<<

Commentary

Creating a niche in the : reorganization by a

Paul A. Janmey*

Institute for Medicine and Engineering, University of Pennsylvania, Philadelphia, PA 19104

ne of the most apparent differences Obetween normal cells and their ma- lignantly transformed counterparts and between immature and fully differentiated cells is the shape of the cells (1). morphology is controlled largely by the of the cytoskeleton, a system of three distinct types of filamentous poly- mers that assemble into networks and bundles of various kinds to link the cell interior physically with the plasma mem- brane and endow the cell with viscoelastic properties. The transformation of cyto- plasm from a liquid to a solid was ob- served by early microscopists to be tightly associated with the cells’ ability to move, and this sol-to-gel transition is now known to be caused by changes in the state of polymerization and organization of the protein actin, one of the three types of

filaments comprising the cytoskeleton (2). COMMENTARY When something goes wrong in the com- plex system of control proteins and mes- sengers that signals for changes in the actin system, as can happen with genetic or by the insertion of viral coding for maliciously altered control pro- teins, cells may migrate and divide inap- propriately, because the signals for divi- sion or cannot be stopped. One such control factor is the protooncogene Fig. 1. Three types of proteins that form bundles of actin filaments. In addition to specific protein– protein Abl. In addition to its implication protein bonds formed by the Top and Middle examples, release of divalent and other small ions (small blue in leukemias, the kinase dots) by multivalent cations with surfaces compatible with the actin filament contour also contribute to Abl also is an important regulator of the stabilization of actin bundles. the actin system involved in neuronal cell and early neuronal develop- ment (3, 4). additional actin that allows it unambiguously that a particular actin- Abl is part of the signaling pathways to reorganize cellular actin and to form containing structure has been directed to that control the actin cytoskeleton and actin filament bundles by a mech- form by a specific protein. Although there associates with large actin-containing cel- anism that does not require its kinase are only a handful of proteins that bind lular , because it has two actin activity. The finding that a member of the actin , there are over 100 that binding sites at its C-terminal domain (6). Abl family can act directly bind polymeric (F)-actin, and many of Abl’s ability to control cell function re- as a cellular actin bundling factor suggests them such as like Arg cause F-actin to quires its actin binding function, presum- a new mechanism for this class of cell form bundles (8). The dozens of actin ably because such targeting to the cy- regulatory proteins. bundling factors fall into the three classes toskeleton directs its kinase activity to The report by Wang et al. highlights shown in Fig. 1. Some such as ␣- appropriate targets (6). Most cells, includ- several features of the actin cytoskeleton appear to have only one actin binding site ing neurons, also express other proteins that are central to understanding its bio- but can link two actin filaments together, similar to Abl including most prominently logical roles. One challenge to under- because they form homodimers with the a protein called the Abl-related (Arg; standing the functions of actin-binding ref. 3). Although Abl and Arg share many proteins is that there are so many of them, features in common, Wang et al. (7) report and they can be coexpressed at high levels See companion article on page 14865. in this issue of PNAS that Arg contains an in most cells, making it difficult to show *E-mail: [email protected].

www.pnas.org͞cgi͞doi͞10.1073͞pnas.011601598 PNAS ͉ December 18, 2001 ͉ vol. 98 ͉ no. 26 ͉ 14745–14747 Downloaded by guest on September 30, 2021 actin binding sites exposed on opposite binding site in Arg resides within a cat- properties, because remodeling the corti- faces of the dimer. A common actin bind- ionic region of the protein. cal actin network can both generate inter- ing motif, present in several bundling fac- In principle the actin bundles formed by nal forces and resist those imposed from tors, is formed by two repeats of a calpo- the first two classes of proteins should outside the cell. This is a difficult problem, nin domain, also present in Arg. have the filaments arranged in a specific in part because there is no complete the- Other actin bundling factors are mono- polarity. Actin filaments are asymmetric oretical model for the elasticity of meric polypeptides, but they possess two both in structure and dynamics due to the networks formed by the open meshworks distinct actin binding sites, each of which asymmetric structure of the monomers of relatively stiff as are found in binds a separate actin filament. Arg falls and the of ATP that follows the the cytoskeleton. Unlike most synthetic into this class, although it also may self- addition of monomers to the faster- gels that are formed by highly flexible assemble into dimers and higher oli- growing end of the filament (11). There- coils where network elasticity can be pre- dicted accurately from the polymer con- gomers. However, many proteins reported fore, the polarity of the filaments in the centration and the density of crosslinks to bundle actin in vitro have no obvious bundles will, for example, depend on between the strands, the situation with to other actin binding whether the homodimeric crosslinkers are arranged in parallel or antiparallel ar- F-actin is different. The fact that the poly- sites but have in common a large positive mers in the network are so stiff that on the that interacts strongly with rangements and on the spatial relationship of the two classes of actin binding sites in length scale of the distance between fila- the highly anionic actin filament. The first proteins such as Arg. Experimentally, a ment contacts (the mesh size) they are two classes of actin bundlers rely on highly mixed polarity of filaments is found usu- approximately straight makes the macro- specific protein–protein binding inter- ally in actin bundles formed in vitro, and scopic stiffness dependent on not only the faces such as those characteristic of bind- even in the cell actin-containing bundles number of crosslinks but also their geom- ing between two protein monomers (9). that look similar in micrographs dis- etry and flexibility. Therefore, it is not The third type of , while not play a range of polarities when examined easy to predict whether areas of the actin- excluding of two complementary by electron (12). Such a mixed rich at which Arg has induced protein interfaces, does not require it, and polarity may reflect either the flexibility actin to bundle become stiffer. The me- the binding depends on the en- of bundling proteins or the contribution of chanical effect will depend on factors such tropy gained when small ions, mostly di- electrostatic mechanisms that are only as the kinds of crosslinks already formed 2ϩ valent Mg , are released as the higher weakly affected by the polar of the by other proteins and the total polymer valence polycation binds the side of the filament. The polarity of the bundles made concentration. filament (10). This type of bundling is by Arg has not been determined yet, but Proteins such as Arg, which can link closely analogous to the this feature would affect the possible func- actin filaments together, do so in a variety effects driving condensation of DNA by tions of such bundles for allowing contrac- of geometries as shown in Fig. 2. The first multivalent counterions and is a topic of tion or other directed motions. actin crosslinker reported, A, is a much current theoretical work. Of course A next step in understanding the cellu- large homodimer with actin binding re- the two kinds of bundling mechanisms are lar function of Arg-mediated actin re- gions at the ends of two 90-nm-long flex- not mutually exclusive, and it is possibly structuring is to relate the morphology of ible arms (13). The crosslinks it forms tend to have angles distributed around 90°, and relevant that the newly reported actin actin-rich networks to their mechanical this finding has led to the idea that filamin A is especially potent at forming orthog- onal networks that maximize elasticity at a minimum of polymer mass. A second class of crosslinkers is exemplified by the Arp2͞3 complex, a relatively compact multisubunit complex that drives the for- mation of dendritic structures at which filaments are more prone to branch at angles centered around 70° than they are to form uniform three-dimensional net- works (14). These types of crosslinkers have distinct mechanical effects in cells, as shown in a recent report that the flat lamellae required for motility of mela- noma cells requires the crosslinking by filamin A, a function that cannot be taken over by Arp2͞3 (15). The effect of actin bundlers on the elasticity of actin is likely to be more subtle than the effects of linkers that form more open networks. In the absence of other crosslinks, the formation of actin bundles has only a small effect on elasticity, based for example on the smooth scaling of elastic modulus with concentration as ac- Fig. 2. Structures formed by three types of F-actin linking proteins. Actin filaments ( lines) are joined at different angles by proteins that form links with various degrees of flexibility. Note that although tin undergoes a transition from isotropic the shapes of filamin A and Arp2͞3 are based on electron micrographs, their sizes are not drawn to scale. to nematic liquid crystalline phases (16). Then contour length of filamin A is near 180 nm, whereas the length of the Arp2͞3 complex is Ϸ10 nm. However, in crosslinked networks the ef- The dimensions of Arg are not known, but its molecular weight is Ϸ1͞3 that of the filamin A dimer. fects may be more pronounced. In one Domains that contact actin filaments are depicted as light green patches. model of the elasticity of orthogonal actin

14746 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.011601598 Janmey Downloaded by guest on September 30, 2021 networks linked by rigid 90° bonds, redis- filament strands, may be crucial to deter- cellular effect of Arg and other cytoskel- tribution of some filaments into bundles mining how specific crosslinking proteins etal-bound will be a combination leads to an Ϸ50% increase in stiffness alter the macroscopic stiffness of materi- of both direct and kinase-mediated effects (17). In other models where the networks als such as the cell cortex. on the actin system. are relatively sparse and heterogeneous The finding by Wang et al. (7) that Arg The tight association between a critical such as those of actin formed by Dictyo- may directly alter the actin cytoskeleton regulatory protein such as Arg and the stelium gelation factor, formation of a rather than doing so by exerting its protein cytoskeleton emphasizes the importance small fraction of actin bundles may have a kinase activity is bolstered by their argu- of spatial localization for signal transduc- greater stiffening effect, because these ment that this relatively scarce protein is tion processes. In many cases the spatial larger bundles take up more of the stress concentrated enough in particular struc- sequestration of signaling proteins may be imposed on the network (18). tures to compete with more abundant as important as their total concentration The importance of the geometry of actin-binding proteins. Although the total or the fraction of proteins in their active specific crosslinkers in determining the cytosolic concentration of cellular Arg (20 state in processes where cells commit elasticity of actin networks is only begin- nM) would seem to be too low to have a themselves to a fate that is dictated by the ning to be characterized. In flexible gels, significant impact on actin structures con- convergence of multiple messages. The the crosslink can be well modeled as a taining Ͼ100 ␮M actin, the localization of rigid point connecting two network Arg to small structures such as dendritic cytoskeleton, in addition to serving me- strands, because the elastic response to spines increases its effective concentration chanical roles, also is the main spatial deformation comes from pulling out the to near micromolar levels that are suffi- organizer of the and is tied two ends of the coils between crosslinks. cient to affect actin assembly in vitro. intimately to the structure of cellular For a network of stiff polymers such as Likewise, the finding that Arg binds to membrane interfaces. The current report actin, imposition of shear or elongational actin in a highly cooperative manner im- of Arg’s ability to manipulate its own deformations can be resisted only slightly plies that selected regions of the cytoskel- cytoskeletal environment suggests that by extending the two ends of the actin eton will be particularly enriched in Arg. there is much to learn about how the strand, because it is already nearly straight Of course these same arguments for lo- cytoplasmic space is partitioned to achieve in the resting state (19, 20). Therefore, calization apply to models in which the levels of control that allow the cell to make flexibility and movement of crosslinks, as kinase activity of Arg is targeted to spe- sense of the numerous simultaneous mes- well as the angles they form between cific cellular sites, and it is likely that the sages that it constantly receives.

1. Porter, K. R. (1975) Adv. Pathobiol. 1, 29–47. 8. dos Remedios, C. G. & Thomas, D. D. (2001) 15. Flanagan, L. A., Chou, J., Falet, H., Neujahr, R., 2. Stossel, T. P. (1994) Sci. Am. 271, 54–63. Results Probl. Cell Differ. 32, 1–7. Hartwig, J. H. & Stossel, T. P. (2001) J. Cell. Biol. 3. Koleske, A. J., Gifford, A. M., Scott, M. L., Nee, 9. Hanein, D., Volkmann, N., Goldsmith, S., Michon, 155, 511–518. M., Bronson, R. T., Miczek, K. A. & Baltimore, D. A. M., Lehman, W., Craig, R., DeRosier, D., Almo, S. 16. Maruyama, K., Kaibara, M. & Fukada, E. (1974) (1998) Neuron 21, 1259–1272. & Matsudaira, P. (1998) Nat. Struct. Biol 5, 787–792. Biochim. Biophys. Acta 371, 20–29. 10. Tang, J. X. & Janmey, P. A. (1996) J. Biol. Chem. 4. Lanier, L. M. & Gertler, F. B. (2000) Curr. Opin. 17. Satcher, R. & Dewey, C. (1996) Biophys. J. 71,

271, 8556–8563. COMMENTARY Neurobiol. 10, 80–87. 109–118. 11. Wegner, A. (1976) J. Mol. Biol. 108, 139–150. 5. McWhirter, J. R. & Wang, J. Y. (1993) EMBO J. 18. Ruddies, R., Goldmann, W. H., Isenberg, G. & 12. Cramer, L. P., Siebert, M. & Mitchison, T. J. 12, Eur. Biophys. J. 22, 1533–1546. (1997) J. Cell Biol. 136, 1287–1305. Sackmann, E. (1993) 309– 6. Van Etten, R. A., Jackson, P. K., Baltimore, D., 13. Hartwig, J. & Stossel, T. (1981) J. Mol. Biol. 145, 321. Sanders, M. C., Matsudaira, P. T. & Janmey, P. A. 563–581. 19. MacKintosh, F. C. (1998) Biol. Bull. (Woods Hole, (1994) J. Cell Biol. 124, 325–340. 14. Volkmann, N., Amann, K. J., Stoilova-McPhie, S., Mass.) 194, 351–352. 7. Wang, Y., Miller, A. L., Mooseker, M. S. & Egile, C., Winter, D. C., Hazelwood, L., Heuser, 20. Hinner, B., Tempel, M., Sackmann, E., Kroy, K. Koleske, A. J. (2001) Proc. Natl. Acad. Sci. USA J. E., Li, R., Pollard, T. D. & Hanein, D. (2001) & Frey, E. (1998) Phys. Rev. Lett. 81, 98, 14865–14870. Science 293, 2456–2459. 2614–2617.

Janmey PNAS ͉ December 18, 2001 ͉ vol. 98 ͉ no. 26 ͉ 14747 Downloaded by guest on September 30, 2021