Characterization of angiotensin-converting enzyme 2 ectodomain shedding from mouse proximal tubular cells

Abstract

Angiotensin-converting enzyme 2 (ACE2) is highly expressed in the kidney proximal tubule, where it cleaves angiotensin (Ang) II to Ang-(1-7). Urinary ACE2 levels increase in diabetes, suggesting that ACE2 may be shed from tubular cells. The aim of this study was to determine if ACE2 is shed from proximal tubular cells, to characterize ACE2 fragments, and to study pathways for shedding. Studies involved primary cultures of mouse proximal tubular cells, with ACE2 activity measured using a synthetic substrate, and analysis of ACE2 fragments by immunoblots and mass spectrometry. The culture media from mouse proximal tubular cells demonstrated a time-dependent increase in ACE2 activity, suggesting constitutive ACE2 shedding. ACE2 was detected in media as two bands at ∼ 90 kDa and ∼ 70 kDa on immunoblots. By contrast, full-length ACE2 appeared at ∼ 100 kDa in cell lysates or mouse kidney cortex. Mass spectrometry of the two deglycosylated fragments identified peptides matching mouse ACE2 at positions 18-706 and 18-577, respectively. The C-terminus of the 18-706 peptide fragment contained a non-tryptic site, suggesting that Met(706) is a candidate ACE2 cleavage site. Incubation of cells in high D-glucose (25 mM) (and to a lesser extent Ang II) for 48-72 h increased ACE2 activity in the media (p<0.001), an effect blocked by inhibition of a disintegrin and metalloproteinase (ADAM)17. High D-glucose increased ADAM17 activity in cell lysates (p<0.05). These data indicate that two glycosylated ACE2 fragments are constitutively shed from mouse proximal tubular cells. ACE2 shedding is stimulated by high D-glucose, at least partly via an ADAM17-mediated pathway. The results suggest that proximal tubular shedding of ACE2 may increase in diabetes, which could enhance degradation of Ang II in the tubular lumen, and increase levels of Ang-(1-7).

Figure 1. Increased ACE2 activity in media from mouse proximal tubular cells.

A time-dependent increase in ACE2 activity occurred in the cell media. Results are means ± SE, corrected for total cell protein from culture dishes set up in parallel to those where media was analyzed for ACE2 activity on three successive days. *P<0.001 vs day 4 and day 5, and **P<0.01 vs day 4, n = 4.

Figure 2. Immunoblot analysis of ACE2 protein in media and cell lysates from mouse PT cells.

(A) Representative immunoblot for ACE2 protein in concentrated media (Lanes 1–3) and cell lysates (Lanes 4–6) from mouse PT cells. Lanes 1 and 4: wildtype cells, Lanes 2 and 5: ACE2 knockout (KO) cells, Lanes 3 and 6: ACE2 KO cells transfected with a human ACE2 expression vector, Lane 7: mouse kidney cortex showing a band at ∼100 kDa, used as a positive control. Lane 1 shows two bands in the media at ∼90 kDa and ∼70 kDa for mouse ACE2. Lane 3 shows two bands in the media for human ACE2 in transfected cells, at ∼110 kDa and ∼95 kDa. Lanes 4 and 6 show a single band in cell lysates at ∼100 kDa for mouse ACE2, and ∼120 kDa for human ACE2, respectively. Lanes 2 and 5 show no ACE2 bands detected on immunoblots of both media and cell lysates from untransfected ACE2 KO cells. (B) Increased ACE2 activity in the media from ACE2 KO cells transfected with a human ACE2 expression vector (HA-hACE2, 3.75 µg on 35 mm culture dishes). Untransfected cells and cells transfected with an empty pcDNA3 vector had no detectable ACE2 activity in the media. Numbers in parentheses represent mean values for ACE2 activity. *P<0.001 vs untransfected control or empty pcDNA3 vector, n = 4.

Figure 3. Deglycosylation of ACE2 protein in media and cell lysates from mouse PT cells.

Representative immunoblot for ACE2 treated without (−) or with (+) deglycosylation with PNGase F in the media (Lanes 1–2) and cell lysates (Lanes 3–4). Lanes 1 and 3: wildtype PT cells, Lanes 2 and 4: ACE2 knockout (KO) PT cells transfected with a human ACE2 vector, Lane 5: mouse kidney cortex. Lanes 1+ and 2+ show a reduction in the sizes of ACE2 fragments in media fractions to ∼75 kDa and ∼60 kDa for mouse ACE2, and to ∼80 kDa and ∼65 kDa for human ACE2, respectively. Lanes 3+ and 4+ show a reduction in the sizes of ACE2 in cell lysates to ∼85 kDa for both mouse and human ACE2 treated with the PNGase F, respectively. Lane 5+ shows a reduction in size of ACE2 in mouse cortex from ∼100 kDa to ∼85 kDa after treatment with PNGase F.

Figure 4. Identification of ACE2 peptides by LC-MS/MS in the 75 kDa deglycosylated band from the media of mouse PT cells.

(A) Mass spectrometry identified peptides matched with the mouse ACE2 sequence (SWISS-PROT database no. Q8R0I0) at positions 18-706. The identified peptide sequences significantly matched with mouse ACE2 precursor in the database are underlined and shown in bold. The peptides matched with mouse ACE2 sequence, but not statistically significant, are underlined only (see Table S2 for detailed analyses of peptides). The overall matched sequences cover 32% of mouse ACE2 sequences. (B) Tandem mass spectrum of the C-terminal peptide (SEVEDAIRM 698–706). The b and y ions result from the cleavage of peptide bonds and correspond to N-terminal and C-terminal fragments of the peptide, respectively. The detected b and y ions are consistent with the peptide sequence shown on the top part of the panel. Ions score: 50, Observed ion: 533.25, Mr(expt): 1064.48, Mr(calc): 1064.48. Expect value: 0.002 (see Table S2 for details). The amino acid Met⁷⁰⁶ detected is from non-tryptic cleavage, suggesting a cleavage site for the ectodomain shedding of mouse ACE2 into the media.

Figure 5. Identification of ACE2 peptides by LC-MS/MS in the 60 kDa deglycosylated band from the media of mouse PT cells.

(A) Mass spectrometry identified peptides matched with the mouse ACE2 sequence (SWISS-PROT database no. Q8R0I0) at positions 18-577. The identified peptides significantly matched with mouse ACE2 sequences are underlined and shown in bold. The peptides matched with mouse ACE2 sequence, but not statistically significant, are underlined only (see Table S3 for detailed analyses of peptides). The overall sequence coverage is 24%. (B) Tandem mass spectrum of the C-terminal peptide (ALENVVGAR 569-577). The b and y ions result from the cleavage of peptide bonds and correspond to N-terminal and C-terminal fragments of the peptide, respectively. The detected b and y ions are consistent with the peptide sequence shown on the top part of the panel. Ions score: 59, Observed ion: 464.77, Mr(expt): 927.52, Mr(calc): 927.51. Expect value: 0.0002 (see Table S3 for details).

Figure 6. Diagram of 2 shed ACE2 fragments and putative C-terminal cleavage sites determined by LC-MS/MS.

A) Partial amino acid sequence of 75 kDa shed ACE2 fragment (deglycosylated) is shown, beginning with N-terminal amino acid Gln¹⁸ (Q). The first and last amino acids of the C-terminal and N-terminal fragments are underlined, where numbers below the line indicate their amino acid positions. The dashed line represents amino acids between position 40 and 701 (sequence not shown). The most C-terminal amino acid observed is shown in bold. Vertical black bars represent sites of tryptic cleavage. Red vertical bar is non-tryptic cleavage site at Met⁷⁰⁶ (M), a putative ACE2 cleavage site. B) Partial amino acid sequence of 60 kDa shed ACE2 fragment (deglycosylated) is shown, beginning with N-terminal amino acid Gln¹⁸ (Q). The first and last amino acids of the C-terminal and N-terminal fragments are underlined, where numbers below the line indicate their amino acid positions. The dashed line represents amino acids between position 40 and 571 (sequence not shown). The most C-terminal amino acid observed is shown in bold. Vertical black bars indicate sites of tryptic cleavage, and include Arg⁵⁷⁷ (R) and Lys⁵⁹⁶ (K). Cleavage site for this fragment may occur at or C-terminal to Arg⁵⁷⁷.

Figure 7. ACE2 shedding is stimulated by high D-glucose or Ang II in mouse PT cells.

Primary cultures of mouse PT cells were incubated for 24, 48 and 72(C, 7.8 mM D-glucose), with Ang II (10⁻⁷ M) or in high D-glucose (D-G, 25 mM) media. As a control for osmolality, some cells were incubated with L-glucose (L-G, 25 mM). ACE2 activity in the media was assayed at each time point. *p<0.001 vs L-G, p<0.004 vs C, p<0.025 vs Ang II, all at 48 hrs, n = 7–9. **p<0.001 vs all 3 other groups at 72 hrs, n = 14–18. ^#p< 0.04 vs C at 72 hrs, n = 14–18.

Figure 8. Role of matrix metalloproteinases (MMP) and ADAM17 in high glucose- or Ang II-stimulated ACE2 shedding.

(A) Effect of MMP inhibitor GM6001 on ACE2 activity in the media. Mouse PT cells were incubated for 72 hrs in high D-glucose (D–G, 25 mM) media in the presence or absence of GM6001 (GM, 5×10⁻⁵ M). *p<0.001 vs C and GM, **p<0.025 vs C and GM, n = 5. (B) Effect of the ADAM17 inhibitor TAPI-1 on high glucose-stimulated ACE2 activity in the media. Mouse PT cells were incubated for 72 hrs in high D-glucose (D–G, 25 mM) media in the presence or absence of TAPI-1 (10⁻⁷–10⁻⁵ M). *p<0.001 vs C, TAPI-1 10⁻⁵ M, and D-G+TAPI-1 10⁻⁵ M. **p<0.001 vs C and TAPI-1 10⁻⁵ M, p>0.05 vs D-G,D-G+TAPI-1 10⁻⁷ M and D-G+TAPI-1 10⁻⁵ M. n = 4–9. (C) Effect of TAPI-1 on Ang II-stimulated ACE2 activity in the media. Mouse PT cells were incubated for 72 hrs with Ang II (10⁻⁷ M) in the presence or absence of TAPI-1 (10⁻⁵ M). *p<0.015 vs C, p<0.03 vs TAPI-1, n = 8. (D) Effect of TAPI-2 on high D-glucose-stimulated ACE2 activity in the media. Mouse PT cells were incubated for 72 hrs in high D-glucose (D-G, 25 mM) media in the presence or absence of TAPI-2 (10⁻⁶– 5×10⁻⁵ M). *p<0.001 vs C and TAPI-2 (5×10⁻⁵ M), p<0.015 vs D-G+TAPI-2 (5×10⁻⁵ M). **p<0.005 vs C, p<0.015 vs TAPI-2 (5×10⁻⁵ M), p>0.05 vs D-G and D-G+TAPI-2 (5×10⁻⁵ M). n = 4.

Figure 9. Effect of high D-glucose or Ang II on ADAM17 activity in cell lysates.

Mouse PT cells were incubated for 72(10⁻⁷ M) or in high D-glucose (D-G, 25 mM) media. *p<0.05 vs C, n = 7-10.

Figure 10. Immunoblot analysis of ACE2 protein in the media from high glucose- or Ang II-stimulated mouse PT cells.

(A) Mouse PT cells were incubated for 72 hrs in normal media (C, 7.8 mM D-glucose), with or without Ang II (10⁻⁷ M), high D-glucose (D-G, 25 mM), or high L-glucose (25 mM). Above graph is representative immunoblot for ACE2 in the media, showing bands at ∼90 kDa and ∼70 kDa. (B) Graphical representation of densitometry analysis of two ACE2 bands on immunoblots. For the ∼90 kDa band, *p<0.05 vs C, **p<0.001 vs C, **p<0.003 vs L-G; n = 5. For the ∼70 kDa band, *p<0.04 vs C; **p<0.001 vs C or L-G, **p<0.03 vs Ang II; n = 5.