Meprin β induces activities of A disintegrin and metalloproteinases 9, 10, and 17 by specific prodomain cleavage

Abstract

Meprin β is a membrane-bound metalloprotease involved in extracellular matrix assembly and inflammatory processes in health and disease. A disintegrin and metalloproteinase (ADAM)10 and ADAM17 are physiologic relevant sheddases of inactive promeprin β, which influences its substrate repertoire and subsequent biologic functions. Proteomic analysis also revealed several ADAMs as putative meprin β substrates. Here, we demonstrate specific N-terminal processing of ADAM9, 10, and 17 by meprin β and identify cleavage sites within their prodomains. Because ADAM prodomains can act as specific inhibitors, we postulate a role for meprin β in the regulation of ADAM activities. Indeed, prodomain cleavage by meprin β caused increased ADAM protease activities, as observed by peptide-based cleavage assays and demonstrated by increased ectodomain shedding activity. Direct interaction of meprin β and ADAM proteases could be shown by immunofluorescence microscopy and immunoprecipitation experiments. As demonstrated by a bacterial activator of meprin β and additional measurement of TNF-α shedding on bone marrow–derived macrophages, meprin β/ADAM protease interactions likely influence inflammatory conditions. Thus, we identified a novel proteolytic pathway of meprin β with ADAM proteases to control protease activities at the cell surface as part of the protease web.—Wichert, R., Scharfenberg, F., Colmorgen, C., Koudelka, T., Schwarz, J., Wetzel, S., Potempa, B., Potempa, J., Bartsch, J. W., Sagi, I., Tholey, A., Saftig, P., Rose-John, S., Becker-Pauly, C. Meprin β induces activities of A disintegrin and metalloproteinases 9, 10, and 17 by specific prodomain cleavage.

Meprin β is a zinc-dependent metalloprotease belonging to the astacins of the metzincin superfamily (1). Known substrates of meprin β indicate functions of the protease in physiologic and pathologic processes (2–5). For example, membrane-bound meprin β cleaves the amyloid precursor protein (APP) at the β-secretase cleavage site, thereby generating neurotoxic amyloid-β species (6). In its soluble shed form, the enzyme is important for mucus detachment in the small intestine, preventing bacterial overgrowth and invasion (7). We recently showed that ectodomain shedding of meprin β is mediated by A disintegrin and metalloproteinase (ADAM)10 and 17. Importantly, this process is restricted to the inactive proform of meprin β because meprin β devoid of its propeptide cannot be shed by ADAM10 and 17 (8).

Employing mass spectrometry (MS)–based proteomic approaches, we previously identified more than 100 putative substrates of meprin β and additionally observed a striking cleavage specificity with a preference for negatively charged amino acids (9, 10). Besides growth factors, cytokines and adhesion molecules as well as several ADAM proteases were identified among those substrates, namely ADAM9/10/17 and ADAM-TS1 (9). Interestingly, most predicted cleavage sites were located within the ADAM prodomains. For several ADAMs, it has been shown that their prodomains require cleavage by proprotein convertases during the secretory pathway to gain catalytic activity (11, 12). However, isolated but noncovalently bound prodomains can act as specific inhibitors for respective ADAM proteases (13–15). Previously, we demonstrated that cleavage of truncated recombinant ADAM10 by meprin β resulted in increased proteolytic activity in vitro (9).

ADAM proteases are multidomain proteins and were identified as major ectodomain sheddases, which release biologically active factors from membrane-bound proteins (16). Several post-translational modifications and interactions with regulatory molecules have been shown to influence ADAM activity (17–20). The probably best-characterized members of the ADAM family are ADAM10 and ADAM17. ADAM10 was identified as the main α-secretase of APP, thus preventing neurotoxic amyloid-β formation (21). ADAM17 is also known as TNF-α converting enzyme, indicating its contribution to inflammation (22). Both ADAM10- and ADAM17-deficient mice are embryonically lethal, demonstrating their important biologic functions in developmental processes (23, 24). Another member of this group, namely, ADAM9, does not have such a severe impact on development and survival. However, it has been identified as sheddase for ADAM10, thereby regulating ADAM10 activity at the cell surface (25).

Meprin β and certain ADAMs have several common substrates, such as APP (26) or the IL-6 receptor (4). Furthermore, these proteases are part of a complex protease web that regulates their localization and back-and-forth activity. Here, we investigated the cleavage of ADAM9, 10, and 17 by meprin β and its impact on ADAM activity. We validated several predicted meprin β cleavage sites and revealed that processing occurs within ADAM prodomains upstream of the reported furin cleavage sites, which resulted in increased protease activity. Thus, we suggest a general mechanism of sequential ADAM processing leading to further enzyme activation, which is consistent with previous observations in prodomain cleavage of ADAM proteases (27).

MATERIALS AND METHODS

Chemicals

All chemicals were of analytical grade and obtained from Merck (Darmstadt, Germany), MilliporeSigma (Burlington, MA, USA), Thermo Fisher Scientific (Waltham, MA, USA), or Carl Roth (Karlsruhe, Germany) if not stated otherwise. Primers were synthesized by MilliporeSigma.

Expression and purification of recombinant proteins

Recombinant meprin β and ADAM10/17 pro- and catalytic domain were expressed as previously described (9, 28, 29). For recombinant ADAM9, a truncated version was cloned lacking regions C-terminal of the ectodomain, and in which the signal peptide was exchanged by the corresponding meprin β sequence using the following primers:

• ADAM9 sense: 5′-ATTTGCGGCCGCATGGATTTATGGAATCTGTCTTGGTTTCTGTTCTTGGATGCTCTTCTCGTGATTTCTGGCTTGGCAACTCCACATCACCATCACCATCACGGGCGACCAGACTTGGAACA-3′

• ADAM9 antisense: 5′-ACATGCATGCTTAGTCCCTCAGTGCTGTGC-3′

The construct was ligated into pFastBac (Thermo Fisher Scientific) and correct cDNA was verified by sequencing (GATC Biotech, Konstanz, Germany).

Baculovirus amplification and heterologous expression of recombinant proteins was performed according to the Bac-To-Bac Baculovirus Expression System (Thermo Fisher Scientific). Purification was done via N-terminal His-tag, and identity and purity of recombinant ADAM9 was analyzed by Coomassie staining and immunoblotting and confirmed by MS (Institute for Experimental Medicine, Christian-Albrechts-Universität Kiel, Kiel, Germany). The recombinant prodomain of ADAM17 was expressed and purified as previously described (30).

Fluorogenic peptide–based activity assay

To analyze ADAM activity upon meprin β incubation, a fluorogenic peptide–based activity assay was performed. ADAM10 activity was measured using the substrate MOCAc-KPLGLA₂pr(Dnp)AR-NH₂ (Peptide Institute, Osaka, Japan), whereas ADAM9 and ADAM17 activity was analyzed with the TNF-α–based substrate Mca-PLAQAV(Dpa)RSSSR-NH₂ (R&D Systems, Minneapolis, MN, USA). Recombinant ADAM proteases (2 µM ADAM9 ectodomain, 5 µM ADAM10 pro- and catalytic domain or 1 µM ADAM17 pro- and catalytic domain, respectively) were incubated with 15 nM recombinant meprin β for 30 min at 37°C in 20 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES; pH 7.5). For inhibition studies, recombinant ADAM17 was incubated with 250 nM isolated soluble ADAM17 prodomain for 30 min at 37°C. Additionally, recombinant ADAM17 prodomain was preincubated with recombinant meprin β for 1 h at 37°C and subsequently added to ADAM17 enzyme. As control, the activity of 15 nM meprin β toward the respective fluorogenic substrate was also determined. Immediately before measurement, 10 µM of the quenched fluorogenic peptide substrate specific for ADAM10 or ADAM9/17 was added. Fluorescence at λ_em = 405 nm and λ_ex = 320 nm was measured in duplicates every 30 s for a duration of 2 h at 37°C with a spectrophotometer (Infinite F200Pro; Tecan, Männedorf, Switzerland). After deduction of meprin β activity alone, the increase in relative fluorescent units (RFUs) was normalized to the initial point of measurement and proportional to ADAM activity. For bar graph presentation of data, the RFUs after 2 h of analysis were used as far as they were in a linear range. Otherwise the time point of 30 min was used for bar graph representation.

Cell culture, stimulation, and transient transfection

HEK293T and ADAM10^−/−;17^−/− HEK293T cells were maintained in DMEM (Thermo Fisher Scientific), supplemented with 10% (v/v) fetal bovine serum (FBS), 100 U/ml penicillin, and 100 µg/ml streptomycin. Cells were cultured under a humidified atmosphere (5% CO₂) at 37°C.

Transient transfection was performed at 80–90% cell confluence. For 10-cm dishes 6 µg of total cDNA was mixed with polyethylenimine transfection reagent (1:3) in serum-free medium and incubated for 30 min at room temperature. For transfection, plasmid-cDNA for murine ADAM9, murine and human ADAM10, murine ADAM17, human meprin β, alkaline phosphatase (AP)-tagged betacellulin (BTC), AP-tagged epiregulin (EREG) (31), empty vector (pcDNA3.1), or plasmids in different combinations were added together with transfection reagent to the cells. After 5 h, incubation medium was exchanged with fresh medium.

For stimulation experiments, cells were washed with PBS (Thermo Fisher Scientific) 24 h after transfection and subsequently incubated with different concentrations of gingipain R (RgpB) for 5 h in serum-free DMEM, which was purified as previously described (32, 33). Prior to stimulation, RgpB was activated for 15 min at 37°C in buffer (50 mM Tris, 150 mM NaCl, 5 mM CaCl₂, pH 7.6) containing 20 mM l-cysteine. For protein interaction analysis, transfected cells were also maintained in serum-free DMEM for 24 h to avoid meprin β inhibition (34).

Cell lysis, SDS-PAGE, and Western blot analysis

Cells were harvested immediately after stimulation or 48 h after transfection. When analyzing C-terminal fragments, cells were treated with 1 µM γ-secretase inhibitor N-[N-(3,5-difluorophenacetyl)-l-alanyl]-S-phenylglycine t-butyl ester overnight before cell lysis. Cells were harvested with a cell scraper in ice-cold PBS and centrifuged at 1100 g for 5 min at 4°C. Remaining cell pellets were washed with PBS 3 times and afterwards resuspended in lysis buffer [Complete Protease Inhibitor Cocktail (Roche, Basel, Switzerland), 1% (v/v) Triton X-100, PBS, pH 7.4] and incubated for 45 min at 4°C. Cell lysates were centrifuged for 15 min at 15,000 g and 4°C. The protein amount was determined using the Bicinchoninic Acid Protein Assay Kit (Thermo Fisher Scientific) following the manufacturer’s instructions, and the remaining lysates were heated in sample buffer containing DTT for 10 min at 95°C. Protein samples were analyzed by SDS-PAGE (7.5% or 10%, 120 V, 90 min) and blotted onto a PVDF membrane (TANK Blot, 0.8 A, 2 h, 4°C). Depending on the primary antibody, membranes were incubated with 5% dry milk or 3% bovine serum albumin diluted in Tris-buffered saline for 1 h at room temperature. Primary antibodies anti-His-tag antibody (34660; Qiagen, Hilden, Germany); anti-ADAM9-GST-cyto (polyclonal rabbit serum, generated against the cytoplasmic part of human ADAM9); Anti-Strep (34850; Qiagen); anti-ADAM10 prodomain (polyclonal serum, generated against the N terminus); anti-ADAM10 (polyclonal serum, generated against the C terminus); anti-ADAM17 (polyclonal serum, generated against ectodomain); anti-Flag (F1804; MillieporeSigma); anti-actin (A2066; MilliporeSigma); anti-Myc (2276; Cell Signaling Technology, Danvers, MA, USA); anti–TNF-α (3707; Cell Signaling Technology), anti-AP (ab133602; Abcam, Cambridge, MA, USA), or anti-meprin β (polyclonal serum, generated against the peptide CGMIQSSGDSADWQRVSQ) were incubated overnight at 4°C. Afterwards, membranes were incubated in horseradish peroxidase–conjugated secondary antibodies (Thermo Fisher Scientific) diluted in 10% dry milk/Tris-buffered saline for 1 h at room temperature. Chemiluminescence was detected in the Intelligent Dark Box (Fujifilm, Tokyo, Japan) using the Super Signal West Femto Kit (Thermo Fisher Scientific) according to the manufacturer’s instructions.

Immunoprecipitation

N-terminal Strep-tagged ADAMs were expressed in HEK293T cells, proteins were harvested and resuspended in lysis buffer [150 mM NaCl, 50 mM Tris/HCl, 0.1% Triton X-100, Complete protein inhibitor cocktail, EDTA free (Roche); pH 7.4]. Equal amounts of cell lysate were incubated with 50 µl Strep-Tactin Sepharose (IBA, Göttingen, Germany) at 4°C overnight. Afterwards, beads were centrifuged for 3 min at 1000 g and 4°C and then washed with lysis buffer several times. Precipitated protein was incubated with 100 µl elution buffer (100 mM Tris/HCl, 50 mM NaCl, 0.5 mg/ml desthiobiotin; pH 8.0) for 1 h at 4°C. Half of the lysate was incubated with 15 nM recombinant meprin β for 30 min at 37°C. Reaction was stopped by addition of DTT containing sampling buffer and heating at 95°C for 10 min. Protein was analyzed by SDS-PAGE and Western blot or Coomassie stained for subsequent MALDI/MS analysis.

Coimmunoprecipitation of C-terminal Flag-tagged meprin β and C-terminal Myc-tagged ADAM proteases was performed in HEK293T cells. Transfected cells were harvested and lysed in lysis buffer (120 mM NaCl, 50 mM Tris, 0.5% NP-40, Complete Protein Inhibitor Cocktail; pH 7.4). Afterwards, 1 µl anti-myc antibody (2276; Cell Signaling Technology) was incubated with cell lysate overnight at 4°C. The next day, 50 µl Pierce Protein G Agarose (Thermo Fisher Scientific) was added to each sample and incubated for 30 min at 4°C. Samples were centrifuged for 3 min at 1000 g and 4°C and then washed with lysis buffer for several times. Immunoprecipitated proteins were separated from agarose by adding DTT that contains sampling buffer and heating at 60°C for 30 min. Proteins were analyzed by SDS-PAGE and Western blot.

MS

N-terminal Strep-tagged ADAM9 and ADAM10 incubated with or without meprin β were separated by SDS-PAGE as described above. Gel bands corresponding to ADAM9 and ADAM10 were excised, cut into ∼1 mm³ and destained and washed using standard protocols. Samples were reduced using DTT (10 mM) at 56°C for 30 min, alkylated using iodoacetamide (50 mM) at 25°C for 30 min in HEPES buffer (100 mM, pH 7.5) and then washed in HEPES buffer, dehydrated using acetonitrile (ACN), and finally dried using vacuum centrifugation. The α-amino groups (protein N terminus and lysine residues) were subsequently labeled with sodium cyanoborohydride (40 mM) and formaldehyde (40 mM) in HEPES buffer (100 mM, pH 7.5) overnight at 25°C. Labeling solution was removed and the reaction quenched with 100 mM ammonium bicarbonate (ABC) for 1 h and then by the addition of 1% formic acid (FA). The gel bands were washed with ABC, dehydrated with ACN, and then dried and incubated with 100 ng of chymotrypsin at 37°C overnight in 10 mM HEPES buffer, 2 mM CaCl₂ (pH 8.0). Peptides were extracted by subsequent sonication in 1% FA, 50% ACN in 1% FA, and in ACN, respectively. The samples were dried in a SpeedVac and the peptides were resuspended in running buffer (3% ACN, 0.1% TFA) for further measurements with liquid chromatography/MS.

Samples were also labeled by N-terminal dimethylation prior to SDS-PAGE analysis in a second experiment. Briefly, proteins were denaturated with 8 M urea in HEPES buffer (pH 7.5), reduced with DTT (5 mM) then alkylated with iodoacetamide (12.5 mM) and reductively dimethylated (as above). Samples were treated with Laemmli buffer prior to running on SDS-PAGE (12%). Here, samples were treated with trypsin (50 ng) in 20 mM ABC buffer and extracted from the gel as mentioned previously.

Samples were analyzed on a Dionex Ultimate 3000 nano-UHPLC coupled to a Q Exactive Plus MS (Thermo Fisher Scientific). The samples were concentrated and washed for 5 min (Acclaim Pepmap 100 C18, 10 mm × 300 μm, 3 μm, 100 Å; Dionex, Sunnyvale, CA, USA) with 3% ACN/0.1% TFA at a flow rate of 30 μl/min prior to peptide separation using an Acclaim PepMap 100 C18 analytical column (50 cm × 75 μm, 3 μm, 100 Å; Dionex). A flow rate of 300 nl/min using eluent A (0.05% FA) and eluent B (80% ACN/0.04% FA) was used for gradient separation. Spray voltage applied on a metal-coated PicoTip Emitter (10-μm tip size; New Objective, Woburn, MA, USA) was 1.6 kV, with a source temperature of 250°C. Full scan MS spectra were acquired between 300 and 2000 m/z at a resolution of 70,000 at m/z 400. The 10 most intense precursors with charge states >2+ were selected with an isolation window of 3.0 m/z and fragmented by higher-energy collisional dissociation with normalized collision energies of 25 and at a resolution of 17,500. Lock mass (445.120025) and dynamic exclusion (15 s) was enabled.

The MS raw files were processed by Proteome Discover 1.4.0.288 (DBversion:79; Thermo Fisher Scientific). MS/MS spectra were searched using either SEQUEST or mascot search algorithm against a database containing common contaminants, human ADAM10, and the canonical murine database (total of 16,741 sequences). For chymotrypsin-treated samples, a semichymotrypsin search was performed with 3-missed cleavages allowed, whereas for trypsin-treated samples, the data were searched with semiargC specificity with 2-missed cleavages allowed. A MS1 tolerance of 5 ppm and a MS2 tolerance of 0.02 Da were implemented. Oxidation (15.995 Da) of methionine residues was set as a variable modification along with peptide N-terminal demethylation (28.031 Da) while dimethylation of lysine residues was set as a static modification (28.031 Da). Results are shown in Supplemental Data.

Sequence analysis and homology modeling

For secondary structure analysis and potential disordered regions, PSI-blast based secondary structure PREDiction (PSIPRED) (35) and Disopred Prediction (DISOPRED3) (36) were used. Additionally, HHpred (37, 38) was used for remote homology detection and 3D structure prediction. The multiple sequence alignment was generated manually, guided by pairwise alignments from ClustalΩ (39). The model for murine ADAM9 was built on the basis of the experimental structure of the ADAM22 [Protein Data Bank (PDB) code 3G5C] ectodomain and fragilysin-3 (Fra3; PDB code 3P24) with regard to the orientation of the interdomain linker region. In the same manner, models for human ADAM10 and murine ADAM17 were generated, in which the ectodomain of human ADAM10 (PDB code 6BE6) served as template. For structural based homology modeling, MODELER (40) was used. The prodomain model of murine ADAM17 (Leu⁶⁹-Ile¹⁷²) was built based on Fra3 (PDB code 3P24). Figures were prepared with Pymol (Schrödinger, New York, NY, USA).

Immunofluorescence microscopy

Cells were seeded into 6-well plates containing small cover slips and transfected as previously described. Thirty-six hours after transfection, cells were washed with PBS (Thermo Fisher Scientific) several times and fixed with 4% (w/v) paraformaldehyde in PBS for 10 min at room temperature. Afterwards, cells were washed with PBS again and permeabilized [0.12% (w/v) glycine, 0.2% (w/v) saponin in PBS] for 10 min at room temperature. Repeated washing steps with PBS were performed before incubation of cells in blocking solution [10% FBS, 0.2% (w/v) saponin in PBS] for 1 h at room temperature. Subsequently, primary antibody anti-myc (2276; Cell Signaling Technology), anti-meprin β (polyclonal serum, generated against the peptide CGMIQSSGDSADWQRVSQ), anti-ADAM9 cyto-49-1 (polyclonal serum, generated against the C terminus), or anti-Flag (F1804; MilliporeSigma) diluted in blocking solution was added and incubated in a humid chamber overnight at 4°C. The next day, cover slips were washed with 0.2% (w/v) saponin/PBS several times and incubated with fluorophore-coupled secondary antibody (Thermo Fisher Scientific) for 1 h at room temperature in a dark and humid chamber. Washing with 0.2% (w/v) saponin/PBS and _ddH₂O was repeated, and cover slips were fixed with 15 µl of mowiol/1,4-diazabicyclo[2.2.2]octane (Dabco) solution [17% (w/v) mowiol, 33% (v/v) glycerol, 50 mg/ml Dabco, 1 µg/ml DAPI], preheated to 60°C. Analysis was performed using a confocal laser-scanning microscope FV1000 (Olympus, Tokyo, Japan).

ELISA

For analysis, cell supernatants of ADAM10^−/−;17^−/− HEK293T were collected and ultracentrifuged at 186,000 g for 2 h. Cells were harvested as previously described, and protein expression was analyzed by SDS-PAGE and Western blot. Subsequently, cell supernatants were analyzed for soluble TNF-α using the TNF-α Mouse ELISA Kit (Thermo Fisher Scientific) for mouse TNF-α following the manufacturer’s instructions. Optical density was measured in a microplate reader (Tecan) at 450 nm with wavelength correction set to 570 nm.

Mice and differentiation of bone marrow–derived macrophages

All procedures performed in this study involving animals were in accordance with the ethical standards set by the National Animal Care Committee of Germany. Animals were maintained in a conventional animal facility and kept under specific pathogen-free conditions in individually ventilated cages under controlled temperature, humidity, and 12-h light/dark cycle. Mep1b^−/− male and female mice (41) and corresponding C57/BL6N wild-type (WT) animals at the age of 8–10 wk were euthanized to isolate and differentiate bone marrow–derived macrophages (BMDMs). Femur and tibia were dissected and bones were opened under sterile atmosphere to flush the bone marrow through a cell strainer (40 µm) with sterile DMEM. Cells were centrifuged for 10 min at 1000 g at 4°C and resolved in 7 ml macrophage medium [DMEM supplemented with 10% (v/v) FBS, 100 U/ml penicillin and 100 µg/ml streptomycin, 1 mM sodium pyruvate, 2 mM l-glutamine, 10 mM HEPES] including 40 ng/ml recombinant mouse macrophage colony-stimulating factor (rm M-CSF; ImmunoTools, Friesoythe, Germany). Cells were seeded on a 10-cm cell culture dish and cultivated under a humidified atmosphere (5% CO₂) at 37°C. The next day, monocytes containing cell supernatants were collected and centrifuged for 10 min at 100 g and 4°C. The 2 × 10⁷ cells were seeded per uncoated 10 cm dish in 7 ml macrophage medium containing rm M-CSF. After 48 h, 7 ml fresh medium including rm M-CSF was added to cells, and on d 7, macrophages were detached by Accutase solution and 1 × 10⁶ cells per well were seeded in 6-well plates. The next day, macrophages were stimulated with 1 µg/µl LPS in serum-free medium for 3 h, and cell supernatants were collected and analyzed for soluble TNF-α using the DuoSet ELISA Development Kit (R&D Systems) as described above.

AP assay

For BTC and EREG shedding experiments, the medium of transiently transfected ADAM10^−/−;17^−/− HEK293T cells was 24 h post-transfection replaced with fresh serum-free medium. After incubation at 37°C for 5 h, cell supernatants were ultracentrifuged at 186,000 g for 2 h at 4°C. For a comparable analysis of possible shedding events, supernatants were normalized to the protein content of the respective cell lysates and analyzed by immunoblotting and photometric quantification of AP activity at 405 nm for 6 h, as previously described (31, 42). Results were normalized to the empty vector control and the amount of product was quantified after 4 h. For concentration of respective supernatants for immunoblotting, proteins were precipitated with trichloracetic acid [10% (w/v)]. After incubation on ice for 60 min, proteins were pelleted (15,000 g, 15 min, 4°C), washed with acetone, and dissolved in SDS-PAGE loading buffer.

Statistical analysis

RFUs of the fluorogenic peptide-based activity assay were normalized to the control sample. Western blots were quantified with ImageJ (National Institutes of Health, Bethesda, MD, USA), and the protein of interest was normalized to actin levels. All statistical analyses were performed with Prism (GraphPad Software, La Jolla, CA, USA) by unpaired Student’s t test or 1-way ANOVA, followed by Tukey’s post hoc test. Normalized values were presented as means ± sd.

RESULTS

N-terminal cleavage of ADAM9, 10, and 17 by meprin β induces activity of recombinant ADAMs

To validate ADAMs as substrates of meprin β as previously identified by proteomics (9), we transiently transfected HEK293T cells with N-terminal Strep-tagged ADAM proteases. ADAMs immunoprecipitated by Strep-Tactin Sepharose were incubated with recombinant meprin β to analyze proteolytic processing in vitro. Immunoprecipitation of murine ADAM9 resulted in purified pro-ADAM9 of ∼120 kDa (Fig. 1A, B). Incubation of immunoprecipitated ADAM9 with recombinant meprin β generated a 25 kDa N-terminal fragment (NTF), which was detected with an anti-Strep-antibody (Fig. 1B, asterisk). A corresponding 95 kDa C-terminal fragment (CTF) was visualized by a polyclonal ADAM9 antibody recognizing the C terminus and confirmed specific N-terminal cleavage of ADAM9 by meprin β (Fig. 1A, B). To determine the impact of meprin β cleavage on ADAM activity in vitro, we generated soluble recombinant ADAM9 ectodomain carrying an N-terminal His-tag. Using a fluorogenic peptide-based cleavage assay, we observed significantly increased ADAM9 activity after preincubation of the recombinant protease with meprin β (Fig. 1C).

Figure 1.

N-terminal processing of ADAM9, 10, and 17 by meprin β induces ADAM activity in vitro. A) Domain structure of ADAM9 with indicated meprin β cleavage area. B) HEK293T cells were transfected with N-terminal Strep-tagged murine ADAM9 (mADAM9). Immunoprecipitated ADAM9 was incubated with recombinant meprin β and analyzed by immunoblotting. C) Incubation of recombinant N-terminal His-tagged ADAM9 ectodomain with recombinant meprin β resulted in significantly increased ADAM9 activity (n = 5). D) Domain structure of ADAM10 with indicated meprin β cleavage area. E) N-terminal Strep-tagged human ADAM10 (hADAM10) was expressed in HEK293T cells. Immunoprecipitated ADAM10 was incubated with recombinant meprin β and analyzed by immunoblotting. F) Incubation of recombinant N-terminal His-tagged ADAM10 pro- and catalytic domain with recombinant meprin β resulted in significantly increased ADAM10 activity (n = 3). G) Domain structure of ADAM17 with indicated meprin β cleavage site inside its prodomain. H) HEK293T cells were transfected with N-terminal Strep-tagged murine ADAM17 (mADAM17) or isolated ADAM17 prodomain (mADAM17_PD), respectively. Immunoprecipitated ADAM17 was incubated with recombinant meprin β and analyzed by immunoblotting. I) Incubation of recombinant N-terminal His-tagged ADAM17 pro- and catalytic domain with recombinant meprin β resulted in significantly increased ADAM17 activity. Isolated ADAM17 prodomain (PD) inhibits ADAM17 activity, which was significantly restored upon meprin β incubation (n = 4). CAT, catalytic domain; CYS, cysteine-rich domain; DIS, disintegrin domain; PRO, prodomain; TM, transmembrane domain.Data are presented as means ± sd. *P < 0.05, **P < 0.01, ***P < 0.001 (unpaired Student’s t test).

Immunoprecipitated N-terminal Strep-tagged human ADAM10 was detected at 95 kDa by immunoblotting, which disappeared upon incubation with recombinant meprin β using an anti-Strep antibody (Fig. 1D, E). On the other hand, an NTF of 25 kDa was generated, visualized by a specific ADAM10 prodomain antibody (Fig. 1E, asterisk). Furthermore, the corresponding CTF of ∼70 kDa was detected using an antibody recognizing the ADAM10 C terminus, thus again indicating N-terminal cleavage of ADAM10 by meprin β (Fig. 1D, E). Recombinant N-terminal His-tagged ADAM10 pro- and catalytic domain revealed significantly increased ADAM10 activity upon meprin β incubation (Fig. 1F).

We additionally immunoprecipitated N-terminal Strep-tagged full-length murine ADAM17 as well as the isolated Strep-tagged murine ADAM17 prodomain (Fig. 1G). After incubation with recombinant meprin β, the signal for full-length ADAM17 at 130 kDa and the signal for its isolated prodomain disappeared, but a 15 kDa NTF was detected for both proteins using an anti-Strep antibody (Fig. 1H, asterisks). Regarding the similar size of the NTFs observed for ADAM17 full-length protein and its isolated prodomain, but differing from ADAM9/10, a more N-terminal meprin β cleavage within the ADAM17 prodomain sequence as compared with the other ADAMs can be assumed (Fig. 1G, H). Incubation of N-terminal His-tagged recombinant ADAM17 pro- and catalytic domain with soluble recombinant meprin β resulted in significantly increased ADAM17 activity (Fig. 1I). Isolated ADAM17 prodomain can act as a specific inhibitor for catalytically active ADAM17 (30). Hence, we speculated whether meprin β cleavage inside ADAM17 prodomain can prevent its inhibitory capacity. Indeed, upon preincubation of recombinant ADAM17 prodomain with meprin β, inhibition of catalytically active ADAM17 was significantly reduced (Fig. 1I).

Taken together, we confirmed specific N-terminal cleavage of ADAM9, ADAM10, and ADAM17 by the metalloprotease meprin β. We showed that meprin β–mediated cleavage of ADAM proteases results in increased ADAM activity. Because furin-cleaved noncovalently bound ADAM prodomains can still inhibit respective ADAM activity (13–15), we hypothesize a regulatory function of meprin β to induce ADAM activity by cleaving their inhibitory prodomains.

Meprin β specifically cleaves ADAM9, 10, and 17 N-terminally of the furin cleavage site

To identify the exact meprin β cleavage sites within the ADAM amino acid sequence, immunoprecipitated N-terminal Strep-tagged ADAM9, 10, and 17 were incubated with recombinant meprin β (Fig. 1), and generated cleavage fragments were further analyzed by MS. A multiple sequence alignment of all 3 ADAM prodomains of human and murine species shows the conserved proprotein convertase motif separating ADAM prodomains from their respective catalytic domains (Fig. 2A, black triangle). Furthermore, a second conserved regulatory proprotein convertase cleavage motif (27) and previously identified meprin β cleavage sites (9) are indicated (Fig. 2A, gray triangle and red boxes). Employing MS, we identified 2 additional meprin β cleavage sites within the ADAM9 prodomain between amino acids Gly¹⁸⁹/Asp¹⁹⁰ and Glu¹⁹¹/Glu¹⁹², which are located directly N-terminally of the already described furin cleavage site (Fig. 2A, red triangle, Supplemental Data). Also, within the ADAM10 prodomain 2, cleavage sites for meprin β were identified between amino acids Gln¹⁹⁸/Glu¹⁹⁹ and Glu¹⁹⁹/Glu²⁰⁰, which again are located N-terminally of the furin cleavage site (Fig. 2A, red triangle). For ADAM17, a shift of ∼10 kDa in size was observed for the meprin β cleavage fragment compared with the ADAM9 and ADAM10 NTFs (Fig. 1), indicating a different meprin β processing step more N-terminally of the furin cleavage site in the ADAM17 prodomain. Although we did not obtain sufficient material to identify the exact meprin β cleavage site within ADAM17, multiple sequence alignment of the different ADAM prodomains suggests potential meprin β cleavage sites within the ADAM17 prodomain, which are in agreement with the observed NTF size (Fig. 2A, black boxes). In contrast to the conserved furin cleavage pattern, meprin β cleavage sites within ADAM prodomains are thus located differently when the prodomains of ADAM9 and ADAM10 were compared with that of ADAM17 (Fig. 2A).

Figure 2.

Activation of ADAM9, 10, and 17 by furin and meprin β. A) Multiple sequence alignment of murine and human ADAM9 (Uniprot accession: Q61072, Q13443), ADAM10 (O35598, O14672), and ADAM17 (Q9Z0F8, P78536). Conserved residues are shown in bold. The conserved furin cleavage motif is underlined and cleavage sites are marked by a black triangle. A proposed secondary proprotein convertase motif (27) is marked by a gray triangle. Red triangles show MS identified meprin β cleavage sites for ADAM9 and 10, whereas black boxes with a yellow background show potential meprin β cleavage sites in ADAM17. In contrast, red boxes with a yellow background highlight meprin β cleavage sites identified by TAILS (9). The gray background marks regions with homology to Fra3 (PDB code 3p24). B) Cartoon representation of a murine ADAM9 (mADAM9) homology model based on the ectodomain of ADAM22 (PDB code 3G5C). For human ADAM10 and murine ADAM17 (mADAM17), the ectodomain of ADAM10 (PDB code 6BE6) served as template. The catalytic domain is further shown in a transparent surface. For simplification, the disintegrin domains (Dis) in the back of the catalytic domain of ADAM10/17 were omitted. The interdomain linker region is shown in red, the catalytic zinc is displayed as a yellow sphere, and interacting residues are depicted as sticks. C) Close-up views of the furin cleavage sites. D) Close-up view of the identified meprin β cleavage sites for mADAM9 and human ADAM10 (hADAM10). In the case of ADAM17, a cartoon representation of a homology model of the prodomain based on Fra3 (PDB code 3P24) shows potential meprin β cleavage sites as sticks. Cys, cysteine-rich domain; EGF, EGF-like domain.

Because no structural information for any isolated ADAM prodomain nor of any pro-ADAM protease is available, we used secondary structure prediction tools and HHpred (37, 38) for remote homology modeling and 3D structure prediction for ADAM prodomains. We used the already known structures of the ADAM10 (43) and ADAM22 (44) ectodomains to obtain structural insight in the postulated activation mechanism. In accordance with a previous report (27), our analysis characterized the prodomains of ADAM9, 10, and 17 to have a central folded core region, which is C-terminally flanked by an unstructured linker region. Besides low sequence identity (∼10%), all 3 ADAM prodomains were identified by the pairwise comparison of profile hidden Markov models–based HHpred (37, 38), with probabilities of >90% to be homologs to the metalloprotease Fra3, a specific virulence factor of enterotoxigenic Bacteroides fragilis. So far, Fra3 homology to ADAM proteases was thought to be solely restricted to the catalytic domain fold (45). Although Fra3 consists only of a pro- and catalytic domain, ADAM proteases have additional C-terminal domains, which make the overall fold much more complex. However, we hypothesize the interdomain linker region between the catalytic domain and the folded prodomain core of ADAM proteases to be important for the inhibitory capacity and association of the prodomain in a similar manner as observed in Fra3. Because of the limited available structural information, it is hard to estimate how the prodomain is orientated within pro- and mature ADAMs. Nevertheless, based on the already solved ADAM10 structure (43), we assume the linker to run in parallel orientation to the substrate into the active site (Fig. 2B). In this scenario, the access to the active site cleft is still partially blocked by the linker of the noncovalently attached prodomain upon initial furin cleavage (Fig. 2C). In contrast, meprin β cleavage N-terminally of the furin cleavage site would rather lead to a destabilization of the prodomain (Fig. 2D), thereby enabling substrates to obtain access to the active site. Indeed, for ADAM9 and ADAM10, polar amino acids Asn⁶¹⁸ and Lys³⁵⁶, respectively, were found in the structural model to interact with amino acids located in the identified meprin β cleavage site, thus stabilizing the inhibitory capacity of ADAM prodomains. In the case of ADAM17, meprin β cleavage likely occurs within the folded core region of the prodomain and hence might result likewise in a destabilization of the inhibitory interaction. Overall, processing of ADAM prodomains by metalloprotease meprin β appears to be an additional regulatory event to control the activity of ADAM9, 10, and 17.

Coimmunoprecipitation confirmed direct interaction of meprin β with ADAM proteases

To further characterize the interaction of meprin β and ADAM proteases in a cellular environment, we transiently transfected ADAM10^−/−;17^−/− HEK293T cells with meprin β and ADAM9, 10, and 17 cDNAs, respectively. Performing immunofluorescence microscopy, we observed colocalization of meprin β and ADAM9, 10, or 17 along the secretory pathway and at the cell surface in transfected cells (Fig. 3A–C). Additionally, C-terminal Flag-tagged meprin β was detected by immunoblotting upon immunoprecipitation of C-terminal Myc-tagged ADAM10/17, thereby confirming direct interaction of both proteases in HEK293T cells (Fig. 3D). Conclusively, proteolytic interaction of ADAM proteases with meprin β is possible in a cell-based system and takes place during the secretory pathway and at the cell surface.

Figure 3.

Interaction of meprin β and ADAM proteases in transfected cells. A) Immunofluorescence microscopy showed colocalization of ADAM9 (green) and C-terminal Flag-tagged meprin β (red) at the cell surface of transfected ADAM10^−/−;17^−/− HEK293T cells. B, C) Colocalization was also observed for C-terminal Myc-tagged ADAM10 (green) (B) or C-terminal Myc-tagged ADAM17 (green) (C) and meprin β (red). D) HEK293T cells were transfected with C-terminal Flag-tagged meprin β and C-terminal Myc-tagged ADAM10 or ADAM17, respectively. ADAMs were immunoprecipitated via Myc-tag, and proteins were analyzed by immunoblotting.

Increased ADAM10 shedding upon coexpression with meprin β

ADAM9 has previously been identified as major sheddases of ADAM10 (25). To further investigate the impact of endogenous ADAM9 activation by meprin β, we analyzed ADAM10 shedding in ADAM10^−/−;17^−/− HEK293T cells (Fig. 4). ADAM10 shedding was already observed upon single transfection, as judged by the presence of ADAM10 CTFs at 14 kDa using an anti-Flag antibody. In presence of meprin β, ADAM10 shedding was indeed increased. Upon cotransfection with the catalytically inactive meprin β variant E153A, this effect was reduced down to the endogenous level of ADAM10 shedding (Fig. 4B, C). In order to increase this further, murine ADAM9 was also transiently transfected, which resulted in a significant increase of ADAM10 shedding in presence of meprin β, which was blocked by the inactive meprin β mutant E153A (Supplemental Fig. 3A, B). The cotransfection of murine ADAM9 and human ADAM10 resulted in CTF levels comparable to the endogenous one’s (Supplemental Fig. 3B). Possible explanations might be that overexpressed ADAM9 was not further activated at the cell surface and consequently was not able to shed ADAM10 or that murine ADAM9 cannot shed human ADAM10. To address the latter point, we also coexpressed murine ADAM9 and ADAM10 in presence or absence of meprin β in ADAM10^−/−;17^−/− HEK293T cells (Supplemental Fig. 3C, D). Here, meprin β co-expression resulted at least in a tendency of increased murine ADAM9-mediated murine ADAM10 shedding.

Figure 4.

Increased ADAM10 shedding upon coexpression with meprin β. A) Model of the proteolytic interaction between ADAM9, ADAM10, and meprin β. B) ADAM10^−/−;17^−/− HEK293T cells were transfected with C-terminal Flag-tagged human (h)ADAM10 and meprin β or its inactive variant E153A. Cells were treated with N-[N-(3,5-difluorophenacetyl)-l-alanyl]-S-phenylglycine t-butyl ester to avoid degradation of ADAM10 CTFs by γ-secretase. C) Western blot quantification of ADAM10 CTFs as shown in B from 2 biologic replicates. Data are presented as means ± sd. *P < 0.05 (1-way ANOVA followed by Tukey’s post hoc test).

Proteolytic activation of ADAM17 by meprin β

We have previously shown that the pathogenic bacterial cysteine protease RgpB secreted from Porphyromonas gingivalis activates meprin β at the cell surface (8). We thus cotransfected ADAM10^−/−;17^−/− HEK293T cells with meprin β and ADAM17 and subsequently stimulated cells with recombinant RgpB to activate membrane-bound meprin β. Importantly, with increasing concentrations of RgpB, enhanced processing of ADAM17 was observed (Fig. 5A). Of note, RgpB alone did not cleave ADAM17, hence maturation of ADAM17 was mediated by increased activity of meprin β induced by the pathogenic protease RgpB.

Figure 5.

Meprin β–mediated activation of ADAM17. A) ADAM10^−/−;17^−/− HEK293T cells were transfected with meprin β and ADAM17. Activation of meprin β was induced by different concentrations of the pathogenic soluble cysteine protease RgpB. B) Protein levels were detected by immunoblotting; ADAM10^−/−;17^−/− HEK293T cells were transfected with TNF-α, WT ADAM17, ADAM17^RVNG, meprin β, or its inactive mutant. Immunoblotting of cell lysates revealed that meprin β mediated the processing of WT ADAM17 and ADAM17^RVNG. C, D) Ultracentrifuged cell supernatants of transfected cells in B were analyzed for soluble TNF-α levels by ELISA. Data are presented as means ± sd from 2 biologic replicates. E) BMDMs were isolated from meprin β–deficient (n = 6) and corresponding WT mice (n = 4). Differentiated macrophages were stimulated with 1 µg/ml LPS for 3 h, and supernatants were analyzed for soluble TNF-α levels by ELISA. H, human; m, murine. Data are presented as means ± sd. *P < 0.05 (unpaired Student’s t test).

Based on this, we asked whether the meprin β–mediated activation of ADAM17 at the plasma membrane has direct impact on the shedding capacity of ADAM17. To address this, we analyzed the activity of WT ADAM17 and the furin-resistent mutant ADAM17^RVNG (46) toward TNF-α release, upon overexpression in ADAM10^−/−;17^−/− HEK293T cells, and compared it with the activity in presence of transiently cotransfected meprin β or its proteolytic inactive form E153A (Fig. 5B–D). Interestingly, for both ADAM17 variants, a meprin β–specific cleavage of pro-ADAM17 was observed in the cell lysates, even though RgpB as a meprin β activator was missing (Fig. 5B). Taking the different transfection efficiencies of WT ADAM17 and ADAM17^RVNG into account, generated soluble TNF-α measured in ultracentrifuged cell supernatants via ELISA was analyzed for both variants separately (Fig. 5C, D). Surprisingly, cotransfection of TNF-α with WT ADAM17 or ADAM17^RVNG revealed for both variants a tendency (not significant) of increased soluble TNF-α in comparison to the TNF-α single transfection (Fig. 5C, D). A similar effect was also observed upon coexpression of TNF-α and active or inactive meprin β. However, active as well as inactive meprin β showed a trend of TNF-α shedding increase for WT ADAM17 (Fig. 5C), which was also seen for the triple-transfection of TNF-α/ ADAM17^RVNG and meprin E153A (Fig. 5D). However, we additionally investigated the possibility of meprin β–mediated activation of ADAM17 under more physiologic conditions. Therefore, BMDMs were isolated from Mep1b^−/− mice (41) and corresponding WT animals. TNF-α shedding was measured by ELISA upon LPS stimulation of differentiated macrophages. Significantly increased soluble TNF-α levels were detected in supernatants of WT BMDMs compared with meprin β–deficient BMDMs (Fig. 5E). Thus, meprin β stimulates ADAM17 activity in macrophages because ADAM17-mediated TNF-α shedding was diminished in the absence of meprin β.

As the BMDM experiments more strongly support a meprin β–mediated activation of ADAM17 in comparison with the overexpression studies, the results stimulated us to investigate whether the different effects observed are dependent on the cellular context or the substrate itself. In order to elucidate that in more detail, the influence of meprin β on ADAM17 activity was further investigated for 2 additional ADAM17 substrates. In the same experimental overexpression setup as for TNF-α, the ADAM17-dependent shedding of the AP-tagged epidermal growth factor receptor (EGFR) ligands BTC and EREG was analyzed (Fig. 6). In comparison with the TNF-α experiments (Fig. 5A–D), the meprin β–dependent maturation of ADAM17 resulted in a significant increase of both BTC and EREG shedding, as judged by Western blotting and AP activity measurements of ultracentrifuged cell supernatants (Fig. 6).

Figure 6.

Meprin β–mediated activation of ADAM17 increases BTC and EREG shedding. A–D) ADAM10^−/−;17^−/− HEK293T cells were either transfected with ADAM17, meprin β, AP-tagged BTC, or EREG. Immunoblotting of cell lysates and supernatants revealed cleavage increased meprin β-dependent shedding of BTC (A) and EREG (D). BTC (B) and EREG (C) shedding was further analyzed by photometric quantification of AP activity after 4 h of incubation. Data are presented as means ± sd, and statistical analysis was assessed by 1-way ANOVA followed by Tukey’s post hoc test from 3 biologic replicates. *P < 0.05, **P < 0.01, ***P < 0.001. Open arrow: BTC in microvesicles; black arrow: shed BTC or EREG, EREG* immunoblot of ultracentrifuged supernatant; gray arrow: cleaved ADAM17. E) Cartoon summarizes the interaction of RgpB, meprin β, and ADAM17 with respect to substrate shedding.

In summary, we identified a complex proteolytic interaction between ADAM9, 10 and 17 and the metalloprotease meprin β, which represents a prototypic example of the distinct protease interactions within the protease web (Fig. 7). Besides ADAM10- and ADAM17-mediated shedding of inactive promeprin β (8), we revealed that meprin β can boost the activity of ADAM proteases 9, 10, and 17. Virulence factors, such as RgpB, can activate membrane-bound meprin β, thereby leading to pathologic conditions by changing its substrate repertoire as previously shown for the intestine (8). Additionally, increased ADAM17 activation by meprin β might result in increased TNF-α shedding, subsequently increasing the proinflammatory response while enhanced ectodomain release of the EGFR ligands BTC and EREG could potentially influence cell proliferation and differentiation (Fig. 7). Conclusively, we hypothesize a regulatory feedback mechanism that controls protease activity at the cell surface.

Figure 7.

Schematic summary of the proteolytic interaction between ADAMs and meprin β. On the one hand, meprin β is shed by ADAM10 and 17 from the cell surface exclusively in its inactive proform. Shed meprin β can subsequently be activated by soluble tryptic proteases. On the other hand, ADAM9, 10, and 17 are cleaved by meprin β within their inhibitory prodomains, leading to enzyme activation. Consequently, proteolytic interaction controls enzyme activity at the cell surface and might play a crucial regulatory role (e.g., under inflammatory conditions).

DISCUSSION

Meprin β has been reported to cleave mucin 2 in the small intestine, which drives intestinal mucus detachment and prevents bacterial overgrowth (7). Importantly, only soluble but not membrane-bound meprin β obtains access to the mucin 2 cleavage site, demonstrating the importance of regulating meprin β shedding and activation. We previously showed that meprin β activation and its shedding by ADAM10 and 17 are mutually exclusive events (8). This strict regulation of meprin β activity can be overridden by bacterial pathogens as demonstrated for the bacterial protease Arg- RgpB, which causes severe inflammation in the intestine (8). On the other hand, meprin β was shown to induce the production of proinflammatory cytokines in macrophages, including IL-1β and IL-18 (47, 48). Several other putative meprin β substrates were identified by a proteomics-based approach, including different ADAM proteases (9). Besides the importance of meprin β–mediated ADAM regulation, meprin β itself was suggested to be an important regulator for inflammatory processes. Here, we show that meprin β cleaves ADAM9, 10, and 17 within their prodomains in vitro, leading to increased enzyme activity. Immunoprecipitation of full-length ADAM proteases and digestion with recombinant meprin β revealed cleavage sites within ADAM prodomains as determined by MS. These identified cleavage sites were all located in the N terminal of the canonical furin consensus sequence and nicely fit to the unique cleavage specificity of meprin β with a preference for acidic amino acids (10). Of note, the in vitro determined that meprin β cleavage sites did not confirm the Terminal Amine Isotopic Labeling of Substrates (TAILS)-predicted sites (9) most likely because of the different experimental conditions. In this context, it is worth pointing out that the previously determined cleavage sites for ADAM9, 10, and 17 were not consistently identified in all analyzed cell lineages within the TAILS approach (9). This might already be a hint for a more complex regulation of meprin β–mediated activation of the ADAMs dependent on other so far unknown cellular factors. Therefore, it cannot be excluded that the immunoprecipitation of the ADAMs altered their accessibility for recombinant meprin β because of the lack of other, for example, complexing/regulatory proteins. Additionally, the prodomain cleavage of the ADAMs was primarily observed for membrane-bound meprin β, which also may influence accessibility of the cleavage sites.

As ADAM proteases are maturated by proprotein convertases on the secretory pathway, which is mandatory to gain catalytic activity (11, 12), additional meprin β cleavage sites might be used subsequent to furin cleavage as a first maturation event. Recently, a second upstream proprotein convertase cleavage site within ADAM prodomains was identified for several ADAMs, which induces ADAM activity (27). Interestingly, at least for ADAM17, the described motif correlates with the previously identified meprin β cleavage site by TAILS, thus supporting a role for meprin β in the regulation of ADAM activity. However, ADAM17 cleavage by meprin β most likely occurs at the plasma membrane and not on the secretory pathway as known for proprotein convertases. Isolated ADAM prodomains can act as specific inhibitors for respective ADAM proteases in vitro (13–15) by binding noncovalently to their catalytic domains. Homology models suggest interactions of the protease peptide backbone with amino acids located in the meprin β cleavage sites of ADAM9 and ADAM10 prodomains. This interaction might stabilize noncovalent binding of the prodomain within the catalytic site cleft. Therefore, we propose loss of the inhibitory function of ADAM prodomains upon meprin β–mediated processing N-terminally of the furin cleavage site. This is in line with a significant decrease in the inhibitory capacity of isolated recombinant ADAM17 prodomain upon preincubation with meprin β.

Thus, we analyzed the impact of ADAM activation by meprin β for already well-described ADAM substrates. ADAM9 was previously identified as the main ectodomain sheddase of ADAM10, thereby regulating ADAM10 activity at the cell surface (25). Indeed, we observed increased accumulation of ADAM10 CTFs in ADAM10/meprin β cotransfected compared with ADAM10 single-transfected ADAM10^−/−;17^−/− HEK293T cells. Of note, ADAM10^−/−;17^−/− HEK293T cells also express ADAM15, another described ADAM15 sheddase of ADAM10 (25, 49). Interestingly, also for ADAM15, a potential cleavage site likewise located within the prodomain was identified within the TAILS analysis (9). Therefore, we cannot explicitly state whether the endogenous ADAM10 shedding is a result of shedding by ADAM9, ADAM15, or even another protease.

Prototypically, we also analyzed the impact of meprin β on ADAM17 shedding for the well-characterized substrate TNF-α and the 2 EGFR ligands BTC and EREG. Interestingly, canonical furin cleavage of WT ADAM17 on the secretory pathway as well as a furin-resistent ADAM17 mutant did not exclude further processing of ADAM prodomains by meprin β. This suggests that meprin β may be an additional regulator of ADAM activity at the plasma membrane. However, ADAM17-dependent TNF-α shedding was not significantly increased in the overexpression setup. Additionally, active as well as inactive meprin β showed a trend of increased shedding activity for WT ADAM17, which was also seen for the triple-transfection of TNF-α/ADAM17^RVNG and meprin E153A. Overall, the data imply, next to an activation function of meprin β, a kind of stabilizing effect of the protease toward TNF-α. The latter is supported by the effects of meprin E153A, which suggest a scenario in which the half-life and accessibility of TNF-α at the plasma membrane is increased because of the binding within a protein complex of meprin β E153A and ADAM17/ ADAM17^RVNG. In contrast, significant decreased TNF-α shedding was observed in LPS-stimulated BMDMs from meprin β–deficient mice as compared with WT controls. This provides evidence for meprin β–mediated activation of ADAM17 in primary cells, thereby releasing proinflammatory TNF-α. The additional analysis of the EGFR ligand BTC and EREG further support a meprin β–mediated activation of ADAM17, whereas nonetheless a substrate stabilization via complex formation in dependence of the cellular context is likely and cannot be excluded.

Meprin β has previously been shown to be expressed in intestinal lymph nodes, thus contributing to leukocyte migration and demonstrating its role in intestinal immune responses (50). As mentioned above, we recently showed that meprin β itself can be activated in its membrane-bound form by the virulence factor RgpB secreted from P. gingivalis (8), which, as shown here, leads to increased ADAM17 maturation. Hence, inflammatory stimuli that induce activity of membrane-bound meprin β can result in increased ADAM17 activation, which in turn triggers inflammation.

Taken together, we identified ADAM proteases as meprin β substrates in vitro and in cellulo. Because ADAM10 and ADAM17 play crucial roles in development and cell survival, induction of ADAM activity by meprin β most likely has consequences for several physiologic and pathologic processes. This could represent an additional post-translational mechanism to regulate ADAM activity, in addition to other well-described factors, such as tetraspanins for ADAM10 or iRhoms for ADAM17 (51, 52). In contrast to ADAM10- and ADAM17-deficient mice, which are embryonically lethal (23, 24), meprin β–deficient mice are viable (41). Therefore, meprin β is apparently not required for basal ADAM activation, but is rather involved in the fine tuning of ADAM activity at the plasma membrane under certain conditions, such as inflammation. Previous studies also mentioned proprotein convertase cleavage at the furin consensus sequence as a mandatory event for ADAM activation and further N-terminal processing as additional induction of ADAM activity (27). Conclusively, we identified a complex proteolytic interaction between meprin β and ADAM proteases within the proteolytic web that controls protease activity at the cell surface, thereby influencing conversion of further substrates. Because ADAM proteases and meprin β also share common substrates, such as APP (6, 21) or the IL-6 receptor (4, 53), it will be interesting to further investigate those cleavage events under physiologic and pathologic conditions.

Supplementary Material

This article includes supplemental data. Please visit http://www.fasebj.org to obtain this information.

Glossary

ABC

ammonium bicarbonate

ACN

acetonitrile

ADAM

A disintegrin and metalloproteinase

AP

alkaline phosphatase

APP

amyloid precursor protein

BMDM

bone marrow–derived macrophage

BTC

betacellulin

CTF

C-terminal fragment

EGFR

epidermal growth factor receptor

EREG

epiregulin

FA

formic acid

FBS

fetal bovine serum

Fra3

fragilysin-3

HEPES

4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

MS

mass spectrometry

NTF

N-terminal fragment

PDB

Protein Data Bank

RFU

relative fluorescent unit

RgpB

gingipain R

rm M-CSF

recombinant mouse macrophage colony-stimulating factor

TAILS

terminal amine isotopic labeling of substrates

WT

wild type

AUTHOR CONTRIBUTIONS

R. Wichert, F. Scharfenberg, C. Colmorgen, and C. Becker-Pauly designed and performed experiments; R. Wichert, F. Scharfenberg, and C. Becker-Pauly analyzed the data and wrote the manuscript; F. Scharfenberg designed structural models; T. Koudelka and A. Tholey performed MS and data analysis; and J. Schwarz, S. Wetzel, B. Potempa, J. Potempa, J. W. Bartsch, I. Sagi, P. Saftig, and S. Rose-John contributed essential reagents.