SIRT6 Lysine-Demyristoylates ATF2 to Ameliorate Vascular Injury via PRKCD/VE-Cadherin Pathway Regulating Vascular Endothelial Barrier

Abstract

Cardiovascular diseases (CVDs) progression is significantly modulated by epigenetic mechanisms, particularly through Sirtuin 6 (SIRT6), a key NAD⁺-dependent deacetylase in the sirtuin family. Though essential for cardiovascular homeostasis, the effects of SIRT6-mediated lysine myristoylation on CVDs progression remain largely unexplored due to detection limitations. This study developes an innovative lysine-myristoylated peptide enrichment technique, identifying mutant SIRT6 (H133Y) with high myristoyl affinity but deficient demyristoylase activity. This advancement enables identification of 15 previously unrecognized human lysine-myristoylated proteins. Further study demonstrates that SIRT6 demyristoylates activating transcription factor 2 (ATF2) at K296 and regulates its nucleoplasmic translocation. Through 4D label-free mass spectrometry and molecular approaches, it is revealed that decreased nuclear localization of ATF2 results in reduced Protein kinase C delta type (PRKCD) expression, establishing a SIRT6/Myr-ATF2/PRKCD/VE-Cadherin pathway that enhances endothelial barrier integrity under high myristate conditions. These findings are validated in vitro (gene overexpression/knockdown cells) and in vivo (SIRT6 knockout/double-transgenic mice). The study provides both a novel method for identifying lysine-myristoylated proteins and critical insights into SIRT6 demyristoylation biology. Modulating the SIRT6 pathway might yield therapies to strengthen endothelial integrity and mitigate vascular dysfunction in CVDs, offering promising clinical translation avenues.

Keywords: ATF2; SIRT6; endothelial barrier function; epigenetic regulation; myristoylation; nucleoplasmic translocation.

Figure 1

Identification and characterization of dSIRT6 mutants for myristoylated peptide enrichment. A) The procedure for enriching myristoylated peptides using dSIRT6 for mass spectrometry (MS) analysis, highlighting the experimental setup. B) Analysis of purified SIRT6 and its mutants (H133Y, G60A, S56Y, R65A) across a concentration gradient, using SDS‐PAGE and Coomassie Blue Staining to evaluate their purity and concentration. C) Design of biotinylated and arginine‐rich peptides (B‐H3K9‐5R, B‐H3K9^Myr‐5R, B‐H3K9Ac‐5R) for the study of SIRT6 binding and activity. D) Utilizes High‐Performance Liquid Chromatography (HPLC) to assess the deacetylase and demyristoylase activities of SIRT6 mutants, providing insight into their enzymatic functions. E) Employs Biolayer Interferometry (BLI) to quantify the binding affinity of SIRT6 and its mutants to myristoylated peptides, demonstrating the specificity and strength of interactions. F) Compiles affinity data for the interaction of SIRT6 and its mutants with myristoylated and non‐myristoylated peptides, summarizing the specificity of binding. G,H) Illustrates the use of Co‐Immunoprecipitation (Co‐IP) for optimizing the enrichment of myristoylated peptides by dSIRT6, including system selection and reaction conditions. I) The optimization of elution conditions for myristoylated peptide enrichment, using acetonitrile to enhance recovery. J) Schematic overview of the H133Y‐based enrichment method for myristoylated peptides, detailing the procedure and its applications.

Figure 2

Exploration of lysine myristoylation in SIRT6 KO Cells. A,B) Investigate global myristoylation levels in SIRT6 knockout (KO) 293T cells, using western blot analysis to determine the effects of ALK14 treatment across a concentration gradient. C) A mass spectrum of peptides enriched by the H133Y mutant, highlighting the precision of the enrichment method. D) The conditions and results from MS analyses, providing data on the identification process. E) Features Gene Ontology (GO) enrichment analysis of identified myristoylated proteins, indicating their biological significance and functional categories. F) The subcellular localization of identified myristoylated proteins, offering insights into their roles within the cell. G) Depicts a protein interaction network centered on SIRT6, illustrating the connections and potential regulatory mechanisms involving myristoylated proteins. H) Analyzes identified myristoylation motifs using bioinformatics tools, contributing to the understanding of myristoylation patterns and preferences. I) A mass spectrum of ATF2 highlighting a specific myristoylation site, exemplifying the method's ability to pinpoint post‐translational modifications. J) Outlines the comprehensive approach of combining H133Y‐based immunoprecipitation with mass spectrometry (IP‐MS) for the identification of myristoylated proteins in cells treated with ALK14, elucidating the methodological framework and its application to studying protein myristoylation.

Figure 3

Biological function of myristoylation at K296 site of ATF2. A) Western blot analysis of ATF2 level in 293T under ALK14 treatment in a concentration gradient (0–20 µg mL⁻¹) for 12 h (n = 3). B) Immunoprecipitation (IP) and CLICK IT assays were performed in the lysates of SIRT6 KO 293T to detect the myristoylation level of ATF2 (n = 3). Before lysis, cells were transfected with ATF2‐FLAG (WT) and ATF2 K296R‐FLAG (KR) plasmids treated with or without ALK14 (10 µg mL⁻¹) for 12 h. C) IP and CLICK IT assays were performed in the lysates of SIRT6KO and WT cells to detect the myrsitoylation of ATF2 (n = 3). Before lysis, cells were transfected with ATF2‐FLAG (WT) and ATF2 K296R‐FLAG (KR) plasmids under ALK14 (10 µg mL⁻¹) treatment for 12 h. D) Nucleoplasmic separation and western blot assays were performed in SIRT6 KD HMEC cells to detect the level of ATF2 in cytoplasm and nucleus. Before lysis, cells were transfected with ATF2 WT and KR plasmids under ALK14 (10 µg mL⁻¹) treatment for 12 h or not. Immunofluorescence staining of ATF2 (red) in each group under ALK14 treatment or not. DAPI was used for counterstaining cellular nuclei (blue) (n = 10, scale bars = 20 µm). The histograms show the relative ATF2 level (%) in nuclear (n = 3). E) Nucleoplasmic separation and western blot assays were performed in SIRT6 KD HMEC cells to detect the level of ATF2 in cytoplasm and nucleus. Before lysis, cells were transfected with G60A, ATF2 WT and KR plasmids under ALK14 (10 µg mL⁻¹) treatment for 12 h. Immunofluorescence staining of ATF2 (red) in each group under ALK14 treatment. DAPI was used for counterstaining cellular nuclei (blue) (n = 10, scale bars = 20 µm). The histograms show the relative ATF2 level (%) in nuclear (n = 3). F) Nucleoplasmic separation and western blot assays were performed in SIRT6 KD HMEC cells to detect the level of ATF2 in cytoplasm and nucleus. Before lysis, cells were transfected with G60A and ATF2 WT plasmids under ALK14 (10 µg mL⁻¹) treatment for 12 h or not. Immunofluorescence staining of ATF2 (red) in each group under ALK14 treatment or not. DAPI was used for counterstaining cellular nuclei (blue) (n = 10, scale bars = 20 µm). The histograms show the relative ATF2 level (%) in nuclear (n = 3). Date is represented as means ± S.E.M. p value by two‐tailed t‐test (A‐F).

Figure 4

Signal pathways and targeted genes regulated by ATF2 myritoylation. A) The heatmaps of the identified proteins derived from 4D‐Label‐Free proteomics analyses of WT versus Ctrl (left), KR versus Ctrl (middle) and KR versus WT (right) groups (n = 3). Pearson correlation coefficients were performed to measure the distance and average of sample clustering. B–D) The venn's diagrams show the upregulated GO biological processes, cellular components, and molecular functions pathways in WT versus Ctrl (yellow) and KR versus Ctrl (purple). The histograms show GO biological processes, cellular components, and molecular functions pathways in the intersected and KR versus ctrl ∖WT versus Ctrl. The x‐axis shows the number of proteins. E) The histograms showing GO biological processes, cellular components and molecular functions pathways in the upregulated KR versus WT and x‐axis show the number of proteins. F) The venn's diagram shows the overlapped differentially expressed proteins (DEPs) in KR upregulated proteins (KR versus Ctrl, KR versus WT) and SIRT6 upregulated proteins (SIRT6 overexpression versus control). (The data of DEPs in SIRT6 comes from our previous published data^{[ 35 ]}). The histogram shows the relative expression level of PRKCD and PLEKHA2. G) The volcano plots of DEPs derived from KR versus WT groups with PRKCD and PLEKHA2 annotated on the diagram (|log2 (Fold change) |≥0.2, and P <0.05). H) The protein interaction network of the detailed(left) and brief (right) cluster from KR versus WT DEPs, constructed by String (https://cn.string‐db.org/) and metascape (www.metascape.org/). I) The expression of PRKCD in different tissues (heart, tibial artery, coronary artery, ascending aorta) from donors based on data obtained from the GTEx database. The x‐axis shows the tissues from which the 1104 samples were derived, and the y‐axis indicated the log2 (TPM + 1) value.

Figure 5

SIRT6 demyristoylated ATF2 to improve the endothelial permeability through PRKCD/VE‐Cadherin signaling pathway. A) Western blot analysis of VE‐Cadherin, FLAG, PRKCD level in HMEC‐1 transfected with ATF2 WT and KR plasmids treated with or without ALK14 (10 ug mL⁻¹) for 12h. The histograms show the relative protein level (%) (n = 3). (B) Western blot analysis of VE‐Cadherin, PRKCD, SIRT6 level in HMEC‐1 transfected with G60A plasmids with or without ALK14 treatment. The histograms show the relative protein level (%) (n = 3). C) Western blot analysis of VE‐Cadherin, PRKCD, SIRT6 level in HMEC‐1 transfected with siPRKCD and G60A with ALK14 treatment. The histograms show the relative protein level (%) (n = 3). D) A brief schematic diagram of the role of siPRKCD in ATF2/PRKCD/VE‐Cadherin signaling pathway. E) A schematic diagram of the transwell permeability model. F) The histograms represented the quantitative assays of endothelial permeability in ATF2 WT and KR HMEC‐1 with or without ALK14 treatment (n = 3). G) The histograms represented the quantitative assay of endothelial permeability in HMEC‐1 transfected with G60A plasmids under ALK14 treatment (n = 3) or not. H) The histogram represented the quantitative assay of endothelial permeability of HMEC‐1 transfected with siPRKCD and G60A under ALK14 treatment (n = 3). Date is represented as means ± S.E.M. p value by two‐tailed t‐test (A, B, C, E, F, G, H).

Figure 6

SIRT6 demyristoylated ATF2 to improve the endothelial permeability through PRKCD/VE‐Cadherin signaling pathway. A) Immunofluorescence staining analyses of VE‐Cadherin (green) in HMEC‐1 transfected with ATF2 WT and KR plasmids under ALK14 (10 µg mL⁻¹) treatment for 12 h or not. Nuclei were counterstained with DAPI (blue) (n = 8–10, scale bars = 50 µm). The histograms show the distribution of the intercellular 5um or 10um gaps of each group (n = 8–10). B) Immunofluorescence staining analyses of VE‐Cadherin (green) in HMEC‐1 transfected with G60A plasmids under ALK14 treatment (n = 8–10, scale bars = 50 µm) or not. The histograms show the distribution of intercellular 5um or 10um gaps of each group (n = 8–10). C) Immunofluorescence staining analyses of VE‐Cadherin (green) in HMEC‐1 transfected with siPRKCD and G60A under ALK14 treatment (n = 8–10, scale bars = 50 µm). The histograms show the distribution of the intercellular 5 um or 10um gaps of each group (n = 8–10). Data were presented as mean ± S.E.M and analyzed using a two‐tailed t‐test A,B,C). ns indicates no significant difference.

Figure 7

ATF2/PRKCD/VE‐Cadherin signaling pathway was regulated by SIRT6 demyristoylase activity. A) Immunoprecipitation (IP) and CLICK IT assays were performed in the lysates of 293T SIRT6KO cells to detect the myristoylation level of ATF2 (n = 3). Before lysis, cells were transfected with SIRT6 mutants plasmids treated with ALK14 (10 µg mL⁻¹) for 12 h. The histograms show the relative ATF2 myristoylation level (%) (n = 3). B) Western blot analysis of VE‐Cadherin, PRKCD level in HUVECs transfected with SIRT6 and its mutants plasmids treated with ALK14 (10 ug mL⁻¹) for 12h. The histograms show the relative protein level (%) (n = 3). C) Immunofluorescence staining analyses of VE‐Cadherin (red) and ATF2(green) in HUVECs transfected with SIRT6 and its mutants treated with ALK14 (10 µg mL⁻¹) for 12 h. The histograms (up) show the distribution of the intercellular 5 or 10 um gaps of each group. The histograms (down) show the ATF2 nuclear proportion. scar bar = 100 µm (n = 6–8) Data were presented as mean ± S.E.M and analyzed using a two‐tailed t‐test. ns indicates no significant difference.

Figure 8

SIRT6 demyristoylated ATF2 to stabilize endothelial cells barrier in vivo. A) Immunofluorescence staining analyses of CD31 (red) αSMA (green) and ATF2 (white) in the aortic section from mice post AAV injection. Nuclei were counterstained with DAPI (blue) (n = 6–8, scale bars = 30 µm). B) Immunofluorescence staining analyses of CD31 (red) αSMA (green) and SIRT6 (white) in the aortic section from mice post AAV injection. Nuclei were counterstained with DAPI (blue) (n = 6–8, scale bars = 30 µm). C) The permeation of the Evans blue dye into the descending aortas of each group. (D) The histograms showed the quantification of Evans blue dye in the aortas, scar bar = 2.5 mm (n = 6). E) En face immunofluorescence staining images of VE‐Cadherin in the descending aortas of each group. F) The histogram showed the quantification of VE‐Cadherin mean junctions continuity in the aortas, scar bar = 25 µm (n = 6). G) The permeation of the Evans blue dye into the descending aortas of each group. H) The histograms showed the quantification of Evans blue dye in the aortas, scar bar = 2.5 mm (n = 6). I) En face immunofluorescence staining images of VE‐Cadherin in the descending aortas of each group. J) The histogram showed the quantification of VE‐Cadherin mean junctions continuity in the aortas, scar bar = 25 µm (n = 6). K) Western blot analyses of VE‐Cadherin, Prkcd expression in each group (n = 6). L)The histogram showed the relative ATF2, SIRT6 expression (n = 6). Date is represented as means ± S.E.M. p value by two‐tailed t‐test (C‐L).

Figure 9

In vivo validation of the SIRT6/ATF2/PRKCD axis A) En face immunofluorescence staining images of Prkcd (green), VE‐Cadherin (red) in the descending aortas of each group. Nuclei were counterstained with DAPI (blue). B) The histogram (up) showed the quantification of VE‐Cadherin mean junctions continuity in the aortas. The histogram (down) showed the relative fluorescent intensity (fold change%) of Prkcd in the aortas, scar bar = 25 µm (n = 6–8). C) Immunofluorescence staining analyses of Atf2(green) and VE‐Cadherin (red) in primary aortic endothelial cells. Nuclei were counterstained with DAPI (blue). D) The histogram (up) showed the quantification of VE‐Cadherin mean junctions continuity in the aortas. The histogram (down) showed the Atf2 nuclear proportion in the aortas, scar bar = 15 µm (n = 6–8). E) Immunofluorescence staining analyses of VE‐Cadherin (red) and Prkcd (green) on sections of mouse descending aorta. DAPI was used for counterstaining cellular nuclei (blue). F) The histogram showed the relative fluorescent intensity levels of VE‐Cadherin and Prkcd in the aortas. scar bar = 30 µm (n = 6–8). Data were presented as mean ± S.E.M and analyzed using a two‐tailed t‐test. ns indicates no significant difference.

Figure 10

Clinical transformation (A) En face immunofluorescence staining images of VE‐Cadherin in the descending aortas of each group. NMN (100 mg kg⁻¹ day⁻¹) and DDD85646 (100 mg kg⁻¹ day⁻¹). The histogram showed the quantification of VE‐Cadherin mean junctions continuity in the aortas, scar bar = 25 µm(n = 6–8). B) The schematic representation for the role of SIRT6 in maintaining endothelial barrier via ATF2 K296/PRKCD/VE‐Cadherin signaling. C) Structure of Wild‐Type ATF2 and K296R ATF2. Top: detailed structure of 296K and 8Å around; Middle: detail structure of 296R and 8Å around; Down: Overall structural comparison of wild‐type ATF2 and K296R ATF2 predicted by AlphaFold2. Date is represented as means ± S.E.M. p value by two‐tailed t‐test (A).