Skip to main content
NCBI home page
The FASEB Journal logoLink to The FASEB Journal
. 2008 Dec;22(12):4281–4295. doi: 10.1096/fj.08-113852

Ectodomain cleavage of the EGF ligands HB-EGF, neuregulin1-β, and TGF-α is specifically triggered by different stimuli and involves different PKC isoenzymes

Andreas Herrlich *,†, Eva Klinman *,‡, Jonathan Fu *,§, Cameron Sadegh *,§,1, Harvey Lodish *,‡,§,2
PMCID: PMC2614613  PMID: 18757500

Abstract

Metalloproteinase cleavage of transmembrane proteins (ectodomain cleavage), including the epidermal growth factor (EGF) ligands heparin-binding EGF-like growth factor (HB-EGF), neuregulin (NRG), and transforming growth factor-alpha (TGF-α), is important in many cellular signaling pathways and is disregulated in many diseases. It is largely unknown how physiological stimuli of ectodomain cleavage—hypertonic stress, phorbol ester, or activation of G-protein-coupled receptors [e.g., by lysophosphatidic acid (LPA)]—are molecularly connected to metalloproteinase activation. To study this question, we developed a fluorescence-activated cell sorting (FACS) -based assay that measures cleavage of EGF ligands in single living cells. EGF ligands expressed in mouse lung epithelial cells are differentially and specifically cleaved depending on the stimulus. Inhibition of protein kinase C (PKC) isoenzymes or metalloproteinase inhibition by batimastat (BB94) showed that different regulatory signals are used by different stimuli and EGF substrates, suggesting differential effects that act on the substrate, the metalloproteinase, or both. For example, hypertonic stress led to strong cleavage of HB-EGF and NRG but only moderate cleavage of TGF-α. HB-EGF, NRG, and TGF-α cleavage was not dependent on PKC, and only HB-EGF and NRG cleavage were inhibited by BB94. In contrast, phorbol 12-myristate-13-acetate (TPA) -induced cleavage of HB-EGF, NRG, and TGF-α was dependent on PKC and sensitive to BB94 inhibition. LPA led to significant cleavage of only NRG and TGF-α and was inhibited by BB94; only LPA-induced NRG cleavage required PKC. Surprisingly, specific inhibition of atypical PKCs zeta and iota [not activated by diacylglycerol (DAG) and calcium] significantly enhanced TPA-induced NRG cleavage. Employed in a high-throughput cloning strategy, our cleavage assay should allow the identification of candidate proteins involved in signal transduction of different extracellular stimuli into ectodomain cleavage.—Herrlich, A., Klinman, E., Fu, J., Sadegh, C., Lodish, H. Ectodomain cleavage of the EGF ligands HB-EGF, neuregulin1-β, and TGF-α is specifically triggered by different stimuli and involves different PKC isoenzymes.

Keywords: osmotic stress, G-protein, EGF receptor

Metalloproteinase cleavage of the extracellular domain (ECD) of type 1 transmembrane proteins (ectodomain cleavage) has been linked to the regulation of many signaling pathways in a number of important diseases, including cardiovascular disease, cancer, and neurodegenerative diseases (1,2,3). Metalloproteinases are widely expressed throughout the body and are involved in a large variety of physiological processes (4). The cleaved substrates are very diverse and include epidermal growth factor (EGF) ligands and some of their receptors, cytokines and their receptors, notch ligands and receptors, and amyloid precursor protein.

A particularly important example of ectodomain cleavage is the involvement of the EGF ligand heparin-binding EGF-like growth factor (HB-EGF) in the development of heart failure. In the heart, cleavage of HB-EGF is mediated by stimulation of the beta-adrenergic G-protein-coupled receptor (GPCR) and leads to cardiac hypertrophy. In mice, metalloproteinase inhibition can reverse this effect (5). Ectodomain cleavage has also been implicated in the biology of many cancers (6, 7). Disregulated signaling via EGF receptors (EGFRs) is a hallmark of many breast cancers, and therapeutic strategies have largely focused on EGFR inhibition. Many breast cancers escape this treatment, and mounting evidence suggests that this may be the result of autocrine loops involving increased production of EGF ligands via metalloproteinase-dependent cleavage of their precursors (6, 8). As evidence, expression of the metalloproteinase ADAM17 and of the EGF ligand transforming growth factor-alpha (TGF-α) is highly correlated with poor prognosis in breast cancer (9). In fact, GPCR ligands in many cells appear to exert their proliferative response via stimulation of metalloproteinase cleavage of HB-EGF and potentially of other EGF ligands, followed by mitogen-activated protein kinase (MAPK) activation (4, 10). Alzheimer’s disease (AD) and the cleavage of amyloid precursor protein (APP) represent another important example of a disease that is driven by protein ectodomain shedding. The ectodomain of APP can be cleaved in 2 different ways: In the amyloidogenic pathway, β-secretase (BACE) -initiated ectodomain cleavage of APP in the Golgi apparatus is followed by γ-secretase cleavage of the intracellular domain (ICD); this leads to the generation of toxic cleavage products, mainly Aβ42. The deposition of Aβ42 into extracellular plaques within the brain correlates with AD progression. In the nonamyloidogenic pathway, APP is cleaved by α-secretase, a cell surface metalloproteinase, at a different site within the ectodomain (which is located within the Aβ sequence). In this case, the cleavage products generated by γ-secretase are nontoxic (11,12,13).

Inhibition of metalloproteinase activity has shown beneficial effects in humans and in several rodent disease models, yet its use in humans has been hampered by side effects. Most of the side effects can be attributed to the lack of specificity of the inhibitors leading to inhibition of several different metalloproteinases at the same time (14). Moreover, some metalloproteinases have tumor suppressive effects, a fact that complicates the therapeutic use of broad spectrum metalloproteinase inhibitors (15). To improve therapeutic approaches to cellular processes involving ectodomain cleavage, we therefore need a detailed understanding of how cleavage of various substrates by metalloproteinases is regulated.

Ectodomain cleavage of many substrates can be activated by hypertonic stress, phorbol ester-mediated protein kinase C (PKC) activation, or activation of certain GPCRs. Little is known about how any of these stimuli are molecularly coupled to ectodomain cleavage of EGF ligands or of other substrates (1).

Recently, we showed that the EGF-ligand neuregulin (NRG)1-β is cleaved by a metalloproteinase in response to hypertonic stress in mouse lung epithelial (MLE) cells. The mature hormone subsequently activates the EGFR dimer HER2/3 in an autocrine fashion leading to MAPK activation followed by enhanced expression of genes encoding water channels (aquaporins) (16). Signaling intermediates upstream of hypertonic stress-induced NRG cleavage are unknown. In COS-7 and TCCsup carcinoma cells, inhibitor data suggest that hypertonic-stress-induced HB-EGF cleavage requires the activation of p38 stress kinase upstream of the metalloproteinase (17).

Phorbol ester [phorbol 12-myristate-13-acetate (TPA)] -mediated ectodomain cleavage of HB-EGF, NRG, and TGF-α is predominantly mediated by activation of A-disintegrin-and-metalloproteinases (ADAM)17 (18,19,20). How the metalloproteinase is activated by TPA is not known; activation possibly involves PKC isoforms, as PKCδ, a PKC isoform belonging to the novel PKC family [activated by diacylglycerol (DAG) and calcium], associates with and phosphorylates the ICD of ADAM9 previous to cleavage of HB-EGF (21).

GPCR-mediated metalloproteinase activation can involve Gαi, Gαq, or Gβγ G-protein subunits depending on the situation studied. Stimulation of the heterologously expressed β-adrenergic receptor in COS-7 cells leads to ectodomain cleavage of HB-EGF via a signal that involves Gβγ subunits released from the Gi-protein and activation of c-Src kinase (22). Lysophophatidic acid (LPA) -mediated shedding of amphiregulin, another EGF ligand, involves ADAM17 and is partially pertussis toxin sensitive, suggesting the involvement of Gi-proteins (23). On the other hand, overexpression of a Gq-inhibitory minigene blocks shedding of HB-EGF mediated by the angiotensin-1 GPCR (AT1 receptor) (24). Signaling pathways leading to activation of a metalloproteinase downstream of the G-protein are not well established and the evidence for involvement of PKC or other kinases is limited. PKC-dependent and Src-mediated metalloproteinase activation appears to link activation of the gonadotropin-releasing hormone receptor and AT1 receptor to HB-EGF cleavage in C9 cells but not in human embryonic kidney (HEK) 293 cells (25, 26). LPA-induced shedding of HB-EGF in Vero-H cells is not dependent on PKC but on the activation of Rac and the Ras/Raf/Mek kinase cascade (27). Another potential signaling intermediate linking GPCRs to metalloproteinase activation is calcium. For example, ADAM10-mediated CD44 cleavage can be stimulated by Ca2+ influx into the cell and association of ADAM10 with calmodulin, or alternatively by ADAM17 in response to TPA-induced PKC and Rac activation (28).

To study regulation of ectodomain cleavage in more detail, we have developed a fluorescence-activated cell sorting (FACS) -based assay that allows quantification of EGF ligand ectodomain cleavage in single living MLE cells. We demonstrate that 3 EGF ligands, HB-EGF, NRG, and TGF-α, expressed in MLE cells, are cleaved to various extents and with various kinetics by osmotic stress, phorbol ester (TPA), or GPCR activation by LPA. More importantly, we show that ectodomain cleavage of the studied EGF ligands is dependent to various degrees on metalloproteinase inhibition and the inhibition of protein kinase C isoforms. Our results suggest that multiple signaling pathways that involve different PKC isoforms regulate ectodomain cleavage; these may act on the level of the metalloproteinase, the substrate, or both. In addition, cleavage appears to be carried out not only by metalloproteinases that are sensitive to hydroxamate-based inhibitors (like batimastat, BB94), but possibly other metalloproteinases and/or enzymes are involved. Employed in a high-throughput cloning strategy, our cleavage assay should allow the identification of proteins involved in intracellular signal transduction of various extracellular stimuli into ectodomain cleavage.

MATERIALS AND METHODS

Antibodies

Polyclonal anti-GFP antibody for Western blot 1:500 dilution (Invitrogen, Carlsbad, CA, USA), monoclonal anti-FLAG antibody M2 for FACS (1:100 dilution) and Western blot (1:1000 dilution; Sigma, St. Louis, MO, USA), monoclonal anti-MYC 9E10 for FACS (1:100 dilution; Covance, Princeton, NJ, USA), monoclonal anti-MYC 9B11 for Western blot (1:500–1:1000 dilution; Cell Signaling, Danvers, MA, USA), monoclonal anti-HA11.1 antibody (1:100 dilution for FACS; Covance), APC-coupled goat anti-mouse antibody (1:100 dilution for FACS; BD Biosciences, San Jose, CA, USA).

Reagents

TPA (Sigma), LPA (Sigma), sorbitol (Sigma), polybrene (Sigma), Fugene 6 (Roche, Indianapolis, IN, USA), RPMI (Life Technologies, Inc.), DMEM (Life Technologies, Inc.), FCS (Life Technologies, Inc., Gaithersburg, MD, USA), propidium iodine (Sigma), BB94 (batimastat; British Biotech, Oxford, UK; provided by Klaus Maskos, Max Plank Institute for Biochemistry, Munich, Germany), bisindolylmaleimide (Calbiochem, La Jolla, CA, USA; #203293), myristoylated PKZ zeta/iota pseudosubstrate inhibitor (Calbiochem #539624).

Cell lines

MLE cells and HEK 293T cells were from American Type Culture Collection (ATCC; Manassas, VA, USA).

Cloning

Epitope-tagged cDNAs for NRG1-β (FLAG tag) were obtained from Dr. Cary Lai (The Scripps Research Institute, La Jolla, CA, USA), the cDNA for HB-EGF (MYC tag) was from Dr. Michael Klagsbrun (Children’s Hospital, Boston, MA, USA) (29) and the TGF-α (HA tag) cDNA was from Dr. Joaquin Arribas (Valle D'Hebron University, Spain) (30). The respective epitope tags were inserted into the respective EGF ligand cDNAs proximal to the EGF signaling domain of the ligand, yet distal to the signal peptide such that the epitope tag was not cleaved off during processing of the pro-EGF ligand. Human GFP was PCR-amplified such that its start codon was eliminated and cloned into the retroviral vector pB (31). Subsequently, cDNAs encoding the ectodomain epitope-tagged versions of NRG (FLAG), HB-EGF (MYC) or TGF-α (HA) were PCR-amplified such that their stop codons were eliminated and then cloned inframe into pB proximal to GFP.

Generation or reporter cell lines

Retroviral EGF-ligand reporter constructs were cotransfected with pCLEco (31) into 293T cells, and resultant retrovirus was used to infect MLE cells at 50% density with 4 μg/ml polybrene.

FACS assay

For the inhibitor experiments, cells were preincubated with either 10 μM BB94 or 1μM bisindolylmaleimide for 30 min or with 10 μM PKC zeta/iota inhibitor for 60 min before treatment. Reporter cells were either control treated or stimulated with 400 mosmol sorbitol (400 mosmol final gradient between extracellular and intracellular space), 1 μM TPA, or 20 μM LPA for times indicated in the individual figures. Cells were washed with cold PBS/3% fetal calf serum (FCS) and subsequently incubated for 1 h at 4°C with the respective anti-epitope primary antibody at 1:100 dilution. After washing 3× with cold PBS/3% FCS, cells were incubated for 1 h at 4°C with anti-mouse APC-coupled secondary antibody at 1:100 dilution. Finally, cells were again washed 3× with cold PBS/3% FCS and then incubated with a PBS-based enzyme-free proprietary cell dissociation solution (S-014-B; Millipore, Billerica, MA, USA) containing 2 μg/ml propidium iodine. FACS analysis was performed with a BD Biosciences Calibur FACS machine using an HTS 96-well robot.

Western blots

Reporter cells were either control treated or stimulated with 400 mosmol sorbitol, 1 μM TPA, or 20 μM LPA. Cells were then washed with cold PBS and lysed on ice with lysis buffer containing 1% Triton. Lysates were harvested, sonicated, and subjected to SDS-PAGE and Western blotting. After transfer to a nitrocellulose membrane and blocking for 1 h at room temperature with 5% milk/Tris buffered saline (TBST; 50 mM Tris-HCL, pH 7.4; 150 mM NaCl with 0.1% Triton), antibodies were incubated overnight at 4°C in 5% milk/TBST. Membranes were then washed 3× with TBST and incubated for 1 h at room temperature with secondary antibody in 5% milk/TBST.

RESULTS

We have developed a FACS-based high-throughput assay that measures cleavage of the EGF ligands HB-EGF, NRG, or TGF-α in living single cells. It detects cleavage of the chosen EGF-ligand in a FACS-based assay using cells stably expressing prohormone ligands tagged at the N terminus in the ECD with one of several epitope tags (MYC, FLAG, or HA). At its cytosol-facing C terminus, the proteins are fused with GFP (Fig. 1A). The extracellular epitope of the transmembrane prohormone ligand can be detected with a fluorochrome-coupled (e.g., “red”) antibody, whereas the ICD GFP-fusion is detectable by green fluorescence. This allows us to track ECD and ICD individually by FACS, Western blot, or immunofluorescence. We inserted the modified EGF ligands into the retroviral vector pB and stably infected them individually into a cleavage-competent cell line, MLE cells. MLE cells were originally isolated from lung tumors of SV40 antigen-positive mice (32). They express various markers of lung epithelial cells, including mucin genes, and also express the water channel aquaporin5 (AQP5) in response to hypertonic stress. Both cellular responses require the activation of the EGF receptor. In the case of hypertonic AQP5 induction, this involves cleavage of NRG (16). MLE cells also express a variety of ADAM metalloproteinases, including ADAM9, 10, 12 and 17, as determined by quantitative PCR (data not shown).

Figure 1.

Figure 1.

Open in a new tab

Design and cellular expression of EGF ligand cleavage reporter constructs in MLE cells. A) The intracellular C termini of NRG, HB-EGF, and TGF-α (epitope-tagged with either FLAG, MYC, or HA-epitopes in their ectodomain as indicated) were fused to GFP. The modified cDNAs were inserted into a retroviral vector (pB; not shown) and expressed in MLE cells. The ECD and ICD (GFP fusion) can be detected separately by FACS as well as by an anti-epitope-tag or anti-GFP Western blot. The ECDs of NRG, HB-EGF, and TGF-α are released after cleavage by an ADAM family metalloproteinase within the juxtamembrane region, as indicated by the arrow. Following metalloproteinase cleavage of the ectodomain, γ-secretase cleaves and releases the ICD. The ICD acts as a separate signaling unit and translocates to the nucleus. EGF domain, signaling domain of the released ectodomain that binds to EGF (HER) receptors; TM, transmembrane domain. B) Anti-FLAG Western blot of FLAG-NRG-GFP overexpressing MLE cells and MLE wild-type (MLE wt.) cells detects only in the overexpressing cells full-length FLAG-NRG-GFP at 140 kDa (open star, top panel) and a protein of ∼70 kDa, (large left arrow, top panel) representing the ectodomain. An unspecific band is visible at 97 kDa (small left arrow, top panel). Anti-GFP Western blot detects the 140 kDa full-length protein (open star, middle panel) and the ICD-GFP fusion (NRG-ICD-GFP), a cleavage product, at 80 kDa (closed star, middle panel). Unspecific bands are visible at 97 and 64 kDa (small left arrows, middle panel). Anti-ERK Western blot as a loading control (bottom panel). C) Detection of endogenous NRG with a C-terminal anti-NRG antibody in MLE wt. cells. Full-length endogenous NRG is visible at ∼100 kDa (open star, top panel) and a major cleavage product representing the C-terminal domain at 50 kDa (closed star, top panel). D) In MYC-HB-EGF-GFP overexpressing MLE cells, anti-MYC antibody detects 2 isoforms at 39 kDa and 1 at 35 kDa (open stars, top panel). On the GFP Western blot, the top 2 isoforms are visualized as 1 band (open stars, middle panel). Free GFP is visible at 28 kDa (small left arrow, middle panel). E) We can detect only the ectodomain of HA-TGF-α-GFP by anti-HA FACS immunostaining (see Fig. 5) but not by anti-HA Western blot. Anti-GFP Western blot detects 2 isoforms around 37–40 kDa (open stars, top panel). A cleavage product representing the TGF-α-ICD-GFP is visible at 35 kDa (closed star, top panel). Free GFP is seen at 28 kDa (left arrow, top panel).

The expression of the doubly tagged HB-EGF, NRG, and TGF-α proteins and their ability to be cleaved in response to hypertonic osmotic stress (400 mosmol sorbitol = 400 mosmol final gradient between extracellular and intracellular space), phorbol ester (1 μM TPA), or G-protein stimulus (20 μM LPA) were verified both by Western blotting and FACS. Figure 1B shows the expression of full-length FLAG-NRG-GFP in MLE cells compared with MLE wild-type cells not expressing the construct. Anti-FLAG Western blot detects 2 major bands. Full-length FLAG-NRG-GFP is detected at ∼140 kDa (Fig. 1B, top panel, band at open star). As determined by its size, a smaller band at ∼70 kDa (Fig. 1B, top panel, band at large left arrow) likely represents the cleaved ectodomain that has been internalized after binding to the EGF receptor. Anti-GFP Western blot detects full-length FLAG-NRG-GFP at 140 kDa (Fig. 1B, middle panel, band at open star) and the cleaved NRG-ICD-GFP at ∼80 kDa (Fig. 1B, middle panel, band at closed star). Unspecific bands are detected in overexpressing cells and control cells at 97 kDa (anti-FLAG) and 97 and 64 kDa (anti-GFP; Fig. 1B, top and middle panels, small left arrows). The bottom panels always show a Western blot with anti-ERK antibody as a loading control. Figure 1C shows the expression of endogenous NRG. The C-terminal anti-NRG antibody detects a full-length band at ∼100 kDa and a cleavage product representing the C-terminal tail at 50 kDa.

Figure 1D shows the expression of MYC-HB-EGF-GFP in MLE cells compared with control cells detected by anti-MYC and anti-GFP Western blot. Two MYC-HB-EGF-GFP isoforms are detected at 39 kDa by the anti-MYC and the anti-GFP antibody. A third isoform is detected by both antibodies at 35 kDa (Fig. 1D, top and middle panels, bands at open stars). These isoforms have been reported previously and are thought to result from differences in glycosylation (33). The top 2 isoforms form 1 large band at 39 kDa in the anti-GFP Western blot (Fig. 1D, middle panel, bands at open stars). The anti-GFP Western blot also detects free GFP at 28 kDa, a result of spontaneous cleavage of GFP from the fusion construct that we can also detect for HA-TGF-α-GFP (see below; Fig. 1D, E, left arrows).

Although we were able to detect background proteins expressed in MLE cells with various anti-HA antibodies, we were unable to detect the ectodomain of HA-TGF-α-GFP by anti-HA Western blot using any antibody. Expression of the HA-TGF-α-GFP protein, however, was detectable by anti-HA stain using FACS (Fig. 5A) and by anti-GFP Western blot (Fig. 1E). Anti-GFP (top panel) Western blot detects 2 isoforms around 39 kDa (open stars) and a third species that is visible at 35 kDa (closed star; Fig. 1E, top panel). This third species could represent a full-length version of the protein or a cleavage product. The latter explanation is more likely because this band increases in strength with induced cleavage (Fig. 5H–J, top panel, closed star). Free GFP is detected at 28 kDa (Fig. 1E, top panel, left arrow).

Figure 2.

Figure 2.

Open in a new tab

FACS analysis detects cleavage of FLAG-NRG-GFP in uncloned cells: MLE cells infected with FLAG-NRG-GFP were preincubated for 30 min with dimethyl sulfoxide (A–D) or the metalloproteinase inhibitor BB94 (10 μM) (E–H). This was followed by stimulation for the indicated time points with either control medium (control; A, E), 400 mosmol sorbitol (B, F), 1 μM phorbol ester (TPA; C, G), or 20 μM LPA (D, H). Cells were then stained with anti-FLAG antibody (vertical scale) to assess presence of the neuregulin ectodomain (secondary antibody coupled to APC, red). GFP fluorescence (abscissa) was used to detect presence of the NRG-ICD (green). Osmotic stress and TPA lead to strong cleavage of the ECD (loss of APC staining; B, C), whereas LPA was much less effective (D). The mean green GFP fluorescence remained largely unchanged. Cleavage induced by any of the stimuli was blocked by metalloproteinase inhibition with BB94 (E–H).

Figure 3.

Figure 3.

Open in a new tab

FACS and Western blot analysis of FLAG-NRG-GFP cleavage in clonal MLE cells: MLE cells expressing FLAG-NRG-GFP were either control treated or stimulated with 400 mosmol sorbitol, 1 μM TPA, or 20 μM LPA for the indicated time points. Cells were then either stained for FACS analysis (A–G; 45 min treatment for A–D) or lysed for Western blot analysis (H–J). For FACS, samples were measured in triplicate. Cells were stained with anti-FLAG ectodomain tag antibody (red), and GFP fluorescence (abscissa) was used to detect presence of the NRG ICD (green; A–G). Western blots were performed with either anti-FLAG (H–J, top panels) or anti-GFP antibodies (H–J, middle panels). Anti-ERK stains were performed as a loading control (H–J, bottom panels). FACS and Western blots show that hypertonic stress and phorbol ester induce strong cleavage of FLAG-NRG-GFP (B, C, E, F; also see bands at open stars in top and middle panels of H, I). In comparison, LPA induces only moderate cleavage (D, G; also see bands at open stars, top and middle panels, J). A major cleavage product (80 kDa; closed star) and a minor cleavage product running just above (small right arrow) are visible only on GFP Western blots (H–J, middle panels; compare bands at closed stars and small right arrow). Breakdown of this major cleavage product is significantly retarded following hypertonic stress cleavage of NRG compared with TPA and LPA.

Figure 4.

Figure 4.

Open in a new tab

FACS and Western blot analysis of MYC-HB-EGF-GFP cleavage in clonal MLE cells: MLE cells expressing MYC-HB-EGF-GFP were either control treated or stimulated with 400 mosmol sorbitol, 1 μM TPA, or 20 μM LPA for the indicated time points. Cells were then either stained for FACS analysis (A–G; 45 min treatment for A–D) or lysed for Western blot analysis (H–J). For FACS, samples were measured in triplicate. Cells were stained with anti-MYC ectodomain tag antibody (red), and GFP fluorescence (abscissa) was used to detect presence of the HB-EGF ICD (green; A–G). Western blots were performed with either anti-MYC (H–J, top panels) or anti-GFP antibodies (H–J, middle panels). Anti-ERK stains were performed as a loading control (H–J, bottom panels). FACS and Western blot show that hypertonic stress and phorbol ester induce very strong cleavage of MYC-HB-EGF-GFP. Two MYC-HB-EGF-GFP isoforms (different glycosylation) are seen at 39 kDa and 1 at 35 kDa (open stars). At the exposure chosen to allow visualization of HB-EGF cleavage over time, these isoforms are difficult to distinguish at time 0 min (B, C, E, F; see bands at open stars, top and middle panels, H, I). In comparison, LPA induces little cleavage of MYC-HB-EF-GFP overall (D, G; also see J, top and middle panels, bands at open stars). Free GFP is seen at 28 kDa (H, I, middle panels, small left arrow). The small right arrow denotes either a cleavage product or possibly newly synthesized full-length protein (see small right arrow in middle panels of H, I).

Figure 5.

Figure 5.

Open in a new tab

FACS and Western blot analysis of HA-TGF-α-GFP cleavage in clonal MLE cells: MLE cells expressing HA-TGF-α-GFP were either control treated or stimulated with 400 mosmol sorbitol, 1 μM TPA, or 20 μM LPA for the indicated time points. Cells were then either stained for FACS analysis (A–G; 45 min treatment for panels A–D) or lysed for Western blot analysis (H–J). For FACS, samples were measured in triplicate. Cells were stained with anti-HA ectodomain tag antibody (red), and GFP fluorescence (abscissa) was used to detect presence of the TGF-α ICD (green; A–G). Western blots were performed with anti-GFP antibodies (H–J, top panels), and anti-ERK stains were performed as a loading control (H–J, bottom panels). We are unable to detect the HA-TGF-α ectodomain by Western blot. FACS and Western blot show that HA-TGF-α-GFP is cleaved only moderately (compared with NRG and HB-EGF) by hypertonic stress, phorbol ester, or LPA. Two full-length isoforms are detected around 37–40 kDa (different glycosylations; see bands at open stars, top panels, H–J). Band at 35 kDa represents the ICD of cleaved HA-TGF-α-GFP, which increases as cleavage proceeds (see bands at closed star, top panel, H–J). Free GFP is visible at 28 kDa (small left arrows, H–J).

All 3 reporter ligands undergo cleavage in response to hypertonic osmotic stress, phorbol ester, and the G-protein stimulus LPA. This is readily quantifiable by studying cell populations expressing reporter ligands by FACS, in which the geometric mean fluorescence of the ECD epitope stain (red) and the green fluorescence of the ICD can be monitored separately. Figure 2 shows a representative FACS experiment using uncloned MLE cells expressing FLAG-NRG-GFP. As Fig. 2A shows, the population of MLE cells infected with FLAG-NRG-GFP stably expresses the fusion protein over the expected ∼100-fold range, exemplified by GFP fluorescence. At a cell-by-cell level, expression of the FLAG-NRG-GFP fusion protein on the surface, measured by anti-FLAG antibody, is proportional to that of GFP; that is, all cells fall on a diagonal with a slope of 1.

Pretreatment with dimethyl sulfoxide or the metalloproteinase inhibitor BB94 (10 μM) has little effect on this profile. This suggests that there is only a small amount of basal cleavage in control treated cells (Fig. 2A, E). However, after hypertonic osmotic stress, FLAG staining is reduced 5-fold, whereas GFP staining is mostly unchanged (Fig. 2B). Thus, osmotic stress causes each cell to cleave ∼80% of the FLAG-NRG-GFP fusion protein; this cleavage is blocked by simultaneous treatment with BB94 (Fig. 2F). Cleavage in response to TPA (Fig. 2C) was equally effective as osmotic stress, whereas the response to maximal stimulation with LPA was much less pronounced (Fig. 2D). The effect of both stimuli on FLAG-NRG-GFP cleavage was also blocked by BB94 (Fig. 2G, H). Similar FACS experiments were carried out for MYC-HB-EGF-GFP and for HA-TGF-α-GFP. Depending on the stimulus tested, metalloproteinase inhibition with BB94 also inhibited cleavage of these reporter constructs in most cases (details below). These results taken together suggested that there might be differential regulation of ectodomain cleavage of the same protein in the same cellular context, depending on the cleavage stimulus used.

To better compare cleavage of the 3 reporter genes in the same cell type, we isolated clonal populations of MLE cells infected with FLAG-NRG-GFP, MYC-HB-EGF-GFP, and HA-TGF-α-GFP, which express approximately the same amount of reporter protein, as judged by GFP fluorescence. Then we analyzed cleavage of our 3 reporter constructs by both FACS and Western blot. We carried out time course experiments using all 3 stimuli at time points from 0 to 45 min. As detailed below, results obtained by FACS and Western blot were essentially comparable. Small differences can be explained by differences in sensitivity of detection inherent to both methods.

In the case of FLAG-NRG-GFP, the FACS time course experiments revealed similar results to the experiment shown in Fig. 2. Hypertonic osmotic stress induced ∼80% ectodomain cleavage at 45 min (80% reduction in mean red fluorescence vs. control) whereas GFP fluorescence remained mostly unchanged (Fig. 3A, B, E). Western blot analysis with the anti-FLAG antibody revealed strong cleavage of full-length FLAG-NRG-GFP by 30 min (Fig. 3H, fading 140 kDa band at open star, top panel, lanes 2–4). In control treated cells, anti-GFP Western blot demonstrates the presence of full-length FLAG-NRG-GFP at 140 kDa and a cleavage product representing the ICD of FLAG-NRG-GFP (Fig. 3H, open star and closed star, respectively, first lane of middle panel). When cells are treated with hypertonic stress, the full-length FLAG-NRG-GFP is progressively cleaved and the cleavage product representing the NRG-ICD-GFP increases (Fig. 3H, middle panel, bands at open star and closed star, lanes 2–4). A second, slightly larger cleavage product is seen just above the NRG-ICD-GFP (Fig. 3H, small arrow in middle panel, lanes 2–4). The anti-GFP Western blot suggests that the overall GFP fluorescence as measured by FACS is a composite of the full-length FLAG-NRG-GFP and the cleaved NRG-ICD-GFP. In the case of hypertonic stress, the NRG-ICD-GFP at 80 kDa is relatively resistant to further cleavage. This explains the largely maintained GFP fluorescence as measured by FACS (Fig. 3A, B, E).

TPA cleavage of FLAG-NRG-GFP shows a similar overall extent of cleavage of ∼80% of the ectodomain, compared with hypertonic stress. The rate of cleavage, however, was different, with TPA being more efficient at earlier time points. In addition, reduction of GFP fluorescence was more extensive over time, suggesting stronger cleavage overall and subsequent breakdown of the NRG-ICD-GFP (Fig. 3A–C; compare E, F). This is corroborated by Western blot studies. Compared with hypertonic stress, TPA-induced ectodomain cleavage of FLAG-NRG-GFP is stronger at earlier time points as detected by anti-FLAG and anti-GFP Western blots. Full-length FLAG-NRG-GFP is completely cleaved by 15 min (Fig. 3I, bands at open stars, top and middle panels, lanes 1–4), and there is complete breakdown of the NRG-ICD-GFP (Fig. 3I, band at closed star, middle panel, lanes 1–4). The second less abundant cleavage product increases significantly over time (Fig. 3I, small arrow, middle panel, lanes 2–4).

Finally, LPA stimulation resulted in only ∼50% ectodomain cleavage at 45 min compared with control, and the rate of cleavage was slower overall compared with the other stimuli (Fig. 3A, D; compare E–G). GFP fluorescence was also reduced over time (Fig. 3G). Anti-FLAG and anti-GFP Western blots show moderate cleavage of full-length FLAG-NRG-GFP in response to LPA, following a similar pattern as TPA stimulation (Fig. 3I, J).

In summary, our results for FLAG-NRG-GFP suggest that there is differential regulation of metalloproteinase-dependent ectodomain cleavage and also of further cleavage and breakdown of the ICD depending on the cleavage-stimulus used. FLAG-NRG-GFP in MLE cells is predominantly and strongly cleaved by TPA and hypertonic stress and to a lesser degree, by LPA.

A similar set of experiments was carried out for MYC-HB-EGF-GFP and HA-TGF-α-GFP expressed in clonal cell lines. Hypertonic stress and TPA led to ∼80% to 90% ectodomain cleavage of MYC-HB-EGF-GFP as measured by FACS (Fig. 4A–C; compare E, F), whereas LPA led to very little cleavage of MYC-HB-EGF-GFP overall (Fig. 4A, D; compare G with E, F). As in the case of FLAG-NRG-GFP, TPA stimulation induced ectodomain cleavage at the quickest rate and also led to the strongest reduction in GFP fluorescence. (Fig. 4E–G). These findings were again paralleled by Western blot results. The anti-MYC Western blot revealed strong cleavage of the ectodomain after hypertonic stress and TPA stimulation. As in Fig. 1, anti-MYC antibodies react with full-length isoforms of MYC-HB-EGF-GFP at 39 kDa (seen as 1 band at this exposure) and 35 kDa (Fig. 4H, I, top panel, bands at open stars, lanes 1–4). Anti-GFP Western blots detect cleavage of all 3 isoforms of MYC-HB-EGF-GFP in the case of hypertonic stress and TPA stimulation. The largest isoform at 39 kDa appears to be most cleavage resistant whereas the 2 smaller isoforms undergo complete cleavage (Fig. 4H, I, compare bands at top open star with bands at 2 bottom open stars in middle panel, lanes 1–4). Cleavage products of these smaller 2 isoforms cannot be detected by anti-GFP Western blot. An additional band is detected at 30–45 min that could represent a cleavage product or newly synthesized full-length protein (Fig. 4H, I, middle panels, right arrows). When ectodomain cleavage is quick and extensive, as in the case of TPA, there is also free GFP detectable ∼28 kDa (starting at 15 min). This likely corresponds to GFP cleaved off the HB-EGF-ICD-GFP when further breakdown occurs (Fig. 4I, middle panel, left arrow). Western blots performed after LPA stimulation confirmed the minor amount of ectodomain cleavage seen by FACS (Fig. 4E, F, G, and bands at open stars in J, lanes 1–4). Hence, like FLAG-NRG-GFP, MYC-HB-EGF-GFP is strongly cleaved by hypertonic stress and TPA. Yet, in contrast to FLAG-NRG-GFP, LPA does not lead to extensive cleavage of MYC-HB-EGF-GFP in MLE cells.

In the case of HA-TGF-α-GFP, all stimuli, hypertonic stress, TPA, and LPA, led to only moderate ectodomain cleavage of ∼50% at 45 min as measured by FACS (Fig. 5E–G). Stimulation with all cleavage stimuli led to an increase of GFP fluorescence as measured by FACS at 15 min (compared with control) that decreased back to control levels over time. This observation is most pronounced with LPA-stimulated cleavage of HA-TGF-α-GFP (Fig. 5A–G). At the same time, anti-GFP Western blots show moderate cleavage of the full-length isoforms of HA-TGF-α-GFP at 39 kDa (Fig. 5H–J, top panels, bands at open stars) and the concurrent appearance of TGF-α-ICD-GFP at 35 kDa. (Fig. 5 H–J, top panels, bands at closed star). Free GFP is detected at 28 kDa (Fig. 5 H–J, top panels, bands at left arrow). We have also seen this increase in FACS-measured GFP fluorescence in the absence of cleavage in Ba/f3 cells expressing FLAG-NRG-GFP or MYC-HB-EGF-GFP. (In Ba/f3 cells, only TGF-α undergoes any significant ectodomain cleavage; data not shown). This observation suggests that this effect may be independent of ectodomain cleavage and could be related to a physical change of the cell, leading to changes in its fluorescent properties, after cleavage stimuli are applied. In the case of hypertonic stress, it is possible that cell shrinkage causes a change in the fluorescence property of the cell. Figure 6 summarizes the results on overall cleavage of the 3 reporter ligands in MLE cells as measured by FACS and Western blot experiments.

Figure 6.

Figure 6.

Open in a new tab

Comparison chart of the effects of different stimuli on ectodomain cleavage of NRG, HB-EGF, and TGF-α in MLE cells.

In an attempt to dissect signaling pathways that may act on metalloproteinase, substrate, or both, we used several inhibitors, including the hydroxamate-based metalloproteinase inhibitor BB94, which inhibits ADAMs (Fig. 7, red lines), bisindolylmaleimide 1 (Fig. 7, blue lines), a diacylglycerol mimic that inhibits phorbol ester binding to and activation of PKC isoenzymes that are activated by DAG and calcium (classic and novel PKCs), and a myristoylated peptide pseudosubstrate inhibitor that specifically binds to PKC zeta and iota (Fig. 7, green lines), both atypical PKCs not activated by DAG or calcium.

Figure 7.

Figure 7.

Open in a new tab

Effect of metalloproteinase and protein kinase C inhibition on basal and induced ectodomain cleavage of EGF ligands in MLE cells: MLE cells expressing the indicated EGF reporter ligand were either control treated (black) or preincubated for 30 min with a metalloproteinase inhibitor, BB94 (10 μM, red); a classic and novel PKC inhibitor, bisindolylmaleimide (1 μM, blue); or a PKC zeta/iota inhibitor (10 μM, green). At the indicated time points, cells were assayed by FACS after staining with the respective anti-epitope ectodomain antibody (FLAG, MYC, or HA) for presence of the ectodomain and for GFP fluorescence. Samples were measured in triplicate. Average mean geometric fluorescent values of the triplicates were plotted as red:green ratio (increase = inhibition of cleavage; decrease = enhanced cleavage). Time point 0 was set as 100%; ses were estimated using the linear model and the delta method. HB-EGF showed little to no detectable baseline cleavage as measured by change in red:green ratio (A), whereas NRG showed decreased cleavage over time, a process further enhanced by BB94 and opposed by bisindolylmaleimide (red and blue lines, B). TGF-α was cleaved moderately over time in control treated cells. This cleavage was inhibited by BB94 (red line, C). HB-EGF and NRG cleavage by hypertonic stress was strongly inhibited by the metalloproteinase inhibitor BB94 (red lines, D, E). Bisindolylmaleimide only partially inhibited NRG cleavage, and PKC zeta/iota inhibition had no effect (blue and green lines, E). None of the inhibitors had any significant effect on hypertonic stress-induced TGF-α cleavage (F). TPA-induced cleavage of HB-EGF and NRG was strongly dependent on metalloproteinase inhibition, whereas TGF-α cleavage was only moderately inhibited by BB94 (G–I, compare red lines vs. black control line). In contrast to hypertonic stress, TPA-induced cleavage of all ligands was strongly dependent on PKC activity. Inhibition of classic and novel PKC isoenzymes by bisindolylmaleimide inhibited cleavage of all ligands (compare blue lines vs. black control lines, A–C vs. D–F). Surprisingly, PKC zeta/iota inhibition enhanced TPA-induced cleavage of NRG significantly (green line, H). LPA induced only very little cleavage of HB-EGF that was sensitive to metalloproteinase inhibition by BB94 and was moderately enhanced by PKC inhibition (classic, novel, and atypical PKCs; red, blue, and green lines, J). In contrast, NRG cleavage was strongly inhibited by BB94 and by bisindolylmaleimide inhibition of classic and novel PKCs. Inhibition of PKZ zeta/iota had little effect (red, blue, and green lines, K). LPA-induced TGF-α cleavage, in contrast to cleavage by hypertonic stress and TPA, was strongly inhibited by the metalloproteinase inhibitor BB94, yet PKC inhibition had no significant effect (compare red, blue, and green lines, D–I vs. J–L).

We also carried out similar experiments (not shown) with a variety of other inhibitors, including inhibitors of calmodulin-dependent protein kinase, casein kinase 2, src kinase, and the small G-proteins Rac and Rho, which did not show any measurable effect on ectodomain cleavage of the studied ligands.

To detect effects on basal cleavage, we first incubated MLE cells expressing 1 of the 3 EGF ligands with inhibitors in the absence of any stimulus. As measured by the red:green fluorescence ratio, in the case of HB-EGF there is little to no measurable baseline cleavage activity, and the inhibitors are without effect. (Fig. 7A). In the case of NRG, there is a mild to moderate increase in the fluorescence red:green ratio in control treated cells, suggesting accumulation of the full-length protein over time, either by decrease in basal cleavage or increased production and insertion of protein into the plasma membrane. Because metalloproteinase inhibition increased the red:green ratio further, we suspect that both processes contribute to the increase in cell surface fusion protein in the absence of cleavage. Inhibition of classic and novel PKCs by bisindolylmaleimide had the opposite effect, whereas PKC zeta inhibition had no effect on baseline cleavage of NRG (Fig. 7B). TGF-α showed little basal cleavage, but cleavage was strongly inhibited by BB94. PKC inhibition had no effect (Fig. 7C).

As measured by the red:green fluorescence ratio, hypertonic stress-induced cleavage of HB-EGF and NRG was strongly inhibited by the metalloproteinase inhibitor BB94. Bisindolylmaleimide only partially inhibited NRG cleavage, and PKZ zeta/iota inhibition had no effect (Fig. 7D, E; compare red, blue, and green lines with black control). None of the inhibitors had any significant effect on hypertonic stress-induced TGF-α cleavage (Fig. 7F).

TPA-induced cleavage of HB-EGF and NRG was strongly dependent on metalloproteinase inhibition (Fig. 7G, H; compare red lines and black control) but TGF-α cleavage was only moderately inhibited by the BB94 metalloproteinase inhibitor (Fig. 7I; compare red line to black control). In contrast to hypertonic stress, TPA-induced cleavage of all ligands was strongly dependent on PKC activity. Inhibition of classic and novel PKC isoenzymes by bisindolylmaleimide inhibited cleavage of all ligands. Surprisingly, PKC zeta/iota inhibition enhanced TPA-induced cleavage of NRG significantly (Fig. 7 G–I, compare green and blue lines with black control).

As reported above, LPA induced very little cleavage of HB-EGF (see Fig. 4). Cleavage was sensitive to metalloproteinase inhibition by BB94 and was moderately enhanced by PKC inhibition (classic, novel, and atypical PKCs; Fig. 7J). In contrast, NRG cleavage was strongly inhibited by BB94 and by inhibition of classic and novel PKCs by bisindolylmaleimide. Inhibition of PKZ zeta/iota had little effect. (Fig. 7K; compare red, blue, and green lines with black control). Finally, LPA-induced TGF-α cleavage, in contrast to cleavage by hypertonic stress and TPA, was strongly inhibited by the metalloproteinase inhibitor BB94, yet PKC inhibition had no significant effect (Fig. 7L; compare red, blue, and green lines with black control). Figure 8 summarizes the effect of the inhibitors on cleavage of the respective EGF ligands in response to all stimuli used. Different stimuli acting on different EGF ligands have different dependencies on metalloproteinases and PKC isoforms.

Figure 8.

Figure 8.

Open in a new tab

Comparison chart of inhibitor effect on ectodomain cleavage of NRG, HB-EGF, and TGF-α in MLE cells.

DISCUSSION

Here we report a FACS-based assay that measures cleavage of EGF ligands in single living cells and allows simultaneous tracking of ectodomain and ICD cleavage. As our studies show, it correlates well with established ways of measuring substrate cleavage, like Western blot (Figs. 345), yet it is superior to Western blot in its sensitivity and quantification of ectodomain cleavage. In addition, unlike Western blot, it is accessible to high-throughput screens studying the effect of overexpression or knockdown of genes or the effect of chemical compounds on ectodomain cleavage.

Many previous studies on ectodomain cleavage have been interpreted as supporting predominant, if not exclusive, regulation of ectodomain cleavage on the metalloproteinase level. In contrast, our study supports the conclusion that there is significant regulation also on the substrate level. Our observations point to the involvement of at least several metalloproteinases responding to various stimuli and/or the possibility that ligands are modified in response to stimulation in a way that makes them more or less susceptible to cleavage.

Principally, we envisage 4 possible modes of regulating ectodomain cleavage (1). In a metalloproteinase-centric model, 1 given stimulus would regulate 1 particular metalloproteinase, for example, by modification of its C-terminal tail. Once activated, the metalloproteinase would cleave 1 or several accessible substrates expressed in the same cell (2). In a substrate-centric model, a given stimulus would modify a given substrate, likely on its cytosolic domain, such that it becomes accessible for ectodomain cleavage by 1 or a set of constitutively active metalloproteinases expressed in the same cell (3). Cleavage stimuli could regulate linker-proteins that allow active metalloproteinases and substrates access to each other (4). There could be combinations of any of these possibilities.

The conclusion that ectodomain cleavage is regulated at least in part at the substrate level is supported by our results in 2 ways:

First, EGF-ligand cleavage in the same cell type, MLE cells, is differentially and specifically regulated by different stimuli. Hypertonic stress and TPA induced strong cleavage of NRG and HB-EGF, whereas TGF-α was cleaved only to a moderate degree by either stimulus. HB-EGF was resistant to LPA-induced cleavage, whereas the same stimulus led to moderate cleavage of NRG and TGF-α (Fig. 6). These observations could not easily be explained with the assumption that a single metalloproteinase is responsible for the transduction of a particular stimulus, irrespective of the substrate cleaved.

Second, PKC inhibition has differential effects on basal and induced EGF ligand cleavage in MLE cells, depending on the stimulus and EGF ligand studied. Inhibition of novel and classic PKCs by bisindolylmaleimide enhanced basal cleavage of NRG, whereas basal cleavage of TGF-α was unchanged. Thus, PKC likely phosphorylates cytoplasmic proteins that normally inhibit cleavage of NRG but not TGF-α. HB-EGF showed no basal cleavage, and PKC zeta/iota inhibition had no effect on basal cleavage overall. In contrast, hypertonic stress-induced cleavage of NRG was partially inhibited by bisindolylmaleimide, whereas cleavage of HB-EGF and TGF-α was independent of PKC activity. TPA-induced cleavage, on the other hand, was strongly dependent on PKC activity because inhibition of classic and novel PKCs by bisindolylmaleimide inhibited the cleavage of all ligands by TPA. Yet, only NRG cleavage by TPA was enhanced when PKC zeta/iota was inhibited. Finally, LPA-induced cleavage of NRG was dependent on classic and novel PKCs but not on PKC zeta and iota. In contrast, LPA-induced TGF-α cleavage was independent of PKC activity (Fig. 8). Taking these results together, it is therefore likely that hypertonic stress, phorbol ester, and LPA induced modifications on the substrate or on linker proteins for a particular metalloproteinase because these changes explain differences in extents and rates of cleavage. This hypothesis does not exclude additional regulation on the metalloproteinase level that could augment or inhibit basal and/or induced cleavage activity and contribute to fine-tuning of ectodomain cleavage. In the following sections, we discuss previous results by various groups that further support the notion of significant regulation of ectodomain cleavage on the substrate level vs. the metalloproteinase level.

Regulation of substrate cleavage at the metalloproteinase level

One possibility is that regulation of substrate cleavage occurs via processing and trafficking of metalloproteinases to the cell surface. Of the 2 main classes of metalloproteinases, matrix metalloproteinases (MMPs) are mostly secreted whereas ADAMs are mostly membrane-bound. The ADAM-type metalloproteinases (inhibited by BB94) are predominantly involved in ectodomain shedding (4). Proprotein convertases (PCs) process and activate ADAMs, predominantly during trafficking to the cell surface, and limited evidence suggests that some metalloproteinases may undergo autocatalytic activation. Whether and how processing is regulated is unknown; many studies have shown that cell-surface ADAMs are proteolytically processed and catalytically active (4) and that overexpression of ADAMs and substrates often leads to significant basal cleavage in unstimulated samples [for example, in the study by Horiuchi et al. (34)]. These observations suggest that ADAMs are constitutively active once on the cell surface and may not require further activation steps. Evidence for regulated trafficking from an intracellular compartment to the cell surface has been described for ADAM12 (35, 36) only and appears to be dependent on PKC phosphorylation of its C-terminal tail (37, 38). In summary, these observations suggest that ADAMs are not predominantly regulated by processing or trafficking to the cell surface and in fact are constitutively active when on the cell surface.

However, there could be additional regulation of constitutively active ADAMs once on the cell surface. Many proteins interact with the C-terminal tail of ADAMs and have also been described as signaling intermediates in osmotic stress-, PKC-, or G-protein-dependent signaling pathways. In most cases, however, these potential interactions had no regulatory effect on ADAM activity (4). As examples that are particularly important in light of our observations, in Vero cells the cytoplasmic tail of ADAM9 associates with PKCδ in vivo and can be phosphorylated by PKCδ in vitro. Moreover, overexpression of constitutively active PKCδ induces HB-EGF shedding and overexpression of a kinase-deficient mutant inhibits TPA-induced shedding, suggesting that PKC may be involved in trafficking of ADAM9 and/or its activation. However, overexpression of ADAM9 and of an ADAM9 mutant lacking the PKC binding domain also resulted in shedding of HB-EGF even in the absence of TPA stimulation, supporting alternative explanations that involve regulation of substrate availability to the metalloproteinase by PKCδ. TPA stimulation also did not appear to change processing of ADAM9 (21). Eve-1 and PACSIN3 interact with the tail of ADAM12 and appear to mediate or potentiate phorbol-ester activation of ADAM12 (38, 39).

ADAM activity could be regulated via interactions with tissue inhibitors of metalloproteinases (TIMPs). The 4 known vertebrate TIMPs all inhibit MMPs with high potency, but TIMPs are not exclusively selective for MMPs. TIMP-3 inhibits ADAM17, and ADAM12 and ADAM10 are inhibited by both TIMP-1 and TIMP-2. ADAM inhibition likely occurs via binding of the TIMP N-terminal domain to the extracellular catalytic site, whereas the TIMP C terminus may impart binding specificity to certain ADAMs (4, 40). Whether TIMP interaction with ADAMs also affects substrate selectivity is unknown. Thus, although a variety of proteins bind to the ADAMs, there is little evidence showing that these interactions have an influence on ADAM activity.

Some studies have linked certain stimuli to activation of a particular metalloproteinase. In a recent study, ADAM17 was linked to phorbol ester-induced shedding of TGF-α, whereas ADAM10 appeared to cleave HB-EGF and betacellulin in a calcium-dependent manner. Whether the stimulus used regulated the metalloproteinase or the substrate is unclear. Interestingly, neither the C-terminal tail of ADAM17 nor ADAM10 was required to rescue TPA-stimulated TGF-α shedding or calcium-ionophore-dependent EGF and betacellulin shedding. Moreover, in experiments using chimeras of TGF-α and betacellulin, an EGF ligand not cleaved by ADAM17, the C terminus of TGF-α could be exchanged for the C terminus of betacellulin without affecting TPA-induced ADAM17-catalyzed shedding of the TGF-α-chimera. Whether TGF-α shedding occurred without any C-terminal tail present was not tested (34). These findings suggest that signaling input via modifications or interactions on the metalloproteinase C terminus does not predominantly regulate TPA-induced ectodomain shedding.

Regulation of substrate cleavage at the substrate level

Regulation of induced ectodomain cleavage on the substrate level could occur either by regulation of its availability on the cell surface or by regulation of its availability to metalloproteinase cleavage once on the cell surface. For example, covalent modifications to the C-terminal cytosolic segment of the protein, such as phosphorylation or ubiquitination, or the binding of an adaptor protein, could make a substrate more or less susceptible to ADAM cleavage. Several groups have reported results that support this hypothesis and are in line with our observations regarding differential cleavage and the involvement of PKC isoenzymes. The cytoplasmic domains of NRG, HB-EGF, and TGF-α contain potential phosphorylation sites for tyrosine- and serine/threonine-kinases that could serve as sites for regulatory input. Inhibitors of PKC, p38MAPkinase, and src-kinase indeed inhibit cleavage of certain substrates in some studies (4, 41, 42). BAG-1 interacts with the cytoplasmic tail of pro-HB-EGF and not only regulates cell adhesion and increased cellular resistance to apoptosis but also increases the secretion of soluble HB-EGF (43, 44). In neuronal cells, the NRG C terminus binds LIM kinase, but it is unknown whether this affects metalloproteinase-dependent NRG cleavage (45). Nardilysin, a metalloendopeptidase and linker protein, potentially regulates HB-EGF ectodomain cleavage. In cotransfection assays in COS7 cells, nardilysin binds to ADAM17 and induces HB-EGF cleavage. Elimination of nardilysin endopeptidase activity did not alter this effect, suggesting that its linker function was involved (46). Nardilysin also enhances metalloproteinase cleavage of amyloid precursor protein (47), but whether cleavage of other substrates is also regulated by nardilysin has not been studied. Linker proteins could help position substrate and metalloproteinase in close proximity for cleavage to occur. For instance, γ-secretase cleavage of notch and amyloid precursor proteins is facilitated by nicastrin, a large transmembrane protein that binds to both the C-terminal intracellular tail of the protease and to the short extracellular stalk (representing the remainder of the ECD after metalloproteinase-dependent ectodomain cleavage) of the substrate. Whether similar mechanisms position an uncleaved EGF-family protein close to the metalloproteinase is unknown.

In summary, our results support the conclusion that different cleavage stimuli activate different signaling pathways involving different PKC isoenzymes; these signals likely act on the EGF ligand substrate to regulate ectodomain cleavage but could also act on the metalloproteinase or both the substrate and metalloproteinase. The system we described here allows detection of cleavage of 3 EGF family members in single living cells. Currently, we are using it in several high-throughput screens of cDNA overexpression libraries and shRNA knockdown libraries to identify the missing signal transduction proteins that regulate ectodomain cleavage.

Acknowledgments

We thank Dr. Cary Lai (The Scripps Research Institute, La Jolla, CA, USA), Dr. Michael Klagsbrun (Children’s Hospital, Boston, MA, USA), and Dr. Joaquin Arribas (Valle D'Hebron University, Spain) for the provision of cDNA constructs. We thank Klaus Maskos (Max Plank Institute for Biochemistry, Martinsried, Germany) for provision of BB94. A.H.’s research was supported by funds from the Fidelity Research Foundation and a K99 National Institutes of Health grant.

References

  1. Sternlicht M D, Werb Z. How matrix metalloproteinases regulate cell behavior. Annu Rev Cell Dev Biol. 2001;17:463–516. doi: 10.1146/annurev.cellbio.17.1.463. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Vu T H, Werb Z. Matrix metalloproteinases: effectors of development and normal physiology. Genes Dev. 2000;14:2123–2133. doi: 10.1101/gad.815400. [DOI] [PubMed] [Google Scholar]
  3. Egeblad M, Werb Z. New functions for the matrix metalloproteinases in cancer progression. Nat Rev Cancer. 2002;2:161–174. doi: 10.1038/nrc745. [DOI] [PubMed] [Google Scholar]
  4. Seals D F, Courtneidge S A. The ADAMs family of metalloproteases: multidomain proteins with multiple functions. Genes Dev. 2003;17:7–30. doi: 10.1101/gad.1039703. [DOI] [PubMed] [Google Scholar]
  5. Asakura M, Kitakaze M, Takashima S, Liao Y, Ishikura F, Yoshinaka T, Ohmoto H, Node K, Yoshino K, Ishiguro H, Asanuma H, Sanada S, Matsumura Y, Takeda H, Beppu S, Tada M, Hori M, Higashiyama S. Cardiac hypertrophy is inhibited by antagonism of ADAM12 processing of HB-EGF: metalloproteinase inhibitors as a new therapy. Nat Med. 2002;8:35–40. doi: 10.1038/nm0102-35. [DOI] [PubMed] [Google Scholar]
  6. Arribas J, Bech-Serra J J, Santiago-Josefat B. ADAMs, cell migration and cancer. Cancer Metastasis Rev. 2006;25:57–68. doi: 10.1007/s10555-006-7889-6. [DOI] [PubMed] [Google Scholar]
  7. Mochizuki S, Okada Y. ADAMs in cancer cell proliferation and progression. Cancer Sci. 2007;98:621–628. doi: 10.1111/j.1349-7006.2007.00434.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Hynes N E, Lane H A. ERBB receptors and cancer: the complexity of targeted inhibitors. Nat Rev Cancer. 2005;5:341–354. doi: 10.1038/nrc1609. [DOI] [PubMed] [Google Scholar]
  9. Kenny P A, Bissell M J. Targeting TACE-dependent EGFR ligand shedding in breast cancer. J Clin Invest. 2007;117:337–345. doi: 10.1172/JCI29518. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Fischer O M, Hart S, Gschwind A, Ullrich A. EGFR signal transactivation in cancer cells. Biochem Soc Trans. 2003;31:1203–1208. doi: 10.1042/bst0311203. [DOI] [PubMed] [Google Scholar]
  11. Lammich S, Kojro E, Postina R, Gilbert S, Pfeiffer R, Jasionowski M, Haass C, Fahrenholz F. Constitutive and regulated α-secretase cleavage of Alzheimer’s amyloid precursor protein by a disintegrin metalloprotease. Proc Natl Acad Sci U S A. 1999;96:3922–3927. doi: 10.1073/pnas.96.7.3922. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Citron M. Beta-secretase inhibition for the treatment of Alzheimer’s disease–promise and challenge. Trends Pharmacol Sci. 2004;25:92–97. doi: 10.1016/j.tips.2003.12.004. [DOI] [PubMed] [Google Scholar]
  13. Postina R, Schroeder A, Dewachter I, Bohl J, Schmitt U, Kojro E, Prinzen C, Endres K, Hiemke C, Blessing M, Flamez P, Dequenne A, Godaux E, van Leuven F, Fahrenholz F. A disintegrin-metalloproteinase prevents amyloid plaque formation and hippocampal defects in an Alzheimer disease mouse model. J Clin Invest. 2004;113:1456–1464. doi: 10.1172/JCI20864. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Peterson J T. Matrix metalloproteinase inhibitor development and the remodeling of drug discovery. Heart Fail Rev. 2004;9:63–79. doi: 10.1023/B:HREV.0000011395.11179.af. [DOI] [PubMed] [Google Scholar]
  15. Lopez-Otin C, Matrisian L M. Emerging roles of proteases in tumour suppression. Nat Rev Cancer. 2007;7:800–808. doi: 10.1038/nrc2228. [DOI] [PubMed] [Google Scholar]
  16. Herrlich A, Leitch V, King L S. Role of proneuregulin 1 cleavage and human epidermal growth factor receptor activation in hypertonic aquaporin induction. Proc Natl Acad Sci U S A. 2004;101:15799–15804. doi: 10.1073/pnas.0406853101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Fischer O M, Hart S, Gschwind A, Prenzel N, Ullrich A. Oxidative and osmotic stress signaling in tumor cells is mediated by ADAM proteases and heparin-binding epidermal growth factor. Mol Cell Biol. 2004;24:5172–5183. doi: 10.1128/MCB.24.12.5172-5183.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Peschon J J, Slack J L, Reddy P, Stocking K L, Sunnarborg S W, Lee D C, Russell W E, Castner B J, Johnson R S, Fitzner J N, Boyce R W, Nelson N, Kozlosky C J, Wolfson M F, Rauch C T, Cerretti D P, Paxton R J, March C J, Black R A. An essential role for ectodomain shedding in mammalian development. Science. 1998;282:1281–1284. doi: 10.1126/science.282.5392.1281. [DOI] [PubMed] [Google Scholar]
  19. Montero J C, Yuste L, Diaz-Rodriguez E, Esparis-Ogando A, Pandiella A. Differential shedding of transmembrane neuregulin isoforms by the tumor necrosis factor-α-converting enzyme. Mol Cell Neurosci. 2000;16:631–648. doi: 10.1006/mcne.2000.0896. [DOI] [PubMed] [Google Scholar]
  20. Sahin U, Weskamp G, Kelly K, Zhou H M, Higashiyama S, Peschon J, Hartmann D, Saftig P, Blobel C P. Distinct roles for ADAM10 and ADAM17 in ectodomain shedding of six EGFR ligands. J Cell Biol. 2004;164:769–779. doi: 10.1083/jcb.200307137. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Izumi Y, Hirata M, Hasuwa H, Iwamoto R, Umata T, Miyado K, Tamai Y, Kurisaki T, Sehara-Fujisawa A, Ohno S, Mekada E. A metalloprotease-disintegrin, MDC9/meltrin-γ/ADAM9 and PKCdelta are involved in TPA-induced ectodomain shedding of membrane-anchored heparin-binding EGF-like growth factor. EMBO J. 1998;17:7260–7272. doi: 10.1093/emboj/17.24.7260. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Pierce K L, Tohgo A, Ahn S, Field M E, Luttrell L M, Lefkowitz R J. Epidermal growth factor (EGF) receptor-dependent ERK activation by G protein-coupled receptors: a co-culture system for identifying intermediates upstream and downstream of heparin-binding EGF shedding. J Biol Chem. 2001;276:23155–23160. doi: 10.1074/jbc.M101303200. [DOI] [PubMed] [Google Scholar]
  23. Gschwind A, Hart S, Fischer O M, Ullrich A. TACE cleavage of proamphiregulin regulates GPCR-induced proliferation and motility of cancer cells. EMBO J. 2003;22:2411–2421. doi: 10.1093/emboj/cdg231. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Mifune M, Ohtsu H, Suzuki H, Nakashima H, Brailoiu E, Dun N J, Frank G D, Inagami T, Higashiyama S, Thomas W G, Eckhart A D, Dempsey P J, Eguchi S. G protein coupling and second messenger generation are indispensable for metalloprotease-dependent, heparin-binding epidermal growth factor shedding through angiotensin II type-1 receptor. J Biol Chem. 2005;280:26592–26599. doi: 10.1074/jbc.M502906200. [DOI] [PubMed] [Google Scholar]
  25. Shah B H, Farshori M P, Catt K J. Neuropeptide-induced transactivation of a neuronal epidermal growth factor receptor is mediated by metalloprotease-dependent formation of heparin-binding epidermal growth factor. J Biol Chem. 2004;279:414–420. doi: 10.1074/jbc.M309083200. [DOI] [PubMed] [Google Scholar]
  26. Shah B H, Baukal A J, Shah F B, Catt K J. Mechanisms of extracellularly regulated kinases 1/2 activation in adrenal glomerulosa cells by lysophosphatidic acid and epidermal growth factor. Mol Endocrinol. 2005;19:2535–2548. doi: 10.1210/me.2005-0082. [DOI] [PubMed] [Google Scholar]
  27. Umata T, Hirata M, Takahashi T, Ryu F, Shida S, Takahashi Y, Tsuneoka M, Miura Y, Masuda M, Horiguchi Y, Mekada E. A dual signaling cascade that regulates the ectodomain shedding of heparin-binding epidermal growth factor-like growth factor. J Biol Chem. 2001;276:30475–30482. doi: 10.1074/jbc.M103673200. [DOI] [PubMed] [Google Scholar]
  28. Nagano O, Murakami D, Hartmann D, De Strooper B, Saftig P, Iwatsubo T, Nakajima M, Shinohara M, Saya H. Cell-matrix interaction via CD44 is independently regulated by different metalloproteinases activated in response to extracellular Ca(2+) influx and PKC activation. J Cell Biol. 2004;165:893–902. doi: 10.1083/jcb.200310024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Gechtman Z, Alonso J L, Raab G, Ingber D E, Klagsbrun M. The shedding of membrane-anchored heparin-binding epidermal-like growth factor is regulated by the Raf/mitogen-activated protein kinase cascade and by cell adhesion and spreading. J Biol Chem. 1999;274:28828–28835. doi: 10.1074/jbc.274.40.28828. [DOI] [PubMed] [Google Scholar]
  30. Borrell-Pages M, Rojo F, Albanell J, Baselga J, Arribas J. TACE is required for the activation of the EGFR by TGF-α in tumors. EMBO J. 2003;22:1114–1124. doi: 10.1093/emboj/cdg111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Bogan J S, Hendon N, McKee A E, Tsao T S, Lodish H F. Functional cloning of TUG as a regulator of GLUT4 glucose transporter trafficking. Nature. 2003;425:727–733. doi: 10.1038/nature01989. [DOI] [PubMed] [Google Scholar]
  32. Wikenheiser K A, Vorbroker D K, Rice W R, Clark J C, Bachurski C J, Oie H K, Whitsett J A. Production of immortalized distal respiratory epithelial cell lines from surfactant protein C/simian virus 40 large tumor antigen transgenic mice. Proc Natl Acad Sci U S A. 1993;90:11029–11033. doi: 10.1073/pnas.90.23.11029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Takenobu H, Yamazaki A, Hirata M, Umata T, Mekada E. The stress- and inflammatory cytokine-induced ectodomain shedding of heparin-binding epidermal growth factor-like growth factor is mediated by p38 MAPK, distinct from the 12-O-tetradecanoylphorbol-13-acetate- and lysophosphatidic acid-induced signaling cascades. J Biol Chem. 2003;278:17255–17262. doi: 10.1074/jbc.M211835200. [DOI] [PubMed] [Google Scholar]
  34. Horiuchi K, Le Gall S, Schulte M, Yamaguchi T, Reiss K, Murphy G, Toyama Y, Hartmann D, Saftig P, Blobel C P. Substrate selectivity of epidermal growth factor-receptor ligand sheddases and their regulation by phorbol esters and calcium influx. Mol Biol Cell. 2007;18:176–188. doi: 10.1091/mbc.E06-01-0014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Hougaard S, Loechel F, Xu X, Tajima R, Albrechtsen R, Wewer U M. Trafficking of human ADAM 12-L: retention in the trans-Golgi network. Biochem Biophys Res Commun. 2000;275:261–267. doi: 10.1006/bbrc.2000.3295. [DOI] [PubMed] [Google Scholar]
  36. Kadota N, Suzuki A, Nakagami Y, Izumi T, Endo T. Endogenous meltrin α is ubiquitously expressed and associated with the plasma membrane but exogenous meltrin α is retained in the endoplasmic reticulum. J Biochem (Tokyo) 2000;128:941–949. doi: 10.1093/oxfordjournals.jbchem.a022845. [DOI] [PubMed] [Google Scholar]
  37. Zhang Q, Thomas S M, Lui V W, Xi S, Siegfried J M, Fan H, Smithgall T E, Mills G B, Grandis J R. Phosphorylation of TNF-α converting enzyme by gastrin-releasing peptide induces amphiregulin release and EGF receptor activation. Proc Natl Acad Sci U S A. 2006;103:6901–6906. doi: 10.1073/pnas.0509719103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Mori S, Tanaka M, Nanba D, Nishiwaki E, Ishiguro H, Higashiyama S, Matsuura N. PACSIN3 binds ADAM12/meltrin α and up-regulates ectodomain shedding of heparin-binding epidermal growth factor-like growth factor. J Biol Chem. 2003;278:46029–46034. doi: 10.1074/jbc.M306393200. [DOI] [PubMed] [Google Scholar]
  39. Tanaka M, Nanba D, Mori S, Shiba F, Ishiguro H, Yoshino K, Matsuura N, Higashiyama S. ADAM binding protein Eve-1 is required for ectodomain shedding of epidermal growth factor receptor ligands. J Biol Chem. 2004;279:41950–41959. doi: 10.1074/jbc.M400086200. [DOI] [PubMed] [Google Scholar]
  40. Brew K, Dinakarpandian D, Nagase H. Tissue inhibitors of metalloproteinases: evolution, structure and function. Biochim Biophys Acta. 2000;1477:267–283. doi: 10.1016/s0167-4838(99)00279-4. [DOI] [PubMed] [Google Scholar]
  41. Wang X, Mizushima H, Adachi S, Ohishi M, Iwamoto R, Mekada E. Cytoplasmic domain phosphorylation of heparin-binding EGF-like growth factor. Cell Struct Funct. 2006;31:15–27. doi: 10.1247/csf.31.15. [DOI] [PubMed] [Google Scholar]
  42. Ohtsu H, Dempsey P J, Eguchi S. ADAMs as mediators of EGF receptor transactivation by G protein-coupled receptors. Am J Physiol Cell Physiol. 2006;291:C1–C10. doi: 10.1152/ajpcell.00620.2005. [DOI] [PubMed] [Google Scholar]
  43. Higashiyama S, Nanba D. ADAM-mediated ectodomain shedding of HB-EGF in receptor cross-talk. Biochim Biophys Acta. 2005;1751:110–117. doi: 10.1016/j.bbapap.2004.11.009. [DOI] [PubMed] [Google Scholar]
  44. Lin J, Hutchinson L, Gaston S M, Raab G, Freeman M R. BAG-1 is a novel cytoplasmic binding partner of the membrane form of heparin-binding EGF-like growth factor: a unique role for proHB-EGF in cell survival regulation. J Biol Chem. 2001;276:30127–30132. doi: 10.1074/jbc.M010237200. [DOI] [PubMed] [Google Scholar]
  45. Falls D L. Neuregulins: functions, forms, and signaling strategies. Exp Cell Res. 2003;284:14–30. doi: 10.1016/s0014-4827(02)00102-7. [DOI] [PubMed] [Google Scholar]
  46. Nishi E, Hiraoka Y, Yoshida K, Okawa K, Kita T. Nardilysin enhances ectodomain shedding of heparin-binding epidermal growth factor-like growth factor through activation of tumor necrosis factor-α-converting enzyme. J Biol Chem. 2006;281:31164–31172. doi: 10.1074/jbc.M601316200. [DOI] [PubMed] [Google Scholar]
  47. Hiraoka Y, Ohno M, Yoshida K, Okawa K, Tomimoto H, Kita T, Nishi E. Enhancement of α-secretase cleavage of amyloid precursor protein by a metalloendopeptidase nardilysin. J Neurochem. 2007;102:1595–1605. doi: 10.1111/j.1471-4159.2007.04685.x. [DOI] [PubMed] [Google Scholar]

Articles from The FASEB Journal are provided here courtesy of The Federation of American Societies for Experimental Biology

RESOURCES