Analysis of the Conditions That Affect the Selective Processing of Endogenous Notch1 by ADAM10 and ADAM17

Abstract

Notch signaling is critical for controlling a variety of cell fate decisions during metazoan development and homeostasis. This unique, highly conserved signaling pathway relies on cell-to-cell contact, which triggers the proteolytic release of the cytoplasmic domain of the membrane-anchored transcription factor Notch from the membrane. A disintegrin and metalloproteinase (ADAM) proteins are crucial for Notch activation by processing its S2 site. While ADAM10 cleaves Notch1 under physiological, ligand-dependent conditions, ADAM17 mainly cleaves Notch1 under ligand-independent conditions. However, the mechanism(s) that regulate the distinct contributions of these ADAMs in Notch processing remain unclear. Using cell-based assays in mouse embryonic fibroblasts (mEFs) lacking ADAM10 and/or ADAM17, we aimed to clarify what determines the relative contributions of ADAM10 and ADAM17 to ligand-dependent or ligand-independent Notch processing. We found that EDTA-stimulated ADAM17-dependent Notch1 processing is rapid and requires the ADAM17-regulators iRhom1 and iRhom2, whereas the Delta-like 4-induced ligand-dependent Notch1 processing is slower and requires ADAM10. The selectivity of ADAM17 for EDTA-induced Notch1 processing can most likely be explained by a preference for ADAM17 over ADAM10 for the Notch1 cleavage site and by the stronger inhibition of ADAM10 by EDTA. The physiological ADAM10-dependent processing of Notch1 cannot be compensated for by ADAM17 in Adam10-/- mEFs, or by other ADAMs shown here to be able to cleave the Notch1 cleavage site, such as ADAMs9, 12, and 19. Collectively, these results provide new insights into the mechanisms underlying the substrate selectivity of ADAM10 and ADAM17 towards Notch1.

Keywords: ADAM10; ADAM17; Notch pathway; Notch receptor; Notch1; cell signaling; intercellular signaling; juxtacrine signaling; proteolysis; regulation.

Figure 1

The Notch1 receptor S2 cleavage product is transient and rapidly converted into the S3 cleavage product after treatment with the calcium-chelating agent ethylenediaminetetraacetic acid EDTA. (A) Schematic of the Notch1 receptor and (B) of its negative regulatory region (NRR, schematic adapted from Gordon et. al, [8,18]). (C,D) Wild-type mouse embryonic fibroblasts (WT mEFs) were treated with phosphate buffered saline (PBS, vehicle control, Ctrl) for 30 min or with 5 mM EDTA, which activates Notch1 [22] for 5, 10, 15, or 20 min, in the absence (C) or in the presence of 5 µM of gamma-secretase inhibitor N-[N-(3,5-Difluorophenacetyl)-l-alanyl]-S-phenylglycine t-butyl ester( DAPT) (D). The Western blots were probed with an antibody against the cytoplasmic domain of the Notch1 receptor. A fragment migrating close to the expected size of the Notch1 S2 product (~115 kDa) appeared after 10 min of EDTA treatment without DAPT (C) but started to disappear after 15 min of EDTA treatment, suggesting that the S2 product is transient in nature. A more stable and smaller fragment migrating close to the expected size of the S3 product appeared after 10 min of EDTA treatment without DAPT and persisted (C). Note that since the furin-dependent S1 cleavage is a constitutive event, the S1 product is present even in the absence of Notch1 activation with EDTA. (D) When WT mEFs were treated with 5 mM of EDTA in the presence of DAPT for 5, 10, 15, or 20 min, the S2 product began to accumulate after 5 min and appeared to reach maximum accumulation after 10 min. The Western blots shown are representative of at least 3 independent experiments. Abbreviations: EGF, Epidermal Growth Factor; LNR, Lin-12/Notch repeat; HD-N, HD-C, N- and C-terminal heterodimerization domain; PEST, peptide sequence that is rich in proline (P), glutamic acid (E), serine (S), and threonine (T) residues.

Figure 2

ADAM17 is the primary S2 protease involved in EDTA-induced endogenous Notch1 processing. Cells were treated with PBS (vehicle control, Ctrl), 5 mM of EDTA, or 5 mM of EDTA in the presence of the inhibitors DAPT (5 µM) or marimastat (5 µM MM, metalloprotease inhibitor that blocks ADAMs10 and 17) for 15 min. MM blocks the S2 cleavage, whereas DAPT blocks S3 processing. All the inhibitors were pre-incubated for 2 h in Opti-MEM and Western blots were performed using an anti-Notch1 cytoplasmic domain antibody. (A) Samples of WT mEFs treated with 5 mM of EDTA showed a band around the expected size of S3, which was not produced in the presence of DAPT. Instead, a fragment around the expected size of the Notch1 S2 product accumulated upon DAPT treatment. WT samples treated with EDTA and MM showed no S2 and S3 products. In contrast, mEFs lacking ADAM17 (A17-/- mEFs) showed no S2 or S3 products under the conditions where these bands were observed for WT mEFs. (B) ADAM10-deficient mEFs (A10-/- mEFs) showed the accumulation of an S3 product after EDTA treatment, which was blocked by MM, just like in WT mEFs. (C) Double-knockout mEFs deficient in ADAM10 and ADAM17 (A10/17-/- mEFs) resembled A17-/- mEFs in that they showed no S3 production in response to EDTA treatment. (D) mEFs deficient in the ADAM17 regulators iRhoms 1 and 2 (iR1/2-/- mEFs) also produced no S2 fragment in the presence of EDTA and DAPT, whereas the inactivation of either iRhom1 or iRhom2 (iR1-/- or iR2-/- mEFs) did not prevent the generation of Notch1 S2 (indicated by an asterisk). All Western blots shown are representative of at least 3 independent experiments.

Figure 3

ADAM10, and not ADAM17, cleaves the S2 site in Dll4 ligand-induced endogenous Notch1 processing. Cells were plated on tissue culture dishes coated with 1 µg/mL of the Notch1 ligand Dll4 in bovine serum albumin (BSA) or with BSA alone. Cells were starved in Opti-MEM for 2 h prior to treatment with DAPT or incubated with Opti-MEM alone for an additional 30 min or 120 min. (A) WT mEFs plated on Dll4 alone showed the minimal accumulation of an S2 product. Treatment with DAPT for 30 min resulted in very little accumulation of S2 compared to the accumulation after 120 min of DAPT treatment, when the S2 accumulation was substantial. (B) In WT mEFs, S2 product accumulated in cells plated on Dll4 that were treated with DAPT for 2 h. A similar accumulation of S2 product in DAPT-treated cells was apparent in A17-/- mEFs plated on Dll4. (C) In contrast, A10-/- mEFs, similar to A10/17-/- mEFs, plated on Dll4 showed no accumulation of S2 in the presence of DAPT. The Western blots are representative of at least 3 independent experiments.

Figure 4

Processing of the endogenous S2 site cannot be induced by the stimulation of ADAM10 or ADAM17. (A) WT mEFs were treated with 25 ng/mL of PMA, a strong activator of ADAM17, or (B) with 2.5 µM ionomycin, which strongly activates both ADAM10 and 17, with 5 mM EDTA treatment for 30 min included as a positive control for the processing of Notch1. The Western blots shown are representative of at least 3 independent experiments.

Figure 5

The constitutively exposed Notch1 S2 site behaves like an ADAM17 substrate. (A) WT and A17-/- mEFs were transfected with a truncated Notch1 receptor, consisting of the C-terminal portion of the receptor starting from amino acid residue 1687 within the C-terminal part of the heterodimerization domain, with an alkaline phosphatase tag attached to its N-terminus (Notch1-AP or N1-AP), and were stimulated with the ADAM17-selective stimulus PMA [52]. WT mEFs stimulated with 25 ng/mL of PMA had increased shedding of N1-AP over untreated WT mEFs. There was no significant increase in the PMA-induced shedding of N1-AP in A17-/- mEFs. (B) N1-AP shedding thus behaved similarly to the shedding of TGFα-AP, a known ADAM17 substrate [35,52], which could be stimulated by PMA in WT mEFs, but not in A17-/- mEFs. In contrast, PMA treatment did not result in the shedding of an established ADAM10 substrate, betacellulin-AP (BTC-AP) [52]. (C) In A17-/- mEFs treated with 2.5 µM of ionomycin, which strongly stimulates ADAM10 and ADAM17, N1-AP shedding was increased over unstimulated mEFs, whereas there was no statistically significant increase in N1-AP shedding in A10/17-/- mEFs treated with ionomycin, suggesting that ADAM10 can shed N1-AP in the absence of ADAM17. All the data are shown as mean ± standard error of the mean (SEM) for n ≥ 3 independent experiments. An * indicates p < 0.05 and n.s. indicates no statistically significant difference using the unpaired two-tailed Student’s t-test.

Figure 6

ADAM10 activity is more strongly inhibited by EDTA than ADAM17. (A,B) After 6 h of starvation in Opti-MEM, WT mEFs were treated with 4-amino-phenylmercuric acetate (APMA, 200 µM) in the presence or absence of 5 mM of EDTA for 30 min. APMA treatment induced the shedding of the ADAM17 substrate TGFα-AP (A) and the ADAM10 substrate BTC-AP (B). APMA-induced BTC-AP shedding, which depends on ADAM10, was more strongly reduced by treatment with 5 mM of EDTA than APMA-induced TGFα-AP shedding, which depends on ADAM17. In addition, PMA-induced TGFα-AP shedding, which depends on ADAM17, was not significantly reduced in the presence of EDTA (C). Data are shown as mean ± SEM for n ≥ 3 independent experiments. * indicates p < 0.05 and n.s. indicates no statistically significant difference using one-way ANOVA followed by Tukey’s multiple comparisons test.

Figure 7

ADAMs 9, 12, and 19 can cleave N1-AP in A10/17-/- mEFs. Overexpression of ADAMs 9, 12, 15, 17, and 19 in A10/17-/- mEFs shows that all but ADAM15 can increase the processing of N1-AP compared to the catalytically inactive ADAM17E>A control. Data are shown as mean ± SEM for n ≥ 3 independent experiments (n = 3 for ADAM17E>A mutant, n = 6 for all other samples). * indicates p < 0.05 using the unpaired 2-tailed Student’s t-test. The difference between the control sample, in which only N1-AP was expressed, and the sample in which the inactive ADAM17E>A was co-expressed with N1-AP was not statistically significant (not indicated on the figure), and the difference between N1-AP shedding in the presence of overexpressed inactive ADAM17E>A and ADAM15 also did not reach statistical significance (n.s.).