Structural basis for dimerization of the Dictyostelium gelation factor (ABP120) rod.

Gelation factor (ABP120) is one of the principal actin-cross-linking proteins of Dictyostelium discoideum. The extended molecule has an N-terminal 250-residue actin-binding domain and a rod constructed from six 100-residue repeats that have an Ig fold. The ability to dimerize is crucial to the actin cross-linking function of gelation factor and is mediated by the rod in which the two chains are arranged in an antiparallel fashion. We report the 2.2 A resolution crystal structure of rod domains 5 and 6, which shows that dimerization is mediated primarily by rod domain 6 and is the result of a double edge-to-edge extension of beta-sheets. Thus, contrary to earlier proposals, the chains of the dimeric gelation factor molecule overlap only within domain 6, and domains 1-5 do not pair with domains from the other chain. This information allows construction of a model of the gelation factor molecule and suggests how the chains in the related molecule filamin (ABP280) may interact.     

The Dictyostelium gelation factor shares a putative actin binding site with alpha-actinins and dystrophin and also has a rod domain containing six 100-residue motifs that appear to have a cross-beta conformation

The 120-kD gelation factor and alpha-actinin are among the most abundant F-actin cross-linking proteins in Dictyostelium discoideum. Both molecules are homodimers and have extended rod-like configurations that are respectively approximately 35 and 40 nm long. Here we report the complete cDNA sequence of the 120-kD gelation factor which codes for a protein of 857 amino acids. Its calculated molecular mass is 92.2 kD which is considerably smaller than suggested by its mobility in SDS-PAGE. Analysis of the sequence shows a region that is highly homologous to D. discoideum alpha-actinin, chicken fibroblast alpha-actinin, and human dystrophin. This conserved domain probably represents an actin binding site that is connected to the rod-forming part of the molecule via a highly charged stretch of amino acids. Whereas the sequence of alpha-actinin (Noegel, A., W. Witke, and M. Schleicher. 1987. FEBS [Fed. Eur. Biochem. Soc.] Lett. 221:391-396) suggests that the extended rod domain of the molecule is based on four spectrin-like repeats with high alpha-helix potential, the rod domain of the 120-kD gelation factor is constructed from six 100-residue repeats that have a high content of glycine and proline residues and which, in contrast to alpha-actinin, do not appear to have a high alpha-helical content. These repeats show a distinctive pattern of regions that have high beta-sheet potential alternating with short zones rich in residues with a high potential for turns. This observation suggests that each 100-residue motif has a cross-beta conformation with approximately nine sheets arranged perpendicular to the long axis of the molecule. In the high beta-potential zones every second residue is often hydrophobic. In a cross-beta structure, this pattern would result in one side of the domain having a surface rich in hydrophobic side chains which could account for the dimerization of the 120-kD gelation factor subunits.     

Differential mechanical stability of filamin A rod segments

Prompted by recent reports suggesting that interaction of filamin A (FLNa) with its binding partners is regulated by mechanical force, we examined mechanical properties of FLNa domains using magnetic tweezers. FLNa, an actin cross-linking protein, consists of two subunits that dimerize through a C-terminal self-association domain. Each subunit contains an N-terminal spectrin-related actin-binding domain followed by 24 immunoglobulinlike (Ig) repeats. The Ig repeats in the rod 1 segment (repeats 1–15) are arranged as a linear array, whereas rod 2 (repeats 16–23) is more compact due to interdomain interactions. In the rod 1 segment, repeats 9–15 augment F-actin binding to a much greater extent than do repeats 1–8. Here, we report that the three segments are unfolded at different forces under the same loading rate. Remarkably, we found that repeats 16–23 are susceptible to forces of ∼10 pN or even less, whereas the repeats in the rod 1 segment can withstand significantly higher forces. The differential force response of FLNa Ig domains has broad implications, since these domains not only support the tension of actin network but also interact with many transmembrane and signaling proteins, mostly in the rod 2 segment. In particular, our finding of unfolding of repeats 16–23 at ∼10 pN or less is consistent with the hypothesized force-sensing function of the rod 2 segment in FLNa.     

A mechanical unfolding intermediate in an actin-crosslinking protein

Many F-actin crosslinking proteins consist of two actin-binding domains separated by a rod domain that can vary considerably in length and structure. In this study, we used single-molecule force spectroscopy to investigate the mechanics of the immunoglobulin (Ig) rod domains of filamin from Dictyostelium discoideum (ddFLN). We find that one of the six Ig domains unfolds at lower forces than do those of all other domains and exhibits a stable unfolding intermediate on its mechanical unfolding pathway. Amino acid inserts into various loops of this domain lead to contour length changes in the single-molecule unfolding pattern. These changes allowed us to map the stable core of approximately 60 amino acids that constitutes the unfolding intermediate. Fast refolding in combination with low unfolding forces suggest a potential in vivo role for this domain as a mechanically extensible element within the ddFLN rod.     

Filamins: promiscuous organizers of the cytoskeleton

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 approximately 96 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 filamin-binding 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.     

Structural basis of filamin A functions

Filamin A (FLNa) can effect orthogonal branching of F-actin and bind many cellular constituents. FLNa dimeric subunits have N-terminal spectrin family F-actin binding domains (ABDs) and an elongated flexible segment of 24 immunoglobulin (Ig) repeats. We generated a library of FLNa fragments to examine their F-actin binding to define the structural properties of FLNa that enable its various functions. We find that Ig repeats 9–15 contain an F-actin–binding domain necessary for high avidity F-actin binding. Ig repeats 16–24, where most FLNa-binding partners interact, do not bind F-actin, and thus F-actin does not compete with Ig repeat 23 ligand, FilGAP. Ig repeats 16–24 have a compact structure that suggests their unfolding may accommodate pre-stress–mediated stiffening of F-actin networks, partner binding, mechanosensing, and mechanoprotection properties of FLNa. Our results also establish the orientation of FLNa dimers in F-actin branching. Dimerization, mediated by FLNa Ig repeat 24, accounts for rigid high-angle FLNa/F-actin branching resistant to bending by thermal forces, and high avidity F-actin binding and cross-linking.     

Filamin-regulated F-actin assembly is essential for morphogenesis and controls phototaxis in Dictyostelium

Dictyostelium strains lacking the F-actin cross-linking protein filamin (ddFLN) have a severe phototaxis defect at the multicellular slug stage. Filamins are rod-shaped homodimers that cross-link the actin cytoskeleton into highly viscous, orthogonal networks. Each monomer chain of filamin is comprised of an F-actin-binding domain and a rod domain. In rescue experiments only intact filamin re-established correct phototaxis in filamin minus mutants, whereas C-terminally truncated filamin proteins that had lost the dimerization domain and molecules lacking internal repeats but retaining the dimerization domain did not rescue the phototaxis defect. Deletion of individual rod repeats also changed their subcellular localization, and mutant filamins in general were less enriched at the cell cortex as compared with the full-length protein and were increasingly present in the cytoplasm. For correct phototaxis ddFLN is only required at the tip of the slug because expression under control of the cell type-specific extracellular-matrix protein A (ecmA) promoter and mixing experiments with wild type cells supported phototactic orientation. Likewise, in chimeric slugs wild type cells were primarily found at the tip of the slug, which acts as an organizer in Dictyostelium morphogenesis.     

Molecular architecture of the rod domain of the Dictyostelium gelation factor (ABP120).

The Dictyostelium discoideum gelation factor is a two-chain actin-cross-linking protein that, in addition to an N-terminal actin-binding domain, has a rod domain constructed from six tandem repeats of a 100-residue motif that has an immunoglobulin fold. To define the architecture of the rod domain of gelation factor, we have expressed in E. coli a series of constructs corresponding to different numbers of gelation factor rod repeats and have characterised them by chemical crosslinking, ultracentrifugation, column chromatography, matrix-assisted laser desorption ionisation (MALDI) mass spectrometry and NMR spectroscopy. Fragments corresponding to repeats 1-6 and 5-6 dimerise, whereas repeats 1-5 and single repeats 3 and 4 are monomeric. Repeat 6 interacts weakly and was present as monomer and dimer when analysed by analytical ultracentrifugation. Proteolytic digestion of rod5-6 resulted in the generation of two polypeptides that roughly corresponded to rod5 and part of rod6. None of these polypeptides formed dimers after chemical crosslinking. Stable dimerisation therefore appears to require repeats 5 and 6. Based on these data a model of gelation factor architecture is presented. We suggest an arrangement of the chains where only the carboxy-terminal repeats interact as was observed for filamin/ABP280, the mammalian homologue of gelation factor.     

Structural basis for vertebrate filamin dimerization

Filamins are essential in cell motility and many developmental processes. They are large actin cross linking proteins that contain actin binding domains in their N termini and a long rod region constructed from 24 tandem Ig domains. Dimerization is crucial for the actin crosslinking function of filamins and requires the most C-terminal Ig domain. We describe here the crystal structure of this 24th Ig domain (Ig24) of human filamin C and show how it mediates dimerization. The dimer interface is novel and quite different to that seen in the Dictyostelium discoideum filamin analog. The sequence signature of the dimerization interface suggests that the C-terminal domains of all vertebrate filamins share the same dimerization mechanism. Furthermore, we show that point mutations in the dimerization interface disrupt the dimer and that the dissociation constant for recombinant Ig24 is in the micromolar range.     

Utrophin lacks the rod domain actin binding activity of dystrophin

We previously identified a cluster of basic spectrin-like repeats in the dystrophin rod domain that binds F-actin through electrostatic interactions (Amann, K. J., Renley, B. A., and Ervasti, J. M. (1998) J. Biol. Chem. 273, 28419-28423). Because of the importance of actin binding to the presumed physiological role of dystrophin, we sought to determine whether the autosomal homologue of dystrophin, utrophin, shared this rod domain actin binding activity. We therefore produced recombinant proteins representing the cluster of basic repeats of the dystrophin rod domain (DYSR11-17) or the homologous region of the utrophin rod domain (UTROR11-16). Although UTROR11-16 is 64% similar and 41% identical to DYSR11-17, UTROR11-16 (pI = 4. 86) lacks the basic character of the repeats found in DYSR11-17 (pI = 7.44). By circular dichroism, gel filtration, and sedimentation velocity analysis, we determined that each purified recombinant protein had adopted a stable, predominantly alpha-helical fold and existed as a highly soluble monomer. DYSR11-17 bound F-actin with an apparent K(d) of 7.3 +/- 1.3 microM and a molar stoichiometry of 1:5. Significantly, UTROR11-16 failed to bind F-actin at concentrations as high as 100 microM. We present these findings as further support for the electrostatic nature of the interaction of the dystrophin rod domain with F-actin and suggest that utrophin interacts with the cytoskeleton in a manner distinct from dystrophin.     

Structure and function of laminin: anatomy of a multidomain glycoprotein

Laminin is a large (900 kDa) mosaic protein composed of many distinct domains with different structures and functions. Globular and rodlike domains are arranged in an extended four-armed, cruciform shape that is well suited for mediating between distant sites on cells and other components of the extracellular matrix. The alpha-helical coiled-coil domain of the long arm is involved in the specific assembly of the three chains (A, B1, B2, and possible variants) of laminin and is the only domain composed of multiple chains. It is terminated by a large globular domain composed of five homologous subdomains formed by the COOH-terminal part of the A chain. Sites for receptor-mediated cell attachment and promotion of neurite outgrowth reside in the terminal region of the long arm. A second cell attachment site, a cell signaling site with mitogenic action, binding sites for the closely associated glycoprotein nidogen/entactin, and regions involved in calcium-dependent aggregation are localized in the short arms. These domains, which to a large extent are composed of Cys-rich repeats with limited homology to EGF, are the most highly conserved regions in laminins of different origin. At present, most structural and functional data have been collected for a laminin expressed by a mouse tumor, which can be readily isolated in native form and dissected into functional fragments by limited proteolysis. Increasing information on laminins from different species and tissues demonstrates considerable variations of structure. Isoforms of laminin assembled from different chains are focally and transiently expressed and may serve distinct functions at early stages of development even before being laid down as major components of basement membranes.     

Structural study of long arm fragments of laminin. Evidence for repetitive C-terminal sequences in the A-chain, not present in the B-chains

The outer segments of the long arm of laminin have recently been shown to mediate attachment of many cell types and to stimulate neurite outgrowth. For a structural characterization of this part of the molecule we prepared, by limited elastase digestion of laminin, fragments E3 and E8, previously identified as a globular heparin-binding domain and as a 35-nm-long rod with a terminal globule, respectively. Fragment E3 is a domain adjacent to fragment E8. Both structures together comprise the complete terminal half of the long arm. Our data confirm current models, which predict that the C-terminal segments from all three chains contribute to its structure. The B chains terminate at the end of the rod like domain, while the large terminal globule is formed by A-chain structures only. In addition to fragment E3, two new fragments T1 and T2 obtained by tryptic cleavage of fragment E8 were characterized as substructures of the globular domain. Screening of a mouse cDNA library with synthetic oligonucleotides allowed isolation of an 1.8-kb cDNA clone encoding 547 C-terminal amino acids of the A chain and some 196 nucleotides of the 3'-untranslated region including a single polyadenylation site. The clone contained portions of domain T2 and the complete heparin binding domain E3 which was thus identified as the most C-terminal domain of the A chain. Sequence alignment indicated that the terminal globule is formed by homologous repeats of some 140 residues having no counterpart in the B chains.     

Differences in the ionic interaction of actin with the motor domains of nonmuscle and muscle myosin II.

Changes in the actin-myosin interface are thought to play an important role in microfilament-linked cellular movements. In this study, we compared the actin binding properties of the motor domain of Dictyostelium discoideum (M765) and rabbit skeletal muscle myosin subfragment-1 (S1). The Dictyostelium motor domain resembles S1(A2) (S1 carrying the A2 light chain) in its interaction with G-actin. Similar to S1(A2), none of the Dictyostelium motor domain constructs induced G-actin polymerization. The affinity of monomeric actin (G-actin) was 20-fold lower for M765 than for S1(A2) but increasing the number of positive charges in the loop 2 region of the D. discoideum motor domain (residues 613-623) resulted in equivalent affinities of G-actin for M765 and for S1. Proteolytic cleavage and cross-linking approaches were used to show that M765, like S1, interacts via the loop 2 region with filamentous actin (F-actin). For both types of myosin, F-actin prevents trypsin cleavage in the loop 2 region and F-actin segment 1-28 can be cross-linked to loop 2 residues by a carbodiimide-induced reaction. In contrast with the S1, loop residues 559-565 of D. discoideum myosin was not cross-linked to F-actin, probably due to the lower number of positive charges. These results confirm the importance of the loop 2 region of myosin for the interaction with both G-actin and F-actin, regardless of the source of myosin. The differences observed in the way in which M765 and S1 interact with actin may be linked to more general differences in the structure of the actomyosin interface of muscle and nonmuscle myosins.     

Domain analysis of α-actinin reveals new aspects of its association with F-actin during cytokinesis

α-actinin is a rod-shaped actin cross-linking protein composed of actin binding domain, spectrin-like repeats of the central rod domain and the EF-hand domain. Cytokinesis in mammalian cells involves remodeling of equatorial actin filaments (F-actin) mediated by α-actinin. However, it remains unknown how α-actinin interacts with F-actin at the cleavage furrow. To address this question, we have conducted functional analysis of the mutant that either lacks the ability to cross-link F-actin (ABD) or to bind to F-actin (ΔABD). We found that equatorial localization of α-actinin requires both its F-actin binding and cross-linking activities. Unexpectedly, we also found that overexpression of ΔABD-GFP but not ABD-GFP frequently caused accelerated cytokinesis and ectopic furrowing similar to those observed in cells depleted of α-actinin. Immunofluorescence revealed that overexpression of ΔABD-GFP caused displacement of endogenous α-actinin and a decrease in the density of F-actin throughout the entire cortex. Biochemical experiments showed that ΔABD was able to form heterodimers with endogenous α-actinin. These results suggest that the central rod spectrin-like repeats of α-actinin is sufficient for its dimerization in vivo. Our findings uncover previously unappreciated functions of the α-actinin domains in a cell.     

Electron Microscopy and 3D Reconstruction Reveals Filamin Ig Domain Binding to F-Actin

Filamin A (FLNa) is an actin-binding protein that cross-links F-actin into networks of orthogonally branched filaments. FLNa also directs the networks to integrins while responding to mechanochemical signaling pathways. Flexible, 160-nm-long FLNa molecules are tail-to-tail dimers, each subunit of which contains an N-terminal calponin homology (CH)/actin-binding domain connected by a series of 24 immunoglobulin (Ig) repeats to a dimerization site at their C-terminal end. Whereas the contribution of the CH domains to F-actin affinity is weak (apparent K_a ~ 10⁵), the binding of the intact protein to F-actin is strong (apparent K_a ~ 10⁸), suggesting involvement of additional parts of the molecule in this association. Indeed, previous results indicate that Ig repeats along FLNa contribute significantly to the strength of the actin filament interaction. In the current study, we used electron microscopy and three-dimensional reconstruction to elucidate the structural basis of the Ig repeat–F-actin binding. We find that FLNa density is clearly delineated in reconstructions of F-actin complexed either with a four-Ig-repeat segment of FLNa containing Ig repeat 10 or with immunoglobulin-like filamin A repeat (IgFLNa)10 alone. The mass attributable to IgFLNa10 lies peripherally along the actin helix over the N-terminus of actin subdomain 1. The IgFLNa10 interaction appears to be specific, since no other individual Ig repeat or fragment of the FLNa molecule examined, besides ones with IgFLNa10 or CH domains, decorated F-actin filaments or were detected in reconstructions. We conclude that the combined interactions of CH domains and the IgFLNa10 repeat provide the binding strength of the whole FLNa molecule and propose a model for the association of IgFLNa10 on actin filaments.     

Molecular substructure of a viral receptor-recognition protein. The gp17 tail-fiber of bacteriophage T7

The bacteriophage T7 tail complex consists of a conical tail-tube surrounded by six kinked tail-fibers, which are oligomers of the viral protein gp17 (Mr 61,400). We have derived a molecular model for the tail-fiber by integrating secondary structure predictions with ultrastructural information obtained by correlation averaging of electron micrographs of negatively stained tail complexes. This model has been further refined by high-resolution scanning transmission electron microscopy of purified fibers, both negatively stained and unstained. Mass measurements made from the latter images establish that the fiber is a trimer of gp17. The proximal half-fiber is a uniform rod, about 2.0 nm in diameter and 16.4 nm long, which we infer to be a triple-stranded coiled-coil, containing three copies of an alpha-helical domain of about 117 residues, starting at Phe151. The distal half-fiber is 15.5 nm long, and is made up of four globules, 3.1 to 4.8 nm in diameter, in rigid linear array: it contains the carboxy-terminal halves (residues approximately 268 to 553) of the constituent gp17 chains, arranged with 3-fold symmetry around its long axis. The amino-terminal domains (residues 1 to 149) link the fiber to the tail-tube. We conclude that the three gp17 chains are quasi-equivalent in the proximal half-fiber, equivalent in the distal half-fiber, and non-equivalent in the kink region that separates the two half-fibers: such localized non-equivalence may represent a general mechanism for the formation of kinked joints in segmented homo-oligomeric proteins.     

Bundling of actin filaments by alpha-actinin depends on its molecular length

Cross-linking of actin filaments (F-actin) into bundles and networks was investigated with three different isoforms of the dumbbell-shaped alpha-actinin homodimer under identical reaction conditions. These were isolated from chicken gizzard smooth muscle, Acanthamoeba, and Dictyostelium, respectively. Examination in the electron microscope revealed that each isoform was able to cross-link F-actin into networks. In addition, F-actin bundles were obtained with chicken gizzard and Acanthamoeba alpha-actinin, but not Dictyostelium alpha-actinin under conditions where actin by itself polymerized into disperse filaments. This F-actin bundle formation critically depended on the proper molar ratio of alpha-actinin to actin, and hence F-actin bundles immediately disappeared when free alpha-actinin was withdrawn from the surrounding medium. The apparent dissociation constants (Kds) at half-saturation of the actin binding sites were 0.4 microM at 22 degrees C and 1.2 microM at 37 degrees C for chicken gizzard, and 2.7 microM at 22 degrees C for both Acanthamoeba and Dictyostelium alpha-actinin. Chicken gizzard and Dictyostelium alpha-actinin predominantly cross-linked actin filaments in an antiparallel fashion, whereas Acanthamoeba alpha-actinin cross-linked actin filaments preferentially in a parallel fashion. The average molecular length of free alpha-actinin was 37 nm for glycerol-sprayed/rotary metal-shadowed and 35 nm for negatively stained chicken gizzard; 46 and 44 nm, respectively, for Acanthamoeba; and 34 and 31 nm, respectively, for Dictyostelium alpha-actinin. In negatively stained preparations we also evaluated the average molecular length of alpha-actinin when bound to actin filaments: 36 nm for chicken gizzard and 35 nm for Acanthamoeba alpha-actinin, a molecular length roughly coinciding with the crossover repeat of the two-stranded F-actin helix (i.e., 36 nm), but only 28 nm for Dictyostelium alpha-actinin. Furthermore, the minimal spacing between cross-linking alpha-actinin molecules along actin filaments was close to 36 nm for both smooth muscle and Acanthamoeba alpha-actinin, but only 31 nm for Dictyostelium alpha-actinin. This observation suggests that the molecular length of the alpha-actinin homodimer may determine its spacing along the actin filament, and hence F-actin bundle formation may require "tight" (i.e., one molecule after the other) and "untwisted" (i.e., the long axis of the molecule being parallel to the actin filament axis) packing of alpha-actinin molecules along the actin filaments.     

Structural predictions for the central domain of dystrophin

The amino acid sequence of dystrophin indicates that the molecule has globular N- and C-terminal domains separated by a long central rod domain. The central rod contains multiple repeats, about 100 amino acids long and of variable length. These diverge sufficiently in sequence that, in previous studies, only 14 of the most similar repeats have been aligned and analysed in any detail. We show here that a heptad pattern of hydrophobic residues is preserved across all repeats. Using the heptad pattern together with a consensus sequence template, we identified and aligned 25 repeats in the dystrophin rod sequence. Each repeat consists of a constant-length core helix of 54 residues, coupled via a short linker to a weakly conserved variable-length helix, and then via a second linker to the next core. The variable-length helix appears truncated in repeats 10 and 13 and extended in repeats 4 and 20. The extension of repeat 20 is particularly interesting since it corresponds to a hotspot of dystrophy-inducing mutations. Detailed modelling suggests that the classical Speicher-Marchesi [(1984) Nature 311, 177-180] model for spectrin may not be appropriate to dystrophin without some modification. We propose that whilst the repeating structural motif in dystrophin is probably a bead of triple coiled coil, this bead is twice as massive as, and out of phase with, those proposed for spectrin. Our model raises the possibility that the rod domain of dystrophin may confer elasticity on the molecule. Deletions which truncate this region would then reduce the extensibility of the molecule without affecting actin crosslinking, consistent with their typically producing the relatively benign Becker phenotype of muscular dystrophy.     

The nonlinear structure of the desmoplakin plakin domain and the effects of cardiomyopathy-linked mutations

Desmoplakin is a cytoplasmic desmosomal protein that plays a vital role in normal intercellular adhesion. Mutations in desmoplakin can result in devastating skin blistering diseases and arrhythmogenic right ventricular cardiomyopathy, a heart muscle disorder associated with ventricular arrhythmias, heart failure, and sudden death. The desmoplakin N-terminal region is a 1056-amino-acid sequence of unknown structure. It mediates interactions with other desmosomal proteins, is found in a variety of plakin proteins, and spans what has been termed the “plakin domain,” which includes residues 180-1022 and consists of six spectrin repeats (SRs) and an Src homology 3 domain. Herein we elucidate the architecture of desmoplakin's plakin domain, as well as its constituent tandem SRs. Small-angle X-ray scattering analysis shows that the entire plakin domain has an “L” shape, with a long arm and a short arm held at a perpendicular angle. The long arm is 24.0 nm long and accommodates four stably folded SRs arranged in tandem. In contrast, the short arm is 17.9 nm in length and accommodates two independently folded repeats and an extended C-terminus. We show that mutations linked to arrhythmogenic right ventricular cardiomyopathy (K470E and R808C) cause local conformational alterations, while the overall folded structure is maintained. This provides the first structural and mechanistic insights into an entire plakin domain and provides a basis for understanding the critical role of desmoplakin in desmosome function.     

Mapping the functional surface of domain 2 in the gelsolin superfamily

The crystal structure of the F-actin binding domain 2 of severin, the gelsolin homologue from Dictyostelium discoideum, has been determined by multiple isomorphous replacement and refined to 1.75 A resolution. The structure reveals an alpha-helix-beta-sheet sandwich similar to the domains of gelsolin and villin, and contains two cation-binding sites, as observed in other domain 1 and domain 2 homologues. Comparison of the structures of several gelsolin family domains has identified residues that may mediate F-actin binding in gelsolin domain 2 homologues. To assess the involvement of these residues in F-actin binding, three mutants of human gelsolin domain 2 were assayed for F-actin binding activity and thermodynamic stability. Two of the mutants, RRV168AAA and RLK210AAA, demonstrated a lowered affinity for F-actin, indicating a role for those residues in filament binding. Using both structural and biochemical data, we have constructed a model of the gelsolin domain 1-domain 2-F-actin complex. This model highlights a number of interactions that may serve as positive and negative determinants of filament end- and side-binding.