Phosphorylation of iRhom2 Controls Stimulated Proteolytic Shedding by the Metalloprotease ADAM17/TACE

Abstract

Cell surface metalloproteases coordinate signaling during development, tissue homeostasis, and disease. TACE (TNF-α-converting enzyme), is responsible for cleavage ("shedding") of membrane-tethered signaling molecules, including the cytokine TNF, and activating ligands of the EGFR. The trafficking of TACE within the secretory pathway requires its binding to iRhom2, which mediates the exit of TACE from the endoplasmic reticulum. An important, but mechanistically unclear, feature of TACE biology is its ability to be stimulated rapidly on the cell surface by numerous inflammatory and growth-promoting agents. Here, we report a role for iRhom2 in TACE stimulation on the cell surface. TACE shedding stimuli trigger MAP kinase-dependent phosphorylation of iRhom2 N-terminal cytoplasmic tail. This recruits 14-3-3 proteins, enforcing the dissociation of TACE from complexes with iRhom2, promoting the cleavage of TACE substrates. Our data reveal that iRhom2 controls multiple aspects of TACE biology, including stimulated shedding on the cell surface.

Keywords: 14-3-3; ADAM metalloproteases; ADAM17/TACE; EGFR; MAP kinases; TNF; ectodomain shedding; iRhom2.

Graphical abstract

Figure 1

iRhom2 Is Phosphorylated in Response to TACE-Activating Stimuli (A) Schematic envisaging iRhom2 phosphorylation as a signal integrator for TACE-activating stimuli. (B and C) Phos-tag gels show that endogenous iRhom2 is phosphorylated in bone marrow-derived macrophages (BMDMs) stimulated with LPS (B) and (C) poly(I:C). (D) The mobility shift in iRhom2 is reverted by phosphatase treatment. (E) Endogenous iRhom2 is phosphorylated in BMDMs stimulated with the G protein-coupled receptor ligand gastrin-releasing peptide (GRP). (F) Endogenous iRhom2 is phosphorylated in response to PMA and LPS in RAW264.7 macrophages. (G) HA-tagged mouse iRhom2, stably expressed in HEK293ET cells, is phosphorylated in response to PMA. Throughout, a white arrowhead denotes phosphorylated iRhom2 and a black arrowhead non-phosphorylated iRhom2. p97 immunoblots are a loading control and pERK (phosphorylated ERK) a positive control for stimulation. Here and throughout, cells were stimulated for 15 min with LPS (1 μg/mL), GRP (200 ng/mL), and PMA (1 μM) and for 60 min with poly(I:C) (1 μg/mL). See also Figure S1.

Figure 2

iRhom2 Phosphorylation Depends on MAP Kinases Downstream of Toll-like Receptor Signaling (A) Schematics illustrating how signaling downstream of TLR3/4, GPCRs, and PMA converges on MAPK activation. Inhibitors are highlighted in red. Adapted from Goldsmith and Dhanasekaran, 2007, Steinberg, 2008, and Oda and Kitano (2006). (B) A combination of ERK- and p38-MAPK inhibitors prevent endogenous iRhom2 phosphorylation in LPS-stimulated BMDMs. (C) MAPK inhibitors (MEK1/2, JNK, and p38) prevent endogenous iRhom2 phosphorylation in poly(I:C)-stimulated BMDM. Throughout, cells were pre-treated for 30 min with inhibitors (2.5 μM PD184352 [MEK1/2 inhibitor], 5 μM SP600125 [JNK inhibitor], 1 μM SB 202190 [p38 inhibitor]) followed by LPS (15 min) or poly(I:C) (60 min) treatment.

Figure 3

iRhom2 Cytoplasmic Tail Phosphorylation Is Required for Rapid Induction of TACE-Dependent Shedding (A) iRhom2 Dead mutant stably expressed in HEK293ET cells is not phosphorylated in response to PMA. (B) PMA-stimulated TGF-α-AP shedding requires iRhom2 phosphorylation. iRhom1/2 DKO MEFs were transduced with iRhom2 WT, iRhom2 Dead mutant, or empty vector (EV) retrovirus. (C) Marimastat (MM; 5 μM) was used to inhibit shedding by metalloproteases. (D) The ADAM10-specific inhibitor GI254023X (GI; 1 μM, 60 min) did not affect TGF-α-AP shedding in rescue assays in iRhom DKO MEFs expressing WT iRhom2 versus the Dead mutant or empty vector. (E) Ionomycin (IO; 2.5 μM, 60 min) stimulated shedding of betacellulin-AP (BTC-AP) was unaffected by iRhom2 phosphorylation. (F) iRhom2 KO BMDMs transduced with iRhom2 Dead retrovirus are defective in the shedding of TNF after 3 hr of LPS stimulation. Data are presented as mean ± SEM. See also Table S1.

Figure 4

Phosphorylation of iRhom2 at Serine 83 Is Essential for TACE Shedding Activity (A) Schematic of the putative iRhom2 phosphorylated residues that were mutated to alanine. (B) The putative phosphorylated residues located within amino acids 1–130 of iRhom2 are required for TACE shedding of TGF-α-AP in rescue assays in iRhom1/2 DKO MEFs stably expressing the specified mutants. PMA (60 min) was used to stimulate shedding. EV, empty vector. (C) The iRhom2 non-phosphorylatable mutants (used in B) had equivalent expression levels in iRhom1/2 DKO MEFs. (D) Mass spectrometry analysis identified six phosphorylation sites in mouse iRhom2-HA, stably expressed in HEK293ET cells. Five of those identified are PMA inducible (15 min PMA). FC, fold change relative to WT cells treated with DMSO. n = 2, Student’s t test. (E and F) Mass spec analysis of iRhom2 phosphorylation at S60 (E) or S83 (F). (G) iRhom1/2 DKO MEFs transduced with iRhom2 retrovirus encoding the indicated serine-to-alanine (S/A) single-point mutations PMA (30 and 60 min) was used to stimulate shedding of TGF-α-AP. (H) The iRhom2 S/A non-phosphorylatable mutants (used in G) had equivalent expression levels in iRhom1/2 DKO MEFs. Data are presented as mean ± SEM.

Figure 5

Phosphorylation-Dependent Recruitment of 14-3-3 to iRhom2 Induces Increased TACE Activity (A) iRhom2-interacting proteins with altered binding upon PMA stimulation, detected by mass spectrometry, in iRhom2-HA immunoprecipitates; 14-3-3 proteins are highlighted in blue. See Figure S1B for a more comprehensive view of this dataset (n = 2). (B) Validation of PMA-inducible 14-3-3 binding to iRhom2. HEK293ET cells expressing an empty vector (EV) control, WT, or iRhom2-Dead-HA were stimulated (PMA, 30 min). iRhom2-HA was immunoprecipitated and endogenous 14-3-3 binding detected with a pan-14-3-3 antibody. (C) S83 is the major phosphorylated residue that recruits 14-3-3 proteins. HEK293ET cells expressing iRhom2-HA serine-to-alanine (S/A) mutants at the three putative 14-3-3 binding sites (S60, S83, and S359) were stimulated with PMA (30 min), iRhom2-HA was immunoprecipitated and endogenous 14-3-3 binding detected. (D) Schematic of the R18-iRhom2 Dead mutant. The R18 peptide (PHCVPRDLSWLDLEANMCLP) was fused to the N terminus of iRhom2 Dead-HA mutant, to confer phosphorylation independent binding of 14-3-3 proteins. (E) The R18-iRhom2 Dead-HA mutant binds constitutively to endogenous 14-3-3 proteins. HEK293ET cells expressing the indicated plasmids were stimulated (PMA, 30 min). (F) 14-3-3 binding induces TACE cell surface proteolytic activity. HEK293ET cells overexpressing the indicated plasmids were left untreated. TACE proteolytic activity on the cell surface was determined by measuring its ability to cleave a fluorogenic substrate. RFU, relative fluorescent units. Student’s t test. (G) TACE maturation is not affected by 14-3-3 binding. Lysates from iRhom1/2 DKO MEFs stably expressing the indicated plasmids were ConA enriched and immunoblotted for TACE. Data are presented as mean ± SEM. See also Figure S1.

Figure 6

iRhom2 Phosphorylation Regulates TACE Beyond the ER (A) TACE maturation is not affected by iRhom2 phosphorylation in iRhom1/2 DKO MEFs stably expressing the indicated iRhom2 mutants. (B) Densitometric scans illustrating the proportion of mature TACE in iRhom DKO MEFs expressing the indicated iRhom2 mutants. (C) Flow cytometric detection of overexpressed cell surface iRhom2-HA in non-permeabilized RAW264.7 cells. Student’s t test. EV, empty vector. (D) Cell surface biotinylation assays detecting endogenous cell surface iRhom2 in RAW264.7 macrophages. (E) Consistent with cell surface localization, endogenous iRhom2 degradation upon treatment with a protein synthesis inhibitor (CHX, 50 μg/mL) is rescued by the inhibitor of dynamin-dependent endocytosis dynasore (80 μM, co-treatment with CHX). (F) Densitometry of (D). Data are presented as mean ± SEM. See also Figure S2 and Table S2.

Figure 7

iRhom2 Phosphorylation Regulates TACE Proteolytic Activity at the Cell Surface Independently of Substrate Delivery (A) iRhom2-HA WT overexpression in HEK293ET cells, but not iRhom2-HA Dead, increases the ability of TACE to cleave a fluorogenic substrate added to the culture media of PMA-stimulated cells. The metalloprotease inhibitor Batimastat (BB94) demonstrates the component of TACE activity conferred by iRhom2 overexpression. AU, arbitrary units. (B) TACE cell surface proteolytic activity expressed as the rate of peptide cleavage (relative fluorescent units [RFU] per minute). Student’s t test. (C) 14-3-3 binding induces dissociation of the iRhom2/TACE complex. HEK293ET cells expressing the indicated plasmids were stimulated (PMA, 30 min). Dissociation of iRhom2/TACE was assessed by α-HA immunoprecipitates. Bottom: densitometric quantification of the interaction between TACE and iRhom2, as a percentage of binding to iRhom2 WT under non-stimulated conditions. (D) Plasmids encoding WT V5 tagged TACE, ADAM10, or the indicated domain swap chimeras of TACE (red) and ADAM10 (blue) were transfected into HEK293ET together with iRhom2-HA (+) or empty vector (−). Binding of TACE, ADAM10, or chimeric constructs to iRhom2-HA was detected using anti-V5 antibody on anti-HA IPs. Data are presented as mean ± SEM. See also Figure S3.