Protection of extraribosomal RPL13a by GAPDH and dysregulation by S-nitrosylation

Abstract

Multiple eukaryotic ribosomal proteins (RPs) are co-opted for extraribosomal "moonlighting" activities, but paradoxically, RPs exhibit rapid turnover when not ribosome-bound. In one illustrative case of a functional extraribosomal RP, interferon (IFN)-γ induces ribosome release of L13a and assembly into the IFN-gamma-activated inhibitor of translation (GAIT) complex for translational control of a subset of inflammation-related proteins. Here we show GAPDH functions as a chaperone, shielding newly released L13a from proteasomal degradation. However, GAPDH protective activity is lost following cell treatment with oxidatively modified low density lipoprotein and IFN-γ. These agonists stimulate S-nitrosylation at Cys(247) of GAPDH, which fails to interact with L13a, causing proteasomal degradation of essentially the entire cell complement of L13a and defective translational control. Evolution of extraribosomal RP activities might require coevolution of protective chaperones, and pathological disruption of either protein, or their interaction, presents an alternative mechanism of diseases due to RP defects, and targets for therapeutic intervention.

Figure 1. LDL_ox suppresses GAIT activity by selective degradation of ribosomal protein L13a

(A) Effect of IFN-γ and LDL_ox on GAIT target gene expression. Human PBM were incubated with IFN-γ (500 unit/ml) in the presence of LDL_ox (50 μg/ml). Cell lysates (40 μg protein) were immunoblotted with anti-Cp, anti-VEGF-A and anti-β-actin antibodies (upper 3 panels). PBM were subjected to metabolic labeling for up to 24 h in the presence of [³⁵S]Met/Cys and IFN-γ plus LDLox. Conditioned media and lysates were analyzed by SDS-PAGE and autoradiography (lower panel). (B) Lack of effect of IFN-γ and LDL_ox on GAIT target mRNA expression. Human U937 monocytic cells were treated with IFN-γ in presence of unmodified LDL or LDL_ox, and lysates immunoblotted as above (top 3 panels). Cp, VEGF-A, and β-actin mRNA were determined by quantitative RT-PCR with appropriate primers; results were normalized to β-actin mRNA and expressed as mean ± SEM (n = 3 experiments). (C) LDL_ox suppresses translational silencing activity of the GAIT pathway. U937 cells were incubated with IFN-γ (500 unit/ml) in the presence of LDL or LDL_ox (50 μg/ml) for 8 or 24 hr. Cp GAIT element-bearing Luc reporter and T7 gene 10 RNA (200 ng of each) were subjected to in vitro translation in rabbit reticulocyte lysate (RRL) in the presence of cell lysate (1 μg) and ³⁵S-methionine. (D) LDL_ox represses GAIT complex assembly. After cell treatment for 24 hr with IFN-γ or IFN-γ plus LDL_ox, lysates (1 mg) were immunoprecipitated with anti-NSAP1 antibody and immunoblotted with anti-NSAP1, -EPRS, -L13a and -GAPDH antibodies. (E) IFN-γ plus LDL_ox represses L13a expression. U937 cells were incubated with IFN-γ and LDL or LDL_ox for up to 24 hr. Expression of GAIT complex proteins was determined by immunoblot analysis (left). L13a mRNA (right) was determined by quantitative RT-PCR, normalized to expression in untreated cells (mean ± SEM, n = 3 experiments). (F) LDL_ox by itself does not degrade L13a. Cells were treated as in (A) and L13a and β-actin determined by immunoblot. (G) Proteasome degradation of L13a. Cells were incubated with IFN-γ plus LDL_ox in presence or absence of MG132 (5 μM); MG132 was added after 4 and 12 hr for the 8- and 24-hr incubations, respectively. Cell lysates were immunoblotted with anti-L13a or anti-β-actin antibodies. (H) Poly-ubiquitinylation of L13a. U937 cells were transfected with plasmid encoding HA-ubiquitin and then treated with IFN-γ, LDL_ox, MG132, and UBEI-41 (25 μM) as indicated. Lysates were subjected to immunoprecipitation with anti-L13a antibody and immunoblot analysis with anti-HA tag and -L13a antibodies.

Figure 2. Phosphorylation-dependence of L13a ubiquitinylation and degradation

(A) L13a degradation is phosphorylation-dependent. U937 cells were transfected with plasmids encoding Flag-tagged wild-type (WT) or phospho-null (S77A) L13a. After recovery, cells were treated with IFN-γ plus LDL_ox in the presence of cycloheximide (CHX) to inhibit de novo synthesis. Lysates were immunoblotted with anti-Flag, -L6, and -L28 antibodies (top 3 panels). Cells were transfected with ZIP kinase (or scrambled) siRNA and then incubated and immunoblotted as above (bottom 3 panels). (B) L13a poly-ubiquitinylation is phosphorylation-dependent. U937 cells were co-transfected with plasmids encoding HA-ubiquitin and Flag-tagged wild-type or S77A mutant L13a. After recovery, cells were treated with IFN-γ plus LDL_ox and with MG132 and UBEI-41 as in Figure 1H. L13a in cell lysates was immunoprecipitated with anti-Flag antibody, and immunoblotted with anti-HA, -L13a, and -phospho-Ser antibodies. (C) Poly-ubiquitinylation requires ZIP kinase (ZIPK). U937 cells were co-transfected with plasmids encoding HA-ubiquitin and Flag-L13a, and with ZIPK (or scrambled) siRNA. Cells were treated as above, immunoprecipitated with anti-Flag antibody, and immunoblotted with anti-HA (top), anti-L13a (second panel), or anti-phospho-Ser (third panel) antibodies. ZIPK was detected in 16 hr lysates by immunoblot with anti-ZIPK antibody (bottom). (D) L13a phosphorylation is required for E3 ligase recognition. Flag-tagged L13a was incubated with biotin-ubiquitin and a reconstituted ubiquitinylation system containing recombinant E1 and E2 and lysate from cells treated with IFN-γ for 24 hr (or 0 hr as negative control) to provide E3 ligase and L13a kinase activity.

Figure 3. GAPDH binds phospho-L13a and protects it from degradation

(A) GAPDH binds phospho-L13a in absence of other GAIT components. Unmodified and phosphorylated L13a (P-L13a, prepared by treatment with ZIPK) were incubated with GST-GAPDH or GST immobilized to glutathione (GSH)-agarose beads. After washing, binding was detected by immunoblot analysis with anti-L13a, -phospho-Ser and -GAPDH antibodies. (B) LDL_ox blocks GAPDH binding to phospho-L13a. Cells were treated with IFN-γ, LDL_ox, and MG132 as shown. Cell lysates were immunoprecipitated with anti-GAPDH antibody, and immunoblotted with anti-L13a and -GAPDH antibodies. Total L13a was detected by immunoblot with anti-L13a antibody. (C) GAPDH protects L13a from degradation. U937 cells were transfected with GAPDH (or scrambled) siRNA. After recovery, cells were treated with IFN-γ, and lysates immunoblotted with anti-GAPDH, -L13a, -L6, and -L28 antibodies. (D) GAPDH prevents ubiquitinylation of phospho-L13a in cells. U937 cells were co-transfected with GAPDH (or scrambled) siRNA and with plasmids encoding HA-ubiquitin and Flag-tagged L13a (wild-type or S77A mutant). After recovery, cells were incubated with IFN-γ and MG132 for 24 hr. Lysates were immunoprecipitated with anti-Flag antibody, and immunoblotted with anti-HA, -L13a, and -GAPDH antibodies. Total GAPDH was determined by immunoblot with anti-GAPDH antibody. (E) LDL_ox induces S-nitrosylation of GAPDH. Cells were treated with IFN-γ, LDL, and LDL_ox. Lysates were immunoprecipitated with anti-GAPDH antibody and immunoblotted with anti-GAPDH, -S-nitrosocysteine (SNO-Cys), and -acetyl-Lys antibodies. (F) Determination of GAPDH S-nitrosylation by biotin-switch assay. S-nitrosylation of endogenous GAPDH and L13a degradation were determined in human PBM treated with IFN-γ in presence of LDL or LDL_ox for 24 hr. GAPDH was immunoprecipitated and S-nitrosylation detected with avidin-HRP in the biotin-switch assay. Efficiency of immunoprecipitation was shown by immunoblot analysis with anti-GAPDH antibody.

Figure 4. Effect of LDL_ox on GAIT target gene expression and schematic

(A) LDL_ox induces S-nitrosylation of GAPDH at Cys²⁴⁷. Mouse BMDM were transfected with plasmids encoding wild-type HA-GAPDH, or C152S, C156S, or C247S mutants. Cells were treated with IFN-γ and LDL_ox for 24 hr. S-nitrosylation of transfected GAPDH was detected in lysates by immunoprecipitation with anti-HA tag antibody followed by biotin-switch analysis. Total transfected HA-GAPDH was determined by immunoprecipitation with anti-HA tag antibody followed by immunoblot analysis with anti-GAPDH antibody. Total GAPDH and β-actin were determined by immunoblot analysis. (B) Expression of GAPDH C247S mutant restores translational silencing of VEGF-A. Cells were transfected with plasmid encoding HA-tagged C247S mutant GADPH and treated with IFN-γ or IFN-γ plus LDL_ox for up to 24 hr. Expression of HA-GAPDH, L13a, VEGF-A, and β-actin was detected by immunoblot analysis. (C) LDL_ox and LPS Induce GAPDH S-nitrosylation at different sites. Human PBM were transfected with plasmids encoding wild-type HA-GAPDH, or C152S, C156S, or C247S mutants. S-nitrosylation of transfected GAPDH in presence of IFN-γ plus LDL_ox or IFN-γ plus LPS (10 μg/ml) was detected by biotin-switch. Transfected HA-GAPDH and GAPDH-bound L13a were determined by immunoprecipitation with anti-HA tag antibody followed by immunoblot analysis with anti-GAPDH and anti-L13a antibodies. Total GAPDH and β-actin were determined by immunoblot analysis. (D) Three-dimensional structure of GAPDH. Ribbon diagram of tetrameric GAPDH structure with individual subunits shown in different colors. Cys residues are shown in spacefill with sulfur atoms highlighted (yellow). NAD⁺ in active site is shown (stick model). (E) Schematic of GAIT pathway activation and its disruption by LDL_ox. IFN-γ induces GAIT complex assembly that causes translational repression of inflammation-related, GAIT target genes (upper pathway). LDL_ox induces GAPDH S-nitrosylation and consequent L13a degradation, causing GAIT pathway dysregulation and prolonging GAIT target gene expression (lower pathway). See also Figure S1.