- Email: [email protected]
Filamins: promiscuous organizers of the cytoskeleton
Review
TRENDS in Biochemical Sciences
Vol.31 No.7 July 2006
Filamins: promiscuous organizers of the cytoskeleton Grzegorz M. Popowicz1, Michael Schleicher2, Angelika A. Noegel3 and Tad A. Holak1 1
Max-Planck Institut fu¨r Biochemie, 82152 Martinsried, Germany Institut fu¨r Zellbiologie der LMU, 80336 Mu¨nchen, Germany 3 Institut fu¨r Biochemie I, Zentrum Molekulare Medizin Ko¨ln, Medizinische Fakulta¨t, Universita¨t zu Ko¨ln, 50931 Ko¨ln, Germany 2
Filamins are elongated homodimeric proteins that crosslink F-actin. Each monomer chain of filamin comprises an actin-binding domain, and a rod segment consisting of six (Dictyostelium filamin) up to 24 (human filamin) highly homologous repeats of w96 amino acid residues, which adopt an immunoglobulin-like fold. Two hinges in the rod segment, together with the reversible unfolding of single repeats, might be the structural basis for the intrinsic flexibility of the actin networks generated by filamins. There are numerous filaminbinding proteins that associate, in most cases, along the repeats of the rod repeats. This rather promiscuous behaviour renders filamin a versatile scaffold between the actin network and finely tuned molecular cascades from the membrane to the cytoskeleton. Introduction The cytoskeleton provides the foundation for the spatial organization of cells and their movement. Despite our inclination to consider the cell ‘skeleton’ as a rigid structure, the real cytoskeleton is dynamic, undergoing continual reorganization and modification. In a nonmuscle cell, the actin cytoskeleton consists of a pool of globular monomeric actin (G-actin), which can reversibly polymerize into filamentous actin (F-actin) and thus markedly alter the viscosity, elasticity or mechanical resistance of the whole cell. Actin filaments are polymerized, depolymerized, crosslinked, bundled or fragmented with the help of specific proteins that either change the polymerization kinetics or stabilize the existing network by connecting single filaments. The best-studied F-actin crosslinking proteins are spectrin, fimbrin, a-actinin and filamin; these proteins are characterized by a conserved actin-binding domain (ABD), which is frequently followed by an extended rodlike domain. They can form dimers or tetramers, enabling them to bundle or to crosslink filaments. Whereas spectrin, fimbrin and a-actinin are thought to form primarily parallel actin bundles, filamins can crosslink actin filaments to form orthogonal networks to bundles depending on their concentration [1]. In filamins, the rod segment consists of highly homologous repeats of w96 amino acids that are characterized structurally by the Corresponding authors: Holak, T.A. ([email protected]), Noegel, A.A. ([email protected]). Available online 16 June 2006
presence of an immunoglobulin-like fold, in which seven b strands are arranged in an antiparallel way [2,3]. By contrast, the spectrin and a-actinin rod segments consist of a-helical regions [4]. In this review, we focus on recent progress in structural research on filamins and seek to link their physiological functions with this new information. Although filamins are found in many organisms, the best studied are the Dictyostelium discoideum filamin (ddFLN) and mammalian filamin. These two prototypical filamins have an ABD comprising two tandem calponin homology domains and a rod region built by six (in ddFLN) or 24 (in human filamin) repeats [2,5] (Figure 1). The last repeat of the rod is responsible for dimerization. In addition, human filamin has two unique long hinges positioned between repeats 15–16 (27 residues) and 23–24 (35 residues) that are postulated to be flexible [6]. The human filamin family consists of three members, filamins A, B and C, that share 70% sequence homology, except for the less homologous hinges. Recent studies show that filamins not only are mechanical linkers for actin filaments but also act as interaction partners for many proteins of great functional diversity ranging from signal transduction to transcription [7]. Furthermore, recent genetic studies have demonstrated the significance of filamins in several diseases affecting brain [7–9], bone and the cardiovascular system [10,11]. Mechanism of filamin dimerization As mentioned, dimerization of filamin is mediated by the last rod repeat. Three structures of fragments of the rod region including the dimerization region have been solved (Figure 1b): two from ddFLN, which show an identical mode of dimerization; and one from human filamin [12–14]. In ddFLN the structure of the last repeat differs from the other repeats of the rod, whereas in human filamin the repeat responsible for dimerization is similar to the other repeats. In ddFLN, repeat 6 lacks 12 residues at the N terminus and, up to the middle of the second b strand, shares no sequence homology with the other repeats. It also contains an additional strand at the C terminus of the repeat. Repeats 6 and 6 0 in the two monomers form an antiparallel dimer by b-sheet extension. The most extensive interactions are between the first strand and the penultimate strand of repeats 6 and 6 0 . The dimer
www.sciencedirect.com 0968-0004/$ - see front matter Q 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.tibs.2006.05.006
Review
412
TRENDS in Biochemical Sciences
Vol.31 No.7 July 2006
(a)
Dictyostelium discoideum
Homo sapiens Hinge 1
ABD
Hinge 2 ABD
ABD
Rod ABD
Rod
(b)
D. discoideum
Repeat 4
Repeat 5
Repeats 6
Repeat 5
Repeat 4
C
N C
N
Repeat 17
Repeats 24 N
NGPIbα
CFLN17
C
H. sapiens
NFLN17 C CGPIbα
N
(c)
C
CH1
CH2
N Figure 1. Structural properties of filamins. (a) Architecture of human filamin and Dictyostelium filamin (ddFLN). Human filamin comprises an ABD and a rod region built by 24 repeats. The rod is divided by two ‘hinge’ linkers located between repeats 15 and 16, and repeats 23 and 24. Each repeat has an immunoglobulin-like fold. The last repeat is the homodimerization domain. ddFLN has a similar ABD and a rod of six repeats. (b) Structures of fragments of the ddFLN rod segment (repeats 4–6), human filamin dimerizing repeat 24 (the monomers that build the homodimer are shown in different colours) and human filamin repeat 17 (FLN17). The structure of repeat 17 (red) was solved in complex with the receptor GPIba (blue), highlighting a common binding motif of filamin-associated proteins. (c) Ribbon plot of an ABD from a-actinin [19]. The ABD is composed of two calponin homology domains, CH1 and CH2. On the basis of sequence similarity, filamin is expected to have a similar domain at its N terminus.
˚ (of which 69% is the interface is large, comprising w4500A primary dimerization contribution of the repeat and 22% is the contact surface of a buried linker between repeats 5 and 6). Repeats 6 and 6 0 form a large b-sandwich of six strands on one side and eight on the other (Figures 1, 2b). The N-terminal regions of both structural units embrace neighbouring repeats and are also deeply buried between their strand-interconnecting loops. Such organiz ation of www.sciencedirect.com
the linker brings the fifth repeat closer to the sixth repeat of the second molecule than to the sixth repeat of the same chain. This arrangement provides a high degree of rigidity between repeats 5 and 6. The structures of the dimerization repeats of ddFLN and human filamin are similar: the root mean square ˚ (Figure 2). This deviation for the core parts is only 1.2 A similarity refers particularly to the global arrangement of
Review
TRENDS in Biochemical Sciences
(a) Ndd
Chs
Cdd
Chs
Ndd
Cdd
Nhs
Nhs
(b) Nhs
Nhs
Ndd
Ndd
Chs
Chs Cdd
Cdd
Cdd
Cdd
Chs
Chs Ndd
Ndd Nhs
Nhs
Figure 2. Structural comparison of the dimerization repeats of ddFLN and human filamin. (a) Stereo view of monomer superimposition shows that the cores have basically identical folds but there are differences at the N- and C-terminal ends of the repeats. ddFLN (red) has an additional long strand at the C terminus, whereas human filamin (blue) has two additional short strands at the N terminus on the other side of the molecule. (b) Stereo view of homodimer superimposition on the monomer structures shown in (a). Despite the structural similarity in the repeats of human filamin and ddFLN, the dimerization area is different. The interacting surfaces are localized at opposite sides of the repeats. There is no interaction between the N-terminal linker of the dimerization repeat and the second molecule in the dimer in vertebrate filamin.
dimerization; for example, the repeats in human filamin are arranged with an exact two-fold symmetry as they are in ddFLN. Furthermore, dimers of human filamin also form through the extension of a b sheet of each monomer to form an elongated b sandwich. However, the strands responsible for direct interaction differ in human filamin, and the internal organization of the strands also varies from that in ddFLN. Differences affecting dimerization are found at the C and N termini of the repeats: the dimerizing strands of the repeat seem to be the third (C) and fourth (D) b strands in human filamin, and the first (A) and sixth (F) b strand in ddFLN. The lengths of the interacting strands are shorter in human filamin than in ddFLN. In addition, there is no buried linker at the N terminus of the repeat in human filamin. A long hinge region of 35 residues precedes repeat 24 in human filamin. The structure of this region is unknown and shows the greatest sequence divergence. The main function of this region is probably mechanical, enabling a high degree of flexibility and structural adaptability [14]. One study has indicated that this region might have a regulatory function for dimerization [15]. The interface of dimerization is much smaller in human filamin than in ddFLN: the buried surface area ˚ 2, taking up 19% of the molecular surface. The is 1109 A crystallographic model is supported by mutagenesis studies, which exclude the possibility of artefacts and www.sciencedirect.com
Vol.31 No.7 July 2006
413
prove that the dimerization mechanism of the wild-type vertebrate filamin is altered and presumably weaker than that of ddFLN [14]. Overall, the comparison between ddFLN and human filamin reveals that the dimerization interfaces in filamins have changed during evolution, although the same mechanism of a double-sided b-sheet extension of immunoglobulin-like repeats has remained. The rigid and strong interaction in filamins of lower organisms has been replaced by more sophisticated, weaker, and more flexible dimerization interfaces in higher organisms. Future studies on longer constructs of filamins are expected to provide information on the spatial organization of their repeats and their rigidity. Inter-repeat organization The mechanical properties of the whole rod region are essential for both the F-actin crosslinking capabilities of filamin and the characteristics of the F-actin networks that they create. Except for the two hinge regions in human filamin, the linkers between the repeats are also short in vertebrate filamins: they comprise only a few residues and are rich in proline. Additional salt bridges have been identified between neighbouring repeats. The two structures of the ddFLN multirepeat constructs show that the repeats of the rod remain in the same spatial position regardless of crystal packing and crystallization conditions. The structures maintain an exact two-fold symmetry. The root mean square deviation ˚ ; thus, the whole rod between the structures is only 1.3 A seems to be a long, extended, spring-like molecule with an inter-repeat tilt angle of w1158. Such conformations support the proposition that in ddFLN the whole rod is an extended structure of limited flexibility. Of course, some flexibility must be provided to facilitate crosslinking of actin filaments and adaptation of the F-actin network generated. This particular feature might be realized by a long linker between the ABD and the rod [13] (Figure 3). It seems that the primary function of ddFLN in Dictyostelium is to reduce the degree of freedom of the bound actin filaments while facilitating the organization of an orthogonal and highly flexible filamentous network. There are no structural data on the inter-repeat organization of vertebrate filamins; however, the rod segment of human filamin, which is almost five times longer than that of ddFLN, is likely to be organized in the same way. Remarkable sequence similarities support such a model; in addition, the two long hinge regions in human filamin have a larger amino acid diversity and their length would enable the whole rod to be more flexible and to crosslink actin filaments in many orientations. The hinges might also present potential proteolytic cleavage sites. A naturally occurring proteolytic fragment of human filamin (FLNa; repeats 16–24), generated by cleavage at the protease-cleavage site between repeats 15 and 16, translocates to the nucleus and acts as a nuclear transcription modulator [16]. However, a region before repeat 24, which is considered to be flexible [17], has been reported to regulate dimerization [15]. In addition to its mechanical function,
414
Review
TRENDS in Biochemical Sciences
Actin-binding domain
Rod repeats
Repeats 6 building a dimer
Rod repeats
Actin-binding domain
Figure 3. Model of the whole ddFLN molecule as a homodimer. The model is built on the basis of known parts of the structure. Actin-binding domains are green; rod regions are red and blue. The extended molecule can bind actin filaments over a distance of w400 A˚ [13]. Reproduced, with permission, from Ref. [13].
the rod region of vertebrate filamins has been shown to be an important scaffold for binding many proteins involved in a broad range of cellular processes [6,7]. www.sciencedirect.com
Vol.31 No.7 July 2006
The actin-binding domain Members of the a-actinin superfamily of actin crosslinking proteins share the same mechanism of interaction with F-actin: filamins, a-actinin, spectrin, nesprins, plectin, dystrophin and fimbrin use a similar motif of two calponin homology domains to construct their ABDs [18]. A typical ABD has w250 residues and shares 20–60% of sequence identity with other ABDs in the family. Structures of the ABD of a-actinin show that the two subdomains form a fully a-helical globular domain [19] (Figure 1b). The subdomains seem to have either a closed or open conformation, depending on the contact area between them. These conformations present various possible subdomain organizations [20], and it is not known which one is present in the ABDs of filamins. The interface between F-actin and a-actinin has been studied extensively with peptides and antibodies of known specificities. The primary binding site for F-actin in the ABD of a-actinin is located between residues 121 and 147. The interaction is mostly hydrophobic, but the susceptibility of the binding to ionic strength suggests that hydrophilic interactions are also involved. The amino acids of actin that participate in binding are located between residues 112–125 and 360–372 [21]. A recent report suggests that binding of the ABD of FLNa to actin is regulated by calmodulin, an abundant Ca2C-binding protein. Binding is facilitated by Ca2C and calmodulin together: in other words, neither Ca2C ions nor calmodulin alone show regulatory properties. The calmodulin-binding region in the ABD has been mapped to a region between residues 50 and 96 in the first calponin homology domain. An interaction between calmodulin and the whole ABD is not observed, however, which might suggest that calmodulin uses a cryptic binding site in the ABD that is uncovered only after the ABD binds to the actin filament [22]. Unfolding of the ddFLN rod Detailed studies using atomic force microscopy have been recently carried out to obtain more information on the mechanical properties of the rod repeats in ddFLN. These experiments indicate that individual repeats unfold before the dimer is broken. To break the dimer, a force of w200 pN is necessary. Notably, the fourth repeat shows a pattern of unfolding that differs from that of all other repeats. It unfolds the most easily and seems to be the only repeat that has a stable folding intermediate. In the first stage of unfolding, w40 residues are stripped from the molecule and the remaining 60 stay folded. The former 40 residues correspond to the first two b strands [23]. The intermediate is a stable structure that can fold on itself. The two-stage folding of the fourth repeat is also the fastest process of folding observed for the rod region [24]. The biological significance of this feature is potentially interesting. Easy unfolding and fast refolding of repeat 4 would enable the whole rod to nearly double its length and then come back to its native state. It is also possible that the folding speed of an elongated molecule would be increased by the presence of the intermediate. The freeenergy barriers between the unfolded state and the intermediate, and, in the next step, between the
Review
TRENDS in Biochemical Sciences
intermediate and a fully folded protein, are easier to break than the barrier for direct transition to the folded state. Another interesting feature of this unfolding mechanism is linked to proteins that can bind to the rod region in the region of repeat 4. Mechanical stress might cause unfolding of the repeat together with dissociation of the binding partner or association of a new one to the intermediate. The bound or released interacting molecules might be part of a signalling pathway that could function as a cellular ‘sensor’ of mechanical forces present in the cytoskeleton. Experiments investigating vertebrate filamin mechanics have been done on the whole filamin molecule and have shown that different repeats unfold under different force [25,26]. To unfold some of the strongest repeats requires a force more than three times greater than that needed to unfold the weakest ones. It has been
Vol.31 No.7 July 2006
415
also shown that repeats can refold spontaneously without stress. The authors of these studies [25,26] postulated that eventual unfolding of the filamin rod during cytoskeletal stresses is a method to protect the whole architecture from being mechanically separated. Formation of a new complex between filamin and its putative binding partner would be much slower and less likely to occur than a folding cycle. Under these circumstances, the spatial organization of cytoskeletal structures might be preserved under mechanical stress. This hypothesis requires a thorough analysis of the binding affinities between the ABD of filamin and the actin filament. On the one hand, if the on- and off-rates are high, the rod repeats would never unfold. On the other hand, our knowledge about the binding characteristics at the filament – that is, the putative associated proteins in vivo that loosen or strengthen the interaction between filamin
Table 1. Interaction partners of filamina Binding partner
Function of binding partner
Function of filamin in complex with partner
Binding site on filamin
Refs
Dictyostelium filamin FIP
Development
Membrane to actin link, signal transduction
Repeats 2–4
[58]
Filamin degradation or regulation Inflammatory and immune signalling G-protein signalling
Downregulated by FILIP
Unkown
[33]
Anchoring and receptor internalization and recycling Receptor to actin anchoring
Repeats 15–19
[35]
Repeats 14–16
[36]
Unkown
[44]
Repeats 13–16 aa 571–866 aa 867–1154 aa 1779–2284 aa 567–571
[59] [54]
[37,51]
Unknown
[60]
Repeat 24 Repeat 19 Repeats 21–22 Repeats 20–22
[38] [39] [40] [41]
Repeats 16–24 after cleavage Repeats 21–23
[16,57] [61]
Repeats 21–24 in nucleus
[55,56]
Repeat 24
[34]
Repeats 23–24 Repeats 20–23
[42] [62]
Repeats 22–24
[46]
Repeats 1–3; hinge 2 to repeat 24 Repeats 19–24 Repeat 23 Repeat 21
[45] [43,52] [47] [49,50]
Repeats 23–24
[53]
Vertebrate filamins FILIP TRAF1, TRAF2 CaR extracellular Ca2C receptor Furin
Proteolytic maturation of proteins Actin organization Transcription factor, cell differentiation
Sorting, compartmentalization and stabilization Unkown Nuclear scaffold (?)
GPIba
Platelet adhesion receptor
SHIP-2 HCN1 D2/D3 dopamine receptors Glutamate receptor type 7 Calcitonin receptor
Cell adhesion, submembrane actin remodelling Pacemaker channels Pre- or postsynaptic receptors Neurotransmitter receptor Ca2C homeostasis
Transport of GPIba from ER to cell surface Receptor to actin anchoring
Androgen receptor SEK-1
Nuclear transcription factor Kinase
BRCA-2
Tumour suppressor
RalA
GTPase
Kir2.1 Smad
KC channel TGF-b signal protein
Caveolin-1
Membrane protein, caveolae formation Signal transduction
FAP52 FOXC1
Protein kinase Ca Integrin Pak1 Migfilin PEBP2/CBF a
ECM receptor Cytoskeleton reorganization Actin remodelling, cell differentiation Transcription factor
Receptor to actin anchoring Receptor to actin anchoring Receptor to actin anchoring Anchoring and receptor internalization and recycling Downregulates AR in nucleus Tumour necrosis factor-a activation Promotes recovery from G2 arrest after DNA damage Cytoskeleton regulation, filopodia formation Receptor to actin anchoring Anchoring and phosphorylation promotion Anchoring caveolae to cytoskeleton Scaffold for signalling pathway Receptor to actin anchoring Ruffle formation Cell adhesion structure to cytoskeleton binding Retains PEBP2 in cytoplasm inhibiting its nuclear activity
Abbreviations: aa, amino acid residues; AR, androgen receptor; CaR, Ca2C-sensing receptor; ECM, extracellular matrix; ER, endoplasmic reticulum; TGF, transforming growth factor.
www.sciencedirect.com
416
Review
TRENDS in Biochemical Sciences
and F-actin – is currently limited. We cannot exclude the idea that binding proteins divide filamins into different and distinct populations that either adapt dynamically to a permanently changing network of actin filaments or are well anchored in distinct subcellular regions and sense mechanical stress by repeated folding cycles. Geometry and mechanics of F-actin crosslinking Filamin is a potent gelation factor in an actin network; in fact, only one molecule of filamin per filament of actin is needed to induce gelation [27]. Actin gels that are crosslinked by filamin show a significant increase in elasticity and a nearly constant viscosity. Because these properties are similar to those of actin polymerized at high concentrations, it has been postulated that filamin functions as an ‘actin-saving’ protein; in other words, without filamin much more actin would have to be used to build such actin networks. Other actin crosslinking proteins, such as a-actinin, do not develop these properties [28]. Vertebrate filamins have been extensively studied by electron microscopy. Human filamin seems to have a V-shaped organization and is fundamentally more flexible, probably because of its two long hinges [29], than ddFLN. It seems that actin networks built by filamin have mostly a perpendicular organization of actin filaments and that filamin acts as a bracket, holding actin filaments perpendicular to each other. The T-, X- or L-shaped junctions created by human filamin have been shown to have myosin arrows pointing towards the branch points of actin filaments [29,30]. The intensive discussion in the past about the roles of the Arp2/3 complex and filamin during the generation of branched F-actin networks has cooled down considerably with the acknowledgement that the different structures of Arp2/3 and filamin consequently lead to different architectures of the filamentous meshwork. Arp2/3-induced nucleation along an existing actin filament forms strictly oriented daughter filaments, which start polymerization from the pointed ends, are capped very early and easily break at the branching point. By contrast, filamins are frequently found at the intersections of existing filaments, where they facilitate a flexible three-dimensional, orthogonal, fairly resistant meshwork that continuously adapts to morphological changes that are typical of slow migration [31]. The presence of the first hinge region between repeats 15 and 16 has been recently reported to be essential for maintaining the viscoelastic properties of actin networks [32]. The nonlinearity of a stress strain produced at large forces, which is characteristic of living cells, can be reproduced in the in vitro actin networks only when the crosslinking filamin possesses this region. When recombinant filamin lacking this hinge is used in such experiments, the networks show linear stiffening and break at much lower stresses [32]. Filamins are not picky about binding partners In recent publications, filamins have revealed many new roles that are equally as important as their primary function of crosslinking F-actin. Vertebrate filamins have www.sciencedirect.com
Vol.31 No.7 July 2006
been found to interact with O20 proteins; many of these interactions are unexpected (Table 1). Thus, filamins have an important role in signalling towards changes in cytoskeletal rearrangements [33,34] and have a scaffolding function. This latter function has a broad range of applications: first, it anchors membrane receptors to the actin cytoskeleton, thereby facilitating precise localization and transport of the receptors [35–43] and other proteins [44]; second, it acts as a colocalization factor for signalling pathways [45]; and third, it serves as a mechanical element in caveolae and membrane ruffle formation [46,47]. It has been proved that the expression level of filamin is essential for functioning of the Ca2C-sensing receptor (CaR) signalling pathway because the SEK-1 protein downstream of the receptor is also a binding partner of filamin [48]. Thus, filamin facilitates the proper colocalization of signalling pathway proteins. Filamins also have an important role in interactions between cells and the extracellular matrix; there, together with extracellular matrix receptors, they provide an important interface between the cytoskeleton and the exoskeleton [43,49–52]. In addition, a few unusual functions of filamin have been identified. It seems that this protein might hold transcription factors in the cytoplasmic compartment, thereby preventing them from being active in the nucleus [53]. Surprisingly, filamin has been recently reported as a nuclear protein that binds to transcription factors. It has been postulated that an elongated filamin molecule might function as a kind of a nuclear ‘cytoskeleton’ for colocalization of nuclear functional complexes [54–56] participating in the nucleocytoplasmic transport [57]. The cleaved C-terminal part of human filamin, representing repeats 16–24, has been also found in the nucleus, where it downregulates the androgen receptor [16]. Two important structures of single rod repeats of human filamin A bound to cytoplasmic fragments of its Pro788 N FLN
Pro776 C FLN Figure 4. The interaction between human filamin repeat 21 and the cytoplasmic tail of integrin b7. The filamin repeat is shown as a surface representation; the bound peptide is shown in stick representation with carbon atoms coloured yellow, nitrogen atoms blue, and oxygen atoms red. Residues Pro776 to Pro788 of integrin form a b strand that expands the filaminb-sheet comprising strands C, F and G.
Review
TRENDS in Biochemical Sciences
Vol.31 No.7 July 2006
417
(Figure 4). Kiema et al. [52] have also reported that filamin cross-competes with talin for integrin. The fact that filamin has competitors for binding its ligands is important for understanding why filamin has so many interaction partners and how these interactions are regulated. Remarkably, most of the filamin-interacting proteins bind to filamin between repeats 16 and 24. The mapped binding regions of different binding partners overlap but are seldom identical. This observation has led to the hypothesis that the primary function of repeats 16–24 is to bind to large proteins such as receptors, whereas that of repeats 1–15 is to bind to a few smaller proteins that participate in signalling processes. We can thus assume that the ABD is an anchor, repeats 1–15 form a chain, and the region of repeats 16–24 is a capstan. In light of the recently identified functions of filamins, it is clear that the primary function (F-actin crosslinking) of filamins has been complemented during evolution by additional tasks (Figure 5).
interacting receptors have been recently published: namely, repeat 17 of FLNa bound to a peptide of the platelet adhesion glycoprotein GPIba [51], and repeat 21 bound to the cytoplasmic tail of integrin b7 [52]. In both of these structures, filamin uses the same structural region of the repeat domain for binding – that is, the face of the molecule that comprises b-strands C and D. Interestingly, this interface is also used in repeat 24 for dimerization. The peptides assume a fully extended conformation: their b strands are hydrogen-bonded to strand C and have several hydrophobic contacts to residues of both strands C and D of filamin (Figure 4). Although the sequence similarity is low between GPIba and the b7 integrin cytoplasmic tail, the filamin-binding site in both proteins is flanked by proline residues. On the basis of these structural similarities, the C–D interface of filamin repeats has been proposed to be a general ligand-binding area, where specificity is determined by the different hydrogen-bonding capabilities and hydrophobic interactions of the binding proteins Extracellular space Extracellular matrix Calcium receptor
Integrin Migfilin
Cytoplasm Protein kinase Cα
TRAF-2
FILIP Filamin
Caveolin-1
PEBP2 Furin Actin filaments
Nucleus
Androgen receptor FOXC1 C terminus of filamin
SEK-1 BRCA-2
Ti BS
Figure 5. The versatile functions of filamins. The main use of filamins – namely, to crosslink F-actin filaments – is supplemented by various other actions. Filamins function as membrane scaffolds for signal pathways, receptor anchoring and trafficking. They have been also shown to colocalize with transcription factors and nuclear receptors. In full or cleaved forms, they serve as nuclear ‘skeletons’ and regulators. www.sciencedirect.com
418
Review
TRENDS in Biochemical Sciences
So far, only one binding partner other than actin has been identified for ddFLN. The filamin-interacting protein (FIP), a elongated molecule of 224 kDa, is important for the developmental cycle, formation of multicellular aggregates and phototactic behaviour [58]. Its binding region includes rod repeat 4, which might support the hypothesis that unfolding of repeat 4 is important for ligand binding. Concluding remarks Filamins present a wonderful example of multifunctionality. Their primary and primal functions of crosslinking actin filaments have been supplemented during evolution by various additional tasks. Filamin and actin cooperate in the creation of exceptional mechanical and dynamical properties of the cytoskeleton. Filamin also acts as a scaffold for membrane-receptor-associated signalling proteins. The medical relevance of filamins, especially in the field of developmental malfunctions, is indisputable [7]. At the current stage of our knowledge of filamin, it seems necessary to introduce a more holistic approach to this protein. Our information on the structural properties of filamins, and vertebrate filamins in particular, is still too limited to link many of their various biochemical functions to structure. The most logical way forward for structural investigation would seem to be to study filamin complexes with other proteins and to search for a ‘filaminbinding motif or domain’ in its interaction partners. The challenge is to study the structure of filamin in complex with its target proteins and to explain the discrepancy between its structural and mechanical properties. Much can be done in this field; however, other properties of filamins, such as regulation of actin binding and interdomain orientation, must be also studied. Acknowledgements We thank Piotr Lapicz for helping with computer graphics. This research was supported by funds from the Deutsche Forschungsgemeinschaft (SFB 413) to T.A.H. and M.S. and by a grant from the Deutsche Forschungsgemeinschaft and Ko¨ln Fortune to A.A.N.
References 1 Tseng, Y. et al. (2004) The bimodal role of filamin in controlling the architecture of F-actin networks. J. Biol. Chem. 279, 1819–1826 2 Fucini, P. et al. (1997) The repeating segments of the F-actin crosslinking gelation factor (ABP120) have an immunoglobulin-like fold. Nat. Struct. Biol. 4, 223–230 3 van der Flier, A. and Sonnenberg, A. (2001) Structural and functional aspects of filamins. Biochim. Biophys. Acta 1538, 99–117 4 Pollard, T.D. et al. (1994) Structure of actin binding proteins: insights about function at atomic resolution. Annu. Rev. Cell Biol. 10, 207–249 5 Fucini, P. et al. (1999) Molecular architecture of the rod domain of the Dictyostelium gelation factor (ABP120). J. Mol. Biol. 291, 1017–1023 6 Stossel, T.P. et al. (2001) Filamins as integrators of cell mechanics and signaling. Nat. Rev. Mol. Cell Biol. 2, 138–145 7 Feng, Y. and Walsh, A. (2004) The many faces of filamin: a versatile molecular scaffold for cell motility and signaling. Nat. Cell Biol. 6, 1034–1037 8 Fox, J.W. et al. (1998) Mutations in filamin 1 prevent migration of cerebral cortical neurons in human periventricular heterotopia. Neuron 21, 1315–1325 9 Sheen, V.L. et al. (2001) Mutations in the X-linked filamin 1 gene cause periventricular nodular heterotopia in males as well as in females. Hum. Mol. Genet. 10, 1775–1783 www.sciencedirect.com
Vol.31 No.7 July 2006
10 Stefanova, M. et al. (2005) A novel 9 bp deletion in the filamin a gene causes an otopalatodigital-spectrum disorder with a variable, intermediate phenotype. Am. J. Med. Genet. A. 132, 386–390 11 Robertson, S.P. et al. (2003) Localized mutations in the gene encoding the cytoskeletal protein filamin A cause diverse malformations in humans. Nat. Genet. 33, 487–491 12 McCoy, A. et al. (1999) Structural basis for dimerization of the Dictyostelium gelation factor (ABP120) rod. Nat. Struct. Biol. 6, 836–841 13 Popowicz, G.M. et al. (2004) Molecular structure of the rod domain of Dictyostelium filamin. J. Mol. Biol. 342, 1637–1646 14 Pudas, R. et al. (2005) Structural basis for vertebrate filamin dimerization. Structure 13, 111–119 15 Himmel, M. et al. (2003) The limits of promiscuity: isoform-specific dimerization of filamins. Biochemistry 42, 430–439 16 Loy, C.J. et al. (2003) Filamin-A fragment localizes to the nucleus to regulate androgen receptor and coactivator functions. Proc. Natl. Acad. Sci. U. S. A. 100, 4562–4567 17 Gorlin, J.B. et al. (1990) Human endothelial actin-binding protein (ABP-280, nonmuscle filamin): a molecular leaf spring. J. Cell Biol. 111, 1089–1105 18 Gimona, M. et al. (2002) Functional plasticity of CH domains. FEBS Lett. 513, 98–106 19 Franzot, G. et al. (2005) The crystal structure of the actin-binding domain from a-actinin in its closed conformation: structural insight into phospholipid regulation of a-actinin. J. Mol. Biol. 348, 151–165 20 Lehman, W. et al. (2004) An open or closed case for the conformation of calponin homology domains on F-actin? J. Muscle Res. Cell Motil. 25, 351–358 21 Lebart, M.C. et al. (1993) Further characterization of the a-actinin– actin interface and comparison with filamin-binding sites on actin. J. Biol. Chem. 268, 5642–5648 22 Nakamura, F. et al. (2005) Ca2C and calmodulin regulate the binding of filamin A to actin filaments. J. Biol. Chem. 280, 32426–32433 23 Schwaiger, I. et al. (2004) A mechanical unfolding intermediate in an actin-crosslinking protein. Nat. Struct. Mol. Biol. 11, 81–85 24 Schwaiger, I. et al. (2005) The folding pathway of a fast-folding immuno-globulin domain revealed by single-molecule mechanical experiments. EMBO Rep. 6, 46–51 25 Furuike, S. et al. (2001) Mechanical unfolding of single filamin A (ABP-280) molecules detected by atomic force microscopy. FEBS Lett. 498, 72–75 26 Yamazaki, M. et al. (2002) Mechanical response of single filamin A (ABP-280) molecules and its role in the actin cytoskeleton. J. Muscle Res. Cell Motil. 23, 525–534 27 Ito, T. et al. (1992) Regulation of water flow by actin-binding proteininduced actin gelation. Biophys. J. 61, 1301–1305 28 Janssen, K.P. et al. (1996) Viscoelastic properties of F-actin solutions in the presence of normal and mutated actin-binding proteins. Arch. Biochem. Biophys. 325, 183–189 29 Tyler, J.M. et al. (1980) Structural comparison of several actin-binding macromolecules. J. Cell Biol. 85, 489–495 30 Hartwig, J.H. and DeSisto, M. (1991) The cytoskeleton of the resting human blood platelet: structure of the membrane skeleton and its attachment to actin filaments. J. Cell Biol. 112, 407–425 31 Flanagan, L.A. et al. (2001) Filamin A, the Arp2/3 complex, and the morphology and function of cortical actin filaments in human melanoma cells. J. Cell Biol. 155, 511–517 32 Gardel, M.L. et al. (2006) Prestressed F-actin networks cross-linked by hinged filamins replicate mechanical properties of cells. Proc. Natl. Acad. Sci. U. S. A. 103, 1762–1767 33 Nagano, T. et al. (2004) Filamin A and FILIP (filamin A-interacting protein) regulate cell polarity and motility in neocortical subventricular and intermediate zones during radial migration. J. Neurosci. 24, 9648–9657 34 Ohta, Y. et al. (1999) The small GTPase RalA targets filamin to induce filopodia. Proc. Natl. Acad. Sci. U. S. A. 96, 2122–2128 35 Arron, J.R. et al. (2002) Regulation of the subcellular localization of tumor necrosis factor receptor-associated factor (TRAF)2 by TRAF1 reveals mechanisms of TRAF2 signaling. J. Exp. Med. 196, 923–934 36 Awata, H. et al. (2001) Interaction of the calcium-sensing receptor and filamin, a potential scaffolding protein. J. Biol. Chem. 276, 34871–34879
Review
TRENDS in Biochemical Sciences
37 Cranmer, S.L. et al. (2005) Identification of a unique filamin A binding region within the cytoplasmic domain of glycoprotein Ib-a. Biochem. J. 387, 849–858 38 Gravante, B. et al. (2004) Interaction of the pacemaker channel HCN1 with filamin A. J. Biol. Chem. 279, 43847–43853 39 Lin, R. et al. (2001) Dopamine D2 and D3 receptors are linked to the actin cytoskeleton via interaction with filamin A. Proc. Natl. Acad. Sci. U. S. A. 98, 5258–5263 40 Enz, R. (2002) The actin-binding protein filamin-A interacts with the metabotropic glutamate receptor type 7. FEBS Lett. 514, 184–188 41 Seck, T. et al. (2003) Binding of filamin to the C-terminal tail of the calcitonin receptor controls recycling. J. Biol. Chem. 278, 10408–10416 42 Sampson, L.J. et al. (2003) Direct interaction between the actinbinding protein filamin-A and the inwardly rectifying potassium channel, Kir2.1. J. Biol. Chem. 278, 41988–41997 43 Travis, M.A. et al. (2004) Interaction of filamin A with the integrin b7 cytoplasmic domain: role of alternative splicing and phosphorylation. FEBS Lett. 569, 185–190 44 Liu, G. et al. (1997) Cytoskeletal protein ABP-280 directs the intracellular trafficking of furin and modulates proprotein processing in the endocytic pathway. J. Cell Biol. 139, 1719–1733 45 Tigges, U. et al. (2003) The F-actin cross-linking and focal adhesion protein filamin A is a ligand and in vivo substrate for protein kinase Ca. J. Biol. Chem. 278, 23561–23569 46 Stahlhut, M. and van Deurs, B. (2000) Identification of filamin as a novel ligand for caveolin-1: evidence for the organization of caveolin-1associated membrane domains by the actin cytoskeleton. Mol. Biol. Cell 11, 325–337 47 Vadlamudi, R.K. et al. (2002) Filamin is essential in actin cytoskeletal assembly mediated by p21-activated kinase 1. Nat. Cell Biol. 4, 681–690 48 Huang, C. et al. (2006) Silencing of filamin A gene expression inhibits Ca2C-sensing receptor signaling. FEBS Lett. 580, 1795–1800 49 Wu, C. (2005) Migfilin and its binding partners: from cell biology to human diseases. J. Cell Sci. 118, 659–664
Vol.31 No.7 July 2006
419
50 Tu, Y. et al. (2003) Migfilin and Mig-2 link focal adhesions to filamin and the actin cytoskeleton and function in cell shape modulation. Cell 113, 37–47 51 Nakamura, F. et al. (2006) The structure of the GPIb–filamin A complex. Blood 107, 1925–1932 52 Kiema, T. et al. (2006) The molecular basis of filamin binding to integrins and competition with talin. Mol. Cell 21, 337–347 53 Yoshida, N. et al. (2005) Filamin A-bound PEBP2b/CBFb is retained in the cytoplasm and prevented from functioning as a partner of the Runx1 transcription factor. Mol. Cell. Biol. 25, 1003–1012 54 Berry, F.B. et al. (2005) FOXC1 transcriptional regulatory activity is impaired by PBX1 in a filamin A-mediated manner. Mol. Cell. Biol. 25, 1415–1424 55 Meng, X. et al. (2004) Recovery from DNA damage-induced G2 arrest requires actin-binding protein filamin-A/actin-binding protein 280. J. Biol. Chem. 279, 6098–6105 56 Yuan, Y. and Shen, Z. (2001) Interaction with BRCA2 suggests a role for filamin-1 (hsFLNa) in DNA damage response. J. Biol. Chem. 276, 48318–48324 57 Ozanne, D.M. et al. (2000) Androgen receptor nuclear translocation is facilitated by the F-actin cross-linking protein filamin. Mol. Endocrinol. 14, 1618–1626 58 Knuth, M. et al. (2004) A novel partner for Dictyostelium filamin is an a-helical developmentally regulated protein. J. Cell Sci. 117, 5013–5022 59 Nikki, M. et al. (2002) FAP52 regulates actin organization via binding to filamin. J. Biol. Chem. 277, 11432–11440 60 Dyson, J.M. et al. (2003) SHIP-2 forms a tetrameric complex with filamin, actin, and GPIb-IX-V: localization of SHIP-2 to the activated platelet actin cytoskeleton. Blood 102, 940–948 61 Marti, A. et al. (1997) Actin-binding protein-280 binds the stressactivated protein kinase (SAPK) activator SEK-1 and is required for tumor necrosis factor-a activation of SAPK in melanoma cells. J. Biol. Chem. 272, 2620–2628 62 Sasaki, A. et al. (2001) Filamin associates with Smads and regulates transforming growth factor-b signaling. J. Biol. Chem. 276, 17871–17877
Reuse of Current Opinion and Trends journal figures in multimedia presentations It’s easy to incorporate figures published in Trends or Current Opinion journals into your PowerPoint presentations or other imagedisplay programs. Simply follow the steps below to augment your presentations or teaching materials with our fine figures! 1. 2. 3. 4. 5.
Locate the article with the required figure in the Science Direct journal collection Click on the ‘Full text + links’ hyperlink Scroll down to the thumbnail of the required figure Place the cursor over the image and click to engage the ‘Enlarge Image’ option On a PC, right-click over the expanded image and select ‘Copy’ from pull-down menu (Mac users: hold left button down and then select the ‘Copy image’ option) 6. Open a blank slide in PowerPoint or other image-display program 7. Right-click over the slide and select ‘paste’ (Mac users hit ‘Apple-V’ or select the ‘Edit-Paste’ pull-down option). Permission of the publisher, Elsevier, is required to re-use any materials in Trends or Current Opinion journals or any other works published by Elsevier. Elsevier authors can obtain permission by completing the online form available through the Copyright Information section of Elsevier’s Author Gateway at http://authors.elsevier.com/. Alternatively, readers can access the request form through Elsevier’s main web site at http://www.elsevier.com/locate/permissions.
www.sciencedirect.com
TRENDS in Biochemical Sciences
Vol.31 No.7 July 2006
Filamins: promiscuous organizers of the cytoskeleton Grzegorz M. Popowicz1, Michael Schleicher2, Angelika A. Noegel3 and Tad A. Holak1 1
Max-Planck Institut fu¨r Biochemie, 82152 Martinsried, Germany Institut fu¨r Zellbiologie der LMU, 80336 Mu¨nchen, Germany 3 Institut fu¨r Biochemie I, Zentrum Molekulare Medizin Ko¨ln, Medizinische Fakulta¨t, Universita¨t zu Ko¨ln, 50931 Ko¨ln, Germany 2
Filamins are elongated homodimeric proteins that crosslink F-actin. Each monomer chain of filamin comprises an actin-binding domain, and a rod segment consisting of six (Dictyostelium filamin) up to 24 (human filamin) highly homologous repeats of w96 amino acid residues, which adopt an immunoglobulin-like fold. Two hinges in the rod segment, together with the reversible unfolding of single repeats, might be the structural basis for the intrinsic flexibility of the actin networks generated by filamins. There are numerous filaminbinding proteins that associate, in most cases, along the repeats of the rod repeats. This rather promiscuous behaviour renders filamin a versatile scaffold between the actin network and finely tuned molecular cascades from the membrane to the cytoskeleton. Introduction The cytoskeleton provides the foundation for the spatial organization of cells and their movement. Despite our inclination to consider the cell ‘skeleton’ as a rigid structure, the real cytoskeleton is dynamic, undergoing continual reorganization and modification. In a nonmuscle cell, the actin cytoskeleton consists of a pool of globular monomeric actin (G-actin), which can reversibly polymerize into filamentous actin (F-actin) and thus markedly alter the viscosity, elasticity or mechanical resistance of the whole cell. Actin filaments are polymerized, depolymerized, crosslinked, bundled or fragmented with the help of specific proteins that either change the polymerization kinetics or stabilize the existing network by connecting single filaments. The best-studied F-actin crosslinking proteins are spectrin, fimbrin, a-actinin and filamin; these proteins are characterized by a conserved actin-binding domain (ABD), which is frequently followed by an extended rodlike domain. They can form dimers or tetramers, enabling them to bundle or to crosslink filaments. Whereas spectrin, fimbrin and a-actinin are thought to form primarily parallel actin bundles, filamins can crosslink actin filaments to form orthogonal networks to bundles depending on their concentration [1]. In filamins, the rod segment consists of highly homologous repeats of w96 amino acids that are characterized structurally by the Corresponding authors: Holak, T.A. ([email protected]), Noegel, A.A. ([email protected]). Available online 16 June 2006
presence of an immunoglobulin-like fold, in which seven b strands are arranged in an antiparallel way [2,3]. By contrast, the spectrin and a-actinin rod segments consist of a-helical regions [4]. In this review, we focus on recent progress in structural research on filamins and seek to link their physiological functions with this new information. Although filamins are found in many organisms, the best studied are the Dictyostelium discoideum filamin (ddFLN) and mammalian filamin. These two prototypical filamins have an ABD comprising two tandem calponin homology domains and a rod region built by six (in ddFLN) or 24 (in human filamin) repeats [2,5] (Figure 1). The last repeat of the rod is responsible for dimerization. In addition, human filamin has two unique long hinges positioned between repeats 15–16 (27 residues) and 23–24 (35 residues) that are postulated to be flexible [6]. The human filamin family consists of three members, filamins A, B and C, that share 70% sequence homology, except for the less homologous hinges. Recent studies show that filamins not only are mechanical linkers for actin filaments but also act as interaction partners for many proteins of great functional diversity ranging from signal transduction to transcription [7]. Furthermore, recent genetic studies have demonstrated the significance of filamins in several diseases affecting brain [7–9], bone and the cardiovascular system [10,11]. Mechanism of filamin dimerization As mentioned, dimerization of filamin is mediated by the last rod repeat. Three structures of fragments of the rod region including the dimerization region have been solved (Figure 1b): two from ddFLN, which show an identical mode of dimerization; and one from human filamin [12–14]. In ddFLN the structure of the last repeat differs from the other repeats of the rod, whereas in human filamin the repeat responsible for dimerization is similar to the other repeats. In ddFLN, repeat 6 lacks 12 residues at the N terminus and, up to the middle of the second b strand, shares no sequence homology with the other repeats. It also contains an additional strand at the C terminus of the repeat. Repeats 6 and 6 0 in the two monomers form an antiparallel dimer by b-sheet extension. The most extensive interactions are between the first strand and the penultimate strand of repeats 6 and 6 0 . The dimer
www.sciencedirect.com 0968-0004/$ - see front matter Q 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.tibs.2006.05.006
Review
412
TRENDS in Biochemical Sciences
Vol.31 No.7 July 2006
(a)
Dictyostelium discoideum
Homo sapiens Hinge 1
ABD
Hinge 2 ABD
ABD
Rod ABD
Rod
(b)
D. discoideum
Repeat 4
Repeat 5
Repeats 6
Repeat 5
Repeat 4
C
N C
N
Repeat 17
Repeats 24 N
NGPIbα
CFLN17
C
H. sapiens
NFLN17 C CGPIbα
N
(c)
C
CH1
CH2
N Figure 1. Structural properties of filamins. (a) Architecture of human filamin and Dictyostelium filamin (ddFLN). Human filamin comprises an ABD and a rod region built by 24 repeats. The rod is divided by two ‘hinge’ linkers located between repeats 15 and 16, and repeats 23 and 24. Each repeat has an immunoglobulin-like fold. The last repeat is the homodimerization domain. ddFLN has a similar ABD and a rod of six repeats. (b) Structures of fragments of the ddFLN rod segment (repeats 4–6), human filamin dimerizing repeat 24 (the monomers that build the homodimer are shown in different colours) and human filamin repeat 17 (FLN17). The structure of repeat 17 (red) was solved in complex with the receptor GPIba (blue), highlighting a common binding motif of filamin-associated proteins. (c) Ribbon plot of an ABD from a-actinin [19]. The ABD is composed of two calponin homology domains, CH1 and CH2. On the basis of sequence similarity, filamin is expected to have a similar domain at its N terminus.
˚ (of which 69% is the interface is large, comprising w4500A primary dimerization contribution of the repeat and 22% is the contact surface of a buried linker between repeats 5 and 6). Repeats 6 and 6 0 form a large b-sandwich of six strands on one side and eight on the other (Figures 1, 2b). The N-terminal regions of both structural units embrace neighbouring repeats and are also deeply buried between their strand-interconnecting loops. Such organiz ation of www.sciencedirect.com
the linker brings the fifth repeat closer to the sixth repeat of the second molecule than to the sixth repeat of the same chain. This arrangement provides a high degree of rigidity between repeats 5 and 6. The structures of the dimerization repeats of ddFLN and human filamin are similar: the root mean square ˚ (Figure 2). This deviation for the core parts is only 1.2 A similarity refers particularly to the global arrangement of
Review
TRENDS in Biochemical Sciences
(a) Ndd
Chs
Cdd
Chs
Ndd
Cdd
Nhs
Nhs
(b) Nhs
Nhs
Ndd
Ndd
Chs
Chs Cdd
Cdd
Cdd
Cdd
Chs
Chs Ndd
Ndd Nhs
Nhs
Figure 2. Structural comparison of the dimerization repeats of ddFLN and human filamin. (a) Stereo view of monomer superimposition shows that the cores have basically identical folds but there are differences at the N- and C-terminal ends of the repeats. ddFLN (red) has an additional long strand at the C terminus, whereas human filamin (blue) has two additional short strands at the N terminus on the other side of the molecule. (b) Stereo view of homodimer superimposition on the monomer structures shown in (a). Despite the structural similarity in the repeats of human filamin and ddFLN, the dimerization area is different. The interacting surfaces are localized at opposite sides of the repeats. There is no interaction between the N-terminal linker of the dimerization repeat and the second molecule in the dimer in vertebrate filamin.
dimerization; for example, the repeats in human filamin are arranged with an exact two-fold symmetry as they are in ddFLN. Furthermore, dimers of human filamin also form through the extension of a b sheet of each monomer to form an elongated b sandwich. However, the strands responsible for direct interaction differ in human filamin, and the internal organization of the strands also varies from that in ddFLN. Differences affecting dimerization are found at the C and N termini of the repeats: the dimerizing strands of the repeat seem to be the third (C) and fourth (D) b strands in human filamin, and the first (A) and sixth (F) b strand in ddFLN. The lengths of the interacting strands are shorter in human filamin than in ddFLN. In addition, there is no buried linker at the N terminus of the repeat in human filamin. A long hinge region of 35 residues precedes repeat 24 in human filamin. The structure of this region is unknown and shows the greatest sequence divergence. The main function of this region is probably mechanical, enabling a high degree of flexibility and structural adaptability [14]. One study has indicated that this region might have a regulatory function for dimerization [15]. The interface of dimerization is much smaller in human filamin than in ddFLN: the buried surface area ˚ 2, taking up 19% of the molecular surface. The is 1109 A crystallographic model is supported by mutagenesis studies, which exclude the possibility of artefacts and www.sciencedirect.com
Vol.31 No.7 July 2006
413
prove that the dimerization mechanism of the wild-type vertebrate filamin is altered and presumably weaker than that of ddFLN [14]. Overall, the comparison between ddFLN and human filamin reveals that the dimerization interfaces in filamins have changed during evolution, although the same mechanism of a double-sided b-sheet extension of immunoglobulin-like repeats has remained. The rigid and strong interaction in filamins of lower organisms has been replaced by more sophisticated, weaker, and more flexible dimerization interfaces in higher organisms. Future studies on longer constructs of filamins are expected to provide information on the spatial organization of their repeats and their rigidity. Inter-repeat organization The mechanical properties of the whole rod region are essential for both the F-actin crosslinking capabilities of filamin and the characteristics of the F-actin networks that they create. Except for the two hinge regions in human filamin, the linkers between the repeats are also short in vertebrate filamins: they comprise only a few residues and are rich in proline. Additional salt bridges have been identified between neighbouring repeats. The two structures of the ddFLN multirepeat constructs show that the repeats of the rod remain in the same spatial position regardless of crystal packing and crystallization conditions. The structures maintain an exact two-fold symmetry. The root mean square deviation ˚ ; thus, the whole rod between the structures is only 1.3 A seems to be a long, extended, spring-like molecule with an inter-repeat tilt angle of w1158. Such conformations support the proposition that in ddFLN the whole rod is an extended structure of limited flexibility. Of course, some flexibility must be provided to facilitate crosslinking of actin filaments and adaptation of the F-actin network generated. This particular feature might be realized by a long linker between the ABD and the rod [13] (Figure 3). It seems that the primary function of ddFLN in Dictyostelium is to reduce the degree of freedom of the bound actin filaments while facilitating the organization of an orthogonal and highly flexible filamentous network. There are no structural data on the inter-repeat organization of vertebrate filamins; however, the rod segment of human filamin, which is almost five times longer than that of ddFLN, is likely to be organized in the same way. Remarkable sequence similarities support such a model; in addition, the two long hinge regions in human filamin have a larger amino acid diversity and their length would enable the whole rod to be more flexible and to crosslink actin filaments in many orientations. The hinges might also present potential proteolytic cleavage sites. A naturally occurring proteolytic fragment of human filamin (FLNa; repeats 16–24), generated by cleavage at the protease-cleavage site between repeats 15 and 16, translocates to the nucleus and acts as a nuclear transcription modulator [16]. However, a region before repeat 24, which is considered to be flexible [17], has been reported to regulate dimerization [15]. In addition to its mechanical function,
414
Review
TRENDS in Biochemical Sciences
Actin-binding domain
Rod repeats
Repeats 6 building a dimer
Rod repeats
Actin-binding domain
Figure 3. Model of the whole ddFLN molecule as a homodimer. The model is built on the basis of known parts of the structure. Actin-binding domains are green; rod regions are red and blue. The extended molecule can bind actin filaments over a distance of w400 A˚ [13]. Reproduced, with permission, from Ref. [13].
the rod region of vertebrate filamins has been shown to be an important scaffold for binding many proteins involved in a broad range of cellular processes [6,7]. www.sciencedirect.com
Vol.31 No.7 July 2006
The actin-binding domain Members of the a-actinin superfamily of actin crosslinking proteins share the same mechanism of interaction with F-actin: filamins, a-actinin, spectrin, nesprins, plectin, dystrophin and fimbrin use a similar motif of two calponin homology domains to construct their ABDs [18]. A typical ABD has w250 residues and shares 20–60% of sequence identity with other ABDs in the family. Structures of the ABD of a-actinin show that the two subdomains form a fully a-helical globular domain [19] (Figure 1b). The subdomains seem to have either a closed or open conformation, depending on the contact area between them. These conformations present various possible subdomain organizations [20], and it is not known which one is present in the ABDs of filamins. The interface between F-actin and a-actinin has been studied extensively with peptides and antibodies of known specificities. The primary binding site for F-actin in the ABD of a-actinin is located between residues 121 and 147. The interaction is mostly hydrophobic, but the susceptibility of the binding to ionic strength suggests that hydrophilic interactions are also involved. The amino acids of actin that participate in binding are located between residues 112–125 and 360–372 [21]. A recent report suggests that binding of the ABD of FLNa to actin is regulated by calmodulin, an abundant Ca2C-binding protein. Binding is facilitated by Ca2C and calmodulin together: in other words, neither Ca2C ions nor calmodulin alone show regulatory properties. The calmodulin-binding region in the ABD has been mapped to a region between residues 50 and 96 in the first calponin homology domain. An interaction between calmodulin and the whole ABD is not observed, however, which might suggest that calmodulin uses a cryptic binding site in the ABD that is uncovered only after the ABD binds to the actin filament [22]. Unfolding of the ddFLN rod Detailed studies using atomic force microscopy have been recently carried out to obtain more information on the mechanical properties of the rod repeats in ddFLN. These experiments indicate that individual repeats unfold before the dimer is broken. To break the dimer, a force of w200 pN is necessary. Notably, the fourth repeat shows a pattern of unfolding that differs from that of all other repeats. It unfolds the most easily and seems to be the only repeat that has a stable folding intermediate. In the first stage of unfolding, w40 residues are stripped from the molecule and the remaining 60 stay folded. The former 40 residues correspond to the first two b strands [23]. The intermediate is a stable structure that can fold on itself. The two-stage folding of the fourth repeat is also the fastest process of folding observed for the rod region [24]. The biological significance of this feature is potentially interesting. Easy unfolding and fast refolding of repeat 4 would enable the whole rod to nearly double its length and then come back to its native state. It is also possible that the folding speed of an elongated molecule would be increased by the presence of the intermediate. The freeenergy barriers between the unfolded state and the intermediate, and, in the next step, between the
Review
TRENDS in Biochemical Sciences
intermediate and a fully folded protein, are easier to break than the barrier for direct transition to the folded state. Another interesting feature of this unfolding mechanism is linked to proteins that can bind to the rod region in the region of repeat 4. Mechanical stress might cause unfolding of the repeat together with dissociation of the binding partner or association of a new one to the intermediate. The bound or released interacting molecules might be part of a signalling pathway that could function as a cellular ‘sensor’ of mechanical forces present in the cytoskeleton. Experiments investigating vertebrate filamin mechanics have been done on the whole filamin molecule and have shown that different repeats unfold under different force [25,26]. To unfold some of the strongest repeats requires a force more than three times greater than that needed to unfold the weakest ones. It has been
Vol.31 No.7 July 2006
415
also shown that repeats can refold spontaneously without stress. The authors of these studies [25,26] postulated that eventual unfolding of the filamin rod during cytoskeletal stresses is a method to protect the whole architecture from being mechanically separated. Formation of a new complex between filamin and its putative binding partner would be much slower and less likely to occur than a folding cycle. Under these circumstances, the spatial organization of cytoskeletal structures might be preserved under mechanical stress. This hypothesis requires a thorough analysis of the binding affinities between the ABD of filamin and the actin filament. On the one hand, if the on- and off-rates are high, the rod repeats would never unfold. On the other hand, our knowledge about the binding characteristics at the filament – that is, the putative associated proteins in vivo that loosen or strengthen the interaction between filamin
Table 1. Interaction partners of filamina Binding partner
Function of binding partner
Function of filamin in complex with partner
Binding site on filamin
Refs
Dictyostelium filamin FIP
Development
Membrane to actin link, signal transduction
Repeats 2–4
[58]
Filamin degradation or regulation Inflammatory and immune signalling G-protein signalling
Downregulated by FILIP
Unkown
[33]
Anchoring and receptor internalization and recycling Receptor to actin anchoring
Repeats 15–19
[35]
Repeats 14–16
[36]
Unkown
[44]
Repeats 13–16 aa 571–866 aa 867–1154 aa 1779–2284 aa 567–571
[59] [54]
[37,51]
Unknown
[60]
Repeat 24 Repeat 19 Repeats 21–22 Repeats 20–22
[38] [39] [40] [41]
Repeats 16–24 after cleavage Repeats 21–23
[16,57] [61]
Repeats 21–24 in nucleus
[55,56]
Repeat 24
[34]
Repeats 23–24 Repeats 20–23
[42] [62]
Repeats 22–24
[46]
Repeats 1–3; hinge 2 to repeat 24 Repeats 19–24 Repeat 23 Repeat 21
[45] [43,52] [47] [49,50]
Repeats 23–24
[53]
Vertebrate filamins FILIP TRAF1, TRAF2 CaR extracellular Ca2C receptor Furin
Proteolytic maturation of proteins Actin organization Transcription factor, cell differentiation
Sorting, compartmentalization and stabilization Unkown Nuclear scaffold (?)
GPIba
Platelet adhesion receptor
SHIP-2 HCN1 D2/D3 dopamine receptors Glutamate receptor type 7 Calcitonin receptor
Cell adhesion, submembrane actin remodelling Pacemaker channels Pre- or postsynaptic receptors Neurotransmitter receptor Ca2C homeostasis
Transport of GPIba from ER to cell surface Receptor to actin anchoring
Androgen receptor SEK-1
Nuclear transcription factor Kinase
BRCA-2
Tumour suppressor
RalA
GTPase
Kir2.1 Smad
KC channel TGF-b signal protein
Caveolin-1
Membrane protein, caveolae formation Signal transduction
FAP52 FOXC1
Protein kinase Ca Integrin Pak1 Migfilin PEBP2/CBF a
ECM receptor Cytoskeleton reorganization Actin remodelling, cell differentiation Transcription factor
Receptor to actin anchoring Receptor to actin anchoring Receptor to actin anchoring Anchoring and receptor internalization and recycling Downregulates AR in nucleus Tumour necrosis factor-a activation Promotes recovery from G2 arrest after DNA damage Cytoskeleton regulation, filopodia formation Receptor to actin anchoring Anchoring and phosphorylation promotion Anchoring caveolae to cytoskeleton Scaffold for signalling pathway Receptor to actin anchoring Ruffle formation Cell adhesion structure to cytoskeleton binding Retains PEBP2 in cytoplasm inhibiting its nuclear activity
Abbreviations: aa, amino acid residues; AR, androgen receptor; CaR, Ca2C-sensing receptor; ECM, extracellular matrix; ER, endoplasmic reticulum; TGF, transforming growth factor.
www.sciencedirect.com
416
Review
TRENDS in Biochemical Sciences
and F-actin – is currently limited. We cannot exclude the idea that binding proteins divide filamins into different and distinct populations that either adapt dynamically to a permanently changing network of actin filaments or are well anchored in distinct subcellular regions and sense mechanical stress by repeated folding cycles. Geometry and mechanics of F-actin crosslinking Filamin is a potent gelation factor in an actin network; in fact, only one molecule of filamin per filament of actin is needed to induce gelation [27]. Actin gels that are crosslinked by filamin show a significant increase in elasticity and a nearly constant viscosity. Because these properties are similar to those of actin polymerized at high concentrations, it has been postulated that filamin functions as an ‘actin-saving’ protein; in other words, without filamin much more actin would have to be used to build such actin networks. Other actin crosslinking proteins, such as a-actinin, do not develop these properties [28]. Vertebrate filamins have been extensively studied by electron microscopy. Human filamin seems to have a V-shaped organization and is fundamentally more flexible, probably because of its two long hinges [29], than ddFLN. It seems that actin networks built by filamin have mostly a perpendicular organization of actin filaments and that filamin acts as a bracket, holding actin filaments perpendicular to each other. The T-, X- or L-shaped junctions created by human filamin have been shown to have myosin arrows pointing towards the branch points of actin filaments [29,30]. The intensive discussion in the past about the roles of the Arp2/3 complex and filamin during the generation of branched F-actin networks has cooled down considerably with the acknowledgement that the different structures of Arp2/3 and filamin consequently lead to different architectures of the filamentous meshwork. Arp2/3-induced nucleation along an existing actin filament forms strictly oriented daughter filaments, which start polymerization from the pointed ends, are capped very early and easily break at the branching point. By contrast, filamins are frequently found at the intersections of existing filaments, where they facilitate a flexible three-dimensional, orthogonal, fairly resistant meshwork that continuously adapts to morphological changes that are typical of slow migration [31]. The presence of the first hinge region between repeats 15 and 16 has been recently reported to be essential for maintaining the viscoelastic properties of actin networks [32]. The nonlinearity of a stress strain produced at large forces, which is characteristic of living cells, can be reproduced in the in vitro actin networks only when the crosslinking filamin possesses this region. When recombinant filamin lacking this hinge is used in such experiments, the networks show linear stiffening and break at much lower stresses [32]. Filamins are not picky about binding partners In recent publications, filamins have revealed many new roles that are equally as important as their primary function of crosslinking F-actin. Vertebrate filamins have www.sciencedirect.com
Vol.31 No.7 July 2006
been found to interact with O20 proteins; many of these interactions are unexpected (Table 1). Thus, filamins have an important role in signalling towards changes in cytoskeletal rearrangements [33,34] and have a scaffolding function. This latter function has a broad range of applications: first, it anchors membrane receptors to the actin cytoskeleton, thereby facilitating precise localization and transport of the receptors [35–43] and other proteins [44]; second, it acts as a colocalization factor for signalling pathways [45]; and third, it serves as a mechanical element in caveolae and membrane ruffle formation [46,47]. It has been proved that the expression level of filamin is essential for functioning of the Ca2C-sensing receptor (CaR) signalling pathway because the SEK-1 protein downstream of the receptor is also a binding partner of filamin [48]. Thus, filamin facilitates the proper colocalization of signalling pathway proteins. Filamins also have an important role in interactions between cells and the extracellular matrix; there, together with extracellular matrix receptors, they provide an important interface between the cytoskeleton and the exoskeleton [43,49–52]. In addition, a few unusual functions of filamin have been identified. It seems that this protein might hold transcription factors in the cytoplasmic compartment, thereby preventing them from being active in the nucleus [53]. Surprisingly, filamin has been recently reported as a nuclear protein that binds to transcription factors. It has been postulated that an elongated filamin molecule might function as a kind of a nuclear ‘cytoskeleton’ for colocalization of nuclear functional complexes [54–56] participating in the nucleocytoplasmic transport [57]. The cleaved C-terminal part of human filamin, representing repeats 16–24, has been also found in the nucleus, where it downregulates the androgen receptor [16]. Two important structures of single rod repeats of human filamin A bound to cytoplasmic fragments of its Pro788 N FLN
Pro776 C FLN Figure 4. The interaction between human filamin repeat 21 and the cytoplasmic tail of integrin b7. The filamin repeat is shown as a surface representation; the bound peptide is shown in stick representation with carbon atoms coloured yellow, nitrogen atoms blue, and oxygen atoms red. Residues Pro776 to Pro788 of integrin form a b strand that expands the filaminb-sheet comprising strands C, F and G.
Review
TRENDS in Biochemical Sciences
Vol.31 No.7 July 2006
417
(Figure 4). Kiema et al. [52] have also reported that filamin cross-competes with talin for integrin. The fact that filamin has competitors for binding its ligands is important for understanding why filamin has so many interaction partners and how these interactions are regulated. Remarkably, most of the filamin-interacting proteins bind to filamin between repeats 16 and 24. The mapped binding regions of different binding partners overlap but are seldom identical. This observation has led to the hypothesis that the primary function of repeats 16–24 is to bind to large proteins such as receptors, whereas that of repeats 1–15 is to bind to a few smaller proteins that participate in signalling processes. We can thus assume that the ABD is an anchor, repeats 1–15 form a chain, and the region of repeats 16–24 is a capstan. In light of the recently identified functions of filamins, it is clear that the primary function (F-actin crosslinking) of filamins has been complemented during evolution by additional tasks (Figure 5).
interacting receptors have been recently published: namely, repeat 17 of FLNa bound to a peptide of the platelet adhesion glycoprotein GPIba [51], and repeat 21 bound to the cytoplasmic tail of integrin b7 [52]. In both of these structures, filamin uses the same structural region of the repeat domain for binding – that is, the face of the molecule that comprises b-strands C and D. Interestingly, this interface is also used in repeat 24 for dimerization. The peptides assume a fully extended conformation: their b strands are hydrogen-bonded to strand C and have several hydrophobic contacts to residues of both strands C and D of filamin (Figure 4). Although the sequence similarity is low between GPIba and the b7 integrin cytoplasmic tail, the filamin-binding site in both proteins is flanked by proline residues. On the basis of these structural similarities, the C–D interface of filamin repeats has been proposed to be a general ligand-binding area, where specificity is determined by the different hydrogen-bonding capabilities and hydrophobic interactions of the binding proteins Extracellular space Extracellular matrix Calcium receptor
Integrin Migfilin
Cytoplasm Protein kinase Cα
TRAF-2
FILIP Filamin
Caveolin-1
PEBP2 Furin Actin filaments
Nucleus
Androgen receptor FOXC1 C terminus of filamin
SEK-1 BRCA-2
Ti BS
Figure 5. The versatile functions of filamins. The main use of filamins – namely, to crosslink F-actin filaments – is supplemented by various other actions. Filamins function as membrane scaffolds for signal pathways, receptor anchoring and trafficking. They have been also shown to colocalize with transcription factors and nuclear receptors. In full or cleaved forms, they serve as nuclear ‘skeletons’ and regulators. www.sciencedirect.com
418
Review
TRENDS in Biochemical Sciences
So far, only one binding partner other than actin has been identified for ddFLN. The filamin-interacting protein (FIP), a elongated molecule of 224 kDa, is important for the developmental cycle, formation of multicellular aggregates and phototactic behaviour [58]. Its binding region includes rod repeat 4, which might support the hypothesis that unfolding of repeat 4 is important for ligand binding. Concluding remarks Filamins present a wonderful example of multifunctionality. Their primary and primal functions of crosslinking actin filaments have been supplemented during evolution by various additional tasks. Filamin and actin cooperate in the creation of exceptional mechanical and dynamical properties of the cytoskeleton. Filamin also acts as a scaffold for membrane-receptor-associated signalling proteins. The medical relevance of filamins, especially in the field of developmental malfunctions, is indisputable [7]. At the current stage of our knowledge of filamin, it seems necessary to introduce a more holistic approach to this protein. Our information on the structural properties of filamins, and vertebrate filamins in particular, is still too limited to link many of their various biochemical functions to structure. The most logical way forward for structural investigation would seem to be to study filamin complexes with other proteins and to search for a ‘filaminbinding motif or domain’ in its interaction partners. The challenge is to study the structure of filamin in complex with its target proteins and to explain the discrepancy between its structural and mechanical properties. Much can be done in this field; however, other properties of filamins, such as regulation of actin binding and interdomain orientation, must be also studied. Acknowledgements We thank Piotr Lapicz for helping with computer graphics. This research was supported by funds from the Deutsche Forschungsgemeinschaft (SFB 413) to T.A.H. and M.S. and by a grant from the Deutsche Forschungsgemeinschaft and Ko¨ln Fortune to A.A.N.
References 1 Tseng, Y. et al. (2004) The bimodal role of filamin in controlling the architecture of F-actin networks. J. Biol. Chem. 279, 1819–1826 2 Fucini, P. et al. (1997) The repeating segments of the F-actin crosslinking gelation factor (ABP120) have an immunoglobulin-like fold. Nat. Struct. Biol. 4, 223–230 3 van der Flier, A. and Sonnenberg, A. (2001) Structural and functional aspects of filamins. Biochim. Biophys. Acta 1538, 99–117 4 Pollard, T.D. et al. (1994) Structure of actin binding proteins: insights about function at atomic resolution. Annu. Rev. Cell Biol. 10, 207–249 5 Fucini, P. et al. (1999) Molecular architecture of the rod domain of the Dictyostelium gelation factor (ABP120). J. Mol. Biol. 291, 1017–1023 6 Stossel, T.P. et al. (2001) Filamins as integrators of cell mechanics and signaling. Nat. Rev. Mol. Cell Biol. 2, 138–145 7 Feng, Y. and Walsh, A. (2004) The many faces of filamin: a versatile molecular scaffold for cell motility and signaling. Nat. Cell Biol. 6, 1034–1037 8 Fox, J.W. et al. (1998) Mutations in filamin 1 prevent migration of cerebral cortical neurons in human periventricular heterotopia. Neuron 21, 1315–1325 9 Sheen, V.L. et al. (2001) Mutations in the X-linked filamin 1 gene cause periventricular nodular heterotopia in males as well as in females. Hum. Mol. Genet. 10, 1775–1783 www.sciencedirect.com
Vol.31 No.7 July 2006
10 Stefanova, M. et al. (2005) A novel 9 bp deletion in the filamin a gene causes an otopalatodigital-spectrum disorder with a variable, intermediate phenotype. Am. J. Med. Genet. A. 132, 386–390 11 Robertson, S.P. et al. (2003) Localized mutations in the gene encoding the cytoskeletal protein filamin A cause diverse malformations in humans. Nat. Genet. 33, 487–491 12 McCoy, A. et al. (1999) Structural basis for dimerization of the Dictyostelium gelation factor (ABP120) rod. Nat. Struct. Biol. 6, 836–841 13 Popowicz, G.M. et al. (2004) Molecular structure of the rod domain of Dictyostelium filamin. J. Mol. Biol. 342, 1637–1646 14 Pudas, R. et al. (2005) Structural basis for vertebrate filamin dimerization. Structure 13, 111–119 15 Himmel, M. et al. (2003) The limits of promiscuity: isoform-specific dimerization of filamins. Biochemistry 42, 430–439 16 Loy, C.J. et al. (2003) Filamin-A fragment localizes to the nucleus to regulate androgen receptor and coactivator functions. Proc. Natl. Acad. Sci. U. S. A. 100, 4562–4567 17 Gorlin, J.B. et al. (1990) Human endothelial actin-binding protein (ABP-280, nonmuscle filamin): a molecular leaf spring. J. Cell Biol. 111, 1089–1105 18 Gimona, M. et al. (2002) Functional plasticity of CH domains. FEBS Lett. 513, 98–106 19 Franzot, G. et al. (2005) The crystal structure of the actin-binding domain from a-actinin in its closed conformation: structural insight into phospholipid regulation of a-actinin. J. Mol. Biol. 348, 151–165 20 Lehman, W. et al. (2004) An open or closed case for the conformation of calponin homology domains on F-actin? J. Muscle Res. Cell Motil. 25, 351–358 21 Lebart, M.C. et al. (1993) Further characterization of the a-actinin– actin interface and comparison with filamin-binding sites on actin. J. Biol. Chem. 268, 5642–5648 22 Nakamura, F. et al. (2005) Ca2C and calmodulin regulate the binding of filamin A to actin filaments. J. Biol. Chem. 280, 32426–32433 23 Schwaiger, I. et al. (2004) A mechanical unfolding intermediate in an actin-crosslinking protein. Nat. Struct. Mol. Biol. 11, 81–85 24 Schwaiger, I. et al. (2005) The folding pathway of a fast-folding immuno-globulin domain revealed by single-molecule mechanical experiments. EMBO Rep. 6, 46–51 25 Furuike, S. et al. (2001) Mechanical unfolding of single filamin A (ABP-280) molecules detected by atomic force microscopy. FEBS Lett. 498, 72–75 26 Yamazaki, M. et al. (2002) Mechanical response of single filamin A (ABP-280) molecules and its role in the actin cytoskeleton. J. Muscle Res. Cell Motil. 23, 525–534 27 Ito, T. et al. (1992) Regulation of water flow by actin-binding proteininduced actin gelation. Biophys. J. 61, 1301–1305 28 Janssen, K.P. et al. (1996) Viscoelastic properties of F-actin solutions in the presence of normal and mutated actin-binding proteins. Arch. Biochem. Biophys. 325, 183–189 29 Tyler, J.M. et al. (1980) Structural comparison of several actin-binding macromolecules. J. Cell Biol. 85, 489–495 30 Hartwig, J.H. and DeSisto, M. (1991) The cytoskeleton of the resting human blood platelet: structure of the membrane skeleton and its attachment to actin filaments. J. Cell Biol. 112, 407–425 31 Flanagan, L.A. et al. (2001) Filamin A, the Arp2/3 complex, and the morphology and function of cortical actin filaments in human melanoma cells. J. Cell Biol. 155, 511–517 32 Gardel, M.L. et al. (2006) Prestressed F-actin networks cross-linked by hinged filamins replicate mechanical properties of cells. Proc. Natl. Acad. Sci. U. S. A. 103, 1762–1767 33 Nagano, T. et al. (2004) Filamin A and FILIP (filamin A-interacting protein) regulate cell polarity and motility in neocortical subventricular and intermediate zones during radial migration. J. Neurosci. 24, 9648–9657 34 Ohta, Y. et al. (1999) The small GTPase RalA targets filamin to induce filopodia. Proc. Natl. Acad. Sci. U. S. A. 96, 2122–2128 35 Arron, J.R. et al. (2002) Regulation of the subcellular localization of tumor necrosis factor receptor-associated factor (TRAF)2 by TRAF1 reveals mechanisms of TRAF2 signaling. J. Exp. Med. 196, 923–934 36 Awata, H. et al. (2001) Interaction of the calcium-sensing receptor and filamin, a potential scaffolding protein. J. Biol. Chem. 276, 34871–34879
Review
TRENDS in Biochemical Sciences
37 Cranmer, S.L. et al. (2005) Identification of a unique filamin A binding region within the cytoplasmic domain of glycoprotein Ib-a. Biochem. J. 387, 849–858 38 Gravante, B. et al. (2004) Interaction of the pacemaker channel HCN1 with filamin A. J. Biol. Chem. 279, 43847–43853 39 Lin, R. et al. (2001) Dopamine D2 and D3 receptors are linked to the actin cytoskeleton via interaction with filamin A. Proc. Natl. Acad. Sci. U. S. A. 98, 5258–5263 40 Enz, R. (2002) The actin-binding protein filamin-A interacts with the metabotropic glutamate receptor type 7. FEBS Lett. 514, 184–188 41 Seck, T. et al. (2003) Binding of filamin to the C-terminal tail of the calcitonin receptor controls recycling. J. Biol. Chem. 278, 10408–10416 42 Sampson, L.J. et al. (2003) Direct interaction between the actinbinding protein filamin-A and the inwardly rectifying potassium channel, Kir2.1. J. Biol. Chem. 278, 41988–41997 43 Travis, M.A. et al. (2004) Interaction of filamin A with the integrin b7 cytoplasmic domain: role of alternative splicing and phosphorylation. FEBS Lett. 569, 185–190 44 Liu, G. et al. (1997) Cytoskeletal protein ABP-280 directs the intracellular trafficking of furin and modulates proprotein processing in the endocytic pathway. J. Cell Biol. 139, 1719–1733 45 Tigges, U. et al. (2003) The F-actin cross-linking and focal adhesion protein filamin A is a ligand and in vivo substrate for protein kinase Ca. J. Biol. Chem. 278, 23561–23569 46 Stahlhut, M. and van Deurs, B. (2000) Identification of filamin as a novel ligand for caveolin-1: evidence for the organization of caveolin-1associated membrane domains by the actin cytoskeleton. Mol. Biol. Cell 11, 325–337 47 Vadlamudi, R.K. et al. (2002) Filamin is essential in actin cytoskeletal assembly mediated by p21-activated kinase 1. Nat. Cell Biol. 4, 681–690 48 Huang, C. et al. (2006) Silencing of filamin A gene expression inhibits Ca2C-sensing receptor signaling. FEBS Lett. 580, 1795–1800 49 Wu, C. (2005) Migfilin and its binding partners: from cell biology to human diseases. J. Cell Sci. 118, 659–664
Vol.31 No.7 July 2006
419
50 Tu, Y. et al. (2003) Migfilin and Mig-2 link focal adhesions to filamin and the actin cytoskeleton and function in cell shape modulation. Cell 113, 37–47 51 Nakamura, F. et al. (2006) The structure of the GPIb–filamin A complex. Blood 107, 1925–1932 52 Kiema, T. et al. (2006) The molecular basis of filamin binding to integrins and competition with talin. Mol. Cell 21, 337–347 53 Yoshida, N. et al. (2005) Filamin A-bound PEBP2b/CBFb is retained in the cytoplasm and prevented from functioning as a partner of the Runx1 transcription factor. Mol. Cell. Biol. 25, 1003–1012 54 Berry, F.B. et al. (2005) FOXC1 transcriptional regulatory activity is impaired by PBX1 in a filamin A-mediated manner. Mol. Cell. Biol. 25, 1415–1424 55 Meng, X. et al. (2004) Recovery from DNA damage-induced G2 arrest requires actin-binding protein filamin-A/actin-binding protein 280. J. Biol. Chem. 279, 6098–6105 56 Yuan, Y. and Shen, Z. (2001) Interaction with BRCA2 suggests a role for filamin-1 (hsFLNa) in DNA damage response. J. Biol. Chem. 276, 48318–48324 57 Ozanne, D.M. et al. (2000) Androgen receptor nuclear translocation is facilitated by the F-actin cross-linking protein filamin. Mol. Endocrinol. 14, 1618–1626 58 Knuth, M. et al. (2004) A novel partner for Dictyostelium filamin is an a-helical developmentally regulated protein. J. Cell Sci. 117, 5013–5022 59 Nikki, M. et al. (2002) FAP52 regulates actin organization via binding to filamin. J. Biol. Chem. 277, 11432–11440 60 Dyson, J.M. et al. (2003) SHIP-2 forms a tetrameric complex with filamin, actin, and GPIb-IX-V: localization of SHIP-2 to the activated platelet actin cytoskeleton. Blood 102, 940–948 61 Marti, A. et al. (1997) Actin-binding protein-280 binds the stressactivated protein kinase (SAPK) activator SEK-1 and is required for tumor necrosis factor-a activation of SAPK in melanoma cells. J. Biol. Chem. 272, 2620–2628 62 Sasaki, A. et al. (2001) Filamin associates with Smads and regulates transforming growth factor-b signaling. J. Biol. Chem. 276, 17871–17877
Reuse of Current Opinion and Trends journal figures in multimedia presentations It’s easy to incorporate figures published in Trends or Current Opinion journals into your PowerPoint presentations or other imagedisplay programs. Simply follow the steps below to augment your presentations or teaching materials with our fine figures! 1. 2. 3. 4. 5.
Locate the article with the required figure in the Science Direct journal collection Click on the ‘Full text + links’ hyperlink Scroll down to the thumbnail of the required figure Place the cursor over the image and click to engage the ‘Enlarge Image’ option On a PC, right-click over the expanded image and select ‘Copy’ from pull-down menu (Mac users: hold left button down and then select the ‘Copy image’ option) 6. Open a blank slide in PowerPoint or other image-display program 7. Right-click over the slide and select ‘paste’ (Mac users hit ‘Apple-V’ or select the ‘Edit-Paste’ pull-down option). Permission of the publisher, Elsevier, is required to re-use any materials in Trends or Current Opinion journals or any other works published by Elsevier. Elsevier authors can obtain permission by completing the online form available through the Copyright Information section of Elsevier’s Author Gateway at http://authors.elsevier.com/. Alternatively, readers can access the request form through Elsevier’s main web site at http://www.elsevier.com/locate/permissions.
www.sciencedirect.com