Disruption of Ca²⁺/calmodulin:KSR1 interaction lowers ERK activation

Abstract

KSR1, a key scaffold protein for the MAPK pathway, facilitates ERK activation upon growth factor stimulation. We recently demonstrated that KSR1 binds the Ca²⁺‐binding protein calmodulin (CaM), thereby providing an intersection between KSR1‐mediated and Ca²⁺ signaling. In this study, we set out to generate a KSR1 point mutant with reduced Ca²⁺/CaM binding in order to unravel the functional implications of their interaction. To do so, we solved the structural determinants of complex formation. Using purified fragments of KSR1, we showed that Ca²⁺/CaM binds to the CA3 domain of KSR1. We then used in silico molecular modeling to predict contact residues for binding. This approach identified two possible modes of interaction: (1) binding of extended Ca²⁺/CaM to a globular conformation of KSR1‐CA3 via electrostatic interactions or (2) binding of collapsed Ca²⁺/CaM to α‐helical KSR1‐CA3 via hydrophobic interactions. Experimentally, site‐directed mutagenesis of the predicted contact residues for the two binding models favored that where collapsed Ca²⁺/CaM binds to the α‐helical conformation of KSR1‐CA3. Importantly, replacing KSR1‐Phe³⁵⁵ with Asp reduces Ca²⁺/CaM binding by 76%. The KSR1‐F³⁵⁵D mutation also significantly impairs the ability of EGF to activate ERK, which reveals that Ca²⁺/CaM binding promotes KSR1‐mediated MAPK signaling. This work, by uncovering structural insight into the binding of KSR1 to Ca²⁺/CaM, identifies a KSR1 single‐point mutant as a bioreagent to selectively study the crosstalk between Ca²⁺ and KSR1‐mediated signaling.

1. INTRODUCTION

The mitogen‐activated protein kinase (MAPK) pathway is a fundamental signaling cascade that transduces diverse extracellular stimuli into cellular outcomes (Guo et al., 2020). This pathway is initiated by the activation of Ras GTPase, which promotes the Raf/MEK/ERK kinase cascade. Once activated, ERK translocates to the nucleus, where it stimulates the expression of target genes, which ultimately influences key cellular processes like proliferation, differentiation, migration, and apoptosis (Sun et al., 2015). Scaffold proteins such as IQ motif‐containing GTPase‐activating protein 1 (IQGAP1) (Roy et al., 2005) and MEK partner 1 (MP1) (Schaeffer et al., 1998) have been documented to assemble MAPK proteins into signalosomes, thereby facilitating ERK activation. One of the best‐characterized MAPK scaffolds is kinase suppressor of Ras 1 (KSR1) (Raabe & Rapp, 2002). In response to Ras activation, KSR1 is dephosphorylated by protein phosphatase 2A (PP2A), which abrogates its inhibitory interaction with 14‐3‐3 in the cytosol. As a result, KSR1 translocates to the plasma membrane, where it promotes ERK activation by scaffolding Raf, MEK, and ERK (Frodyma et al., 2017; McKay et al., ²⁰¹¹; Yu et al., ¹⁹⁹⁸). Structurally, KSR1 contains five conserved area (CA) domains: an N‐terminal CA1 domain unique to KSR1 that contributes to Raf binding; a proline‐rich CA2 domain with unidentified function; a cysteine‐rich CA3 domain that mediates KSR1 recruitment to the plasma membrane; a serine/threonine‐rich CA4 domain where ERK binds via a F‐X‐F‐P motif; and a C‐terminal CA5 pseudokinase domain where MEK binds (Figure 1a) (Frodyma et al., 2017).

FIGURE 1.

The Ca²⁺/CaM binding site on KSR1 is between amino acid residues 354–377. (a) Schematic representation of KSR1. KSR1 has five conserved area (CA) domains, termed CA1 to CA5. (b) KSR1 constructs were used for the binding analyses. The constructs are KSR1‐M (amino acids 319–433), ‐M1 (319–376), ‐M2 (349–404), and ‐M3 (377–433). The red line encompasses the Ca²⁺/CaM binding region determined based on panels (c and d). (c) GST‐tagged fragments of KSR1 generated in Escherichia coli and purified on glutathione‐Sepharose were incubated with pure CaM in the presence of 1 mM CaCl₂. Pull‐down (PD) with GST‐coated Sepharose beads was the negative control. Input designates pure CaM not subjected to PD. Proteins bound to the beads were eluted and analyzed by SDS‐PAGE. The upper portion of the gel (above ~23 kDa) was stained with Coomassie blue; the lower portion was analyzed by Western blotting and probed for CaM. d. HEK293T cells were transiently transfected with GFP‐tagged wild‐type (WT) KSR1 or the deletion mutant construct KSR1Δ^354–377. Equal amounts of cell lysate were subjected to CaM‐Sepharose PDs in the presence of 1 mM CaCl₂. PDs with protein A‐Sepharose (ProA) beads were carried out in parallel as negative controls. Samples were resolved by Western blotting and probed with anti‐GFP antibodies. Input designates cell lysate not subjected to PD. All data shown in this figure are representative of three independent experiments. The position of migration of the molecular weight markers is shown on the left of the blots and gels. CaM, calmodulin.

Calcium ions are essential second messengers that transmit intracellular signals by binding to Ca²⁺‐binding proteins, including calmodulin (CaM), a highly conserved, ubiquitously expressed protein (Haeseleer et al., 2002). CaM consists of two lobes, each comprising two Ca²⁺‐binding EF‐hand motifs, separated by a flexible central linker. Ca²⁺/CaM generally adopts one of two main conformations when binding to its targets: an extended conformation where the two lobes are separated by the central linker or a compact globular conformation where the linker is collapsed, allowing the two CaM lobes to wrap around α‐helical target peptides (Tidow & Nissen, 2013). Ca²⁺ has been documented to alter MAPK signaling (Wiegert & Bading, 2011), but the detailed mechanisms are incompletely understood. We recently documented that CaM binds to KSR1 in a Ca²⁺‐dependent manner (Parvathaneni et al., 2021). Importantly, the chemical antagonism of CaM prevents KSR1 translocation to the plasma membrane and reduces activation of ERK by epidermal growth factor (EGF) (Parvathaneni et al., 2021). Thus, by binding Ca²⁺/CaM, KSR1 couples Ca²⁺ signaling to the MAPK cascade.

In the present study, we used a combination of in silico molecular modeling, site‐directed mutagenesis, and binding assays to identify essential residues on KSR1 that mediate Ca²⁺/CaM binding. Importantly, we show that the replacement of KSR1‐Phe³⁵⁵ with Asp reduces Ca²⁺/CaM binding by 76%. Moreover, the ability of EGF to activate ERK was significantly reduced in cells expressing KSR1‐F³⁵⁵D mutant protein. Together, this study yields structural insight into the binding properties of KSR1 and Ca²⁺/CaM, and provides a KSR1 mutant as a bioreagent to selectively study the functional effects of Ca²⁺/CaM binding to KSR1.

2. RESULTS

2.1. The Ca²⁺ /CaM binding site on KSR1 is located within amino acid residues 354–377

Our previous publication identifies that Ca²⁺/CaM binds to KSR1 between residues 319 and 433 (Parvathaneni et al., 2021). To narrow the binding site, we evaluated shorter KSR1 fragments that collectively cover residues 319–433. We generated GST fusion fragments of KSR1, termed KSR1‐M (amino acids 319–433), KSR1‐M1 (319–376), KSR1‐M2 (349–404), and KSR1‐M3 (377–433) (Figure 1b). Each construct was expressed in Escherichia coli, purified on a glutathione‐Sepharose column, and then incubated with pure CaM. Because Ca²⁺ is required for the interaction between KSR1 and CaM (Parvathaneni et al., 2021), all the binding assays described in this study were carried out in the presence of 1 mM CaCl₂. Western blotting reveals that Ca²⁺/CaM binds to KSR1‐M (Figure 1c), as previously observed (Parvathaneni et al., 2021). Furthermore, Ca²⁺/CaM binds to KSR1‐M1 and KSR1‐M2, but not to KSR1‐M3 (Figure 1c). The absence of binding to GST alone validates specificity (Figure 1c). These results indicate that Ca²⁺/CaM binds to KSR1 in the overlapping portion of KSR1‐M1 and KSR1‐M2, that is, amino acids 349–376.

To validate the Ca²⁺/CaM binding region, we generated a KSR1 deletion mutant construct lacking residues 354–377, termed KSR1Δ^354–377, and tagged it with GFP. We transiently expressed GFP‐KSR1Δ^354–377 or full‐length GFP‐KSR1 in HEK293T cells. Probing cell lysates for GFP reveals that the levels of expression of full‐length GFP‐KSR1 and GFP‐KSR1Δ^354–377 are approximately equal (Figure 1d, Inputs). The ability to bind Ca²⁺/CaM was assessed by CaM‐Sepharose pull‐downs (PDs) from cell lysates in the presence of CaCl₂. GFP‐tagged, full‐length KSR1 was pulled down by Ca²⁺/CaM, as expected. In contrast, we observed minimal binding of GFP‐KSR1Δ^354–377 to Ca²⁺/CaM (Figure 1d). Specificity is demonstrated by the lack of KSR1 in protein A‐Sepharose control PDs (Figure 1d). Taken together, these results indicate that amino acids 354–377 of KSR1 are necessary for Ca²⁺/CaM binding, suggesting that these residues encompass the binding site.

2.2. Binding simulations of globular KSR1‐CA3 to extended Ca²⁺ /CaM

Residues 354–377 of KSR1 are part of its CA3 domain (Figure 1a). The 3D solution structure of the CA3 domain of KSR1 has been resolved by nuclear magnetic resonance (NMR) (Zhou et al., 2002). KSR1‐CA3 adopts a globular conformation where residues His³³⁴, Cys³⁵⁹, Cys³⁶², and Cys³⁷⁷ coordinate one zinc ion, and Cys³⁴⁶, Cys³⁴⁹, His³⁶⁷, and Cys³⁷⁰ coordinate another zinc ion, both of which stabilize the structure. To gain structural insight into the binding of Ca²⁺/CaM to KSR1‐CA3, we performed in silico protein–protein docking. Analysis was done using the globular conformation of KSR1‐CA3 (PDB: 1KBE), as well as the crystal structure of Ca²⁺‐loaded CaM in its extended conformation (PDB: 1CLL) (Chattopadhyaya et al., 1992).

First, we generated more than 10,000 decoys of Ca²⁺/CaM in complex with KSR1‐CA3, from which we selected the five configurations with the lowest energy score (Figure S1). We then performed molecular dynamics (MD) simulations on these five configurations in an aqueous environment. Interestingly, we observed that KSR1‐CA3 converges to approximately the same linker region of Ca²⁺/CaM in all five configurations, but with different orientations. Clustering of the conformations across the simulations shows that KSR1‐CA3 contacts both lobes and the linker region of Ca²⁺/CaM (Figure S2). We further determined the intermolecular residue‐residue contacts between KSR1‐CA3 and Ca²⁺/CaM in the predicted configurations (Figure S3). Within the Ca²⁺/CaM binding region on KSR1 identified experimentally (residues 354–377) (Figure 1c,d), Thr³⁷¹, Lys³⁷², and Glu³⁷³ are the three most contacting residues. Note, residues involved in coordinating Zn²⁺ were excluded as contact residues. Figure 2 further illustrates that strong salt bridge interactions are a major driving force in stabilizing the complex of KSR1‐CA3 and Ca²⁺/CaM. Indeed, the KSR1 positively charged residues Lys³⁵⁸, Lys³⁶⁰, Arg³⁶³, Lys³⁶⁵, Lys³⁶⁹, and Lys³⁷², as well as negatively charged Glu³⁷³, all part of the Ca²⁺/CaM binding region, form electrostatic interactions with Ca²⁺/CaM. Of these amino acids, Lys³⁷² and Glu³⁷³ appear to be the charged residues with the highest probability of contact with Ca²⁺/CaM.

FIGURE 2.

The relaxed structures of extended Ca²⁺/CaM interacting with globular KSR1‐CA3. Snapshots showing the best representative conformations from the ensemble clusters for selected configurations (left panels). Violin plots represent the intermolecular salt bridge pair distance for the Ca²⁺/CaM:KSR1‐CA3 complex (right panels). In KSR1‐CA3, the transparent surface encompasses the Ca²⁺/CaM binding region of KSR1‐CA3 (residues 354–377). Key residues involved in binding are marked in green for KSR1 and purple for CaM. The dark and light blue spheres represent calcium and zinc ions, respectively. CaM, calmodulin.

2.3. Binding simulations of α‐helical KSR1‐CA3 to collapsed Ca²⁺ /CaM

Ca²⁺/CaM has been reported to bind α‐helical targets in a collapsed conformation (Meador et al., 1992; Tidow & Nissen, ²⁰¹³). We investigated whether such binding configuration could occur for the Ca²⁺/CaM:KSR1‐CA3 complex. This mode of binding implies that globular KSR1‐CA3 has to (i) unfold into a random chain by breaking adjacent β‐sheets and (ii) refold into an α‐helix. Because the time scale of unfolding/refolding of KSR1‐CA3 would exceed the simulation time scale, structural analysis of the full KSR1‐CA3 domain is not feasible by MD simulations. Instead, we constructed nine α‐helices covering KSR1‐CA3. Each peptide contains 20 amino acids, with a two amino acid increment at the start residue. These nine α‐helices, termed H1–H9, were then modeled in complex with Ca²⁺/CaM in its collapsed conformation (PDB: 1CDL) (Figure S4). The root mean square deviation (RMSD) values between the initial (Figure S4) and relaxed (Figure S5) configurations were calculated during MD simulations (Table S1). Among the nine models, H1, H2, and H8 retained both the α‐helical conformation of KSR1‐CA3 and the compact, collapsed conformation of Ca²⁺/CaM. The other models deviated from the α‐helical conformation of KSR1‐CA3 (H4 and H5), the collapsed conformation of Ca²⁺/CaM (H3 and H7), or both (H6 and H9). Calculation of the interaction energies reveals that electrostatic interactions contribute largely to the total interaction energy. Although weaker, hydrophobic interactions driven by van der Waals forces also contribute substantially to the binding affinity of the KSR1 peptides for Ca²⁺/CaM (Figure S6).

To identify the most favorable interaction model among the nine simulations, we calculated the binding free energy of complex formation using molecular mechanics combined with the generalized Born (GB) and surface area continuum solvation (MMGBSA) algorithm (Jang et al., 2019; Liao et al., ²⁰¹⁸; Ozdemir et al., ²⁰¹⁸; Zhang et al., ²⁰¹⁸). H8 had the lowest values of binding free energy, followed by the H7 simulations (Figure 3a), reflecting relatively high binding affinities. Note, the binding free energy of the different systems positively correlates with the van der Waals interaction energy, with a correlation coefficient r of 0.57 (Figure 3b), revealing that hydrophobic interactions contribute to the binding free energy.

FIGURE 3.

The relaxed structures of collapsed Ca²⁺/CaM binding to α‐helical peptides of KSR1‐CA3. (a) Violin plots representing the binding free energy (ΔG _binding) of the α‐helical peptides of KSR1‐CA3 interacting with collapsed Ca²⁺/CaM. (b) Correlation of the binding free energy with the van der Waals (vdW) interactions for the Ca²⁺/CaM:KSR1‐CA3 complex. (c and d) Snapshots showing the most favorable conformations from the ensemble clusters for the H8 (c) and H7 (d) simulations. The α‐helical structures of KSR1‐CA3 are colored gray. Hydrophobic, polar/glycine, positively charged, and negatively charged residues are colored white, green, blue, and red, respectively. For Ca²⁺/CaM, only the C‐lobe is shown. Key residues for interaction are marked in green for KSR1 and purple for CaM. CaM, calmodulin.

Figure 3c,d depicts H7 and H8, the two most representative models of KSR1‐CA3 in complex with Ca²⁺/CaM. In both configurations, KSR1 residues Met³⁵³, Ile³⁵⁴, and Phe³⁵⁵, which form a ³⁵³MIF³⁵⁵ motif, dock into a hydrophobic pocket in the C‐lobe of Ca²⁺/CaM, which strongly contributes to stabilization of the complex. In addition, Leu³⁶⁴ docks to the hydrophobic pocket of the N‐lobe of Ca²⁺/CaM, thereby keeping the two CaM lobes in proximity. Taken together, these observations suggest that amino acids Met³⁵³, Ile³⁵⁴, Phe³⁵⁵, and Leu³⁶⁴ of α‐helical KSR1‐CA3 establish important hydrophobic interactions with collapsed Ca²⁺/CaM.

2.4. Collapsed Ca²⁺ /CaM binds to α‐helical KSR1‐CA3 via hydrophobic interactions

We used experimental approaches to discriminate between the two possible modes of Ca²⁺/CaM binding to KSR1‐CA3 (Figures 2 and 3). Because Zn²⁺ ions stabilize the globular conformation of KSR1, we evaluated the effect of Zn²⁺ supplementation on binding. We incubated the purified KSR1‐M fragment, which contains the CA3 domain, with pure CaM in the presence or absence of ZnCl₂. Western blotting reveals that Zn²⁺ supplementation suppresses Ca²⁺/CaM binding (Figure 4a). Similarly, the addition of Zn²⁺ to HEK293T cell lysates abrogates the binding of endogenous Ca²⁺/CaM to GFP‐tagged KSR1 (Figure 4b). These results favor the model of Ca²⁺/CaM binding to α‐helical KSR1‐CA3.

FIGURE 4.

Binding of Ca²⁺/CaM to KSR1 is mediated by hydrophobic interactions. (a) The GST‐tagged KSR1‐M fragment (amino acids 319–433) was produced in Escherichia coli and purified on glutathione‐Sepharose. GST‐tagged KSR1‐M was then incubated with pure CaM in the presence or absence of 10 μM ZnCl₂. All incubations contained 1 mM CaCl₂. Pull‐down (PD) with GST‐coated Sepharose beads (GST) was the negative control. Proteins bound to the beads were eluted and analyzed by SDS‐PAGE. The gel was cut at ~23 kDa. The upper portion of the gel was stained with Coomassie blue; the lower portion was analyzed by Western blotting and probed for CaM. Input designates pure CaM not subjected to PD. # designates an empty lane. Data are representative of four independent experiments. (b) GFP‐tagged KSR1 was expressed in HEK293T cells. Equal amounts of protein from cell lysates were subjected to immunoprecipitation (IP) using GFP‐Trap beads in the presence of 1 mM CaCl₂ and in the presence or absence of 10 μM ZnCl₂. IP with control (Ctrl) agarose beads was carried out in parallel. Samples were resolved by Western blotting and probed for GFP and calmodulin. Inputs are aliquots of cell lysates prior to IP. Data are representative of three independent repetitions. (c) Protein sequence from amino acid 354–377 of wild‐type (WT) KSR1, KSR1‐I³⁵⁴D;F³⁵⁵D, and KSR1‐K³⁷²A;E³⁷³A. Residues in bold were mutated. Red indicates the introduced mutations. (d) GFP‐tagged KSR1‐WT, KSR1‐I³⁵⁴D;F³⁵⁵D, and KSR1‐K³⁷²A;E³⁷³A were expressed in HEK293T cells. Equal amounts of protein cell lysates were then subjected to CaM‐Sepharose PDs in the presence of 1 mM CaCl₂. PDs with protein A‐Sepharose (ProA) beads were carried out as negative controls. Samples were resolved by Western blotting and probed for GFP. Inputs are equal amounts of protein cell lysate not subjected to PD. Data are representative of three independent repetitions.

To further validate the mode of complex formation, we mutated the predicted contact residues in KSR1 and evaluated possible effects on Ca²⁺/CaM binding. We first assessed whether Ca²⁺/CaM associates with KSR1 via its positively charged Lys³⁷² and negatively charged Glu³⁷³ residues (binding of globular KSR1‐CA3 to extended Ca²⁺/CaM, Figure 2) by replacing these two amino acids with Ala. The construct is termed KSR1‐K³⁷²A;E³⁷³A (Figure 4c). We expressed GFP fusion constructs of this double point mutant or wild‐type KSR1 (KSR1‐WT) in HEK293T cells. CaM‐Sepharose PDs from cell lysates, followed by Western blotting, revealed that Ca²⁺/CaM binds KSR1‐K³⁷²A;E³⁷³A and KSR1‐WT to a similar extent (Figure 4d). The input lanes indicate that the level of protein expression of the KSR1 mutant construct is similar to that of the non‐mutated protein. These data show that Lys³⁷² and Glu³⁷³ of KSR1 are not required for Ca²⁺/CaM binding.

To determine whether Ca²⁺/CaM binds KSR1 via the hydrophobic amino acids Ile³⁵⁴ and Phe³⁵⁵ (binding of α‐helical KSR1‐CA3 to collapsed Ca²⁺/CaM, Figure 3), we replaced these two residues with negatively charged Asp. The construct is termed KSR1‐I³⁵⁴D;F³⁵⁵D (Figure 4c). KSR1‐I³⁵⁴D;F³⁵⁵D binding to Ca²⁺/CaM is minimal (Figure 4d), implying that the hydrophobic properties of Ile³⁵⁴ and Phe³⁵⁵ of KSR1 are required for Ca²⁺/CaM binding. Taken together, these results validate that Ca²⁺/CaM in its collapsed conformation binds to α‐helical KSR1‐CA3 via hydrophobic interactions.

2.5. Single mutation of KSR1‐Phe³⁵⁵ into Asp markedly reduces Ca²⁺ /CaM binding

With the ultimate goal to generate a single‐point mutant of KSR1 that has minimal binding to Ca²⁺/CaM, we individually mutated Ile³⁵⁴ and Phe³⁵⁵ into Asp. Based on the H7 and H8 simulations (Figure 3c,d), we generated an additional KSR1 mutant construct where Leu³⁶⁴ of KSR1 is replaced by Asp. We expressed GFP fusion constructs of KSR1‐WT, KSR1‐I³⁵⁴D, KSR1‐F³⁵⁵D, or KSR1‐L³⁶⁴D in HEK293T cells. We evaluated binding to Ca²⁺/CaM by CaM‐Sepharose PDs from cell lysates. The amount of KSR1‐I³⁵⁴D and KSR1‐F³⁵⁵D that bound to Ca²⁺/CaM was 64% and 79% less, respectively, than that of KSR1‐WT. In contrast, the binding of KSR1‐L³⁶⁴D to Ca²⁺/CaM was similar to that of KSR1‐WT (Figure 5a,b).

FIGURE 5.

Replacement of KSR1‐Phe³⁵⁵ with Asp significantly attenuates binding to Ca²⁺/CaM. (a) GFP‐tagged wild‐type (WT) KSR1, KSR1‐I³⁵⁴D, KSR1‐F³⁵⁵D, and KSR1‐L³⁶⁴D were separately expressed in HEK293T cells. Cell lysates were supplemented with 1 mM CaCl₂ and incubated with CaM‐Sepharose beads. Pull‐downs (PDs) with protein A beads (ProA) served as negative controls. Samples were resolved by Western blotting and probed for GFP. Inputs designate aliquots of cell lysate not subjected to PD. In each panel, images for WT and mutated KSR1 were from the same membrane (irrelevant lanes were omitted). (b) The GFP‐KSR1 signal observed after PD was quantified using LI‐COR Image Studio software (mean ± SD, n = 5). Binding to KSR1‐WT was set as 1. Data were analyzed by one‐way ANOVA followed by Dunnet's test (***, p ≤ 0.001; ns, not significant). (c) GFP‐tagged KSR1‐WT and KSR1‐F³⁵⁵D were expressed separately in HEK293T cells. KSR1 was immunoprecipitated (IP) using GFP‐Trap beads from cell lysates in the presence of 1 mM CaCl₂. Precipitations with control agarose beads (Ctrl) were carried out in parallel. Inputs are cell lysates not subjected to IP. Samples were resolved by Western blotting and probed for GFP, CaM, MEK, and 14‐3‐3. # indicates an empty lane. (d) The CaM, MEK, and 14‐3‐3 bands observed after IP were quantified (mean ± SD, n = 3). Band intensities were divided by the values observed with cells transfected with KSR1‐WT to normalize this condition to 1. Data were analyzed by one‐sample t tests (**, p ≤ 0.01; ns, not significant). All blots shown in this figure are representative of at least three biological replicates. CaM, calmodulin.

The binding analyses described above were carried out with pure CaM. We further determined whether KSR1‐F³⁵⁵D has decreased binding to endogenous CaM. To do so, we separately expressed GFP‐tagged KSR1‐WT and KSR1‐F³⁵⁵D in HEK293T cells and immunoprecipitated the proteins using GFP‐Trap beads. Western blotting revealed that Ca²⁺/CaM co‐immunoprecipitates with KSR1‐WT, as expected (Figure 5c). Consistent with the results obtained with pure CaM, KSR1‐F³⁵⁵D bound 76% less endogenous Ca²⁺/CaM than that bound by KSR1‐WT (Figure 5c,d). This result confirms that the replacement of Phe³⁵⁵ of KSR1 markedly reduces Ca²⁺/CaM binding.

Finally, we verified that replacing Phe³⁵⁵ of KSR1 with Asp does not disrupt binding of other well‐characterized KSR1 binding partners. Analysis was conducted by evaluating the interaction with MEK, which binds constitutively to the CA5 domain of KSR1 (Yu et al., 1998). We compared the amounts of endogenous MEK that co‐immunoprecipitated with KSR1‐WT and KSR1‐F³⁵⁵D. The two KSR1 proteins bind a similar amount of MEK (Figure 5c,d). Similarly, we evaluated whether the F³⁵⁵D mutation within KSR1 affects its binding to 14‐3‐3, which retains KSR1 in the cytosol (Frodyma et al., 2017) and binds to a region of KSR different from that where MEK binds. The data reveal that the amounts of KSR1‐WT and KSR1‐F³⁵⁵D that co‐immunoprecipitated with 14‐3‐3 are equivalent (Figure 5c,d). Taken together, these results suggest that mutating Phe³⁵⁵ of KSR1 into Asp does not alter its structural integrity.

2.6. Replacing Phe³⁵⁵ of KSR1 with Asp reduces EGF‐stimulated activation of ERK

The identification of a KSR1 single‐point mutant with marked reduction of Ca²⁺/CaM binding provides opportunities to selectively study the functional implications of Ca²⁺/CaM:KSR1 binding. KSR1 is known to promote EGF‐stimulated ERK activation (Nguyen et al., 2002). Therefore, we investigated whether impairment of Ca²⁺/CaM binding to KSR1 influences ERK activation by EGF. We separately expressed GFP alone, GFP‐tagged KSR1‐WT, or GFP‐tagged KSR1‐F³⁵⁵D in HEK293T cells. We then activated the MAPK pathway with EGF for 1, 5, 10, or 20 min. We validated by Western blot that EGF activates ERK in cells expressing GFP only, with maximal ERK activation seen 1 min after adding EGF (Figure 6a). EGF also stimulates ERK in cells expressing the KSR1 constructs, with similar activation kinetics (Figure 6a). Cells expressing wild‐type KSR1 had significantly lower EGF‐stimulated ERK activation than cells expressing GFP only (Figure 6b,c), likely because of high KSR1 overexpression (Figure S7). Importantly, the F³⁵⁵D mutation reduces the ability of KSR1 to activate ERK upon EGF stimulation (Figure 6b,c). This observation strongly suggests that Ca²⁺/CaM binding is required for optimal KSR1‐mediated MAPK signaling.

FIGURE 6.

Replacement of KSR1 Phe³⁵⁵ with Asp significantly reduces EGF‐stimulated ERK activation. (a) HEK293T cells were transiently transfected (Tf) with plasmids coding for GFP alone, GFP‐KSR1‐WT, or GFP‐KSR1‐F³⁵⁵D. Cells were starved of serum for 16 h, then incubated with 100 ng/mL EGF for 0, 1, 5, 10, or 20 min. Cells were lysed and equal amounts of protein lysate were resolved by SDS‐PAGE and Western blotting. Blots were probed for phosphorylated ERK (pERK), total ERK, GFP, and HDAC (loading control). Blots are representative of three independent experiments. (b) Aliquots of the same lysates shown in panel A for t0, 1, and 5 min time points were resolved on a separate SDS‐PAGE gel and analyzed by Western blotting as described above. (c) The pERK and total ERK bands observed for 1 and 5 min EGF time points were quantified, and the ratio of their intensities was calculated (mean ± SD, n = 3–4). The pERK/ERK ratio in cells expressing GFP only was set to 1. Significant differences between GFP and KSR1‐WT, as well as between KSR1‐WT and KSR1‐F³⁵⁵D, are shown in the graphs (one‐way ANOVA followed by Tukey's test, *, p ≤ 0.05; **, p ≤ 0.01). ANOVA, analysis of variance; EGF, epidermal growth factor.

3. DISCUSSION

CaM and KSR1 are two proteins with essential functions in intracellular signaling. We have previously shown that these proteins form a complex in cells, and that chemical antagonism of CaM influences MAPK signaling (Parvathaneni et al., 2021). In this study, we investigated the structural determinants of the Ca²⁺/CaM:KSR1 complex and explored the effects of binding on ERK activation. Using short peptides of selected regions of KSR1, we initially determined that binding occurs to the CA3 cysteine‐rich domain of KSR1. This domain contains 1–10, 1–14, and 1–16 canonical Ca²⁺/CaM binding motifs, where hydrophobic anchor amino acids are separated by 8, 12, and 14 residues, respectively (Denesyuk et al., 2023). The major documented function of the KSR1‐CA3 domain is to drive KSR1 recruitment to the plasma membrane upon cell stimulation, via interactions with ceramide lipids (Yin et al., 2009). Our work identifies a new function for this domain, namely to integrate Ca²⁺ signaling with KSR1 via Ca²⁺/CaM binding.

We refined the structural analysis of Ca²⁺/CaM:KSR1 binding using in silico MD simulations. In cells, the flexible central linker that separates the N‐ and C‐terminal lobes of Ca²⁺/CaM allows for transitions between an extended, dumbbell‐shaped conformation and a collapsed conformation, where the distance between the two lobes is reduced from 50 Å to less than 10 Å (Andrews et al., 2021). In the latter conformation, Ca²⁺/CaM preferentially binds α‐helical peptides, with its two lobes wrapping around the target (Gifford et al., 2012; Meador et al., ¹⁹⁹²; Yamauchi et al., ²⁰⁰³). Therefore, we simulated two modes of interaction: (i) binding of extended Ca²⁺/CaM to the resolved NMR structure of globular KSR1‐CA3 (Zhou et al., 2002) and (ii) binding of collapsed Ca²⁺/CaM to α‐helical KSR1‐CA3. Experimental analyses revealed that Zn²⁺ supplementation suppresses Ca²⁺/CaM:KSR1 binding, and that the hydrophobic residues Ile³⁵⁴ and Phe³⁵⁵ of KSR1 are required for Ca²⁺/CaM binding. These data favor the model where collapsed Ca²⁺/CaM binds to α‐helical KSR1‐CA3 (Figure 7). Calculation of the RMSD values illustrates the dynamics of the Ca²⁺/CaM:KSR1‐CA3 complex, with possible deviations of collapsed Ca²⁺/CaM toward a more open conformation upon binding. Our data further reveal that the hydrophobic amino acids Ile³⁵⁴ and Phe³⁵⁵ of KSR1, which form predicted 1–10 and 1–14 canonical Ca²⁺/CaM binding motifs, are critical interacting residues through docking to the hydrophobic pocket of the C‐lobe of Ca²⁺/CaM. Of the two residues, the mutation of Phe³⁵⁵ is the one that causes the more pronounced decrease of Ca²⁺/CaM binding.

FIGURE 7.

Model of Ca²⁺/CaM binding to the CA3 domain of KSR1. (1) Binding of four Ca²⁺ ions (dark blue spheres) to apo‐CaM (red) induces a conformational shift in CaM to its extended Ca²⁺‐loaded conformation. (2) Ca²⁺/CaM in its collapsed conformation then binds to KSR1 by wrapping around the CA3 domain (purple), which stabilizes it in an α‐helical conformation. (3) In cells, the Ca²⁺/CaM:KSR1 complex promotes EGF‐stimulated activation of ERK. Figure generated with BioRender. CaM, calmodulin; EGF, epidermal growth factor.

CaM requires Ca²⁺ ions in order to bind KSR1 (Parvathaneni et al., 2021). In cells, Ca²⁺ concentrations are tightly regulated in time and space through the activation of influx and efflux machineries, both at the cell surface and at organelle membranes (Bagur & Hajnóczky, 2017). During Ca²⁺ transients, Ca²⁺ functions as a second messenger, relaying signals to effector proteins in response to stimuli (Bagur & Hajnóczky, 2017). Our results indicate that Ca²⁺, by binding to CaM, induces the CaM:KSR1 interaction, ultimately modulating KSR1 functions in MAPK signaling. Moreover, our data strongly suggest that Ca²⁺/CaM binds to an α‐helical portion of the CA3 domain of KSR1. This observation implies that binding occurs in the absence of Zn²⁺ coordination by KSR1‐CA3, which would otherwise stabilize its globular conformation (Zhou et al., 2002). The basal free concentration of Zn²⁺ ranges between picomolar and nanomolar values in different subcellular domains (Maret, 2017). Cellular Zn²⁺ concentrations are maintained by the concerted actions of transporters, which mediate Zn²⁺ influx/efflux across membranes, and small cysteine‐rich metallothionein proteins that sequester Zn²⁺ (Maret, 2017). Similar to Ca²⁺, Zn²⁺ functions as a second messenger through transient increases in free Zn²⁺ concentrations in the cytosol (Yamasaki et al., 2007). Moreover, Zn²⁺ binds to defined motifs in numerous cellular proteins, including transcription factors, enzymes, and scaffold proteins, thereby promoting their activity and/or structural stability (Maret, 2017). The binding affinity of KSR1‐CA3 for Zn²⁺ ions has not been determined. Structural Zn²⁺ ions usually bind to proteins with a high affinity (Kds in the picomolar to femtomolar range) (Maret, 2017). This suggests that KSR1 may be predominantly Zn²⁺‐loaded in resting cells. Nevertheless, some KSR1 may exist locally and transiently in a Zn²⁺‐free conformation to which Ca²⁺/CaM may bind. Ca²⁺/CaM, by wrapping around the CA3 domain upon binding, may also reduce the affinity of KSR1 for Zn²⁺, leading to Zn²⁺ release from the CA3 domain and structural destabilization. In turn, Ca²⁺/CaM may stabilize α‐helical re‐folding of KSR1‐CA3, as previously observed for other Ca²⁺/CaM binding partners (Gifford et al., 2012; Meador et al., ¹⁹⁹²; Tidow & Nissen, ²⁰¹³; Yamauchi et al., ²⁰⁰³).

Numerous binding partners bind to KSR1 (McKay et al., 2009) and CaM (Hoeflich & Ikura, 2002). Knockdown, knockout, or chemical inhibition of these two proteins provides clues to their global cellular functions, which result from multiple interactions. In contrast, the functional study of a selected interaction involving scaffold and/or sensor proteins requires advanced structural data, from which point mutations, small molecule inhibitors, or short interfering peptides that abrogate binding can be designed. In this study, we generated a KSR1 point mutant construct, termed KSR1‐F³⁵⁵D, which has markedly impaired binding to Ca²⁺/CaM. This mutant construct is a valuable reagent to selectively study the crosstalk between Ca²⁺/CaM and KSR1‐regulated signaling.

We used the KSR1‐F³⁵⁵D mutant to evaluate the effect of Ca²⁺/CaM binding on the role of KSR1 in ERK activation. We observed that binding of Ca²⁺/CaM to KSR1 is required for EGF to maximally activate MAPK signaling (Figure 6). The molecular mechanism by which Ca²⁺/CaM modulates KSR1 function has not been identified. Importantly, we showed that binding of Ca²⁺/CaM does not influence the constitutive interaction of KSR1 with MEK, the kinase that catalyzes ERK phosphorylation, suggesting that Ca²⁺/CaM does not alter the ability of KSR1 to scaffold MAPK components. Moreover, Ca²⁺/CaM binding does not modulate the interaction of KSR1 with 14‐3‐3, a protein that maintains phosphorylated KSR1 in the cytosol (Frodyma et al., 2017; Jagemann et al., ²⁰⁰⁸). Since ERK activation occurs at the plasma membrane, Ca²⁺/CaM may increase targeting of KSR1 to the cell surface upon growth factor stimulation. Possible molecular mechanisms by which this may occur include stimulation of KSR1 dephosphorylation by PP2A phosphatase or increased affinity for plasma membrane lipids upon Ca²⁺/CaM binding. (Ca²⁺/CaM binds to the KSR1 CA3 domain, which is responsible for binding plasma membrane lipids; Yin et al., ²⁰⁰⁹). Once at the plasma membrane, Ca²⁺/CaM could also promote scaffolding of Raf and ERK kinases through cooperative binding to increase MAPK activation. Finally, Ca²⁺/CaM may facilitate EGF‐stimulated ERK activation by competing with inhibitory proteins for binding to KSR1. For example, Ca²⁺/CaM binding may create steric hindrance that reduces the interaction of KSR1 with C‐TAK1 kinase, which promotes cytoplasmic retention of KSR1 (Müller et al., 2001). Regardless of the exact mode of action, our data show that Ca²⁺ signaling, through CaM, influences KSR1‐modulated signaling.

In conclusion, our simulation data, combined with mutagenesis and binding experiments, yield a realistic model for the dynamic interaction of Ca²⁺/CaM with KSR1. Moreover, the KSR1‐F³⁵⁵D point mutant construct is an effective bioreagent to specifically elucidate the intersection between Ca²⁺/CaM and KSR1‐mediated signaling pathways.

4. MATERIALS AND METHODS

4.1. Materials

Dulbecco's modified Eagle's medium (DMEM), fetal bovine serum (FBS), and Halt protease and phosphatase inhibitor cocktail were purchased from Thermo Fisher Scientific. Glutathione‐Sepharose, CaM‐Sepharose, and protein A‐Sepharose were from GE Healthcare. EGF was purchased from Sigma. GFP‐Trap and control agarose beads were from Chromotek. Pure CaM was from Ocean Biologics. InstantBlue protein stain was purchased from Expedeon. Lipofectamine 2000 was from Invitrogen. The primary antibodies used in this study are listed in Table S2. Infrared dye‐conjugated secondary antibodies were purchased from LI‐COR Biosciences.

4.2. Plasmid construction

The pcDNA3‐Myc‐mKSR1 and pBlueScriptII‐mKSR1 plasmids that encode mouse KSR1 have been previously described (Parvathaneni et al., 2021). GFP‐tagged KSR1 was generated by PCR amplification of mKSR1 from the pBlueScriptII‐mKSR1 template, followed by incorporation into the pEGFP‐C1 vector (NovoPro) between the BglII and EcoRI restriction sites. The DNA fragments corresponding to KSR1‐M, ‐M1, ‐M2, and ‐M3 were obtained by PCR using the pcDNA3‐Myc‐mKSR1 plasmid as a template. The DNA fragments were then incorporated into the pGEX2T‐TEV plasmid (NovoPro) between the BamHI and EcoRI restriction sites. The KSR1 deletion mutant Δ^354–377 was generated by PCR from the pBlueScriptII‐mKSR1 plasmid using the following 5′‐phosphorylated primers: TGCAGGATCACCTTCCTCCCACTGG (forward) and CATGCTCTTCTGGCACACGTTGCAC (reverse). The PCR products were allowed to self‐ligate, and were then inserted in the pEGFP‐C1 vector between the BglII and EcoRI restriction sites. The KSR1 point mutants (I³⁵⁴D;F³⁵⁵D, K³⁷²A;E³⁷³A, I³⁵⁴D, F³⁵⁵D, and L³⁶⁴D) were generated by site‐directed mutagenesis using pBlueScriptII‐mKSR1 as a template. The constructs were then incorporated into the pEGFP‐C1 vector between the BglII and EcoRI restriction sites. We validated the sequences of all constructs by DNA sequencing.

4.3. Binding of pure CaM to GST‐KSR1 constructs

The GST‐KSR1 constructs were expressed in E. coli. Expression was induced by 1.5 μM iso‐propyl‐β‐D‐thiogalactoside at 25°C. After 16 h, bacteria were lysed by sonication in PBS buffer containing 2 mM ethylenediaminetetraacetic acid, 0.2 mM phenylmethylsulfonyl fluoride, and 10 mM dithiothreitol. After removal of cell debris by centrifugation, lysates were loaded on glutathione‐Sepharose columns to purify the GST‐tagged constructs, followed by washing with PBS containing 10 mM dithiothreitol. The integrity and purity of the fragments were verified by Coomassie blue SDS‐PAGE gel.

The GST‐KSR1 constructs on the beads were incubated with 2 μg of pure CaM in 500 μL Buffer A (50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 1% Triton X‐100, and 1 mM CaCl₂) for 3 h at 4°C. Where indicated, ZnCl₂ was added to Buffer A at a concentration of 10 μM. This Zn²⁺ concentration is within the physiological range (Chang et al., 2006) and matches that used by Zhou et al. (2002) for the NMR structural analysis of KSR1‐CA3. Control PDs with GST‐Sepharose were carried out in parallel. After five washes in Buffer A, proteins were eluted from the beads in Laemmli sample buffer at 95°C for 5 min. Samples were analyzed by SDS‐PAGE and Western blotting using fluorescent‐labeled secondary antibodies. Imaging was done with the Odyssey CLx infrared imaging system (LI‐COR).

4.4. Cell culture and treatments

HEK293T cells were purchased from the American Type Culture Collection. Cells were cultured in DMEM medium containing 10% FBS at 37°C in 5% CO₂. The GFP‐tagged KSR1 constructs (wild‐type, Δ^354‐377, I³⁵⁴D;F³⁵⁵D, K³⁷²A;E³⁷³A, I³⁵⁴D, F³⁵⁵D, and L³⁶⁴D) were transiently transfected into HEK293T cells using Lipofectamine 2000 according to the manufacturer's instructions. Cells were collected for analysis 48 h after transfection.

4.5. CaM‐Sepharose PDs

Cells were lysed in 1 mL Buffer A supplemented with protease and phosphatase inhibitors. Cell debris was pelleted by centrifugation (15,000g for 10 min at 4°C). Cell lysates were pre‐cleared with protein A‐Sepharose for 1 h at 4°C, then equal amounts of pre‐cleared cell lysates were incubated with CaM‐Sepharose beads for 3 h at 4°C. Control PDs were carried out in parallel with protein A‐Sepharose beads. The beads were washed five times in Buffer A, followed by protein elution in Laemmli sample buffer at 95°C for 5 min. Samples were analyzed by SDS‐PAGE and Western blotting. The blots were probed with the antibodies indicated in the figure legend.

4.6. GFP‐Trap immunoprecipitation

Cells were lysed as described above. Where indicated, 10 μM ZnCl₂ was added to the cell lysates. Cell lysates were pre‐cleared for 1 h at 4°C using agarose beads. Pre‐cleared cell lysates were then incubated with GFP‐Trap agarose beads or control agarose beads for 3 h at 4°C. Beads were then washed five times with Buffer A. Proteins bound to the beads were eluted in Laemmli sample buffer at 95°C for 5 min. Samples were analyzed by SDS‐PAGE and Western blotting.

4.7. EGF‐stimulated ERK activation

HEK293T cells at 70% confluence were transiently transfected with the GFP‐tagged KSR1‐WT or KSR1‐F³⁵⁵D constructs. After 48 h, cells were starved of serum for 16 h, then incubated with either 100 ng/mL EGF or vehicle (PBS) for 5 min. Cells were lysed in Buffer A supplemented with protease and phosphatase inhibitors. Equal amounts of protein lysates were resolved by SDS‐PAGE and Western blotting. The blots were probed with the antibodies indicated in the figure legend.

4.8. Generation of initial configurations of Ca²⁺ /CaM binding to KSR1‐CA3

We used the Rosetta docking program (Leaver‐Fay et al., 2011) to generate initial configurations of Ca²⁺/CaM binding to KSR1‐CA3. In the first approach (binding of globular KSR1‐CA3 to extended Ca²⁺/CaM), the NMR solution structure of the CA3 domain of KSR1 (PDB: 1KBE) was docked to the crystal structure of extended Ca²⁺/CaM (PDB: 1CLL). Over 10,000 decoys were generated, from which the lowest energy score decoys were selected as initial configurations. In a second approach (binding of α‐helical KSR1‐CA3 to collapsed Ca²⁺/CaM), we modeled nine 20 amino acid‐long α‐helices (H1–H9), covering KSR1‐CA3 from position 338 to 377 with a two amino acid increment in the start residue. Then, we docked the crystal structure of collapsed Ca²⁺/CaM (PDB: 1CDL) to each of these α‐helices, yielding nine initial configurations.

4.9. Atomistic molecular dynamics simulations

The initial configurations of binding were subjected to all‐atom MD simulations in an aqueous environment using CHARMM all‐atom force field, version 36m (Brooks et al., 2009; Huang et al., ²⁰¹⁷). The simulations were performed using the same protocol as in our previous work (Jang et al., 2019; Weako et al., ²⁰²¹; Zhang et al., ²⁰¹⁹). The modified TIP3P water model was used to create the isometric unit cell box containing the Ca²⁺/CaM:KSR1‐CA3 complex. Na⁺ and Cl⁻ were added to generate a final ionic strength of approximately 100 mM and to neutralize the system. During the pre‐equilibrium stage, we performed a series of minimization and dynamic cycles for the solvents, including the ions, with the harmonically restrained protein backbone until the solvent reached 310K. In the final stage of pre‐equilibrium, the harmonic constraints on the protein backbones were progressively removed through dynamic cycles using the long‐range particle mesh Ewald electrostatics calculation. The production runs of the all‐atom MD simulations were implemented using the NAMD parallel computing code (Phillips et al., 2005) on a Biowulf cluster at the National Institutes of Health (Bethesda, MD). In the production runs, the Langevin thermostat maintained a constant temperature of 310K and the Nosé‐Hoover Langevin piston pressure control sustained the pressure at 1 atm with the NPT condition. The SHAKE algorithm was applied to constrain the motion of bonds involving hydrogen atoms. A time step of 2 fs was used for all simulations. A total of 14 μs simulations were performed for the 14 systems, each with 1 μs simulation time. Additional simulations were performed for some systems for reproducibility. CHARMM was used to analyze the simulation trajectories. To obtain the most populated conformational representatives for each configuration of the Ca²⁺/CaM:KSR1‐CA3 complex, Ensemble Cluster in Chimera (Pettersen et al., 2004) was used to cluster all the simulation frames. To determine the binding free energy of Ca²⁺/CaM binding to KSR1‐CA3, the molecular mechanics energies combined with MMGBSA were employed. The GBSW module (Im et al., 2003) of CHARMM was used to calculate the solvation‐free energy.

4.10. Data analyses

GraphPad Prism 9 was used for all statistical analyses. The statistics used to analyze each data set are described in the figure legends. Where indicated, Image Studio 2.0 (LI‐COR) was used to quantify pertinent bands from Western blots.

AUTHOR CONTRIBUTIONS

David B. Sacks: Conceptualization; writing – review and editing; project administration; supervision; funding acquisition; resources. Louise Thines: Writing – original draft; investigation; writing – review and editing; formal analysis. Hyunbum Jang: Formal analysis; writing – original draft; investigation; conceptualization. Zhigang Li: Methodology; investigation; resources. Samar Sayedyahossein: Investigation; formal analysis. Ryan Maloney: Investigation. Ruth Nussinov: Funding acquisition; resources; writing – review and editing; supervision; project administration; conceptualization.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Supporting information

Figure S1. Initial configurations of Ca²⁺/CaM binding to Zn²⁺‐loaded KSR1‐CA3. The upper left panel shows the NMR solution structure and the amino acid sequence of the CA3 domain of mouse KSR1 (mKSR1). In the structure, key residues involved in the coordination of zinc ions are highlighted. In the amino acid sequence, hydrophobic, polar/glycine, positively charged, and negatively charged residues are colored black, green, blue, and red, respectively. The other panels show snapshots representing the five different initial configurations (Configs. 1–5) of the Ca²⁺/CaM:KSR1‐CA3 complex that were subjected to all‐atom MD simulations. The dark and light blue spheres represent calcium and zinc ions, respectively.

Figure S2. Simulated relaxed configurations of Ca²⁺/CaM in complex with Zn²⁺‐loaded KSR1‐CA3. Each configuration (Configs. 1–5) shows the top five representative conformations from the ensemble clusters. The representative conformations were superimposed. The color code for the KSR1‐CA3 five clusters is shown in the upper left panel.

Figure S3. Intermolecular residue‐residue contacts between Ca²⁺/CaM and KSR1‐CA3. The left panels depict snapshots of the most representative conformations from the ensemble clusters for each configuration (Configs. 1–5). In KSR1‐CA3, the surface area delineates the Ca²⁺/CaM binding region (residues 354–377). The dark and light blue spheres represent calcium and zinc ions, respectively. The graphs on the right show the residue‐residue contacts for the Ca²⁺/CaM:KSR1‐CA3 complex. For two intermolecular residues i and j, the probability of contact for the distance between the $C_{\beta }^i-C_{\beta }^j$ (C_α is used for Gly residue), with a cutoff of 10 Å, was calculated. In the calculation, the probability p < 0.1 was omitted.

Figure S4. Initial structures of the α‐helical peptides (H1–H9) of KSR1‐CA3 bound to collapsed Ca²⁺/CaM. For each panel, the sequence of the 20 amino acid‐long α‐helix of KSR1‐CA3 is shown. In the sequence, hydrophobic, polar/glycine, positively charged, and negatively charged residues are colored black, green, blue, and red, respectively. In the α‐helical structures of KSR1‐CA3, the same colors are used, except hydrophobic residues are in white. The dark blue spheres represent calcium ions.

Figure S5. The relaxed structures of the α‐helical peptides of KSR1‐CA3 bound to Ca²⁺/CaM. Snapshots showing the best representative conformations from the ensemble clusters for nine different models of the α‐helical peptides of the KSR1 CA3 domain (H1–H9). For each panel, the sequence of the 20 amino acid‐long α‐helix of KSR1‐CA3 is shown. In the sequence, hydrophobic, polar/glycine, positively charged, and negatively charged residues are colored black, green, blue, and red, respectively. In the α‐helical structures of KSR1‐CA3, the same colors are used, except hydrophobic residues are in white. The dark blue spheres represent calcium ions.

Figure S6. Interaction energy of Ca²⁺/CaM binding to the α‐helical peptides of KSR1‐CA3. Top graph: Averaged total interaction energy, along with the electrostatic and van der Waals (vdW) contributions. Error bars indicate standard deviation. Bottom graph: Violin plots of the vdW interaction energy for the nine different simulations.

Figure S7. Overexpression of GFP‐tagged KSR1 in HEK293T cells. HEK293T cells were transfected (Tf) with plasmids coding for GFP‐tagged wild‐type (WT) KSR1, KSR1‐F³⁵⁵D, or GFP alone. Cell lysates were analyzed by Western blotting and probied with anti‐KSR1 and anti‐tubulin (loading control) antibodies.

Table S1. Root mean square deviation (RMSD) values (Å) for Ca²⁺/CaM and KSR1‐CA3 between the initial and relaxed configurations.

Table S2. Primary antibodies used in this study.

Thines L, Jang H, Li Z, Sayedyahossein S, Maloney R, Nussinov R, et al. Disruption of Ca²⁺/calmodulin:KSR1 interaction lowers ERK activation. Protein Science. 2024;33(5):e4982. 10.1002/pro.4982