Abstract

Filamin A (FLNA) crosslinks F-actin and binds proteins consistent with roles integrating cell signalling and the cytoskeleton. FLNA missense mutations are associated with the otopalatodigital syndrome (OPD) spectrum of skeletal disorders, clustering in discrete domains. One cluster is found in the second calponin homology domain of the FLNA actin-binding domain (ABD), implicating this region as essential for mediating correct function. Here we show that OPD (FLNA E254K) fibroblast lysates have equivalent concentrations of FLNA compared with controls and that recombinant FLNA E254K ABD has increased in vitro F-actin binding (Kd 13 µm) compared with wild type (WT; Kd 48 µm). These observations are consistent with a gain-of-function mechanism for OPD. We have determined the crystal structures of the WT and E254K FLNA ABDs at 2.3 Å resolution, revealing that they adopt similar closed conformations. The E254K mutation removes a conserved salt bridge but does not disrupt the ABD structure. The solution structures are also equivalent as determined by circular dichroism spectroscopy, but differential scanning fluorimetry denaturation showed reduced stability (decreased Tm of 5.6°C) for E254K relative to WT. Ex vivo characterization of E254K OPD patient fibroblasts revealed they have similar motility and adhesion as control cells, implying that many core functions mediated by FLNA are unaffected, consistent with OPD only affecting specific tissues despite FLNA being widely expressed. These data provide the first biochemical evidence for a gain-of-function mechanism for the OPD disorders, and mechanistically distinguishes them from the loss-of-function phenotypes that manifest as disorders of neuronal migration.

INTRODUCTION

Cell motility, adhesion, polarization and division depend on dynamic regulation of the actin cytoskeleton by actin-binding proteins. The filamin family of proteins bind F-actin through the highly conserved N-terminal actin-binding domain (ABD) (1) and to additional protein partners through C-terminal repeat domains (2). X-linked filamin A (FLNA) is expressed ubiquitously encoding a 280 kDa monomer that forms ‘V’-shaped tail-to-tail homodimers (3) that cross-link F-actin into orthogonal networks (4). Each FLNA monomer is composed of the ABD followed by a rod region comprised of 24 immunoglobulin-like repeat domains with dimerization occurring at the C-terminal repeat (5). FLNA is required for membrane stability, providing protection from force-induced cell death in an integrin-dependent manner (6) for which the ABD is essential (7). Human melanoma cell lines that lack FLNA show impaired motility with cytoskeletal defects that can be rescued by transfection of FLNA (8), though mouse fibroblasts null for FLNA show normal motility (9). In addition to a structural role, FLNA integrates cytoskeletal organization and signalling events through interactions mediated through its C-terminal repeats. FLNA's interaction with β-integrins (10) inhibits cell migration and polarization by reducing membrane remodelling (11).

The FLNA ABD, composed of two calponin homology domains (CH1 and CH2), is conserved in other F-actin-binding proteins (Fig. 1) with three actin-binding sequences (ABS1-3) identified (reviewed in 12). Most evidence suggests the ABD functions as a complete module even though CH1 (ABS1-2) contributes more to the F-actin interaction than CH2 (ABS3) (13). Missense mutations in the α-actinin-4 CH2 domain, associated with an autosomal-dominant late-onset kidney disorder focal segmental glomerulosclerosis (FSGS), result in increased F-actin affinity (14,15). CH2 is required for the correct interaction with F-actin and the ABSs are not functionally interchangeable (16). The entire ABD is also required for correct dystrophin function as seen when a mini-dystrophin transgene deleted for CH2 failed to restore normal muscle physiology and force production when microinjected into the mdx (dystrophin null) mouse, in contrast to transgenes that contain CH2 within an intact ABD (17).

Figure 1.

Sequence alignment of ABDs from filamins and homologous proteins showing the distribution of disease-associated missense mutations. Amino acids that are associated with disease are highlighted; red and yellow, PH and OPD spectrum disorders (Filamin_A), respectively; blue, atelosteogenesis disorders (Filamin_B); grey, focal segmental glomerulosclerosis (α_Actinin4); green, hereditary spherocytosis (β_Spectrin); magenta, Duchenne muscular dystrophy (Dystrophin). Residues conserved with the FLNA Lys169-Glu254 salt bridge are shown in red text. The secondary structure elements of FLNA are shown in blue above the alignment and are labelled.

Figure 1.

Sequence alignment of ABDs from filamins and homologous proteins showing the distribution of disease-associated missense mutations. Amino acids that are associated with disease are highlighted; red and yellow, PH and OPD spectrum disorders (Filamin_A), respectively; blue, atelosteogenesis disorders (Filamin_B); grey, focal segmental glomerulosclerosis (α_Actinin4); green, hereditary spherocytosis (β_Spectrin); magenta, Duchenne muscular dystrophy (Dystrophin). Residues conserved with the FLNA Lys169-Glu254 salt bridge are shown in red text. The secondary structure elements of FLNA are shown in blue above the alignment and are labelled.

Mutations that result in loss of FLNA are associated with periventricular nodular heterotopia (PH) (18), a disease typically embryonic lethal in males and manifesting as seizures in females through abnormal radial migration of neocortical neurons during foetal development (19). The neuroradiological appearances in PH include the formation of nodules at the cerebral ventricular surfaces, implicating FLNA in neuronal migration signalling events during brain development (20). Most PH mutations disrupt the FLNA reading frame and completely ablate FLNA expression with only a smaller number of missense mutations identified. PH missense mutations located within the ABD are only found in the CH1 domain and not CH2.

A group of X-linked developmental malformation disorders, clinically distinct from PH, are associated with missense mutations within FLNA. Male patients with the mildest dysplasia, otopalatodigital syndrome 1 (OPD1), have mild skeletal anomalies, cleft palate and deafness. Otopalatodigital syndrome 2 (OPD2) males present with bowed bones, absent or small digits and cleft palate, and also show non-skeletal phenotypes with malformations occurring in the abdominal wall, heart, hindbrain, intestines, central nervous system and kidney. Frontometaphyseal dysplasia (FMD) presents as a severe skeletal dysplasia, with deafness and urogenital defects. The most severe disease is Melnick Needles syndrome (MNS) affecting similar organs as OPD2 causing either pre-natal or early post-natal death in males and a severe but survivable skeletal dysplasia in heterozygous females (21). These disorders do not share the defining brain abnormalities found in PH patients. Mutations associated with OPD are intensely clustered within FLNA domains; OPD1 mutations are only found in CH2, OPD2 only in CH2 and repeats 14–15, and MNS only in repeat 10. FMD mutations are more widespread; CH2, repeats 3, 9–10, 14–15, and 22–23 (21). The location and recurrence of the mutations predicts the severity and type of disease. This clustering of point mutations in particular FLNA domains, the clinical distinction of the OPD phenotypes from the loss of function PH disorder and skewed X-inactivation in leucocytes from female OPD patients suggest a gain or altered function in discrete FLNA domains for OPD.

The FLNA mutation c.760G>A, which translates to the p.E254K substitution in the ABD, is associated with OPD2 and has been observed on multiple independent occasions in unrelated families (21). The biochemical basis underlying OPD is completely unknown. To determine a molecular mechanism for OPD and to further our understanding of FLNA biology, we have compared the structure and function of the FLNA ABD containing the OPD2-associated mutation E254K relative to wild type (WT). We have investigated FLNA expression in fibroblasts from an OPD2 patient (E254K) further defining the clinical distinction of OPD from PH, and compared the motility and adhesion properties of these fibroblasts relative to those from non-affected individuals. The size and flexibility of filamin makes it unlikely to crystallize for X-ray diffraction studies requiring analysis of discrete functional domains. To analyse structural alterations that result from the E254K mutation, we have determined the crystal structures of the WT and E254K FLNA ABDs at 2.3 Å resolution. The in vitro F-actin affinity and thermostability of the FLNA E254K ABD have been compared with WT.

RESULTS

FLNA expression

Western blotting revealed that equivalent levels of FLNA are present in fibroblasts from a male OPD2 patient hemizygous for the E254K mutation compared with three control fibroblast cell lines (Fig. 2A). FLNA protein levels were quantified by densitometry normalized to an α-tubulin loading control, 1.0 (±0.05), and E254K FLNA levels cells measured relative to WT 1.0 (±0.04) (mean ± SD) (Fig. 2B).

Figure 2.

Expression of filamin A from OPD2 (E254K) patient cells. (A) Western blots of fibroblasts from an OPD2 patient (FLNA E254K) and controls. (B) Densitometry of western blots. The FLNA concentration was normalized to the α-tubulin control and E254K is shown relative to the WT level. Error bars indicate SD from triplicate experiments.

Figure 2.

Expression of filamin A from OPD2 (E254K) patient cells. (A) Western blots of fibroblasts from an OPD2 patient (FLNA E254K) and controls. (B) Densitometry of western blots. The FLNA concentration was normalized to the α-tubulin control and E254K is shown relative to the WT level. Error bars indicate SD from triplicate experiments.

FLNA actin binding, solution structure and stability

Co-sedimentation assays show both WT and E254K FLNA ABDs saturated F-actin, binding in the micromolar range. The WT ABD bound F-actin with an apparent dissociation constant (Kd) of 48 µm (±11), whereas the E254K ABD bound with Kd of 13 µm (±5) (mean ± SEM) (Fig. 3A and B). Circular dichroism spectroscopy reveals equivalent solution structures for the WT FLNA and E254K ABDs (Fig. 3C). Differential scanning fluorimetry assays performed with SYPRO orange showed the E254K ABD exhibited reduced stability with Tm of 44.7°C (±0.7) relative to WT, Tm of 50.3°C (±0.8) (Fig. 3D). Lysozyme was used as a standard and denatured with a Tm of 73.7°C as reported (22). Controls with dye or protein omitted exhibited no temperature-dependent change in fluorescence (mean ± SD).

Figure 3.

Comparison of E254K filamin A ABD in vitro properties with wild type (WT). (A) Representative 12% SDS–PAGE gels for WT and E254K ABD F-actin-binding assays. The total concentration of ABD included in the assay, and the supernatant (S) and pellet (P) fractions, are shown. (B) Binding curves for WT (red) and E254K (blue) ABDs with F-actin. The Kd of each protein is indicated with a dashed line; WT, 13 ± 5 µm and E254K, 48 ± 11 µm (±SEM). (C) Circular dichroism spectra of the FLNA WT (red) and E254K (blue) ABDs. (D) Differential scanning fluorimetry of FLNA ABDs. Triplicate melting curves are shown for WT (red) and E254K (blue). Lysozyme is included as a standard (green). The fluorescence of the samples is expressed as a percentage of the maximum signal (normalized to 100%).

Figure 3.

Comparison of E254K filamin A ABD in vitro properties with wild type (WT). (A) Representative 12% SDS–PAGE gels for WT and E254K ABD F-actin-binding assays. The total concentration of ABD included in the assay, and the supernatant (S) and pellet (P) fractions, are shown. (B) Binding curves for WT (red) and E254K (blue) ABDs with F-actin. The Kd of each protein is indicated with a dashed line; WT, 13 ± 5 µm and E254K, 48 ± 11 µm (±SEM). (C) Circular dichroism spectra of the FLNA WT (red) and E254K (blue) ABDs. (D) Differential scanning fluorimetry of FLNA ABDs. Triplicate melting curves are shown for WT (red) and E254K (blue). Lysozyme is included as a standard (green). The fluorescence of the samples is expressed as a percentage of the maximum signal (normalized to 100%).

FLNA ABD X-ray crystal structure

The FLNA ABD structure has been determined to 2.3 Å and refined to an Rfactor of 0.208 and Rfree of 0.261, and good geometry (Table 1). FLNA adopts the canonical compact ABD structure with CH1 and CH2 packed in close proximity (Fig. 4A). Each CH domain contains four major α-helices and two (CH1) or three (CH2) shorter α-helices. Three of the major α-helices (C, E and G) make a triple helical bundle with the N-terminal α-helix (A) lying perpendicular to this bundle forming the core of the domain structure. The N-terminal extension of 36 residues (chain A) and 38 (chain B) are not observed and are assumed to be disordered. For chain A, the CH1–CH2 connecting loop (156–162) is disordered, in contrast to chain B where the loop is ordered by crystal contacts. The CH1–CH2 interface buries 13% (880Å2) of the total accessible surface area and completely excludes solvent. The interface is predominantly formed by the packing of the A and G helices from both CH1 and CH2. When ABS1-3 are mapped to the FLNA ABD structure, they do not form a continuous surface that could simultaneously contact F-actin, implying that not all the ABSs contact F-actin at once or structural rearrangement is required to align all three ABS regions for binding (Fig. 4B). Superposition of the FLNA ABD with the homologous domain from related proteins reveals they are structurally well conserved (Supplementary Material, Table S1).

Figure 4.

The structure of the filamin A actin-binding domain. (A) Cartoon representation of the FLNA ABD structure colour-coded from N-terminus (blue) to C-terminus (red). The CH1 α helices are labelled A–G, and CH2 A′–G'. (B) Surface representation of the FLNA ABD showing the three actin-binding sites: red, ABS1; yellow, ABS2; orange, ABS3. The ABD is shown in the same orientation as (A) (left) and rotated 90° about the x-axis (right).

Figure 4.

The structure of the filamin A actin-binding domain. (A) Cartoon representation of the FLNA ABD structure colour-coded from N-terminus (blue) to C-terminus (red). The CH1 α helices are labelled A–G, and CH2 A′–G'. (B) Surface representation of the FLNA ABD showing the three actin-binding sites: red, ABS1; yellow, ABS2; orange, ABS3. The ABD is shown in the same orientation as (A) (left) and rotated 90° about the x-axis (right).

Table 1.

Crystallographic X-ray data and refinement statistics for filamin A actin-binding domains

 Wild type E254K 
Diffraction data 
 Wavelength (Å) 1.5418 1.5418 
 Space group P212121 P212121 
 Cell parameters 
  a (Å) 57.85 57.66 
  b (Å) 71.22 73.94 
  c (Å) 158.28 154.44 
 α = β = γ (°) 90 90 
 Number of measured reflections 116 874 122 444 
 Number of unique reflections 29 893 29 480 
 Resolution (Å) (outer shell) 44.90–2.30 (2.42–2.30) 57.64–2.30 (2.42–2.30) 
 Redundancy 3.9 (3.9) 4.2 (4.1) 
 Completeness (%) 100 (100) 98.0 (93.9) 
Rmerge 0.10 (0.53) 0.088 (0.53) 
 I/SigI 9.4 (2.8) 12.3 (2.4) 
Refinement 
  Rfactor (%) 20.8 22.4 
  Rfree (%) 26.1 26.3 
 Rmsd from ideal geometry 
  Bond lengths (Å) 0.018 0.006 
  Bond angles (°) 1.65 0.9 
  Number of protein atoms 3623 3375 
  Number of H2O molecules 140 103 
 Ramachandran plota 
  Outliers (%) 0.2 
  Allowed (%) 99.8 100 
  Most favoured (%) 96.4 98.8 
  Wilson B (Å245.3 38.4 
  Overall B (Å237.7 39.5 
 Wild type E254K 
Diffraction data 
 Wavelength (Å) 1.5418 1.5418 
 Space group P212121 P212121 
 Cell parameters 
  a (Å) 57.85 57.66 
  b (Å) 71.22 73.94 
  c (Å) 158.28 154.44 
 α = β = γ (°) 90 90 
 Number of measured reflections 116 874 122 444 
 Number of unique reflections 29 893 29 480 
 Resolution (Å) (outer shell) 44.90–2.30 (2.42–2.30) 57.64–2.30 (2.42–2.30) 
 Redundancy 3.9 (3.9) 4.2 (4.1) 
 Completeness (%) 100 (100) 98.0 (93.9) 
Rmerge 0.10 (0.53) 0.088 (0.53) 
 I/SigI 9.4 (2.8) 12.3 (2.4) 
Refinement 
  Rfactor (%) 20.8 22.4 
  Rfree (%) 26.1 26.3 
 Rmsd from ideal geometry 
  Bond lengths (Å) 0.018 0.006 
  Bond angles (°) 1.65 0.9 
  Number of protein atoms 3623 3375 
  Number of H2O molecules 140 103 
 Ramachandran plota 
  Outliers (%) 0.2 
  Allowed (%) 99.8 100 
  Most favoured (%) 96.4 98.8 
  Wilson B (Å245.3 38.4 
  Overall B (Å237.7 39.5 

aAs defined by MOLPROBITY. Rmerge = Σ|Ii − <Ii>|/ΣIi where Ii is the intensity of a single reflection and <Ii> is the mean intensity of that reflection. Rfactor = Σ|FoFc|/ΣFoRfree is the R factor calculated for the cross-validated test set of reflections. Rmsd = root mean square difference.

FLNA E254K ABD X-ray crystal structure

The structure of the E254K FLNA ABD has been determined to 2.3 Å. The E254K ABD crystallized in an isomorphous cell to WT and the structure has been refined to an Rfactor of 0.224 and an Rfree of 0.263 and is of good stereochemistry (Table 1). Like WT, the N-terminus and CH1–CH2 connecting loop are not ordered. The WT crystal structure shows Glu254 forming a salt bridge with Lys169 (OE-NZ distances between 2.6 and 3.2 Å for both chains). This interaction appears structurally important for linking the N- and C-terminal helices of CH2, and shielding Phe188 and Met258 from solvent. Lys169 is located on helix A′ (the site of ABS3) while Glu254 is found on helix G′. Comparison of E254K with WT shows that the tertiary structure of the E254K ABD remains essentially unchanged with an RMS difference of 0.4 Å (214 Cα atoms) (Fig. 5A). For E254K, the Lys169-Glu254 salt bridge is disrupted; the same rotamer conformation for Lys169 is observed as for the WT but the distance between 169 and 254 side chains, now both lysines, has increased to 4.8 Å. The equivalent filamin B (FLNB) mutation E227K (Fig. 1) is associated with the skeletal disorder Larsen syndrome (23), implying that this salt bridge is also essential for correct FLNB function. A salt bridge, in similar orientation (helix A′ Lys-helix G′ Glu), is found at this position in the CH2 domains of α-actinin, plectin and spectrin (Fig. 5B). Dystrophin and utrophin have a salt bridge at the same position but with the sidechains reversed (helix A′ Glu contacting helix G′ Lys), inferring that the interaction is more likely to be important for protein stability rather than a specific actin-binding motif (24). There is no structurally equivalent salt bridge for the CH1 domain. FLNA mutations Q170P and L172F, associated with OPD2 and OPD1 (21), are located near to Lys169 and are likely to affect this salt bridge indirectly, with Q170P expected to alter the backbone structure around Lys169. Leu172 is also on helix A′ packing against helix G′ in the hydrophobic interior of the structure and is conserved between ABDs (Fig. 1); the mutation to a larger Phe would disrupt the packing between the A′ and G′ helices linked by the Lys169-Glu254 salt bridge. The structure of the K255E α-actinin-4 ABD (25) reveals that FSGS mutations are also found on CH2 helix G′ but lie on the opposite face towards the CH1–CH2 domain interface. For K255E α-actinin-4 the homologous salt bridge between helices A′ and G′ is not disrupted compared with FLNA, inferring a different mechanism for disrupting CH2 function for FSGS compared with OPD2. The FSGS K255E mutation also does not significantly alter the conformation of the α-actinin-4 ABD structure compared with WT, similar to our observations for the FLNA E254K structure.

Figure 5.

Effects of the E254K mutation on the filamin A ABD structure. (A) Superposition of the WT (grey) and E254K (cyan) FLNA ABD structures. Sidechains at positions 169 and 254 are shown in stick representation, and the salt bridge for WT as a dashed line. (B) Superposition of ABDs. Sidechains contributing to the conserved 169–254 salt bridge are colour coded; grey, FLNA; cyan, β-spectrin; green, α-actinin; purple, dystrophin; pink, utrophin. Helices linked by the salt bridge are labelled as A′ and G′.

Figure 5.

Effects of the E254K mutation on the filamin A ABD structure. (A) Superposition of the WT (grey) and E254K (cyan) FLNA ABD structures. Sidechains at positions 169 and 254 are shown in stick representation, and the salt bridge for WT as a dashed line. (B) Superposition of ABDs. Sidechains contributing to the conserved 169–254 salt bridge are colour coded; grey, FLNA; cyan, β-spectrin; green, α-actinin; purple, dystrophin; pink, utrophin. Helices linked by the salt bridge are labelled as A′ and G′.

Fibroblast motility and adhesion

Fibroblasts from a male E254K OPD2 patient (21) and non-affected controls cultured on fibronectin-coated plates show similar motility and adherence properties. The cell monolayer was wounded with a pipette tip and monitored by time-course microscopy to determine the rate at which cells were able to migrate into the wound (Fig. 6A). E254K fibroblasts migrated at a rate of 0.022 mm/h (±0.003), equivalent to control cells that moved at a rate of 0.026 mm/h (±0.008) (Fig. 6B). Adhesion of the E254K OPD2 fibroblasts was compared with control cell lines to investigate whether this mutation affects coordination between the cytoskeleton and extracellular matrix. There was no difference in adhesion between E254K and control fibroblasts (Fig. 6C), indicating that this mutation is not affecting core functions in these cells.

Figure 6.

Comparison of migration and adhesion in OPD2 (FLNA E254K) patient fibroblasts with control fibroblasts. (A) Light microscopy images comparing migration of FLNA E254K fibroblasts relative to control fibroblasts at time 0 (left) and 8 h (right). Scale bars represent 0.5 mm. (B) Quantification of fibroblast migration assays showing the migration rate (in mm/h) of WT and E254K fibroblasts (error bars show 1 SD). (C) Adhesion assay comparing FLNA E254K fibroblasts to control fibroblasts. The absorbance of dye from adhered E254K fibroblasts is plotted normalized to the WT (error bars show 1 SD).

Figure 6.

Comparison of migration and adhesion in OPD2 (FLNA E254K) patient fibroblasts with control fibroblasts. (A) Light microscopy images comparing migration of FLNA E254K fibroblasts relative to control fibroblasts at time 0 (left) and 8 h (right). Scale bars represent 0.5 mm. (B) Quantification of fibroblast migration assays showing the migration rate (in mm/h) of WT and E254K fibroblasts (error bars show 1 SD). (C) Adhesion assay comparing FLNA E254K fibroblasts to control fibroblasts. The absorbance of dye from adhered E254K fibroblasts is plotted normalized to the WT (error bars show 1 SD).

DISCUSSION

Patients with clustered FLNA missense mutations present with the OPD spectrum of conditions that range from a mild bone dysplasia to lethality. Several observations suggest that OPD arises from a gain-of-function mechanism. (i) The loss-of-function disorder PH is clinically distinct from OPD and is typically characterized by a lack of FLNA arising from frameshift and non-sense mutations distributed throughout FLNA. (ii) OPD mutations retain the reading frame and cluster in domains of FLNA that correlate with disease severity. (iii) Skewed X-inactivation is observed in females heterozygous for FLNA OPD mutations in contrast to PH where females show random X-inactivation. (iv) A similar distribution of mutations in the autosomal paralogous gene, FLNB, including an equivalent mutation to FLNA E254K (E227K), is associated with Larsen syndrome, a disease distinct from the loss-of-function disorder spondylocarpotarsal synostosis syndrome (23).

Here we show FLNA expressed in OPD2 patient cells at levels similar to controls and recombinant FLNA E254K ABD binding F-actin with increased affinity, consistent with a gain-of-function mechanism for this disease. The FLNA ABD structures reveal that E254K disrupts a salt bridge between Lys169 and Glu254, an interaction conserved even when the side-chain charges are reversed. Salt bridges are typically responsible for stabilizing tertiary structure, with low conservation unless they are involved in a particular function (26). The exact role of the Lys169-Glu254 salt bridge remains to be determined, but given its conservation, and the resultant disease when it is removed (FLNA and FLNB), it is reasonable to assume that it has an important role in filamin function. Our structural data confirm the domain-specific separation of FLNA ABD disease-associated missense mutations proposed from sequence analysis; PH-associated missense mutations are located in CH1- and OPD-associated mutations in CH2 (Fig. 7). Substitutions that are causative of OPD are spread through the CH2 domain, suggesting that alteration of domain-specific functions are central to the pathogenesis of these disorders although there is no obvious overall correlation with the putative actin-binding sequence (ABS3) (Fig. 7). The E254K ABD exhibits decreased thermal stability, consistent with studies showing activity as being inversely correlated with stability (27), but without affecting total FLNA protein levels as measured by western blot analysis of E254K patient fibroblasts. The absence of a motility or adhesion phenotype for the E254K fibroblasts is consistent with a major role for FLNA in mediating cell–cell contacts (9,28) and with OPD only exhibiting a phenotype in specific tissues even though FLNA is widely expressed.

Figure 7.

Disease-associated mutations mapped to the WT FLNA ABD structure. Cα trace representation of the FLNA WT ABD structure with the ABS (1–3) highlighted in ribbon representation. The positions of disease-associated mutations are shown in stick representation and labelled with colour coding: blue, OPD1; orange, OPD2; green, OPD1/2; red, PH.

Figure 7.

Disease-associated mutations mapped to the WT FLNA ABD structure. Cα trace representation of the FLNA WT ABD structure with the ABS (1–3) highlighted in ribbon representation. The positions of disease-associated mutations are shown in stick representation and labelled with colour coding: blue, OPD1; orange, OPD2; green, OPD1/2; red, PH.

In conclusion, our observations of FLNA expression in an OPD2 patient and increased actin-binding affinity suggest a biochemical basis for a gain-of-function mechanism for E254K that is mechanistically distinct from PH. The FLNA ABD adopts a canonical compact conformation that is not greatly disturbed by the OPD2-associated E254K mutation either in solution or in the crystal structure. Functional analysis of the FLNA E254K ABD is consistent with the CH2 domain acting as a critical modifier of F-actin affinity. Previous analysis of α-actinin-4 and now FLNA show that mutations altering CH2 activity lead to misregulation of the interaction with F-actin without appearing to perturb the ABD structure other than a decrease in stability. The correct interaction of FLNA with F-actin is important for the structural stability of the cytoskeleton and for normal signal transduction. However, for OPD, it remains unclear how this subverts chondrocyte function leading to skeletal developmental disorders and why other cell types appear unaffected.

MATERIALS AND METHODS

Western blotting

Cell lysates from male OPD2 patient (FLNA E254K) fibroblasts and from control fibroblasts were analysed by western blotting. Fibroblasts were grown to 90% confluence in DMEM media, washed with PBS and lysed with buffer (10 mm Tris/HCl, pH 7.4, 100 mm NaCl, 1 mm EDTA, 1 mm EGTA, 1% Triton X-100, 10% glycerol, 0.1% SDS, 0.5% deoxycholate, protease inhibitor). The cell lysate was centrifuged 12 000g, 10 min at 4°C and the supernatant diluted to a protein concentration of 0.5 mg/ml and run on 10% SDS–PAGE. Transfer to nitrocellulose was conducted overnight in transfer buffer (25 mm glycine, 20 mm Tris/HCl, 7.6 mm SDS, 20% methanol) and confirmed by ponceau stain. The membrane was blocked with 5% w/v skim milk powder in TBST (20 mm Tris/HCl, 150 mm NaCl, 0.05% Tween 20, pH 7.5) for 1 h then rinsed with TBST and incubated with the primary antibody for 1.5 h; the FLNA antibody (Chemicon MAB1678) was diluted 1/15 000 and the α-tubulin antibody (Sigma T9026) was diluted 1/4000. After washing, the membranes were incubated with mouse secondary antibody (Sigma) diluted to 1/6000 in TBST, washed thoroughly with TBST, and secondary antibody detected using the SuperSignal West Pico chemiluminescent substrate solution (Pierce) on film (Fujifilm) and an intelligent dark box II (Fujifilm). Densitometry was performed with the LAS-1000 pro software.

Expression and purification of FLNA ABDS

The cDNA encoding residues 2–269 of human FLNA was PCR amplified, using PWO DNA polymerase, and ligated into the NcoI/HindIII sites of pProEX HTb with T4 DNA ligase. The E254K ABD construct was prepared by PCR overlap extension mutagenesis using the WT as template, with the PCR product sub-cloned into pProEX HTb as for the WT ABD. The sequences of both constructs were confirmed by DNA sequencing. FLNA ABDs were over-expressed in E. coli BL21 (DE3) grown at 25°C for 4 h in Luria-Bertani broth containing 0.1 mg/ml ampicillin. The cells were harvested by centrifugation (12 000g 30 min) and resuspended in lysis buffer (50 mm KH2PO4/K2HPO4, pH 8, 500 mm NaCl, 10 mm imidazole, complete protease inhibitor (Roche)) and lysed in a French press. Following removal of cell debris by centrifugation (12 000g for 30 min), the recombinant FLNA ABD was purified from the soluble fraction by elution from Ni2+-NTA resin (HisTrap, GE Healthcare) in the same buffer with 250 mm imidazole. After buffer exchange, the (His)6 tag was cleaved by tobacco etch virus protease at a ratio of 1:100 (w/w) overnight at 10°C. The FLNA ABD was run through fresh Ni2+-NTA resin separating uncleaved protein, TEV and cleaved (His)6 tag. The resulting ABD was further purified by gel filtration on Superdex75 (GE Healthcare) in buffer containing 50 mm KH2PO4/K2HPO4, 150 mm NaCl. During calibrated gel filtration purification, the filamin ABD eluted at the expected monomeric molecular weight of 29 kDa, but also with a dimer of 58 kDa when DTT was absent from the buffers (Supplementary Material text).

Actin-binding assays

The actin-binding affinity was determined by a cosedimentation assay as previously described (29). F-actin (rabbit skeletal; #AKL99, Cytoskeleton, Inc) was prepared in 2 mm MgCl2, 1 mm ATP, 50 mm KCl, 20 mm Tris–HCl, pH 8) and incubated for 1 h at room temperature. The FLNA ABD was spun at 214 000g for 30 min to remove any aggregated protein. The F-actin was added to the FLNA ABD in actin-binding buffer (36 mm Tris–HCl, pH 8, 108 mm NaCl, 2 mm MgCl2, 1 mm ATP, 6.25 mm DTT) and incubated at 25°C for 30 min, followed by centrifugation at 214 000g for 30 min. Initial experiments were performed to optimize the ABD concentration in the assay to obtain sensible spacing of data points; this resulted in differing concentrations for WT and E254K owing to their different affinities. Equal volumes of supernatant and pellet fractions were loaded onto a 12% SDS–PAGE gel, then stained with coomassie brilliant blue, and protein bands were measured by densitometry (Bio-Rad Quantity One). Non-linear regression analysis for single site binding, calculation of Kd and standard errors were performed with GraphPad Prism (GraphPad Software, Inc., USA).

Circular dichroism spectroscopy

Circular dichroism spectra were collected on a Chirascan spectrophotometer (Applied Photophysics) in a 0.1 mm quartz cell at 4°C with baseline correction performed for sample buffer (20 mm K2HPO4/KH2PO4, pH 8, 20 mm NaCl). For each sample, 10 spectra were collected over a wavelength range of 180–260 nm in 1 nm steps and averaged.

Differential scanning fluorimetry

The thermal stability of the WT and E254K ABDs were determined by differential scanning fluorimetry with SYPRO orange (22) in a Rotor-Gene6000 thermocycler (Corbett). ABD at 2 mg/ml in buffer (20 mm Tris/HCl, pH 8, 6.25 mm DTT, 120 mm NaCl) was mixed with SYPRO orange (Sigma) and heated from 30 to 90°C, in 1°C steps over 18 min. Data were collected in triplicate with excitation at 470 nm and detection at 510 nm. Controls of buffer and dye, protein and buffer only and a positive control of hen egg-white lysozyme (Sigma) were also measured. Melting temperatures were calculated from the first derivative using the Rotor-Gene 6000 software.

WT FLNA ABD crystal structure determination

WT FLNA ABD was concentrated to 9.6 mg/ml in 20 mm HEPES, pH 7.4, 100 mm NaCl and rod-like crystals grown from 20% PEG 4000, 0.1 m Tris–HCl, pH 8.5, 0.2 m Li2SO4 over 2 days by hanging drop vapour diffusion at 20°C. Crystals were briefly soaked in cryoprotectant (mother liquor/20% glycerol) then flash frozen in liquid nitrogen. X-ray diffraction data were collected at 120 K with resolution 44.9–2.3 Å and processed using MOSFLM (30) and SCALA (31). The unit cell was orthorhombic, a = 58.07 Å, b = 71.64 Å, c = 158.91 Å, corresponding to a solvent content of 47% v/v assuming two molecules in the asymmetric unit. The structure was determined by molecular replacement with PHASER (32) using α-actinin-3 ABD (33) as a search model. Two clear solutions were obtained providing an initial model with Rfactor of 0.41 and correlation coefficient of 0.75, confirming the presence of two molecules in the asymmetric unit and the space group P212121. Electron density maps calculated from PHASER coefficients were used to rebuild the model using COOT (34) with clear difference density observed for side chains not in the search model. After rounds of rebuilding and refinement with PHENIX (35) and REFMAC (36) water molecules were added to the structure at chemically sensible positions where a 3σ peak was present in the weighted FoFc map. Refinement converged to give a structure with an Rfactor of 0.208 and an Rfree of 0.261 (5% reflections selected randomly) and good geometry (Table 1). Structure geometry was analysed with MOLPROBITY (37). Protein surface calculations were performed with PROTORP (38), and sequences were aligned with CLUSTAL W (39). Structural figures were prepared with PYMOL (40).

E254K FLNA ABD structure determination

The E254K FLNA ABD was concentrated to 4.1 mg/ml in buffer 20 mm HEPES, pH 7.4, 100 mm NaCl, 5 mm DTT and crystallized from 20% PEG 4000, 0.1 m Tris–HCl, pH 8.8, 0.2 m Li2SO4, 5 mm DTT at 21°C. X-ray diffraction data of resolution 57–2.3 Å were collected and processed as for WT. The E254K ABD crystals were isomorphous with WT; a = 57.66 Å, b = 73.94 Å, c = 154.44 Å. The E254K structure was solved from the WT with residues 253–255 initially omitted. Maps calculated from this solution were used for rebuilding and refinement as for WT to give a final model with an Rfactor of 0.224 and an Rfree of 0.263 and good geometry (Table 1).

Fibroblast motility assays

Motility assays were carried out as described previously (41). Fibroblast monolayers from a male FLNA E254K OPD2 patient (21) and non-affected male controls were scratched, and the rate of cell motility into the scratch measured. Fibronectin (10 µg/ml) in PBS was incubated overnight at 4°C coating the plates. Unbound fibronectin was removed, the plates rinsed with PBS and blocked with BSA at 37°C for 1 h. The plates were further rinsed with PBS, and DMEM/20% FCS was then added followed by inoculation with fibroblasts (5 × 104) which were grown for 3 days until 90% confluent when the plate was scratched with a pipette tip. The media were replaced (10% FCS) and cells incubated at 37°C with 5% CO2. The width of the scratch was measured by light microscopy at various time points and captured images analysed using ImageJ (42). The change in width of the scratch divided by the time incubated gave the rate (mm/h). The experiments were conducted in triplicate and the values averaged.

Fibroblast adhesion assays

A crystal violet stain adhesion assay was carried out as described previously (43). Fibroblasts were plated (2 × 104 cells/well) into a 96-well nuclon-coated plate (Nunc) and incubated for 1 h at 5% CO2 and 37°C, a time interval after which 90% of the WT cells have adhered to the fibronectin matrix. Media and non-adherent cells were removed and the plate washed with PBS. The adhered cells were fixed with 95% ethanol washed and then stained with 0.1% crystal violet (Sigma). Excess stain was removed by carefully washing with water. The cells were lysed and stain solubilized with 0.2% Triton X with gentle mixing for 15 min. Absorbance at 570 nm was measured on a Powerwave XS spectrometer (Biotek). Visual inspection of cell number also showed no difference between WT and E254K cells. The experiments were conducted in triplicate and the values obtained were averaged.

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG online.

FUNDING

This work was supported by the Royal Society of New Zealand Marsden Fund (MAU0411 to A.J.S.S. and S.P.R.).

ACKNOWLEDGEMENTS

The authors would like to thank Tony Pernthaner (AgResearch, New Zealand) and Stacey Nelson (bio-strategy, New Zealand) for access and technical assistance with the Rotor-Gene6000.

Conflict of Interest statement. None declared.

REFERENCES

1
Hartwig
J.H.
Tyler
J.
Stossel
T.P.
Actin-binding protein promotes the bipolar and perpendicular branching of actin filaments
J. Cell Biol.
 , 
1980
, vol. 
87
 (pg. 
841
-
848
)
2
Stossel
T.P.
Condeelis
J.
Cooley
L.
Hartwig
J.H.
Noegel
A.
Schleicher
M.
Shapiro
S.S.
Filamins as integrators of cell mechanics and signalling
Nat. Rev. Mol. Cell Biol.
 , 
2001
, vol. 
2
 (pg. 
138
-
145
)
3
Tyler
J.M.
Anderson
J.M.
Branton
D.
Structural comparison of several actin-binding macromolecules
J. Cell Biol.
 , 
1980
, vol. 
85
 (pg. 
489
-
495
)
4
Brotschi
E.A.
Hartwig
J.H.
Stossel
T.P.
The gelation of actin by actin-binding protein
J. Biol. Chem.
 , 
1978
, vol. 
253
 (pg. 
8988
-
8993
)
5
Himmel
M.
Van Der Ven
P.F.
Stocklein
W.
Furst
D.O.
The limits of promiscuity: isoform-specific dimerization of filamins
Biochemistry
 , 
2003
, vol. 
42
 (pg. 
430
-
439
)
6
Glogauer
M.
Arora
P.
Chou
D.
Janmey
P.A.
Downey
G.P.
McCulloch
C.A.
The role of actin-binding protein 280 in integrin-dependent mechanoprotection
J. Biol. Chem.
 , 
1998
, vol. 
273
 (pg. 
1689
-
1698
)
7
Kainulainen
T.
Pender
A.
D'Addario
M.
Feng
Y.
Lekic
P.
McCulloch
C.A.
Cell death and mechanoprotection by filamin A in connective tissues after challenge by applied tensile forces
J. Biol. Chem.
 , 
2002
, vol. 
277
 (pg. 
21998
-
22009
)
8
Cunningham
C.C.
Gorlin
J.B.
Kwiatkowski
D.J.
Hartwig
J.H.
Janmey
P.A.
Byers
H.R.
Stossel
T.P.
Actin-binding protein requirement for cortical stability and efficient locomotion
Science
 , 
1992
, vol. 
255
 (pg. 
325
-
327
)
9
Hart
A.W.
Morgan
J.E.
Schneider
J.
West
K.
McKie
L.
Bhattacharya
S.
Jackson
I.J.
Cross
S.H.
Cardiac malformations and midline skeletal defects in mice lacking filamin A
Hum. Mol. Genet.
 , 
2006
, vol. 
15
 (pg. 
2457
-
2467
)
10
Kiema
T.
Lad
Y.
Jiang
P.
Oxley
C.L.
Baldassarre
M.
Wegener
K.L.
Campbell
I.D.
Ylanne
J.
Calderwood
D.A.
The molecular basis of filamin binding to integrins and competition with talin
Mol. Cell
 , 
2006
, vol. 
21
 (pg. 
337
-
347
)
11
Calderwood
D.A.
Huttenlocher
A.
Kiosses
W.B.
Rose
D.M.
Woodside
D.G.
Schwartz
M.A.
Ginsberg
M.H.
Increased filamin binding to beta-integrin cytoplasmic domains inhibits cell migration
Nat. Cell Biol.
 , 
2001
, vol. 
3
 (pg. 
1060
-
1068
)
12
Norwood
F.L.
Sutherland-Smith
A.J.
Keep
N.H.
Kendrick-Jones
J.
The structure of the N-terminal actin-binding domain of human dystrophin and how mutations in this domain may cause Duchenne or Becker muscular dystrophy
Structure
 , 
2000
, vol. 
8
 (pg. 
481
-
491
)
13
Way
M.
Pope
B.
Weeds
A.G.
Evidence for functional homology in the F-actin binding domains of gelsolin and alpha-actinin: implications for the requirements of severing and capping
J. Cell Biol.
 , 
1992
, vol. 
119
 (pg. 
835
-
842
)
14
Kaplan
J.M.
Kim
S.H.
North
K.N.
Rennke
H.
Correia
L.A.
Tong
H.Q.
Mathis
B.J.
Rodriguez-Perez
J.C.
Allen
P.G.
Beggs
A.H.
, et al.  . 
Mutations in ACTN4, encoding alpha-actinin-4, cause familial focal segmental glomerulosclerosis
Nat. Genet.
 , 
2000
, vol. 
24
 (pg. 
251
-
256
)
15
Weins
A.
Schlondorff
J.S.
Nakamura
F.
Denker
B.M.
Hartwig
J.H.
Stossel
T.P.
Pollak
M.R.
Disease-associated mutant alpha-actinin-4 reveals a mechanism for regulating its F-actin-binding affinity
Proc. Natl Acad. Sci. USA
 , 
2007
, vol. 
104
 (pg. 
16080
-
16085
)
16
Lorenzi
M.
Gimona
M.
Synthetic actin-binding domains reveal compositional constraints for function
Int. J. Biochem. Cell Biol.
 , 
2008
, vol. 
40
 (pg. 
1806
-
1816
)
17
Banks
G.B.
Gregorevic
P.
Allen
J.M.
Finn
E.E.
Chamberlain
J.S.
Functional capacity of dystrophins carrying deletions in the N-terminal actin-binding domain
Hum. Mol. Genet.
 , 
2007
, vol. 
16
 (pg. 
2105
-
2113
)
18
Fox
J.W.
Lamperti
E.D.
Eksioglu
Y.Z.
Hong
S.E.
Feng
Y.
Graham
D.A.
Scheffer
I.E.
Dobyns
W.B.
Hirsch
B.A.
Radtke
R.A.
, et al.  . 
Mutations in filamin 1 prevent migration of cerebral cortical neurons in human periventricular heterotopia
Neuron
 , 
1998
, vol. 
21
 (pg. 
1315
-
1325
)
19
Moro
F.
Carrozzo
R.
Veggiotti
P.
Tortorella
G.
Toniolo
D.
Volzone
A.
Guerrini
R.
Familial periventricular heterotopia: missense and distal truncating mutations of the FLN1 gene
Neurology
 , 
2002
, vol. 
58
 (pg. 
916
-
921
)
20
Sarkisian
M.R.
Bartley
C.M.
Rakic
P.
Trouble making the first move: interpreting arrested neuronal migration in the cerebral cortex
Trends Neurosci.
 , 
2008
, vol. 
31
 (pg. 
54
-
61
)
21
Robertson
S.P.
Twigg
S.R.
Sutherland-Smith
A.J.
Biancalana
V.
Gorlin
R.J.
Horn
D.
Kenwrick
S.J.
Kim
C.A.
Morava
E.
Newbury-Ecob
R.
, et al.  . 
Localized mutations in the gene encoding the cytoskeletal protein filamin A cause diverse malformations in humans
Nat. Genet.
 , 
2003
, vol. 
33
 (pg. 
487
-
491
)
22
Yeh
A.P.
McMillan
A.
Stowell
M.H.
Rapid and simple protein-stability screens: application to membrane proteins
Acta Crystallogr. D Biol. Crystallogr.
 , 
2006
, vol. 
62
 (pg. 
451
-
457
)
23
Krakow
D.
Robertson
S.P.
King
L.M.
Morgan
T.
Sebald
E.T.
Bertolotto
C.
Wachsmann-Hogiu
S.
Acuna
D.
Shapiro
S.S.
Takafuta
T.
, et al.  . 
Mutations in the gene encoding filamin B disrupt vertebral segmentation, joint formation and skeletogenesis
Nat. Genet.
 , 
2004
, vol. 
36
 (pg. 
405
-
410
)
24
Borrego-Diaz
E.
Kerff
F.
Lee
S.H.
Ferron
F.
Li
Y.
Dominguez
R.
Crystal structure of the actin-binding domain of alpha-actinin 1: evaluating two competing actin-binding models
J. Struct. Biol.
 , 
2006
, vol. 
155
 (pg. 
230
-
238
)
25
Lee
S.H.
Weins
A.
Hayes
D.B.
Pollak
M.R.
Dominguez
R.
Crystal structure of the actin-binding domain of alpha-actinin-4 Lys255Glu mutant implicated in focal segmental glomerulosclerosis
J. Mol. Biol.
 , 
2008
, vol. 
376
 (pg. 
317
-
324
)
26
Barlow
D.J.
Thornton
J.M.
Ion-pairs in proteins
J. Mol. Biol.
 , 
1983
, vol. 
168
 (pg. 
867
-
885
)
27
Shoichet
B.K.
Baase
W.A.
Kuroki
R.
Matthews
B.W.
A relationship between protein stability and protein function
Proc. Natl Acad. Sci. USA
 , 
1995
, vol. 
92
 (pg. 
452
-
456
)
28
Feng
Y.
Chen
M.H.
Moskowitz
I.P.
Mendonza
A.M.
Vidali
L.
Nakamura
F.
Kwiatkowski
D.J.
Walsh
C.A.
Filamin A (FLNA) is required for cell–cell contact in vascular development and cardiac morphogenesis
Proc. Natl Acad. Sci. USA
 , 
2006
, vol. 
103
 (pg. 
19836
-
19841
)
29
Moores
C.A.
Kendrick-Jones
J.
Biochemical characterisation of the actin-binding properties of utrophin
Cell Motil. Cytoskeleton
 , 
2000
, vol. 
46
 (pg. 
116
-
128
)
30
Leslie
A.G.W.
Recent changes to the MOSFLM package for processing film and image plate data
Joint CCP4 and ESF-EAMCB Newslett. Protein Crystallogr.
 , 
1992
, vol. 
26
 
31
Evans
P.
Scaling and assessment of data quality
Acta Crystallogr. D Biol. Crystallogr.
 , 
2006
, vol. 
62
 (pg. 
72
-
82
)
32
McCoy
A.J.
Grosse-Kunstleve
R.W.
Adams
P.D.
Winn
M.D.
Storoni
L.C.
Read
R.J.
Phaser crystallographic software
J. Appl. Crystallogr.
 , 
2007
, vol. 
40
 (pg. 
658
-
674
)
33
Franzot
G.
Sjoblom
B.
Gautel
M.
Djinovic Carugo
K.
The crystal structure of the actin binding domain from alpha-actinin in its closed conformation: structural insight into phospholipid regulation of alpha-actinin
J. Mol. Biol.
 , 
2005
, vol. 
348
 (pg. 
151
-
165
)
34
Emsley
P.
Cowtan
K.
Coot: model-building tools for molecular graphics
Acta Crystallogr. D Biol. Crystallogr.
 , 
2004
, vol. 
60
 (pg. 
2126
-
2132
)
35
Adams
P.D.
Grosse-Kunstleve
R.W.
Hung
L.W.
Ioerger
T.R.
McCoy
A.J.
Moriarty
N.W.
Read
R.J.
Sacchettini
J.C.
Sauter
N.K.
Terwilliger
T.C.
PHENIX: building new software for automated crystallographic structure determination
Acta Crystallogr. D Biol. Crystallogr.
 , 
2002
, vol. 
58
 (pg. 
1948
-
1954
)
36
Murshudov
G.N.
Vagin
A.A.
Dodson
E.J.
Refinement of macromolecular structures by the maximum-likelihood method
Acta Crystallogr. D Biol. Crystallogr.
 , 
1997
, vol. 
53
 (pg. 
240
-
255
)
37
Davis
I.W.
Leaver-Fay
A.
Chen
V.B.
Block
J.N.
Kapral
G.J.
Wang
X.
Murray
L.W.
Arendall
W.B.
III
Snoeyink
J.
Richardson
J.S.
, et al.  . 
MolProbity: all-atom contacts and structure validation for proteins and nucleic acids
Nucleic Acids Res.
 , 
2007
, vol. 
35
 (pg. 
W375
-
W383
)
38
Reynolds
C.
Damerell
D.
Jones
S.
ProtorP: a protein–protein interaction analysis server
Bioinformatics
 , 
2009
, vol. 
25
 (pg. 
413
-
414
)
39
Thompson
J.D.
Higgins
D.G.
Gibson
T.J.
CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice
Nucleic Acids Res.
 , 
1994
, vol. 
22
 (pg. 
4673
-
4680
)
40
DeLano
W.L.
The PyMOL Molecular Graphics System
 , 
2002
San Carlos, CA, USA
DeLano Scientific
41
Liang
C.C.
Park
A.Y.
Guan
J.L.
In vitro scratch assay: a convenient and inexpensive method for analysis of cell migration in vitro
Nat. Protoc.
 , 
2007
, vol. 
2
 (pg. 
329
-
333
)
42
Collins
T.J.
ImageJ for microscopy
Biotechniques
 , 
2007
, vol. 
43
 (pg. 
25
-
30
)
43
Gillies
R.J.
Didier
N.
Denton
M.
Determination of cell number in monolayer cultures
Anal. Biochem.
 , 
1986
, vol. 
159
 (pg. 
109
-
113
)

Author notes

Coordinates and structure factors are available in the Protein Data Bank, FLNA ABD: WT, 3HOP; E254K, 3HOC; WT (reduced), 3HOR.