Intermediate Filament Scaffolds Fulfill Mechanical, Organizational, and Signaling Functions in the Cytoplasm
Total Page：16
File Type：pdf, Size：1020Kb
Downloaded from genesdev.cshlp.org on September 28, 2021 - Published by Cold Spring Harbor Laboratory Press REVIEW Intermediate filament scaffolds fulfill mechanical, organizational, and signaling functions in the cytoplasm Seyun Kim1 and Pierre A. Coulombe1,2,3 1Department of Biological Chemistry, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, USA; 2Department of Dermatology, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, USA Intermediate filaments (IFs) are cytoskeletal polymers wise represents the major driving force sustaining self- whose protein constituents are encoded by a large family assembly (for review, see Fuchs and Weber 1994; Herr- of differentially expressed genes. Owing in part to their mann and Aebi 2004; Parry 2005). The rod is flanked, at properties and intracellular organization, IFs provide both ends, by nonhelical sequences that differ in length, crucial structural support in the cytoplasm and nucleus, sequence, substructure, and properties. Variations in the the perturbation of which causes cell and tissue fragility so-called “head” and “tail” domains account for the ∼ and accounts for a large number of genetic diseases in marked heterogeneity in IF protein size (Mr 40–240 humans. A number of additional roles, nonmechanical kDa) (Table 1) and other attributes. A “one gene/one pro- in nature, have been recently uncovered for IF proteins. tein” rule seems to prevail in the family, as relatively These include the regulation of key signaling pathways few IF mRNAs (lamin A/C, GFAP, peripherin, and syn- that control cell survival, cell growth, and vectorial pro- emin) (Table 1) yield distinct protein products via alter- cesses including protein targeting in polarized cellular native splicing. settings. As this discovery process continues to unfold, a Most biomedical researchers’ understanding of fibrous rationale for the large size of this family and the context- cytoskeletal polymers is primarily influenced by the dependent regulation of its members is finally emerging. extraordinary properties of F-actin and microtubules, whose pleiotropic roles tend to be universal and can be investigated in cultured cell lines and simple model eu- Intermediate filaments (IFs), first described by Holtzer karyotes (Alberts et al. 2002). IFs are fundamentally dif- and colleagues (Ishikawa et al. 1968) from studies of ferent, as follows: Functionally, cytoplasmic IF proteins muscle in the late 1960s, serve as ubiquitous cytoskel- are not required for life at the single-cell level, as evi- etal scaffolds in both the nucleus and cytoplasm of denced by their complete absence in yeast, in Drosophila higher metazoans (Erber et al. 1998). In human, mouse, (Erber et al. 1998), and in some mammalian cell lines ∼ and other mammalian genomes, 70 conserved genes en- (Venetianer et al. 1983). Yet, they are clearly functionally code proteins that can self-assemble into 10- to 12-nm- required in a broad range of metazoans ranging from Cae- wide IFs. Apart from three lamin-encoding genes, whose norhabditis elegans (e.g., see Karabinos et al. 2001) to products localize to and function in the nucleus, the mammals (Fuchs and Cleveland 1998; Omary et al. 2004; ∼ other 67 IF genes encode cytoplasmic proteins (Table 1; this review), and possibly in bacteria as well (Ausmees et Hesse et al. 2001). Although heterogeneous in size, pri- al. 2003). Biochemically, they exhibit an apolar (i.e., the mary structure, and regulation, IF proteins share a com- two ends of the fiber are structurally identical) and het- mon tripartite domain structure, with the defining fea- erogeneous substructure (Aebi et al. 1988; Herrmann and ␣ ture being a centrally located, 310-residue-long -helical Aebi 2004). Whereas polymerized IFs exhibit cycles of domain (352 for lamins) containing long-range heptad disassembly and reassembly under “steady-state condi- repeats of hydrophobic/apolar residues (Fig. 1A). These tions” in vivo (e.g., see Vikstrom et al. 1992; Windoffer et conserved features were formalized with the cloning and al. 2004; for review, see Helfand et al. 2004), IF proteins sequencing of the first IF protein-encoding gene, keratin do not directly bind and metabolize nucleotides as actin 14 (Hanukoglu and Fuchs 1982). The central “rod” do- and tubulin do. Several IF proteins assemble into IFs as main mediates coiled-coil dimer formation and other- obligate heteropolymers (e.g., keratin, neurofilaments, nestin), while others (e.g., vimentin, desmin) can do so [Keywords: Keratin; vimentin; cytoskeleton; adhesion; cell polarity; in- either as homo- or hetero-polymers. There currently are tracellular transport] no data consistent with IFs serving as “tracks” for mo- 3Corresponding author. E-MAIL [email protected]; FAX (410) 614-7567. lecular motors, yet they significantly contribute to the Article is online at http://www.genesdev.org/cgi/doi/10.1101/gad.1552107. complex cytoarchitecture of a host of differentiated cell GENES & DEVELOPMENT 21:1581–1597 © 2007 by Cold Spring Harbor Laboratory Press ISSN 0890-9369/07; www.genesdev.org 1581 Downloaded from genesdev.cshlp.org on September 28, 2021 - Published by Cold Spring Harbor Laboratory Press Kim and Coulombe Table 1. IF proteins Size IF name Type (kDa) Cell and tissue distribution Key feature and/or disease association Cytoplasmic Keratinsa I(n = 28) 40–64 K9–K28 (epithelia); Types I and II keratins form obligate 1:1 heteropolymers. K31–K40 (hair/nail) There are 54 functional keratin genes in Keratinsa II (n = 26) 52–68 K1–K8, K71–K74 (epithelia); the human genome. K81–K86 (hair). Mutated in >20 diseases. Vimentin III 55 Mesenchymal Widely expressed in embryos. Desmin III 53 Muscle Mutated in cardiomyopathies. GFAP III 52 Astrocytes/glia Mutated in Alexander disease. Peripherin III 54 Peripheral neurons Induced after neuronal injury. Neurofilaments IV 61–110 CNS neurons NF-L, M, and H form obligatorily heteropolymers (L,M,H chains) with ␣-internexin. ␣-Internexin IV 66 CNS neurons Neuronal IFs are key effectors of axonal radial growth. Nestin IV 177 Neuroepithelial Markers of “early” progenitor (stem) cells in several tissues. Syncoilin IV 54 Muscle Interacts with ␣-dystrobrevin. Synemin IV 182 Muscle ␣ and  isoforms;  form is also known as desmuslin; binds actin-associated proteins. Nuclear Lamins B1, B2 V 66–68 Nuclear lamina Enriched in progenitor cells. Lamins A/C V 62–78 Nuclear lamina Subject to differential splicing; enriched in differentiated cells. Mutated in a progeria condition, muscular dystrophy, and others. Orphan Phakinin (CP49) undefined 47 Lens CP49 and filensin form beaded filaments in lens epithelial Filensin undefined 83 Lens cells. CP49 mutations cause cataracts. aThe keratin nomenclature has recently been revised (see Schweizer et al. 2006). types. This text discusses the recent inroads made to- ter (Strelkov et al. 2001), and thus there is as yet no ward defining the properties and functions of cytoplas- “atomic-level” information for the structure of poly- mic IFs, and identifies some of the key challenges lying merized IFs. Various analyses of filaments, both in vitro ahead. Concurrent progress made for the nuclear lamins (Aebi et al. 1988; Herrmann and Aebi 2004) and in suit- is not discussed but was recently covered by others (Gru- able tissues in vivo (e.g., see Er Rafik et al. 2004; Norlen enbaum et al. 2005; Worman and Courvalin 2005; Broers and Al-Amoudi 2004), have shown that mature IFs et al. 2006; Capell and Collins 2006; Mattout et al. 2006; (Fig. 1C) are comprised of several subfibrils (Fig. 1D). The Navarro et al. 2006). presence of a clear 21- to 22-nm axial repeat in some EM preparations of filaments (Aebi et al. 1983, 1988) im- plies that coiled-coiled dimers are staggered in a consis- General features of IFs tent fashion within the wall of the mature polymer. How subunits interact, laterally and longitudinally, to Structure give rise to mature IFs is being defined with increasing Form defines function. For cytoskeletal polymers, depth (e.g.,see Hess et al. 2004; Bernot et al. 2005; “form” consists of the combination of their structure Sokolova et al. 2006). Compared with the central rod, and functional organization, including assembly sites, the contributions of the N-terminal head and C-terminal dynamics and turnover, and integration with other ele- tail domains to filament assembly vary depending on the ments of the cell. Owing to the presence of long- IF protein considered. Given their exposure at the range heptad repeats (Fig. 1A), cytoplasmic IF proteins polymer surface, the end domains are poised to me- readily form highly stable coiled-coil dimers (42–44 nm diate interactions with other filaments and a host of cel- in length) in which the two participating monomers lular proteins, as well as serve as substrates for post- exhibit a parallel, in-register alignment. Dimers then translational modifications that regulate structure, orga- associate along their lateral surfaces, with an anti- nization, and function (Coulombe and Wong 2004; parallel orientation, to form apolar tetramers (Fig. 1B), Green et al. 2005; Izawa and Inagaki 2006; Omary et al. which are readily obtained in high yield in vitro and have 2006). been isolated as soluble entities from in vivo sources (Soellner et al. 1985; Herrmann and Aebi 2004; Bernot et Organization within the cell al. 2005). IF proteins have been difficult to crystallize, owing to their notorious insolubility, lack of suitable IFs are integrated with other key elements making up assembly inhibitors, and polymerization-prone charac- the cell’s interior. They interact with the other major 1582 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on September 28, 2021 - Published by Cold Spring Harbor Laboratory Press Functions of intermediate filaments Figure 1. Introduction to cytoplasmic IFs. (A) Schematic representation of the tripar- tite domain structure shared by all keratin and other IF proteins. A central ␣-helical “rod” domain acts as the major determi- nant of self-assembly and is flanked by nonhelical “head” and “tail” domains at the N and C termini, respectively.