreciated more widely in the future.
REFERENCES
- Drummond, D.R., Peckham, M., Sparrow, J.C. & White, D.C. Nature 348, 440−442 (1990). | Article | PubMed | ISI | ChemPort |
- Prochniewicz, E. & Yanagida, T. J. Mol. Biol. 216, 761−772 (1990). | Article | PubMed | ISI | ChemPort |
- Schwyter, D.H., Kron, S.J., Toyoshima, Y.Y., Spudich, J.A. & Reisler, E. J. Cell Biol. 111, 465−470 (1990). | Article | PubMed | ISI | ChemPort |
- Kim, E. et al. Biochemistry 37, 17801−17809 (1998). | Article | PubMed | ISI | ChemPort |
- Otterbein, L.R., Graceffa, P. & Dominguez, R. Science 293, 708−711 (2001). | PubMed | ISI | ChemPort |
- Schutt, C.E., Myslik, J.C., Rozycki, M.D., Goonesekere, N.C.W. & Lindberg, U. Nature 365, 810−816 (1993). | Article | PubMed | ISI | ChemPort |
- Bamburg, J.R. Annu. Rev. Cell Dev. Biol. 15, 185−230 (1999). | Article | PubMed | ISI | ChemPort |
- Chik, J.K., Lindberg, U. & Schutt, C.E. J. Mol. Biol. 263, 607−623 (1996). | Article | PubMed | ISI | ChemPort |
- Egelman, E.H., Francis, N. & DeRosier, D.J. Nature 298, 131−135 (1982). | Article | PubMed | ISI | ChemPort |
- McGough, A., Pope, B., Chiu, W. & Weeds, A. J. Cell Biol. 138, 771−781 (1997). | Article | PubMed | ISI | ChemPort |
- Galkin, V.E., Orlova, A., Lukoyanova, N., Wriggers, W. & Egelman, E.H. J. Cell Biol. 153, 75−86 (2001). | Article | PubMed | ISI | ChemPort |
- McLaughlin, P.J., Gooch, J.T., Mannherz, H.G. & Weeds, A.G. Nature 364, 685−692 (1993). | Article | PubMed | ISI | ChemPort |
- Kabsch, W., Mannherz, H.G., Suck, D., Pai, E.F. & Holmes, K.C. Nature 347, 37−44 (1990). | Article | PubMed | ISI | ChemPort |
- Dokland, T. et al. Nature 389, 308−313 (1997). | Article | PubMed | ISI | ChemPort |
- Belmont, L.D., Orlova, A., Drubin, D.G. & Egelman, E.H. Proc. Nat. Acad. Sci. USA 96, 29−34 (1999). | Article | PubMed | ChemPort |
- Orlova, A., Prochniewicz, E. & Egelman, E.H. J. Mol. Biol. 245, 598−607 (1995). | Article | PubMed | ISI | ChemPort |
- Orlova, A. & Egelman, E.H. J. Mol. Biol. 232, 334−341 (1993). | Article | PubMed | ISI | ChemPort |
- Orlova, A. & Egelman, E.H. J. Mol. Biol. 227, 1043−1053 (1992). | Article | PubMed | ISI | ChemPort |
- Owen, C. & DeRosier, D. J. Cell Biol. 123, 337−344 (1993). | Article | PubMed | ISI | ChemPort |
- Kuznetsova, I., Antropova, O., Turoverov, K. & Khaitlina, S. FEBS Lett. 383, 105−108 (1996). | Article | PubMed | ISI | ChemPort |
- DalleDonne, I., Milzani, A. & Colombo, R. Biochemistry 38, 12471−12480 (1999). | Article | PubMed | ISI | ChemPort |
- Khaitlina, S.Y. et al. FEBS Lett. 279, 49−51 (1991). | Article | PubMed | ISI | ChemPort |
- Kim, E. & Reisler, E. Biophys. Chem. 86, 191−201 (2000). | Article | PubMed | ISI | ChemPort |
- Cheung, P., Allis, C.D. & Sassone-Corsi, P. Cell 103, 263−271 (2000). | Article | PubMed | ISI | ChemPort |
- Holmes, K.C., Popp, D., Gebhard, W. & Kabsch, W. Nature 347, 44−49 (1990). | Article | PubMed | ISI | ChemPort |
- Lorenz, M., Popp, D. & Holmes, K.C. J. Mol. Biol. 234, 826−836 (1993). | Article | PubMed | ISI | ChemPort |
- Orlova, A., Yu, X. & Egelman, E.H. Biophys. J. 66, 276−285 (1994). | PubMed | ISI | ChemPort |
- Orlova, A. et al. J. Mol. Biol. in the press (2001).
- Mendelson, R.A. & Morris, E. J. Mol. Biol. 240, 138−154 (1994). | Article | PubMed | ISI | ChemPort |
|
