A composite motif in calcimembrin/C16orf74 dictates multimeric dephosphorylation by calcineurin

Abstract

Calcineurin, the Ca²⁺/calmodulin-activated protein phosphatase, recognizes substrates and regulators via short linear motifs, PxIxIT and LxVP, which dock to distinct sites on calcineurin to determine enzyme distribution and catalysis, respectively. Calcimembrin/C16orf74 (CLMB), an intrinsically disordered microprotein whose expression correlates with poor cancer outcomes, targets calcineurin to membranes where it may promote oncogenesis by shaping calcineurin signaling. We show that CLMB associates with membranes via lipidation, i.e., N-myristoylation and reversible S-acylation. Furthermore, CLMB contains an unusual composite 'LxVPxIxIT' motif, that binds the PxIxIT-docking site on calcineurin with extraordinarily high affinity when phosphorylated, ³³LDVPDIIITPP(p)T⁴⁴. Calcineurin dephosphorylates CLMB to decrease this affinity, but Thr44 is protected from dephosphorylation when PxIxIT-bound. We propose that CLMB is dephosphorylated in multimeric complexes, where one PxIxIT-bound CLMB recruits calcineurin to membranes, allowing a second CLMB to engage via its LxVP motif to be dephosphorylated. In vivo and in vitro data, including nuclear magnetic resonance (NMR) analyses of CLMB-calcineurin complexes, support this model. Thus, CLMB with its composite motif imposes distinct properties to calcineurin signaling at membranes including sensitivity to CLMB:calcineurin ratios, CLMB phosphorylation and dynamic S-acylation.

Fig. 1. Calcimembrin localizes to membranes via protein lipidation.

a CLMB schematic showing N-myristoylation (red); S-acylation (yellow), phosphorylation (green), calcineurin-binding motifs LxVP (pink) and PxIxIT (purple). Sequences of wild-type and mutated regions are shown. b Representative immunoblot showing metabolic labeling of CLMB-Flag (wild-type or indicated mutants) with palmitate (17-ODYA) or myristate (YnMyr) analogs, analyzed via CLICK chemistry with biotin-azide and detection with IRdye 800CW streptavidin (“Biotin”). Inputs and immunoprecipitated (IP) samples are shown. Asterisks indicate CLMB dimers. (n = 3 independent experiments). c Representative immunoblot showing S-acylation of CLMB-FLAG (wild-type or indicated mutants) and endogenous calnexin (positive control) detected via Acyl-PEG exchange. Asterisks show PEGylation events. Input, cleaved (NH₂OH) and PEGylated (mPEG), and non-cleaved PEGylated conditions are indicated (n = 3 independent experiments). d Representative in-gel fluorescence scan showing pulse-chase analysis of CLMB-FLAG using palmitate (17-ODYA) or methionine (L-AHA) analogs detected via CLICK chemistry with Cy5 or AZDye 488, respectively. Incorporation into CLMB-FLAG with no analog (Neg Ctrl), at 0 or 30 min after unlabeled chase with DMSO (control) or PalmB shown (n = 3 independent experiments). e Data from (d) showing mean ± SEM of 17-ODYA signal normalized to L-AHA; *p < 0.05, T0 vs PalmB p = 0.0103, **p < 0.01, DMSO vs PalmB p = 0.007. f Representative immunoblot showing subcellular fractionation of CLMB-FLAG (wild-type or indicated mutants) with input, cytoplasmic (cyto), and membrane (mem) fractions. GM130 and α-tubulin mark mem and cyto fractions, respectively (n = 3 independent experiments). g Data from (f) showing mean ± SEM cyto or mem signals normalized to total (cyto + mem); n.s. not significant, ****p < 0.0001. CLMBC_7,14S vs CLMB_G2A: cytoplasm, p = 0.238, membrane p = 0.238. For (e, g), p-values were calculated by two-way ANOVA corrected by Tukey’s multiple comparison. Source data are provided as a Source Data file and at Mendeley Data [10.17632/6kp379tsv4.1].

Fig. 2. Composite LxVPxIxIT SLiM in calcimembrin mediates binding to and dephosphorylation by calcineurin.

a Representative immunoblot showing co-purification of CLMB-FLAG (wild-type and indicated mutants) with CNA-GFP. Inputs and purified (IP) samples shown (n = 3 independent experiments). Glyceraldehyde-3-Phosphate Dehydrogenase (GAPDH) was used as a loading control. b Data from (a) showing mean ± SEM of co-purified CLMB-FLAG normalized to CLMB-FLAG input levels and precipitated CNA-GFP. Indicated p-values calculated by two-way ANOVA corrected by Dunnett’s multiple comparison test. ****p < 0.0001, ***p < 0.001, *p < 0.05; CLMB_WT vs CLMB_LxVmut p = 0.0325, CLMB_WT vs CLMB_T44A p = 0.0002. c Representative images of MCF7 cells co-transfected with CLMB-FLAG (wild-type or indicated mutants) and CNA-GFP. Fixed cells immunostained with anti-FLAG (magenta), anti-GFP (yellow), and nuclei (DNA) with Hoechst 33342. Scale bar = 50 μm. Arrows indicate some examples of CLMB and CNA colocalization to the plasma membrane. d Data from (c) showing median with 95% CI for Pearson’s correlation coefficient of CLMB-calcineurin colocalization. Each point represents a single cell. CLMB_WT n = 525, CLMB_IxITMut n = 641, CLMB_LxVMut n = 500, CLMB_T44A n = 517. p-values calculated by Kruskal–Wallis with Dunn’s multiple comparisons correction. *p < 0.05 (p = 0.0116), **p < 0.01 (p = 0.0092), ****p < 0.0001. e Dephosphorylation of pCLMB_WT (blue) or pCLMB_LxVMut (pink) phosphopeptides with 20 nM CNA/B and 1 μM CaM. Mean ± SEM of phosphate released displayed for n = 6 replicates at each timepoint with reaction rate as shown. Source data are provided as a Source Data file and at Mendeley Data [10.17632/6kp379tsv4.1].

Fig. 3. Thr44 phosphorylation and the LxVP motif contribute to PxIxIT binding.

a Determination of the affinity between CNA/B and unlabeled pCLMB_WT peptide using isothermal calorimetry. Dissociation constant (K_D) = 37 ± 1.2 nM, reaction stoichiometry (n value) = 1.00 ± 0.02, enthalpy (ΔH) = −4 ± 0.8, and entropy (TΔS) = 21 ± 3 are shown. Data are mean ± s.d. for n = 3 independent experiments. b Binding isotherm obtained by titration of unlabeled pCLMB_WT peptide into CNA/B (40 μM). c Fluorescence polarization (FP) saturation binding curves of FITC-pCLMB_WT and FITC-CLMB_WT (5 nM) incubated with serial dilutions of CNA/B (0 to 5 μM). FP data (mean values from n = 2 independent experiments) were fitted to a one-site binding model. K_D’s are as shown. d Competitive binding assay of unlabeled pCLMB_WT, CLMB_WT, pCLMB_IxITMut, and pCLMB_LxVMut peptides titrated against FITC-PVIVIT peptide (5 nM) in the presence of 2 μM CNA/B. Values are plotted for n = 2 independent experiments. e Competitive binding assay of unlabeled pCLMB_WT, CLMB_WT, pCLMB_IxITMut, and pCLMB_LxVMut peptides titrated against FITC-pCLMB_WT peptide (5 nM) in the presence of 0.08 μM CNA/B. Values are plotted for n = 2 independent experiments. f Table displaying the experimentally obtained IC₅₀ values and converted K_i values from data in (d, e). Source data are provided as a Source Data file.

Fig. 4. CLMB-calcineurin binding extends beyond the LxVP and PxIxIT motifs.

a ¹⁵N HSQC spectra of free CLMB with the resonance assignment. b NMR spectral analysis of CLMB interaction with CNA at 2:1 (CLMB: CNA); c with CNA at 1:1 (CLMB: CNA); d with CNA/B at 2:1 (CLMB: CNA/B); e with CNA/B at 1:1 (CLMB: CNA/B). b–e Graphs display the intensity ratios of CLMB resonances with and without the titrant (CNA or CNA/B) at different molar ratios. The ratios corresponding to the LxV motif (pink) and IxIT (purple) are highlighted. Residues with no access to the intensity ratio are represented with negative numbers, and peaks resulting in complete broadening are shown in green (I/I0 = −0.4). Unassigned residues in apo CLMB are blue (I/I0 = −0.2), prolines (which are not represented in ¹⁵N-HSQC are gray (I/I0 = −0.2), and residues for which the bound resonance cannot be unambiguously assigned are red (I/I0 = −0.2). *Residues with I/I0 higher than 1 due to overlapping resonances. NMR chemical shift data have been deposited in the Biological Magnetic Resonance Bank (BMRB) under accession code [52641].

Fig. 5. Calcineurin dephosphorylates calcimembrin by forming multimers.

a Representative immunoblot showing phosphorylation-induced mobility shifts in CLMB-FLAG (wild-type or indicated mutants) when co-expressed with GFP-CNA and CLMB-V5 (wild-type or indicated mutants). GAPDH was used as a loading control. b Data from (a) showing mean ± SEM phospho-CLMB-FLAG normalized to total CLMB-FLAG. p-values calculated by two-way ANOVA corrected by Dunnett’s multiple comparison test (n = 3 independent experiments). n.s., not significant; without vs with CLMB_IxIxITmut-V5 expression: p = 0.995, 0.748, 0.999 from left to right. CLMB_IxITmut without vs with CNA-GFP p = 0.626. ****p < 0.0001, ***p < 0.001, p = 0.0005. c Cartoon representation of calcineurin (turquoise)-CLMB complex formation and CLMB dephosphorylation (removal of pink “P”) under conditions in (a), i.e., CLMB-FLAG (blue with “F”) overexpression with or without co-expression of Calcineurin-GFP and CLMB-V5 (green with “V”) as indicated. Top row shows CLMB_WT-FLAG, bottom row shows CLMB_IXITMut-FLAG (labeled LDVPxxxx). d (Top) Experimental data showing phosphate released from 100 nM pCLMB_WT (blue) or pCLMB_IxITMut (purple) phosphopeptide dephosphorylated by 20 nM (left), 100 nM (middle) or 200 nM (right) CNA/B and 1 μM calmodulin. Each dephosphorylation rate is displayed, and data are mean ± SEM at each timepoint; n = 6 independent replicates for all conditions except pCLMB_IxITMut with 200 nM CNA/B, where n = 5. (Bottom) Overlay of experimental data (circles) with results of kinetic simulations (solid lines) using the model shown in Fig. S5a, b for each calcineurin concentration. Source data are provided as a Source Data file and at Mendeley Data [10.17632/6kp379tsv4.1].

Fig. 6. Models explaining how calcimembrin may modulate calcineurin signaling.

a (1) CLMB localizes to membranes via dynamic acylation, where it (2) recruits calcineurin (turquoise). At low CLMB levels, calcineurin remains tightly bound to phosphorylated CLMB (purple) regardless of Ca²⁺ signals. (3) At higher CLMB levels, CLMB (purple and green) forms multimers with calcineurin and is dephosphorylated upon calcineurin activation by Ca²⁺ and CaM. (4) Calcineurin is then released and can engage and dephosphorylate membrane-associated substrates (red) that contain PxIxT and LxVP SLiMs. b (1) CLMB localizes to membranes via dynamic acylation, where it (2) recruits calcineurin (turquoise). At low CLMB levels, calcineurin remains tightly bound to phosphorylated CLMB (purple). (3) Ca²⁺ and CaM activate calcineurin, allowing it to engage and dephosphorylate LxVP-containing substrates (red). (4) At high CLMB levels (purple and green), calcineurin-CLMB multimers form. (5) This promotes CLMB dephosphorylation and triggers calcineurin release to terminate signaling.