Skip to main content
NCBI home page
Protein Science : A Publication of the Protein Society logoLink to Protein Science : A Publication of the Protein Society
. 2025 Apr 18;34(5):e70122. doi: 10.1002/pro.70122

Integrated structural model of the palladin–actin complex using XL‐MS, docking, NMR, and SAXS

Rachel Sargent 1, David H Liu 1, Rahul Yadav 1,2, Drew Glennenmeier 1, Colby Bradford 1, Noely Urbina 1, Moriah R Beck 1,✉
PMCID: PMC12006749  PMID: 40248864

Abstract

Palladin is an actin‐binding protein that accelerates actin polymerization and is linked to the metastasis of several types of cancer. Previously, three lysine residues in an immunoglobulin‐like domain of palladin have been identified as essential for actin binding. However, it is still unknown where palladin binds to F‐actin. Evidence that palladin binds to the sides of actin filaments to facilitate branching is supported by our previous study showing that palladin was able to compensate for Arp2/3 in the formation of Listeria actin comet tails. Here, we used chemical crosslinking to covalently link palladin and F‐actin residues based on spatial proximity. Samples were then enzymatically digested, separated by liquid chromatography, and analyzed by tandem mass spectrometry. Peptides containing the crosslinks and specific residues involved were then identified for input to the HADDOCK docking server to model the most likely binding conformation. Small‐angle x‐ray scattering was used to provide further insight into palladin flexibility and the binding interface, and NMR spectra identified potential interactions between palladin's Ig domains. Our final structural model of the F‐actin:palladin complex revealed how palladin interacts with and stabilizes F‐actin at the interface between two actin monomers. Three actin residues that were identified in this study also appear commonly in the actin‐binding interface with other proteins such as myotilin, myosin, and tropomodulin. An accurate structural representation of the complex between palladin and actin extends our understanding of palladin's role in promoting cancer metastasis through the regulation of actin dynamics.

Keywords: actin complex, crosslinking mass spectrometry, docking, Ig domain, NMR, SAXS

1. INTRODUCTION

Actin is a cytoskeletal protein that is abundant in eukaryotic cells and is crucial for cell shape and motility. Actin performs its functions through the dynamic reorganization of monomers into filaments and vice versa through polymerization or depolymerization, respectively. Many proteins are required to modulate actin's functions in the cytoskeleton and muscle, and these proteins are known as actin‐binding proteins (ABPs). Palladin is a widely expressed ABP that is known to bind to actin and promote complex structures known as crosslinks or bundles. Palladin is critical during times of high cell motility, such as embryonic development (Luo et al., 2005) and wound healing (Chang et al., 2015; Ronty et al., 2006), and is also known to be upregulated in several types of metastatic cancers, including breast (Goicoechea et al., 2009), pancreatic (Goicoechea et al., 2014), adult gliomas (Mayer et al., 2022), renal cell carcinoma (Gupta et al., 2011), and non‐small cell lung cancer (Shu et al., 2023).

There are over 150 known ABPs that bind, bundle, cap, sever, and further regulate actin within the cell. Some actin‐binding motifs in ABPs have been identified: calponin homology (CH) (Gimona et al., 2001), Wiskott‐Aldrich syndrome homology (WH2) (Paunola et al., 2001), gelsolin homology (Way et al., 1991), actin depolymerizing factor (ADF‐H) domains (Poukkula et al., 2011), and RPEL domains (containing RPxxxEL repeats) (Guettler et al., 2008). However, several immunoglobulin‐like (Ig) domain‐containing proteins, like palladin and its protein family members, are also known to bind to actin despite not containing these recognized actin‐binding motifs. Although the longest isoform of palladin contains five Ig domains, only the Ig3 domain of palladin directly binds to actin—both the monomeric actin (G‐actin) (Albraiki et al., 2023) and F‐actin forms (Dixon et al., 2008). Interestingly, although the Ig4 domain does not bind to actin, the tandem Ig3‐Ig4 (Ig3‐4) construct binds and bundles F‐actin more efficiently than Ig3 alone (Dixon et al., 2008), perhaps due in part to a unique, lengthier linker region between the Ig3 and Ig4 domains of palladin. Therefore, our experiments presented here focus on the isolated Ig3 and the tandem Ig3‐4 domains to explore possible structural reasons for this enhancement in binding.

Palladin has two solvent‐exposed basic patches on the surface of the Ig3 structure (PDB 2LQR Mus musculus, PDB 2DM2 Homo sapiens), and a previous study has identified several lysine residues that are critical for binding to negatively charged actin filaments (Beck et al., 2013). While the isolated Ig3 and Ig4 (PDB 2DM3 H. sapiens) domain structures have been solved by solution state NMR, no structural information is available to discern how the tandem Ig3‐4 construct is oriented or where the Ig3 domain binds to actin filaments. Interestingly, a previous study of palladin's role in actin comet tail formation of Listeria monocytogenes showed that the ubiquitously expressed isoform of palladin (containing the Ig3, Ig4, and Ig5 domains) was able to compensate for Arp2/3's role in actin branching (Dhanda et al., 2018), suggesting that palladin is likely to bind to the sides of actin filaments or be integrated within the filament as it polymerizes.

To elucidate the molecular mechanism of palladin interaction with F‐actin and the role of this interface in actin dynamics, we used an integrative structural approach involving crosslinking mass spectrometry (XL‐MS), molecular docking, and small angle x‐ray scattering (SAXS) to build the first structural model of the palladin:F‐actin complex (Gimona et al., 2001). To begin with, SAXS data are presented that indicate a high degree of flexibility between Ig3 and Ig4 in the tandem construct and provide the first information regarding the orientation of these two domains in solution. NMR data also indicate an interaction between the two Ig domains. Chemical crosslinking coupled with mass spectrometry has become an increasingly popular way of identifying bound protein interfaces and residues involved in binding (Piersimoni et al., 2022). Covalent chemical crosslinkers vary in composition but share a common format: an active group on either end that binds to specific amino acids connected by a spacer arm. We utilized several chemical crosslinkers in this study with varying head groups and spacer arm lengths. Crosslinked samples of 1:1 actin to palladin complexes were trypsinized, subjected to liquid chromatography (LC), and assessed by tandem mass spectrometry (MS/MS). Interprotein linkages were identified and used as restraints to dock protein structures and generate a bound palladin:F‐actin model using HADDOCK software. Docked structures were further refined based on normalized Xwalk scores, which emphasize the crosslinking restraints (Orban‐Nemeth et al., 2018a). Based on our results, we present the first integrated structural model of palladin's side‐binding interaction with actin and identify actin residues in the interface that are commonly involved in the binding of other ABPs to actin as well. Considering palladin's role in cancer cell metastasis, structural insights into the interactions between these two proteins could be useful in screening and developing chemotherapeutic agents that target this interaction and prevent cancer cell metastasis.

2. RESULTS

2.1. Ig3 and Ig4 domains display conformational plasticity

Both AlphaFold and Rosetta predictions of the structure of the tandem Ig3‐4 domain suggest that the linker region is highly disordered; however, it is unclear how the two Ig domains orient with respect to one another given the implicit uncertainty associated with the flexible linker region. To investigate interdomain dynamics and flexibility of palladin's Ig3‐4 domain, SAXS data was collected for the tandem domains as well as each isolated domain in solution. We monitored the maximum scattering intensities extrapolated to zero angle (Io) for a concentration series of Ig3, Ig4, and Ig3‐4 and observed a minimal concentration dependence for each construct (Supplemental Figure S1). Therefore, two data curves were merged per protein as described in the methods section. Analysis of the Guinier region for each protein showed a linear fit with residuals distributed evenly around zero, indicating monodisperse samples without interparticle interactions (Supplemental Figure S2). The normalized Kratky plots provide information on the compactness and flexibility of each protein (Figure 1a). All proteins have a peak around 0.1–0.15 Å−1 that is indicative of a globular protein. Each data curve also dips around 0.3 Å−1 but does not return to baseline, which implies flexibility. The pair distance distribution function (PDDF) plot shows single peaks for Ig3 and Ig4, indicating a spherical shape or globular protein as expected (Figure 1b). The Ig3‐4 PDDF shows one main peak as well as a shoulder peak around 40–45 Å, which suggests a dumbbell shape that is typical for a multidomain protein with two relatively isolated domains connected by a flexible linker (Figure 1b).

FIGURE 1.

FIGURE 1

Open in a new tab

SAXS structural data and EOM structures. (a) Kratky plot and (b) PDDF of Ig3 (blue), Ig4 (cyan), and Ig3‐4 (red). (c) Goodness‐of‐fit between experimental SAXS scattering (blue) and theoretical scattering from EOM models (orange). (d, e) The two most likely conformations of Ig3‐4 as predicted by EOM with each contributing ~25% fractional occupancy.

The quality of the experimental data was assessed using the FoXS online server, which compares experimental SAXS data with theoretical scattering profiles generated from atomic models (Schneidman‐Duhovny et al., 2013; Schneidman‐Duhovny et al., 2016). FoXS was applied to the solved NMR structures of individual domains (Ig3 and Ig4), producing χ 2 values of 3.17 and 10.6, respectively, indicating good agreement with the experimental data (Supplemental Figure S3a,b). The structure of the Ig3‐4 tandem domain was excised from the AlphaFold‐predicted palladin model (Q9ET54) and compared to the experimental Ig3‐4 scattering data using FoXS (Supplemental Figure S3c) (Jumper et al., 2021). A higher χ2 value of 38.9 suggested that the predicted orientation of Ig3‐4 does not match the experimental results.

Similarly, CRYSOL, a tool for generating theoretical scattering profiles from atomic structures, was used to evaluate the AlphaFold‐predicted Ig3‐4 structure against the experimental data. This analysis produced an even higher χ 2 value of 85.471, further highlighting discrepancies between the model and experimental results. To explore alternative structural arrangements, SASREF rigid body structural modeling was performed using CRYSOL‐generated scattering amplitude files for the individual domains (Petoukhov & Svergun, 2005). However, these models also showed low correlation with the experimental Ig3‐4 scattering curve, emphasizing significant deviations between the predicted and experimental structures and the need for further refinement.

To account for the likely flexibility between Ig3 and Ig4 within the tandem domain, molecular ensemble models were generated by the Ensemble Optimization Modeling (EOM) software (Bernado et al., 2007; Tria et al., 2015). Sub‐program RANCH created an ensemble of 10,000 possible conformations, and sub‐program GAJOE used a genetic algorithm to determine the conformations consistent with the experimental SAXS data. The best fit to the data produced a χ 2 value of 1.077, and the fit between experimental and theoretical scattering as well as distribution plots for D max, N–C distance, and R g distance are included in Figure 1c and Supplemental Figure S4. GAJOE returned an ensemble of 8–12 conformations as the best fit to the experimental data across 10 independent trials. Two components of this ensemble were consistently found to be the predominant contributors (at approximately 25% occupancy each). These models represent two distinct subunit arrangements, with one adopting a linear conformation (Figure 1d) and the other adopting "V‐"like shape (Figure 1e). In both models, the flexible linker is compact. The remaining models each contribute to the ensemble at an occupancy of <8%. These minor models adopt more extended conformations compared to the two major components (Supplemental Figure S4e). While it is possible that this collection of minor models serves to fit noise in the experimental scattering data, we also find it plausible that they represent a collection of more extended but less favored conformations present in solution. Taken together, our results suggest that the linker is typically compact but samples extended conformations at varying levels of elongation approximately half of the time.

2.2. Ig3 and Ig4 domain residues interact within the tandem Ig3‐4 construct

One factor that could enhance the F‐actin binding affinity and bundling efficiency of Ig3‐4 versus the minimal actin binding domain is an interaction between the Ig3 and Ig4 domains. To identify potential interactions, we collected separate 1H‐15N HSQC spectra of Ig3, Ig4, and Ig3‐4 (Figure 2a) and calculated the chemical shift perturbation (CSP) of residues from the individual versus tandem domains (Figure 2b). CSP values above a cutoff of 0.05 ppm were considered significant. While high CSP values near the C‐terminus and central linker region were not considered due to the likely flexibility of these regions, the chemical shifts of several residues within both the Ig3 and Ig4 regions significantly differed between the individual domains and tandem construct as highlighted in Figure 2c,d. Much more significant chemical shift perturbations were observed in Ig3 domain residues in comparison with Ig4 domain. This result is consistent with our previous result in which Ig3‐4 forms an intramolecular crosslinked species (Vattepu et al., 2015). This data also identifies the potential involvement of the following residues in the interdomain interactions between Ig3 and Ig4 as: G22, T26, H54, T56, H68, T70, T73, D76, L97, F144, L168, H198, A221, and F226. Moreover, an overlay of Ig3, Ig4, and Ig3 + Ig4 (mixture of separate domains, not tandem Ig3‐4) 15N‐HSQC spectra reveal no CSP in Ig3 or Ig4 with Ig3 + 4 (Supplemental Figure S5). The CSPs we observed with Ig3, Ig4, and tandem Ig3‐4 occur only when a linker connects both Ig3 and Ig4. This indicates that the linker plays an important role in maintaining a more productive conformation of Ig3‐4, which is reflected by CSPs in Ig3‐4. This is also consistent with our previous results, which show that a mixture of Ig3 and Ig4 does not enhance Ig3's actin binding and bundling (Dixon et al., 2008; Gurung et al., 2016; Vattepu et al., 2015); in contrast, the tandem Ig3‐4 domain substantially enhances palladin's binding and bundling activity. An additional observation is that neither the lipid‐binding nor actin‐binding regions of the Ig3 domain overlap with the Ig4 domain interacting regions (Yadav et al., 2016). This suggests that the interaction between the Ig3 and Ig4 domains would not interfere with lipid or actin binding.

FIGURE 2.

FIGURE 2

Open in a new tab

NMR results showing interaction between Ig3 and Ig4 domain residues. (a) HSQC spectra overlay of Ig3‐4 (red), Ig3 (green), and Ig4 (blue). Labeled residues were found to have a significant CSP between Ig3‐4 and Ig3 (no *) or Ig4 (*). (b) Chemical shift perturbation of residues from comparison of Ig3‐4 with the individual domains. Red dotted line indicates cutoff value over which CSPs were deemed significant. Residues above the cutoff value are highlighted in the (c) Ig3 and (d) Ig4 domains.

2.3. Integrative structural model of palladin:F‐actin complex

To validate and identify potential new contacts between F‐actin and palladin, we performed cross‐linking coupled to mass spectrometry (XL‐MS) experiments on Ig3 and tandem Ig3‐4 domains of palladin bound to F‐actin using a variety of crosslinking agents (BS3, DSSO, DFDNB, and DMTMM). Polymerized actin was incubated with various concentrations of the crosslinker and Ig3 or Ig3‐4 before the reaction was quenched and the samples were run on an SDS‐PAGE gel (Figure 3a, Supplemental Figure S6). Bands at the molecular weight for a palladin:actin heterodimer were excised, trypsinized, and analyzed by LC–MS/MS. Although several different crosslinking agents were used initially (including BSS, DSSO, and DFDNB), the zero‐length crosslinker 4‐(4,6‐dimethoxy‐1,3,5‐triazin‐2‐yl)‐4‐methylmorpholinium chloride (DMTMM), which links carboxylic acids and amine groups, consistently produced crosslinked species as expected, given the likely electrostatic nature of the interaction between palladin and F‐actin (Figure 3b). BS3, DSSO, and DFDNB were ultimately excluded from our analysis as MS/MS data yielded no crosslinks between actin and Ig3 when these crosslinking agents were used. Furthermore, the chemistry of these crosslinks would not be compatible with electrostatically driven binding interactions such as that expected for actin and Ig3. As expected, in both the Ig3 and Ig3‐4 results, most crosslinks were identified between lysine residues (K13, K36, K46, and K51) in the Ig3 domain and aspartic or glutamic acid residues on actin (D24, D25, D51, E99, and E100) (Figure 3b‐e and Table 1). Many of these actin residues have also been found in the binding interface between actin and other ABPs (Table 2). Consistent with previous studies finding no binding between F‐actin and the isolated Ig4 domain, no crosslinks were detected from Ig4 residues. Interestingly, although no interaction between actin and the isolated linker region has been detected previously (Dixon et al., 2008), results from the Ig3‐4 XL‐MS indicate that several residues at the end of the linker region are in close proximity with F‐actin. In contrast to the crosslinks from Ig3 to F‐actin, these link actin residue K328 to four negatively charged residues within the Ig3‐4 linker, D 127SGD 130 E 131NE 133 (Figure 3f). While the crosslinks to the linker region were not used in our current modeling experiments due to a lack of any structural information for this linker or tandem domain, these results could indicate an undiscovered function for this linker region that warrants further investigation.

FIGURE 3.

FIGURE 3

Open in a new tab

Overview of crosslinks and residues involved in palladin to actin binding as determined by XL‐MS. (a) Representative SDS‐PAGE gel lanes showing 1:1 palladin:actin crosslinked species. Bands representing Ig3WT (12 kDa), actin (42 kDa) and XL species at 54 kDa indicated on gel at left and Ig3‐4WT (26.4 kDa), actin (42 kDa), and XL species at 68 kDa on right (entire gels with ladders in Supplemental Figure S6). (b) Interprotein crosslinks shown as lines indicating general positioning and (c) enlarged to show specific linked residues between actin (top row) and Ig3‐4 (bottom row) with lines indicating specific links. (d) Actin dimer and (e) Ig3 showing positions of residues involved in the binding interface. Red residues were included in Pool 2 for HADDOCK trials and yellow residues were also included in Pool 1. Surface exposed lysine residues on Ig3 not found in this study are green. Orange residues were found in crosslinks with the Ig3‐4 linker region. (f) Orientation of palladin Ig3‐4 as predicted by AlphaFold. Labeled residues in the linker were identified in crosslinks by XL‐MS.

TABLE 1.

Interprotein links between actin and the palladin constructs as identified by chemical crosslinking followed by LC–MS/MS and pLink processing.

Actin residue Palladin residue Palladin construct
D24** K36 Ig3 (3/3), Ig3‐4 (4/4)
D24** K51 Ig3 (3/3), Ig3‐4 (4/4)
D25** K36 Ig3 (3/3), Ig3‐4 (4/4)
D25** K46 Ig3 (2/3), Ig3‐4 (4/4)
E99** K36 Ig3 (3/3), Ig3‐4 (4/4)
D24* K13 Ig3 (2/3), Ig3‐4 (2/4)
D51* K46 Ig3 (1/3), Ig3‐4 (3/4)
E99 K46 Ig3 (2/3), Ig3‐4 (1/4)
E100 K36 Ig3 (2/3), Ig3‐4 (1/4)
K328 E133 Ig3‐4 (4/4)
K328 E131 Ig3‐4 (4/4)
K328 D130 Ig3‐4 (4/4)
K328 D127 Ig3‐4 (4/4)
Open in a new tab

Note: Number in parentheses indicates how many replicates per construct identified the link. Links identified by fewer than three total trials are omitted. **Crosslinks included in Pool 2. *Crosslinks included in Pool 1, which include those in Pool 2.

TABLE 2.

List of actin binding proteins that include shared actin residues in the binding interface.

Actin binding protein Function Actin residues involved in binding interface
Myotilin (Kostan et al., 2021) Z‐disc structural component and influences sarcomere assembly D24, D25, E99, K328
Twitchin (Funabara et al., 2007) Elastic sarcomere thick filament that regulates muscle contraction and relaxation Region including D24, D25
Cardiac myosin binding protein C (MyBP‐C) M‐domain (Risi et al., 2022) Modulates cardiac contraction via direct interaction with myosin and actin filaments D24, D25
Tropomodulin (Yamashiro et al., 2010) Caps pointed end of actin filaments, binds monomeric actin, nucleates actin filaments D24, D25, E99
Vinculin (Golji & Mofrad, 2013) Recruits actin to focal adhesions and caps barbed end of actin filaments D25
Tropomyosin (Barua et al., 2011) Regulates actin function and stability in muscle and cytoskeletal actin filaments D25
αE‐catenin (Xu et al., 2020) Links cell–cell adhesion complexes to actin cytoskeleton and indirectly strengthens adherens junction linkage D24, D25
Myosin (Johara et al., 1993; Robert‐Paganin et al., 2021) Influences contraction in muscle and non‐muscle cells D24, D25, E99
Troponin (Lehman et al., 2021) Regulates skeletal and cardiac muscle contraction D24, D25
WASP (Pang et al., 2012) Promotes actin filament nucleation D25
Ciboulot (Hertzog et al., 2004) Promotes barbed end actin filament polymerization D24, D25
Thymosin‐β4 (Hertzog et al., 2004) Sequesters actin monomers D24, D25
Open in a new tab

Prior to incorporating identified crosslinks, HADDOCK 2.4 (Honorato et al., 2024) was used to dock Ig3 to F‐actin with no restraints (as shown in Supplemental Figure S7). To leverage the consistently observed crosslinks between Ig3 (and/or Ig3‐4) and F‐actin in biological and technical replicates, these crosslinks were divided into two sets of restraints for integrative structural modeling (Table 1). One set (Pool 1) included crosslinks observed in at least 6 out of 7 MS trials, while the other set (Pool 2) was limited to crosslinks present in all seven trials.

The two SAXS‐derived ensembles of Ig3‐4 were used independently as the ligand in separate HADDOCK docking experiments. Docking was further complicated by the inability to identify the specific actin subunit bound by the crosslinks in the filament. To address this, actin subunits were sequentially numbered, and all crosslinks were duplicated before macromolecular docking using DisVis (Honorato et al., 2021; van Zundert et al., 2017; van Zundert & Bonvin, 2015) and HADDOCK 2.4 (Honorato et al., 2024). DisVis was used to filter out false positives, and normalized Xwalk scores were calculated for all HADDOCK‐generated structures to validate crosslink consistency following the established workflow for creating structural models using XL‐MS distance restraints (Kahraman et al., 2011).

The top‐scoring HADDOCK structures consistently positioned the Ig3 domain in the cleft between the actin monomers, regardless of which SAXS ensemble structure was used (Supplemental Figure S7b). Restraints based on crosslinks found in all MS trials yielded the highest HADDOCK scores across both ensembles; however, the precise orientation of the Ig3 domain varied slightly, as shown in Supplemental Figure S7b, which depicts the orientations with the best HADDOCK scores. Unexpectedly, the unrestrained docked structures and the top‐scoring HADDOCK structures placed Ig3 on distinctly different faces between actin monomers, rotated by about 90° from each other (Supplemental Figure S7b). The crosslinks were then evaluated based on their fit in the preliminary models using normalized Xwalk scores. Only crosslinks in agreement with these preliminary models were retained for a refined HADDOCK docking to generate refined structures of the complex for both the V‐shaped and linear EOM models, as described in the methods section (Orban‐Nemeth et al., 2018a; Orban‐Nemeth et al., 2018b). This refinement approach converged on a single placement of Ig3 at the interface between actin subunits, as shown in Figure 4, which was consistent with both EOM models. Notably, the crosslinks excluded by the Xwalk scores were inconsistent across HADDOCK runs, and there was no significant correlation between HADDOCK and Xwalk scores (Tables S1 and S2). The resulting integrative structural model strongly suggests that the Ig3 domain resides at the interface between two adjacent actin monomers in the longitudinal direction, near the DNase I‐binding loop (residues 39–51), while the Ig4 domain projects away from the actin filament in both conformations, as shown in Figure 4b,c. This structure contrasts with the model of myotilin, where the tandem Ig domains span subdomain 1 of adjacent actin subunits (Kostan et al., 2021).

FIGURE 4.

FIGURE 4

Open in a new tab

Refined Ig3‐4:F‐actin complex models. (a) The top two structures for each of the EOM models were calculated by HADDOCK using the Xwalk score to provide measure of how well the docked structures align with experimental crosslinking restraints. Yellow and magenta structures represent the linear model while the cyan and green represent the V‐shaped models from ensemble fitting. These refined models converge from initial trials to position the Ig3 domain between two actin subunits. Panels (b) and (c) provide expanded view with experimental crosslinked residue side chains shown in red for actin and blue for Ig3. The DNase‐I binding loop of actin is also indicated with arrow to highlight interaction with palladin.

2.4. Validation of lysine residue involvement in actin binding

While prior research has indicated two basic patches on opposite faces of Ig3 are involved in actin binding (Beck et al., 2013), this new structural model of the complex between palladin's Ig3 domain and actin indicates additional basic residues are involved in this interaction. We set out to test these predictions by generating point mutations at K13, K36, and K46, which were each consistently identified in crosslinks with actin and had not previously been shown to be involved in F‐actin binding. Mutations at K36 and K46 significantly reduced the apparent binding affinity for F‐actin with respect to the WT Ig3 domain, while K13 showed diminished binding affinity, and no binding affinity could be obtained for the K51 binding curve (Figure 5). Collectively, these results validate the structural model and corroborate the involvement of electrostatic interactions involving multiple basic patches on palladin with the acidic surface of actin.

FIGURE 5.

FIGURE 5

Open in a new tab

Binding of WT and mutant forms of Ig3 to F‐actin. Point mutations of lysine residues in the Ig3 domain of palladin that were identified in XL‐MS data reduce interactions with F‐actin. Binding curves generated from in vitro co‐sedimentation assays with varying concentrations of F‐actin (0–30 μM) and a constant concentration of the Ig3 domain (10 μM) reveal significantly reduced apparent binding affinities (K d) as determined by exponential binding curves fitted for each set of data points with the exception of K51. Plotted values represent the mean ± standard deviation from three or more replicates.

3. DISCUSSION

Here, we have defined the molecular basis through which palladin binds to the sides of actin filaments (Figure 4). A combination of SAXS, NMR, XL‐MS, and mutagenesis studies indicates how distinct basic patches within the Ig3 domain of palladin contribute to binding actin. Previous work had established that the minimal domain required for F‐actin binding by palladin is the Ig3 domain (Dixon et al., 2008). Several lysine residues within two basic patches on the surface of palladin's Ig3 domain have been identified as being critical for F‐actin interactions. Initially, mutagenesis studies identified K15 and K18 as one basic patch and K51 on the opposite face of Ig3 to be crucial for F‐actin binding (Beck et al., 2013). However, mutagenesis of individual residues was not enough to eliminate all F‐actin binding, which required double and triple mutants that also resulted in nuclear localization when expressed in cells. A later study found that K38, which is near K51 in the structure, may be part of the second basic patch with K51 and influence actin polymerization and bundling (Yadav et al., 2016). Here, mass spectrometry identification of chemically crosslinked peptides of the Ig3:actin complex revealed that K13, K36, K46, and K51 are in close proximity to D24, D25, and E99, all residues of actin's most negatively charged subdomain 1 (Sutoh & Yin, 1989). Co‐sedimentation studies revealed reduced actin binding for all four lysine mutants, reinforcing the idea of two distinct basic patches interacting at the interface between two actin monomers. This interaction aligns with the orientation observed in several other side‐binding proteins (Figure 6). These structural insights have significant implications for understanding how palladin modulates actin binding, bundling, and polymerization. The presence of distinct basic patches likely allows for versatile regulation of these functions, enabling palladin to fulfill its diverse roles in cytoskeletal organization and dynamics.

FIGURE 6.

FIGURE 6

Open in a new tab

Overlay of docked Ig3 (green) to actin (gray) structure with other ABPs that bind to the same actin residues (cyan). (a) ciboulot, PDB: 1SQK; (b) MyBP‐C C2 domain, PDB: 7LRG; (c) N‐WASP VC domain, PDB: 2VCP; (d) vinculin tail domain, PDB: 3JBI; (e) αE‐catenin, PDB: 6WVT; (f) MyBP‐C M‐domain, PDB: 7TIJ; (g) myosin, PDB: 7JH7; and (h) troponin C, PDB: 8UZX.

Protein regulation works by adjusting the conformational states, which are often regulated by binding interactions. Our previous study of Ig3 and Ig3‐4 with membrane phosphoinositide, PI(4,5)P2, indicated that the binding of PI(4,5)P2 reduces the actin‐binding and polymerizing activity of both Ig3 and Ig3‐4 (Yadav et al., 2016). Moreover, the chemical shift perturbation experiment of Ig3 and inositol 1,4,5‐trisphosphate (IP3) pinpointed a region involving residues K38, Q47, and K51 in this interaction (Yadav et al., 2016). The similarity of the region involved in both interactions (Ig3/actin and Ig3/PI4,5P2) and the effect of PI4,5P2 binding in the reduction of Ig3/Ig3‐4 actin binding and polymerizing activity reinforces an essential role of the region involving K36, K46, and K51 in actin binding and its activity.

We also showed that the chemical shift perturbations observed for the Ig3‐4 tandem domain, when compared with the isolated Ig3 and Ig4 domains, identified a region (residues T26, H54, T56, H68, T70, and T73) that does not overlap with the Ig3:actin interacting residues identified in this manuscript or the Ig3:lipid interacting residues previously identified (Yadav et al., 2016). This suggests that the interaction between the Ig3 and Ig4 domains would not interfere with lipid or actin binding. Furthermore, the refined Ig3‐4 models based on SAXS ensemble modeling shown in Figure 4 substantiate that the interface between the Ig3 and Ig4 domains does not overlap with Ig3's actin‐binding site. In addition, it is also likely that Ig3‐4 could adopt a more extended conformation in the presence of actin or that fractional occupancy of other conformations could change upon actin binding, favoring higher actin binding and bundling activity by Ig3‐4 in comparison to Ig3 alone.

Palladin and actin binding dynamics are further complicated by the formation of a palladin homodimer in the presence of actin, and prior chemical crosslinking studies found that the palladin Ig3‐4 tandem domain forms an intramolecular crosslinked species even in the absence of actin (Vattepu et al., 2015). These intramolecularly crosslinked species showed a reduced ability to bind and bundle actin. Cleavage of the crosslinking agent used partially restored binding and bundling ability despite remaining modifications on the lysine residues, indicating that flexibility and orientation between Ig3 and Ig4 are critical for these functions. In this study, we identified several positions in the linker (D127, D130, E131, and E133) that are involved in crosslinks to actin at K328. While these residues may not be specifically involved in interactions between Ig3 and Ig4, the fact that they were able to be crosslinked indicates that some surface‐exposed residues could be obstructed in this intramolecularly linked species. This also raises the possibility that binding to actin induces a conformational change within Ig3‐4, exposing a dimerization site. Further research into the specific mechanism of palladin dimerization and interactions between Ig3 and Ig4 is clearly needed.

While Ig3 is the only palladin domain known to bind to actin, the enhanced binding and bundling ability of Ig3‐4 warrants further investigation. The mechanism for this amplification could be due to either the presence of the Ig4 domain or the unique, unstructured linker region. Our data presented here suggest that both of these factors contribute. The XL‐MS data provide the first evidence of an interaction between the residues in palladin's linker and residues on the surface of actin. HSQC spectroscopy shows CSPs of residues in both the Ig3 and Ig4 domains, indicating potential binding between the domains. It has previously been found that both Ig3 and Ig3‐4 form homodimers in the presence of actin as well as an intramolecularly crosslinked Ig3‐4 species. XL‐MS of these dimeric and crosslinked species would provide further insight into potential binding between palladin's Ig domains. Little research has been conducted on the linker region, but several factors could impact it, including its extended length, overall charge, and specific composition of amino acids.

Various lines of evidence support the notion that basic residues on the surface of the Ig3 domain of palladin are critical for actin binding, displaying a binding mode similar to several other side‐binding proteins, as shown in Figure 6. Yet, the functions of these actin‐binding proteins vary widely, as highlighted in Table 2. Palladin, as the founding member of a unique family of actin‐binding proteins that are involved in controlling the architecture and dynamics of diverse and versatile actin‐based structures, displays a wide variety of roles. Besides binding and crosslinking actin filaments, palladin has also been shown to promote nucleation of actin polymerization (Gurung et al., 2016) and can functionally replace the Arp2/3 complex in Listeria actin comet tail motility (Dhanda et al., 2018). Based on our results, we present a structural model that is consistent with these diverse roles (Figure 7). In conclusion, our results clearly show that the Ig3‐4 tandem domain of palladin is flexible, given the unusually long linker region, and is thus able to accommodate both an extended, linear and a V‐shaped conformation, with possible implications for different functions in the cell, as indicated in our model (Figure 7). Although the actin‐binding interface and Ig domain dynamics differ from what was observed for myotilin (Kostan et al., 2021; Puz et al., 2017), further studies are required to determine how this family of proteins carries out both overlapping and distinct functions in a wide array of cellular roles.

FIGURE 7.

FIGURE 7

Open in a new tab

Model illustrating multiple roles of palladin in actin dynamics and organization. Palladin interacts with actin to influence various stages of filament assembly and network formation. (a) Actin nucleation: The Ig3 domain of palladin binds directly to actin monomers and/or seeds, stabilizing them and facilitating the nucleation phase of actin polymerization. This interaction occurs through palladin binding between actin protomers, as indicated by our integrative structural model, providing structural support and enhancing the efficiency of filament initiation. (b) Conformational change and dimerization: The Ig3‐4 domains of palladin exhibit two distinct conformations: a V‐shaped conformation, where the two domains form an angle, and a linear conformation, where the domains align along an extended axis. These conformations are connected by a flexible linker, allowing dynamic adjustments to accommodate different actin‐binding or dimerization scenarios. Upon binding to actin, palladin undergoes a conformational change that promotes its dimerization. This structural rearrangement enables palladin to mediate higher‐order interactions with actin filaments, a crucial step in organizing actin‐based structures. (c) Actin bundling: Palladin dimers align parallel actin filaments into tightly packed bundles. This bundling activity contributes to the formation of actin structures such as stress fibers, which provide mechanical strength and support to the cell. The flexibility of the Ig3‐4 linker may allow palladin to adjust its conformation to optimize filament alignment during bundling. (d) Actin Branching: In addition to bundling, palladin dimerization can also facilitate the branching of actin filaments. By serving as a scaffold or interacting with other actin‐regulating proteins, palladin promotes the formation of branched networks, which are essential for cell motility and the generation of protrusive forces, such as those seen in lamellipodia and podosomes. This model underscores the multifunctional nature of palladin, emphasizing its critical role in modulating actin dynamics. The structural flexibility of the Ig3‐4 domains further highlights its adaptability in shaping the cytoskeletal architecture necessary for diverse cellular functions.

4. MATERIALS AND METHODS

4.1. Protein expression and purification

Palladin domains were expressed and purified from BL21(DE3) Escherichia coli cells as previously described (Gurung et al., 2016). Several lysine residue mutants (K13A, K36A, K46A, and K51A) were generated from the pTBSG WT Ig3 construct using the Q5® site‐directed mutagenesis kit (New England Biolabs). Sequences were verified by DNA sequencing (Supplemental Note 1 provides each protein sequence). Briefly, the WT and mutant Ig3 and Ig4 domains in the pTBSG expression vector were expressed in E. coli using autoinduction media (Studier, 2005) and were subsequently lysed by sonication and centrifuged to pellet cell debris. The supernatant containing soluble protein was purified using HisPur Ni‐NTA resin (Thermo Scientific). The His6‐tag was cleaved using tobacco etch virus (TEV) protease and further purified by cation exchange chromatography (SP Sepharose, GE Healthcare Life Sciences). Protein was dialyzed to HEPES storage buffer (20 mM HEPES, pH 7.4, 1 mM DTT, 100 mM NaCl). The Ig3‐4 tandem domain was expressed from the pTBMalE expression vector that, in addition to the His6 tag, contains a maltose binding protein (MBP) tag for increased solubility. Ig3‐4 was purified using the same method as Ig3/Ig4, with the addition of an amylose column (New England Biolabs) chromatography step to remove the MBP tag prior to cation exchange. Actin was purified from rabbit muscle acetone powder (Pel‐Freez Biologicals) using an adapted method from Spudich and Watt (Spudich & Watt, 1971).

4.2. Small angle x‐ray scattering

After dialysis of purified palladin domains to HEPES storage buffer, proteins were further purified by size exclusion chromatography on the Cytiva Superdex 200 10/300 GL column in preparation for SAXS data collection. Ig3 and Ig4 were concentrated and prepared directly for SAXS, while Ig3‐4 was subjected to a final purification step on the Cytiva HiLoad 16/600 Superdex 200 pg. column to remove the remaining contaminants. Concentration series were prepared for each protein: 0.5–4.3 mg/mL for Ig3, 0.5–10 mg/mL for Ig4, and 0.5–7.4 mg/mL for Ig3‐4. SAXS data was collected at 12 frames per sample with 1 s exposures at the Stanford Synchrotron Radiation Library (SSRL) (Smolsky et al., 2007). Resulting data was automatically buffer‐subtracted and analyzed on‐line with SAXSPipe (Anonymous, 2024). The ATSAS software package, including PRIMUS, CRYSOL, and SASREF, was used for further analysis (Franke et al., 2017; Manalastas‐Cantos et al., 2021). Buffer‐subtracted files were compared at every concentration for each protein. Slight concentration dependence was observed for all domains, so the lower q range of a lower concentration sample was merged with the higher q range of a higher concentration sample using PRIMUS software (Manalastas‐Cantos et al., 2021). PRIMUS was further used for each merged data curve to estimate radius of gyration (R g) and the forward scattering intensity I(0) from the Guinier plot and Porod volume and R max from the distance distribution analysis.

Prior to structural modeling, sequence and structure files were optimized as follows. For Ig3, the N‐terminal “SNA” and “GGS” residues were removed from the sequence and PDB files, respectively, for sequence continuity. For Ig4, the C‐terminal “AHK” and “SGPSSG” residues were removed from the sequence and PDB files (2LQR and 2DM3), respectively. Furthermore, the sequence was modified with I179V and N224S to match the structure.

CRYSOL (maximum s value 0.4) (Franke et al., 2017; Svergun et al., 1995) and SASREF (Petoukhov & Svergun, 2005) were used to check data quality for Ig3 and Ig4 with known structures, while the FoXS webserver (Schneidman‐Duhovny et al., 2013; Schneidman‐Duhovny et al., 2016) was used to create and compare theoretical scattering curves to the experimental data and to create an initial rigid‐body model. The AlphaFold predicted structure for Ig3‐4 was used for experimental data comparison with FoXS. SASREF was run with the Ig3‐4 experimental data, and amplitudes for the individual domains were computed by CRYSOL. The molecular ensemble model was generated by the Ensemble Optimization Modeling (EOM) software (Bernado et al., 2007; Tria et al., 2015), a pipeline of SAXS analysis programs. Initially, the RANCH package was used to generate a pool of 10,000 models based on two rigid bodies defined by the known protein structures (PDB: 2LQR and 2DM3) while the missing linker region (residues 97–139 in the tandem Ig3‐4) was allowed to move freely. A theoretical scattering intensity curve based on the ensemble was generated using GAJOE, which fits the experimental SAXS data with theoretical scattering curves generated from weighted‐average sub‐ensembles. Since GAJOE is a genetic algorithm that produces degenerate results across independent trials, we conducted 10 separate runs and assessed the solutions for convergence. Default parameters were used in all cases. EOM is a semi‐quantitative approach to analyze the flexibility and size distribution of possible configurations, and we were able to obtain optimized ensembles with a fit to the experimental scattering data (χ 2 ~ 1.077) (Figure 4c,d).

4.3. 1H‐15N HSQC

All NMR experiments were performed at 298 K on a Bruker AVANCE 800 MHz spectrometer equipped with a triple resonance cryoprobe. The NMR buffer was 20 mM Hepes (pH 6.5), 100 mM NaCl, 2 mM DTT, 0.01% NaN3, and 10% D2O for spectrometer lock. NMR was processed using the NMRPipe program (Delaglio et al., 1995) and analyzed and visualized using NMRViewJ (Johnson & Blevins, 1994). A 1H‐15N HSQC spectrum was recorded separately for samples containing 0.35 mM 15N labeled Ig3, Ig4, and Ig3‐4 domains of palladin. Combined chemical shift perturbations (CSPs) were calculated using the equation Δδ = [((ΔδH)2 + (ΔδN/5)2)/2]∣1/2∣, where ΔδH and ΔδN are chemical shift changes of 1H and 15NH, respectively.

4.4. Chemical crosslinking

Actin was polymerized in 2× polymerization buffer (20 mM Tris at pH 8.0, 200 mM KCl, 2 mM MgCl2, and 4 mM DTT) for at least 30 min at room temperature. F‐actin (final concentration 10 μM) was incubated with varying amounts of DMTMM (2 or 5 mM) for 5 min at room temperature. Purified palladin (10–20 μM) in HEPES storage buffer was added, and the mixture was incubated for a further 40 min at room temperature. Next, 4× Laemmli buffer (200 mM Tris–HCl, pH 6.8, 8% SDS, 40% glycerol, 0.08% bromophenol blue, and 400 mM DTT) was added to quench the reaction, and samples were boiled and separated by SDS‐PAGE. Protein bands at a molecular weight according to a 1:1 ratio of palladin to actin were excised from the gel.

4.5. Gel‐based MS/MS analysis

Each SDS‐PAGE gel band was subjected to in‐gel trypsin digestion as follows. Gel segments were destained in 50% methanol (Fisher), 50 mM ammonium bicarbonate (Sigma‐Aldrich), followed by reduction in 10 mM Tris[2‐carboxyethyl]phosphine (Pierce) and alkylation in 50 mM iodoacetamide (Sigma‐Aldrich). Gel slices were then dehydrated in acetonitrile (Fisher), followed by the addition of 100 ng porcine sequencing grade modified trypsin (Promega) in 50 mM ammonium bicarbonate (Sigma‐Aldrich) and incubation at 37°C for 12–16 h. Peptide products were then acidified in 0.1% formic acid (Pierce). Tryptic peptides were then separated by reverse phase XSelect CSH C18 2.5 um resin (Waters) on an in‐line 150 × 0.075 mm column using an UltiMate 3000 RSLCnano system (Thermo). Peptides were eluted using a 75 min gradient from 98:2 to 65:35 buffer A:B ratio where Buffer A is 0.1% formic acid and 0.5% acetonitrile and Buffer B is 0.1% formic acid and 99.9% acetonitrile. Eluted peptides were ionized by electrospray (2.4 kV) followed by mass spectrometric analysis on an Orbitrap Eclipse Tribrid mass spectrometer (Thermo). MS data were acquired using the FTMS analyzer in profile mode at a resolution of 120,000 over a range of 375–1400 m/z with advanced peak determination. Following HCD activation, MS/MS data were acquired using the FTMS analyzer in profile mode at a resolution of 15,000 over a range of 150–2000 m/z with a stepped collision energy of 27%–33%.

4.6. Analysis and visualization of LC–MS/MS data

pLink 2 version 2.3.11 software (Chen et al., 2019) was used for the identification of crosslinked peptides, with searches performed against the forward and reverse protein sequences. Raw data files were analyzed using the default options, with the following exceptions: EDC‐DE was used as the crosslinker, carbamidomethyl[C] was fixed, oxidation[M] was variable, and E‐values were computed. Results were filtered at a tolerance of ±10 ppm and a false discovery rate (FDR) of 5% at the peptide spectrum matches (PSM) level. The xiNET cross‐link viewer (Combe et al., 2015) was used for the visualization of the crosslinks identified by pLink2. Results were uploaded to the Proxl (Protein Cross‐Linking Database) server (Riffle et al., 2016) and a link to the data is provided in Supplemental Note 2.

4.7. Protein docking with DisVis and HADDOCK

The interprotein crosslinks identified by pLink 2 were sorted into two categories: crosslinks present in at least six of seven trials (Pool 1) and crosslinks present in all trials (Pool 2). Crosslinks present in less than six trials were not utilized. For docking, all steps were performed separately for both Pool 1 and Pool 2. Palladin is hypothesized to bind between two actin monomers within a filament, and because there is no way to differentiate which monomer was actually linked, all crosslinks were duplicated to the second monomer. For actin, a dimer of chains D and F was extracted from PDB 3J8I and reformatted as one chain with sequential residue numbering. The top two EOM models of palladin Ig3‐4 were both used for all integrative modeling with HADDOCK, DisVis, and Xwalk. Links were reformatted as required by the docking software indicating the PDB file chain ID, residue number of the crosslink, atom position of the crosslink, and upper and lower distance restraints (23 and 24 Å for Lys–Asp and Lys–Glu links, respectively). Distance restraints were calculated by summing the Euclidean distance between the α‐carbon and the most distant atom on each residue in each crosslink with 13 Å added for potential conformation changes (Orban‐Nemeth et al., 2018a). DisVis software was used to identify the accessible interaction space and to detect crosslinks that violate the distance restraints (van Zundert et al., 2017; van Zundert & Bonvin, 2015). Crosslinks with a DisVis Z‐score of less than 0.5 were used in HADDOCK 2.4 (Honorato et al., 2021; 2024), and trials were run in duplicate, assigning crosslinks as either ambiguous or unambiguous. Center of mass restraints were selected, and the best models were identified using the workflow described by Orbán‐Németh et al. specifically optimized for structural modeling of protein complexes using distance restraints from XL‐MS data (Orban‐Nemeth et al., 2018a). Key to this workflow is the use of Xwalk (Kahraman et al., 2011) to calculate the Solvent Accessible Surface Distance (SASD) between the atoms of interest on palladin and actin while the surface was defined using the backbone alpha carbon atom coordinates, with a radius of 2 Å. A maximum distance of 50 Å was defined for inter‐molecular distances output. The top scoring cluster from each ensemble, as determined by normal Xwalk, was selected to undergo further refinement. This was accomplished by removing crosslinks with a SASD >28 Å within each HADDOCK cluster as determined by the normalized XL‐score (the area under the curve of a plot of the cumulative number of unique cross‐links up to this distance versus the SASDs of the experimental cross‐links for each predicted structural model divided by the product of the cutoff distance and the number of crosslinks under said cutoff) and docking again in HADDOCK 2.4 with selected interprotein crosslinks as restraints. The final refined complex structures were then ranked again by XL‐score.

4.8. Actin co‐sedimentation assays

Actin binding properties of the three Ig3 variants (K13A, K36A, and K46A) were compared to WT using an actin co‐sedimentation assay (Dixon et al., 2008). Briefly, we incubated 10 μM Ig3 protein with 20 μM F‐actin in actin polymerization buffer (10 mM Tris, pH 8.0, 100 mM KCl, 2 mM MgCl2, and 2 mM DTT) that was centrifuged at high speed (150,000g) for 30 min before separating the supernatant and pellet. Actin and Ig3 proteins in the supernatant and solubilized pellet were separated using 12% SDS‐PAGE, and protein bands were quantified using Fiji software (Schindelin et al., 2012). All binding assays were repeated at least three times, and averaged data with standard deviation error bars are presented.

AUTHOR CONTRIBUTIONS

Rachel Sargent: Writing – original draft; data curation; investigation; formal analysis; visualization; validation; supervision; methodology. David H. Liu: Investigation; writing – review and editing; visualization; validation; data curation; methodology. Rahul Yadav: Investigation; writing – review and editing; validation; formal analysis; methodology. Drew Glennenmeier: Investigation; formal analysis. Colby Bradford: Investigation; formal analysis. Noely Urbina: Investigation. Moriah R. Beck: Conceptualization; funding acquisition; writing – original draft; methodology; validation; visualization; project administration; supervision; resources.

CONFLICT OF INTEREST STATEMENT

The authors declare that they have no conflicts of interest with the contents of this article.

Supporting information

Data S1 Supporting Information.

Supplemental Note S1. Protein sequences.

Supplemental Note S2. Crosslinking data deposition link.

Table S1. Initial HADDOCK and Xwalk scores.

Table S2. Refined Xwalk and HADDOCK scores.

Figure S1. SAXS concentration series.

Figure S2. Guinier region linear fit.

Figure S3. FoXS scattering curves with χ 2 values.

Figure S4. Fit and distribution plots of RANCH/GAJOE Ig3‐4 modeling.

Figure S5. 15N‐HSQC spectra of Ig3, Ig4, and Ig3 + Ig4.

Figure S6. Crosslinked samples of palladin Ig3 and Ig3‐4.

Figure S7. HADDOCK structure of complex with no restraints and initial structures versus refined models.

PRO-34-e70122-s001.pdf (4.3MB, pdf)

ACKNOWLEDGMENTS

Research reported in this publication was supported by NIGMS of the National Institutes of Health under award number R15GM120670 and the Kansas INBRE, P20 GM103418. The authors thank Dr. Haifan Wu (Wichita State University, USA) for assistance with chemical crosslinking knowledge and data analysis software; the UAMS Proteomics Core for XL sample processing and LC/MS–MS; Drs. Thomas Weiss (SSRL, USA) and Allyn Schoeffler (Loyola University New Orleans, USA) for help with SAXS data collection and analysis support; and Dr. Justin Douglas (KU NMR Core, University of Kansas, USA) for assistance with NMR data collection. The FP7 WeNMR (project# 261572), H2020 West‐Life (project# 675858), the EOSC‐hub (project# 777536), and the EGI‐ACE (project# 101017567) European e‐Infrastructure projects are acknowledged for the use of their web portals, which make use of the EGI infrastructure with the dedicated support of CESNET‐MCC, INFN‐LNL‐2, NCG‐INGRID‐PT, TW‐NCHC, CESGA, IFCA‐LCG2, UA‐BITP, TR‐FC1‐ULAKBIM, CSTCLOUD‐EGI, IN2P3‐CPPM, CIRMMP, SURFsara, and NIKHEF, and the additional support of the national GRID Initiatives of Belgium, France, Italy, Germany, the Netherlands, Poland, Portugal, Spain, the UK, Taiwan, and the US Open Science Grid. Use of the Stanford Synchrotron Radiation Lightsource, SLAC National Accelerator Laboratory, is supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences under Contract No. DE‐AC02‐76SF00515. The SSRL Structural Molecular Biology Program is supported by the DOE Office of Biological and Environmental Research and by the National Institutes of Health, National Institute of General Medical Sciences (P30GM133894). The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of NIGMS or NIH.

Sargent R, Liu DH, Yadav R, Glennenmeier D, Bradford C, Urbina N, et al. Integrated structural model of the palladin–actin complex using XL‐MS, docking, NMR, and SAXS . Protein Science. 2025;34(5):e70122. 10.1002/pro.70122

Review Editor: Aitziber L. Cortajarena

DATA AVAILABILITY STATEMENT

The XL‐MS data that support the findings of this study are openly available in Proxl at http://www.yeastrc.org/proxl_public/projectReadProcessCode.do?code=seyd6pmwj5ybwuou17l56tusfalk4v4jowf0or6ni5kveknqjxue9tfgvge8d9rs, reference numbers 2295, 2293, 2292, 2272, 2271, 2270, 2268. SAXS data is available at SASBDB (Small Angle Scattering Biological Data Bank) at https://www.sasbdb.org/, IDs: DVE6 (Ig3), DVD6 (Ig4), and DVF6 (Ig3‐4).

REFERENCES

  1. Albraiki S, Ajiboye O, Sargent R, Beck MR. Functional comparison of full‐length palladin to isolated actin binding domain. Protein Sci. 2023;32:e4638. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Anonymous . SAXSPipe for Auotmated SAXS data processing and analysis. In SAXSPipe . 2024.
  3. Barua B, Pamula MC, Hitchcock‐DeGregori SE. Evolutionarily conserved surface residues constitute actin binding sites of tropomyosin. Proc Natl Acad Sci USA. 2011;108:10150–10155. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Beck MR, Dixon RD, Goicoechea SM, Murphy GS, Brungardt JG, Beam MT, et al. Structure and function of palladin's actin binding domain. J Mol Biol. 2013;425:3325–3337. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bernado P, Mylonas E, Petoukhov MV, Blackledge M, Svergun DI. Structural characterization of flexible proteins using small‐angle X‐ray scattering. J Am Chem Soc. 2007;129:5656–5664. [DOI] [PubMed] [Google Scholar]
  6. Chang EH, Gasim AH, Kerber ML, Patel JB, Glaubiger SA, Falk RJ, et al. Palladin is upregulated in kidney disease and contributes to epithelial cell migration after injury. Sci Rep. 2015;5:7695. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chen ZL, Meng JM, Cao Y, Yin JL, Fang RQ, Fan SB, et al. A high‐speed search engine pLink 2 with systematic evaluation for proteome‐scale identification of cross‐linked peptides. Nat Commun. 2019;10:3404. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Combe CW, Fischer L, Rappsilber J. xiNET: cross‐link network maps with residue resolution. Mol Cell Proteomics. 2015;14:1137–1147. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Delaglio F, Grzesiek S, Vuister GW, Zhu G, Pfeifer J, Bax A. NMRPipe: a multidimensional spectral processing system based on UNIX pipes. J Biomol NMR. 1995;6:277–293. [DOI] [PubMed] [Google Scholar]
  10. Dhanda AS, Vogl AW, Albraiki SE, Otey CA, Beck MR, Guttman JA. Palladin compensates for the Arp2/3 complex and supports actin structures during Listeria infections. MBio. 2018;9(2):e02259‐17. 10.1128/mBio.02259-17 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dixon RD, Arneman DK, Rachlin AS, Sundaresan NR, Costello MJ, Campbell SL, et al. Palladin is an actin cross‐linking protein that uses immunoglobulin‐like domains to bind filamentous actin. J Biol Chem. 2008;283(10):6222–6231. 10.1074/jbc.M707694200 [DOI] [PubMed] [Google Scholar]
  12. Franke D, Petoukhov MV, Konarev PV, Panjkovich A, Tuukkanen A, Mertens HDT, et al. ATSAS 2.8: a comprehensive data analysis suite for small‐angle scattering from macromolecular solutions. J Appl Cryst. 2017;50:1212–1225. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Funabara D, Hamamoto C, Yamamoto K, Inoue A, Ueda M, Osawa R, et al. Unphosphorylated twitchin forms a complex with actin and myosin that may contribute to tension maintenance in catch. J Exp Biol. 2007;210:4399–4410. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Gimona M, Djinovic‐Carugo K, Kranewitter WJ, Winder SJ. Functional plasticity of CH domains. FEBS Lett. 2001;513:98–106. [DOI] [PubMed] [Google Scholar]
  15. Goicoechea SM, Bednarski B, García‐Mata R, Prentice‐Dunn H, Kim HJ, Otey CA. Palladin contributes to invasive motility in human breast cancer cells. Oncogene. 2009;28:587–598. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Goicoechea SM, García‐Mata R, Staub J, Valdivia A, Sharek L, McCulloch C, et al. Palladin promotes invasion of pancreatic cancer cells by enhancing invadopodia formation in cancer‐associated fibroblasts. Oncogene. 2014;33:1265–1273. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Golji J, Mofrad MR. The interaction of vinculin with actin. PLoS Comput Biol. 2013;9:e1002995. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Guettler S, Vartiainen MK, Miralles F, Larijani B, Treisman R. RPEL motifs link the serum response factor cofactor MAL but not myocardin to Rho signaling via actin binding. Mol Cell Biol. 2008;28:732–742. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Gupta V, Bassi DE, Simons JD, Devarajan K, al‐Saleem T, Uzzo RG, et al. Elevated expression of stromal palladin predicts poor clinical outcome in renal cell carcinoma. PLoS One. 2011;6:e21494. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Gurung R, Yadav R, Brungardt JG, Orlova A, Egelman EH, Beck MR. Actin polymerization is stimulated by actin cross‐linking protein palladin. Biochem J. 2016;473:383–396. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hertzog M, van Heijenoort C, Didry D, Gaudier M, Coutant J, Gigant B, et al. The beta‐thymosin/WH2 domain; structural basis for the switch from inhibition to promotion of actin assembly. Cell. 2004;117:611–623. [DOI] [PubMed] [Google Scholar]
  22. Honorato RV, Koukos PI, Jiménez‐García B, Tsaregorodtsev A, Verlato M, Giachetti A, et al. Structural biology in the clouds: the WeNMR‐EOSC ecosystem. Front Mol Biosci. 2021;8:729513. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Honorato RV, Trellet ME, Jiménez‐García B, Schaarschmidt JJ, Giulini M, Reys V, et al. The HADDOCK2.4 web server for integrative modeling of biomolecular complexes. Nat Protoc. 2024;19:3219–3241. [DOI] [PubMed] [Google Scholar]
  24. Johara M, Toyoshima YY, Ishijima A, Kojima H, Yanagida T, Sutoh K. Charge‐reversion mutagenesis of Dictyostelium actin to map the surface recognized by myosin during ATP‐driven sliding motion. Proc Natl Acad Sci USA. 1993;90:2127–2131. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Johnson BA, Blevins RA. NMR view: a computer program for the visualization and analysis of NMR data. J Biomol NMR. 1994;4:603–614. [DOI] [PubMed] [Google Scholar]
  26. Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, et al. Highly accurate protein structure prediction with AlphaFold. Nature. 2021;596:583–589. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kahraman A, Malmstrom L, Aebersold R. Xwalk: computing and visualizing distances in cross‐linking experiments. Bioinformatics. 2011;27:2163–2164. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kostan J, Pavšič M, Puž V, Schwarz TC, Drepper F, Molt S, et al. Molecular basis of F‐actin regulation and sarcomere assembly via myotilin. PLoS Biol. 2021;19:e3001148. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Lehman W, Pavadai E, Rynkiewicz MJ. C‐terminal troponin‐I residues trap tropomyosin in the muscle thin filament blocked‐state. Biochem Biophys Res Commun. 2021;551:27–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Luo H, Liu X, Wang F, Huang Q, Shen S, Wang L, et al. Disruption of palladin results in neural tube closure defects in mice. Mol Cell Neurosci. 2005;29:507–515. [DOI] [PubMed] [Google Scholar]
  31. Manalastas‐Cantos K, Konarev PV, Hajizadeh NR, Kikhney AG, Petoukhov MV, Molodenskiy DS, et al. ATSAS 3.0: expanded functionality and new tools for small‐angle scattering data analysis. J Appl Cryst. 2021;54:343–355. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Mayer O, Bugis J, Kozlova D, Leemann A, Mansur S, Peerutin I, et al. Cytoskeletal protein palladin in adult gliomas predicts disease incidence, progression, and prognosis. Cancers (Basel). 2022;14:5130. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Orban‐Nemeth Z, Beveridge R, Hollenstein DM, Rampler E, Stranzl T, Hudecz O, et al. Author correction: structural prediction of protein models using distance restraints derived from cross‐linking mass spectrometry data. Nat Protoc. 2018b;13:1724. [DOI] [PubMed] [Google Scholar]
  34. Orban‐Nemeth Z, Beveridge R, Hollenstein DM, Rampler E, Stranzl T, Hudecz O, et al. Structural prediction of protein models using distance restraints derived from cross‐linking mass spectrometry data. Nat Protoc. 2018a;13:478–494. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Pang X, Zhou KH, Qin S, Zhou HX. Prediction and dissection of widely‐varying association rate constants of actin‐binding proteins. PLoS Comput Biol. 2012;8:e1002696. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Paunola E, Mattila PK, Lappalainen P. WH2 domain: a small, versatile adapter for actin monomers. FEBS Lett. 2001;513:92–97. [DOI] [PubMed] [Google Scholar]
  37. Petoukhov MV, Svergun DI. Global rigid body modeling of macromolecular complexes against small‐angle scattering data. Biophys J. 2005;89:1237–1250. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Piersimoni L, Kastritis PL, Arlt C, Sinz A. Cross‐linking mass spectrometry for investigating protein conformations and protein–protein interactions horizontal line a method for all seasons. Chem Rev. 2022;122:7500–7531. [DOI] [PubMed] [Google Scholar]
  39. Poukkula M, Kremneva E, Serlachius M, Lappalainen P. Actin‐depolymerizing factor homology domain: a conserved fold performing diverse roles in cytoskeletal dynamics. Cytoskeleton (Hoboken). 2011;68:471–490. [DOI] [PubMed] [Google Scholar]
  40. Puz V, Pavsic M, Lenarcic B, Djinovic‐Carugo K. Conformational plasticity and evolutionary analysis of the myotilin tandem Ig domains. Sci Rep. 2017;7:3993. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Riffle M, Jaschob D, Zelter A, Davis TN. ProXL (protein cross‐linking database): a platform for analysis, visualization, and sharing of protein cross‐linking mass spectrometry data. J Proteome Res. 2016;15:2863–2870. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Risi CM, Villanueva E, Belknap B, Sadler RL, Harris SP, White HD, et al. Cryo‐electron microscopy reveals cardiac myosin binding protein‐C M‐domain interactions with the thin filament. J Mol Biol. 2022;434:167879. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Robert‐Paganin J, Xu XP, Swift MF, Auguin D, Robblee JP, Lu H, et al. The actomyosin interface contains an evolutionary conserved core and an ancillary interface involved in specificity. Nat Commun. 2021;12:1892. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Ronty MJ, Leivonen SK, Hinz B, Rachlin A, Otey CA, Kähäri VM, et al. Isoform‐specific regulation of the actin‐organizing protein palladin during TGF‐beta1‐induced myofibroblast differentiation. J Invest Dermatol. 2006;126:2387–2396. [DOI] [PubMed] [Google Scholar]
  45. Schindelin J, Arganda‐Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, et al. Fiji: an open‐source platform for biological‐image analysis. Nat Methods. 2012;9:676–682. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Schneidman‐Duhovny D, Hammel M, Tainer JA, Sali A. Accurate SAXS profile computation and its assessment by contrast variation experiments. Biophys J. 2013;105:962–974. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Schneidman‐Duhovny D, Hammel M, Tainer JA, Sali A. FoXS, FoXSDock and MultiFoXS: single‐state and multi‐state structural modeling of proteins and their complexes based on SAXS profiles. Nucleic Acids Res. 2016;44:W424–W429. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Shu X, Chen M, Liu SY, Yu L, Sun LX, Sun LC, et al. Palladin promotes cancer stem cell‐like properties in lung cancer by activating Wnt/Beta‐catenin signaling. Cancer Med. 2023;12:4510–4520. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Smolsky IL, Liu P, Niebuhr M, Ito K, Weiss TM, Tsuruta H. Biological small‐angle x‐ray scattering facility at the Stanford Synchrotron Radiation Laboratory. J Appl Cryst. 2007;40:s453–s458. [Google Scholar]
  50. Spudich JA, Watt S. The regulation of rabbit skeletal muscle contraction. I. Biochemical studies of the interaction of the tropomyosin‐troponin complex with actin and the proteolytic fragments of myosin. J Biol Chem. 1971;246:4866–4871. [PubMed] [Google Scholar]
  51. Studier FW. Protein production by auto‐induction in high density shaking cultures. Protein Expr Purif. 2005;41:207–234. [DOI] [PubMed] [Google Scholar]
  52. Sutoh K, Yin HL. End‐label fingerprintings show that the N‐ and C‐termini of actin are in the contact site with gelsolin. Biochemistry. 1989;28:5269–5275. [DOI] [PubMed] [Google Scholar]
  53. Svergun DI, Barberato C, Koch MHJ. CRYSOL ‐ a program to evaluate X‐ray solution scattering of biological macromolecules from atomic coordinates. J Appl Crystallogr. 1995;28(6):768–773. 10.1107/s0021889895007047 [DOI] [Google Scholar]
  54. Tria G, Mertens HD, Kachala M, Svergun DI. Advanced ensemble modelling of flexible macromolecules using X‐ray solution scattering. IUCrJ. 2015;2:207–217. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. van Zundert GC, Bonvin AM. DisVis: quantifying and visualizing accessible interaction space of distance‐restrained biomolecular complexes. Bioinformatics. 2015;31(19):3222–3224. 10.1093/bioinformatics/btv333 [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. van Zundert GC, Trellet M, Schaarschmidt J, Kurkcuoglu Z, David M, Verlato M, et al. The DisVis and PowerFit web servers: explorative and integrative modeling of biomolecular complexes. J Mol Biol. 2017;429(3):399–407. 10.1016/j.jmb.2016.11.032 [DOI] [PubMed] [Google Scholar]
  57. Vattepu R, Yadav R, Beck MR. Actin‐induced dimerization of palladin promotes actin‐bundling. Protein Sci. 2015;24:70–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Way M, Pope B, Weeds A. Molecular biology of actin binding proteins: evidence for a common structural domain in the F‐actin binding sites of gelsolin and alpha‐actinin. J Cell Sci Suppl. 1991;14:91–94. [DOI] [PubMed] [Google Scholar]
  59. Xu XP, Pokutta S, Torres M, Swift MF, Hanein D, Volkmann N, et al. Structural basis of alphaE‐catenin‐F‐actin catch bond behavior. elife. 2020;9:e60878. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Yadav R, Vattepu R, Beck MR. Phosphoinositide binding inhibits actin crosslinking and polymerization by palladin. J Mol Biol. 2016;428:4031–4047. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Yamashiro S, Speicher KD, Speicher DW, Fowler VM. Mammalian tropomodulins nucleate actin polymerization via their actin monomer binding and filament pointed end‐capping activities. J Biol Chem. 2010;285:33265–33280. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Data S1 Supporting Information.

Supplemental Note S1. Protein sequences.

Supplemental Note S2. Crosslinking data deposition link.

Table S1. Initial HADDOCK and Xwalk scores.

Table S2. Refined Xwalk and HADDOCK scores.

Figure S1. SAXS concentration series.

Figure S2. Guinier region linear fit.

Figure S3. FoXS scattering curves with χ 2 values.

Figure S4. Fit and distribution plots of RANCH/GAJOE Ig3‐4 modeling.

Figure S5. 15N‐HSQC spectra of Ig3, Ig4, and Ig3 + Ig4.

Figure S6. Crosslinked samples of palladin Ig3 and Ig3‐4.

Figure S7. HADDOCK structure of complex with no restraints and initial structures versus refined models.

PRO-34-e70122-s001.pdf (4.3MB, pdf)

Data Availability Statement

The XL‐MS data that support the findings of this study are openly available in Proxl at http://www.yeastrc.org/proxl_public/projectReadProcessCode.do?code=seyd6pmwj5ybwuou17l56tusfalk4v4jowf0or6ni5kveknqjxue9tfgvge8d9rs, reference numbers 2295, 2293, 2292, 2272, 2271, 2270, 2268. SAXS data is available at SASBDB (Small Angle Scattering Biological Data Bank) at https://www.sasbdb.org/, IDs: DVE6 (Ig3), DVD6 (Ig4), and DVF6 (Ig3‐4).

Articles from Protein Science : A Publication of the Protein Society are provided here courtesy of The Protein Society

RESOURCES