Delineation of the substrate recognition domain of MARCHF6 E3 ubiquitin ligase in the Ac/N-degron pathway and its regulatory role in ferroptosis

Abstract

Nα-terminal acetylation in eukaryotic proteins creates specific degradation signals (Ac/N-degrons) targeted for ubiquitin-mediated proteolysis via the Ac/N-degron pathway. Despite the identification of key components of the Ac/N-degron pathway over the past 15 years, the precise recognition domain (Ac/N domain) remains unclear. Here, we defined the Ac/N domain of the endoplasmic reticulum MARCHF6 E3 ubiquitin ligase through a systematic analysis of its cytosol-facing regions using alanine-stretch mutagenesis, chemical crosslinking-based co-immunoprecipitation-immunoblotting, and split-ubiquitin assays in human and yeast cells. The Ac/N domain of MARCHF6 exhibits preferential binding specificity to Nα-terminally acetylated proteins and peptides over their unacetylated counterparts, mediating the degradation of Ac/N-degron-bearing proteins, such as the G-protein regulator RGS2 and the lipid droplet protein PLIN2. Furthermore, abolishing the recognition of Ac/N-degrons by MARCHF6 stabilized RGS2 and PLIN2, thereby increasing the resistance to ferroptosis, an iron-dependent lipid peroxidation-mediated cell death. These findings provide mechanistic and functional insights into how MARCHF6 serves as a rheostatic modulator of ferroptosis by recognizing Ac/N-degron substrates via its Ac/N domain and non-Ac/N-degron substrates via distinct recognition sites.

Keywords: MARCHF6; N-terminal acetylation; acetylation/N domain; degradation signal; ferroptosis; proteolysis; ubiquitin.

Figure 1

MARCHF6 specifically identifies Ac/N-degron substrates through its fifth cytosolic region.A, a depiction of 25 alanine-stretch mutations (yellow dots) dispersed across eight cytosolic regions of MARCHF6. Cytosolic residues and alanine-stretch mutation residues for each of the eight cytosolic regions are indicated within parentheses next to Cn and Cn-X, respectively. Cn and Cn-X represent each cytosolic region and the mutations within it. B, a schematic for chemical crosslinking and reciprocal co-immunoprecipitation–immunoblotting analyses with HeLa cells coexpressing 25 Ala-stretch MARCHF6_3f mutants and Nt-acetylatable M-RGS2_ha. See also Fig. S3. C, chemical crosslinking and reciprocal co-immunoprecipitation–immunoblotting analyses using HeLa cells coexpressing C5-17, C5-18, C5-19, or C5-20 MARCHF6_3f mutant with Nt-acetylatable M-RGS2_ha. D, same as in (C) but with Nt-acetylatable A-PLIN2_ha. E, same as in (C) but with non-Nt-acetylatable P-SM_ha.

Figure 2

The Ac/N domain is located within residues 1 to 721 of MARCHF6.A and B, schematic of the split-Ub assay using Nt-acetylatable or non-Nt-acetylatable baits and either wild-type MARCHF6 or C-terminally truncated segments of MARCHF6 as prey. MARCHF6^1–X, where X represents the end residue of each cytosolic segment of MARCHF6. The results of the split-Ub experiments are shown on the right-hand side of (B). C, split-Ub assays of bait Nt-acetylatable M-RGS2 or non-Nt-acetylatable P-RGS2 with prey MARCHF6^1–188, MARCHF6^1–367, MARCHF6^1–467, MARCHF6^1–565, MARCHF6^1–721, MARCHF6^1–810, and MARCHF6^1–910. Yeast cells expressing the indicated baits and preys were cultured on SC media with histidine (+His) or without histidine (−His) media for 3 days. D, split-Ub assays of bait Nt-acetylatable A-PLIN2 or non-Nt-acetylatable P-PLIN2 with prey MARCHF6^1–188, MARCHF6^1–367, MARCHF6^1–467, MARCHF6^1–565, MARCHF6^1–721, MARCHF6^1–810, and MARCHF6^1–910. Yeast cells expressing the indicated baits and preys were cultured on SC media with histidine (+His) or without histidine (−His) media for 3 days.

Figure 3

The fifth cytosolic region (C5) of MARCHF6 contains the Ac/N domain.A, a schematic of the split-Ub assay using either Nt-acetylatable A-PLIN2 or non-Nt-acetylatable P-PLIN2 as bait, and eight cytosol-facing segments of MARCHF6 as preys. Cn represents each cytosol-facing segment of MARCHF6. B, split-Ub assays of the bait A-PLIN2 or P-PLIN2 with the prey C1 (MARCHF6^1–91), C2 (MARCHF6^164–283), C3 (MARCHF6^358–376), C4 (MARCHF6^443–480), C5 (MARCHF6^541–632), C6 (MARCHF6^700–721), C7 (MARCHF6^786–810), and C8 (MARCHF6^870–910). Yeast cells expressing the indicated bait and prey were cultured on selective synthetic complete (SC) media with histidine (+His) or without histidine (-His) media for 3 days. C, chemical crosslinking and reciprocal co-immunoprecipitation–immunoblotting analyses of HeLa cells coexpressing C-terminally triple-flag tagged C5_3f (${\text{MARCHF}6}_{3\mathrm{f}}^{541–632}$) with empty vector only, Nt-acetylatable M-RGS2_ha, or non-Nt-acetylatable P-RGS2_ha. D, same as in (C) but with Nt-acetylatable A-PLIN2_ha or non-Nt-acetylatable P-PLIN2_ha. E, ITC analyses for the binding affinity of GST-C5 (MARCHF6^541-632-GST) to N-terminally acetylated Ac-MQASMD, MQSAMD, Ac-ASVAVD, and ASVAVD peptides (see also Fig. S5). F, same as in (B), but with preys MARCHF6^541–579, MARCHF6^580–632, and MARCHF6^552–600.

Figure 4

The Ac/N domain of MARCHF6 is essential for the specific recognition of Ac/N-degron substrates.A, sequence alignment of the human MARCHF6 Ac/N domain with its counterparts in mouse (Mus musculus), chicken (Gallus gallus), fish (Danio rerio), fly (Drosophila melanogaster), frog (Xenopus laevis), worm (Caenorhabditis elegans), plant (Arabidopsis thaliana), and yeast (Saccharomyces cerevisiae). B and C, chemical crosslinking and co-immunoprecipitation-immunoblotting analyses of wild-type MARCHF6_3f, ${\text{MARCHF}6}_{3\mathrm{f}}^{\mathrm{W}556A}$, ${\text{MARCHF}6}_{3\mathrm{f}}^{\mathrm{L}566\mathrm{A}}$, ${\text{MARCHF}6}_{3\mathrm{f}}^{S568\mathrm{A}}$, ${\text{MARCHF}6}_{3\mathrm{f}}^{\mathrm{L}571\mathrm{A}}$, or ${\text{MARCHF}6}_{3\mathrm{f}}^{\mathrm{P}599\mathrm{A}}$ with RGS2_ha and PLIN2_ha. D–F, same as in (B, C) but with wild-type MARCHF6_3f, ${\text{MARCHF}6}_{3\mathrm{f}}^{\mathrm{L}566\mathrm{A}}$, or ${\text{MARCHF}6}_{3\mathrm{f}}^{\mathrm{L}571\mathrm{A}}$ with SM_ha, p53_ha, and ACSL4_ha, respectively. G and H, CHX-chase experiments of endogenous RGS2 and PLIN2 in MARCHF6-KO HeLa cells expressing empty vector only, wild-type MARCHF6_3f, or ${\text{MARCHF}6}_{3\mathrm{f}}^{\mathrm{L}571\mathrm{A}}$.

Figure 5

MARCHF6 induces ferroptosis by promoting the degradation of Ac/N-degron-bearing substrates.A and B, immunoblotting of HeLa or A549 cells overexpressing the empty vector alone, M-RGS2_ha, or P-RGS2_ha, A-PLIN2_ha, or P-PLIN2_ha. Tubulin was used as a loading control. C and D, levels of lipid ROS in wild-type HeLa and A549 cells overexpressing empty vector alone, Nt-acetylatable M-RGS2_ha and A-PLIN2_ha or non-Nt-acetylatable P-RGS2_ha and P-PLIN2_ha with RSL3 (0.15 μM) treatment for 24 h. The graphs show the ratio of oxidized to total C11- (oxidized + reduced) C11-BODIPY^581/591 signals (as indicated by the bar). E and F, relative viability of wild-type HeLa and A549 cells overexpressing empty vector alone, Nt-acetylatable M-RGS2_ha and A-PLIN2_ha or non-Nt-acetylatable P-RGS2_ha and P-PLIN2_ha under increasing concentrations of RSL3 for 24 h. Quantification data are presented as mean ± SD (n = 3 biologically independent samples); one-way (A and D) or two-way (E and F) ANOVA with Tukey’s HSD post hoc test. Significant p values are presented in the figures.

Figure 6

Blocking the recognition of Ac/N-degrons by MARCHF6 enhances resistance to ferroptosis.A, immunoblotting of endogenous RGS2, PLIN2, p53, ACSL4, SM, GPX4, NRF2, and SLC7A11 in MARCHF6-KO HeLa and A549 cells expressing empty vector alone, MARCHF6_3f, or ${\text{MARCHF}6}_{3\mathrm{f}}^{\mathrm{L}571\mathrm{A}}$. Tubulin was used as a loading control. B, Levels of lipid ROS in MARCHF6-KO HeLa and A549 cells expressing empty vector alone, MARCHF6_3f, or ${\text{MARCHF}6}_{3\mathrm{f}}^{\mathrm{L}571\mathrm{A}}$ with RSL3 (0.15 μM) treatment for 24 h. C, relative viability of MARCHF6-KO HeLa and A549 cells expressing empty vector alone, MARCHF6_3f, or ${\text{MARCHF}6}_{3\mathrm{f}}^{\mathrm{L}571\mathrm{A}}$ under increasing concentrations of RSL3 for 24 h. Quantification data are presented as mean ± SD (n = 3 biologically independent samples); one-way (B) or two-way (C) ANOVA with Tukey’s HSD post hoc test. Significant p values are presented in the figures.

Figure 7

Proposed structure of the Ac/N domain of MARCHF6 and its rheostatic role in the regulation of ferroptosis.A, three-dimensional structure of the Ac/N domain (dark grey) and its docking with Nt-acetylated Ac-ASVAVD peptide (yellow). Arg554 and Asn579 in Ac/N domain, which form hydrogen bonds with the Ac-ASVAVD peptide, are annotated with side chains, and the hydrogen bonds are indicated with red dotted lines. Leu566 and Leu571 (magenta), critical for the formation of the Ac/N domain structure, are also annotated with their side chains. Nitrogen, oxygen, and hydrogen atoms are colored blue, red, and white, respectively. B, MARCHF6 acts as a rheostatic modulator of ferroptosis, exhibiting both pro-ferroptotic and anti-ferroptotic functions. The E3 Ub ligase promotes ferroptosis sensitivity by targeting Ac/N-degron-bearing substrates (RGS2 and PLIN2) via its Ac/N domain. Conversely, it also promotes ferroptosis resistance by targeting pro-ferroptosis effectors (p53, ACSL4, and SM) via distinct substrate-biding sites. Furthermore, MARCHF6 upregulates the expression of anti-ferroptosis effectors, including GPX4, SLC7A11, and NRF2. While MARCHF6 promotes the degradation of p53, leading to increased SLC7A11 expression (23), it also prevents the degradation of GPX4 through chaperone-mediated autophagy by promoting the degradation of cytosol-misoriented pro-hormones, such as POMC (37, 40). However, the mechanisms by which MARCHF6 regulates NRF2 remain unknown. The Nt-catalytic RING domain and Ct-regulatory MRR of MARCHF6 are depicted in the figure.