The metalloprotease ADAM10 (a disintegrin and metalloprotease 10) undergoes rapid, postlysis autocatalytic degradation

Abstract

The transmembrane protein, ADAM10 (a disintegrin and metalloprotease 10), has key physiologic functions-for example, during embryonic development and in the brain. During transit through the secretory pathway, immature ADAM10 (proADAM10) is converted into its proteolytically active, mature form (mADAM10). Increasing or decreasing the abundance and/or activity of mADAM10 is considered to be a therapeutic approach for the treatment of such diseases as Alzheimer's disease and cancer. Yet biochemical detection and characterization of mADAM10 has been difficult. In contrast, proADAM10 is readily detected-for example, in immunoblots-which suggests that mADAM10 is only a fraction of total cellular ADAM10. Here, we demonstrate that mADAM10, but not proADAM10, unexpectedly undergoes rapid, time-dependent degradation upon biochemical cell lysis in different cell lines and in primary neurons, which prevents the detection of the majority of mADAM10 in immunoblots. This degradation required the catalytic activity of ADAM10, was efficiently prevented by adding active site inhibitors to the lysis buffer, and did not affect proADAM10, which suggests that ADAM10 degradation occurred in an intramolecular and autoproteolytic manner. Inhibition of postlysis autoproteolysis demonstrated efficient cellular ADAM10 maturation with higher levels of mADAM10 than proADAM10. Moreover, a cycloheximide chase experiment revealed that mADAM10 is a long-lived protein with a half-life of approximately 12 h. In summary, our study demonstrates that mADAM10 autoproteolysis must be blocked to allow for the proper detection of mADAM10, which is essential for the correct interpretation of biochemical and cellular studies of ADAM10.-Brummer, T., Pigoni, M., Rossello, A., Wang, H., Noy, P. J., Tomlinson, M. G., Blobel, C. P., Lichtenthaler, S. F. The metalloprotease ADAM10 (a disintegrin and metalloprotease 10) undergoes rapid, postlysis autocatalytic degradation.

Keywords: ADAM17; Alzheimer’s; GI254023X; NrCAM; tetraspanin15.

Figure 1

The ADAM10-preferring inhibitor, GI254023X, increases mature ADAM10 levels. After 5 d in vitro, primary murine neurons (E16.5) were treated with TAPI-1 (50 µM), GI254023X (5 µM), or DMSO (control) for 48 h, then supernatants were collected and cells were lysed. A) Shown are representative Western blots. For the detection of ADAM10, an antibody to its C terminus was used. B) Densitometric quantification of proADAM10 and mADAM10 levels normalized to control levels. Two-sided Student’s t test with Welch’s correction and Bonferroni’s correction for multiple hypothesis testing (n = 6–8). ****P < 0.0001.

Figure 2

Specific ADAM10 inhibitors block the postlysis degradation of ADAM10. A) HEK293 cells were lysed with lysis buffer that contained GI254023X (5 µM), 1,10-phenanthroline (Phe.; 10 mM), TAPI-1 (50 µM), MN8 (5 µM), LT4 (5 µM), or DMSO as control. For the detection of ADAM17, lysates were subjected to an overnight ConA pull down. B) SH-SY5Y cells were lysed with lysis buffer that contained GI254023X (5 µM) or DMSO as control. C) Primary murine neurons (E16.5) were lysed with lysis buffer that contained GI254023X (5 µM) or DMSO as control. D) HEK293 cells were lysed with lysis buffer that contained GI254023X (5 µM) or DMSO as control. Cells were kept on ice for 60 min. After the indicated time points (0, 5, 10, 30, or 60 min), GI254023X (5 µM) was added to lysates, which previously did not contain GI254023X, to stop the reaction. Shown are representative Western blots. For the detection of ADAM10, C-terminal antibody was used. E) Quantification of mean proADAM10 and mADAM10 levels (n = 6) relative to those at the starting time point. Shown are means ± sem. Relative log₂ transformed values using 2-sided Student’s t test with Welch’s correction and post hoc Bonferroni’s correction. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 3

mADAM10 degradation creates several fragments, which are precluded by an intact membrane. A) HEK293 cells were lysed with lysis buffer that contained GI254023X (5 µM). Lysates were then subjected to an overnight ConA pull down. ADAM10 was detected under nonreducing conditions with an antibody raised against part of its N-terminal extracellular domain. Arrow indicates the smaller fragment. Asterisks indicate the larger fragments. B) HEK293 cells transfected with wtADAM10 (A10), ADAM10 ecto, ADAM10 Δcyto, or empty vector control. Cells were lysed in the presence of GI254023X (5 µM), with the exception of A10 con. Lysates (30 µl) or supernatants (500 µl) were pulled down by ConA-Sepharose beads. Arrow indicates the smaller fragment. The upper 3 panels are 3 different exposure (exp.) times of the blot with the N-terminal ADAM10 antibody. The lowest panel shows the blot developed with the C-terminal ADAM10 antibody, as indicated. C) HEK293 cells were lysed with lysis buffer that contained GI254023X (5 µM) or DMSO. Cells were kept on ice for 60 min. After the indicated time points (0, 2, 5, 10, 30, or 60 min), GI254023X (5 µM) was added to lysates, which previously did not contain GI254023X, to stop the reaction. Lysates were then subjected to an overnight ConA pull down. Arrow indicates the smaller 50-kDa fragment. Three different exposure times of the blot are shown. D) Membrane pellets were prepared from HEK293 cells. Laemmli buffer (1×; reducing), STET-lysis buffer without Triton X-100, or usual STET-lysis buffer with Triton X-100 was added to pelleted membranes in the presence or absence of GI254023X (5 µM). Samples were then incubated for 15 min on ice, then 1× Laemmli buffer (reducing) was added to the samples not yet containing Laemmli buffer. For the detection of ADAM10 in panel D, C-terminal antibody was used.

Figure 4

mADAM10 degradation requires catalytically active mADAM10. A) wtHEK293 cells and ADAM10 KO HEK293 cells were transfected with ADAM10, inactive hADAM10 mutant (E384A), or the empty vector. After 48 h of incubation, cells were lysed with lysis buffer that contained GI254023X (5 µM) or solvent. B) HEK293 cells were transfected with hADAM10, hADAM10 and tetraspanin15 (T15), or the empty vector. Cells were lysed in the presence of GI254023X (5 µM). Shown are representative Western blots (n = 3–4). For the detection of ADAM10, C-terminal antibody was used.

Figure 5

Determination of the half-life of mADAM10. A) HEK293 cells were treated with 100 µg/ml cycloheximide for indicated times (0, 2, 6, 12, 18, or 24 h). Cells were lysed by using STET-lysis buffer that contained GI254023X (5 µM). B) Quantification of mean proADAM10 and mADAM10 levels. mADAM10 corrected: Subtractions of proADAM10 at time points of 2–24 h from proADAM10 at time point 0 h. These values were then subtracted from mADAM10 values at every time point. Shown are representative Western blots (n = 3–4). For the detection of ADAM10, C-terminal antibody was used.

Figure 6

Scheme of the autocatalytic degradation of mADAM10 after membrane disruption. After mADAM10 is released from the membrane during cell lysis, potential cleavage sites—within the juxtamembrane and intracellular domains (indicated in red)—become accessible to the protease. This leads to an autocatalytic, presumably intramolecular cleavage, which creates at least 2 mADAM10 fragments of different sizes that comprise either the ectodomain (smaller fragment) or the ectodomain plus the transmembrane and part of the cytoplasmic domain. Generation of the smaller fragment occurs via an initial cleavage at the C-terminal end (1), followed by a secondary cleavage (2) in the juxtamembrane domain. The vertical dashed lines indicate the binding sites of the antibodies used in this study (black: anti–N-terminal antibody to the ADAM10 ectodomain; red: anti–C-terminal antibody to the cytoplasmic tail of ADAM10; domains not drawn to scale to emphasize the transmembrane and juxtamembrane domains).