Substrate specificity of human metallocarboxypeptidase D: Comparison of the two active carboxypeptidase domains

Abstract

Metallocarboxypeptidase D (CPD) is a membrane-bound component of the trans-Golgi network that cycles to the cell surface through exocytic and endocytic pathways. Unlike other members of the metallocarboxypeptidase family, CPD is a multicatalytic enzyme with three carboxypeptidase-like domains, although only the first two domains are predicted to be enzymatically active. To investigate the enzymatic properties of each domain in human CPD, a critical active site Glu in domain I and/or II was mutated to Gln and the protein expressed, purified, and assayed with a wide variety of peptide substrates. CPD with all three domains intact displays >50% activity from pH 5.0 to 7.5 with a maximum at pH 6.5, as does CPD with mutation of domain I. In contrast, the domain II mutant displayed >50% activity from pH 6.5-7.5. CPD with mutations in both domains I and II was completely inactive towards all substrates and at all pH values. A quantitative peptidomics approach was used to compare the activities of CPD domains I and II towards a large number of peptides. CPD cleaved C-terminal Lys or Arg from a subset of the peptides. Most of the identified substrates of domain I contained C-terminal Arg, whereas comparable numbers of Lys- and Arg-containing peptides were substrates of domain II. We also report that some peptides with C-terminal basic residues were not cleaved by either domain I or II, showing the importance of the P1 position for CPD activity. Finally, the preference of domain I for C-terminal Arg was validated through molecular docking experiments. Together with the differences in pH optima, the different substrate specificities of CPD domains I and II allow the enzyme to perform distinct functions in the various locations within the cell.

Fig 1. Linear representation of full-length human CPD and recombinant forms showing the location of single point mutations.

The positions indicated in human CPD correspond to key residues essential for the catalytic mechanism: His69, Glu72, Arg145, His198, Tyr248, and Glu270 (according to the bCPA1 numbering). Sp, indicates the location of the endogenous signal peptide on the N-terminus. Recombinant proteins correspond to C-terminal truncated forms of human CPD, which lack the C-terminal transmembrane anchor and the cytoplasmic tail. The mutations (i.e., Glu to Gln) performed to generate single point mutants for the CPD domain I (E350Q), domain II (E762Q) and a double mutant (E350Q/E762Q) are indicated. The black shaded box on the N-term of the recombinant proteins corresponds to the IgM secretion signal sequence and N-terminal Strep-Tag II fusion protein.

Fig 2. Expression and purification of CPD.

(A) Schematic diagram of the strategy for expression and purification of rhCPD and mutants. Protein expression was performed by transient transfection of suspension-growing HEK293F cells. Extracellular medium was collected after 7 days incubation, followed by purification of the recombinant proteins in three steps; (1) hydrophobic interaction chromatography using a Butyl 650-M, (2) affinity chromatography using anti-Strep-tag resin, and (3) gel filtration chromatography. (B) Coomassie-stained SDS–PAGE showing the purity of recombinant CPD proteins.

Fig 3. Effect of pH on the activity of different recombinant CPD forms.

(A) Effect of pH on rhCPD and (B) single point mutants using 200 μM dansyl-Phe-Ala-Arg in a Tris-acetate buffer at the indicated pH for 60 min at 37°C. The activity represented is the average of three independent measures with less than a 10% of variation, expressed as a percentage of the maximal activity, observed at optimal pH.

Fig 4. Quantitative peptidomics scheme for the characterization of rhCPD substrate specificity using the tryptic peptide library and representative spectra.

(A) Quantitative peptidomics scheme. Tryptic peptides were obtained from digestion of five selected proteins (BSA, bovine thyroglobulin, bovine α-lactalbumin and human α and β–hemoglobin) with trypsin. The resultant peptide library was aliquoted and digested with no enzyme or different rhCPD concentrations of 0.1, 1, 10, and 100 nM for 16 h at 37°C. After incubation samples were labeled with one of five stable isotopic TMAB-NHS tags (D0 = 100 nM; D3 = 10 nM; D6 = 1 nM; D9 = 0.1 nM; D12 = No enzyme). Then, samples were pooled and analyzed by LC-MS. Examples of representative data are shown for (B) non-substrates, (C) good substrates, (D) weak substrates and (E) products.

Fig 5. Analysis of the substrate preferences of rhCPD using the tryptic peptide library.

(A) Substrate preferences of rhCPD at C-terminal (P1’) and (B) P1 positions. The number of times each amino acid was present in P1 or P1’ was determined for good substrates, weak substrates and non-substrates and represented. For P1 analysis, only substrates with permissive P1’ residues (Lys or Arg) were considered.

Fig 6. Quantitative peptidomics scheme for the substrate characterization of CPD domains I and II using a tryptic peptide library and representative spectra.

(A) Quantitative peptidomics scheme. Tryptic peptides were obtained as described above. The resultant peptide library was aliquoted and digested with 100 nM rhCPD, 100 nM rhCPD E350Q, 100 nM rhCPD E762Q or 100 nM rhCPD E350Q/E762Q for 16 h at 37°C. Then, samples were labeled with one of the isotopic TMAB-NHS tags. For the representative data shown in this figure, D3 = rhCPD; D6 = rhCPD E762Q; D9 = rhCPD E350Q; and D12 = rhCPD E350Q/E762Q. Finally, samples were pooled and analyzed by LC-MS. Examples of representative data are shown for (B) good substrates of both domains I and II, (C) preferential substrates of domain II and (D) preferential substrates of domain I. The entire experiment was performed three times and the various enzyme digests were labelled with different tags to control for variability of the individual reagents.

Fig 7. Analysis of the substrate preferences of CPD domains I and II using a tryptic peptide library.

(A,B) Substrate preferences of CPD domain I (i.e., rhCPD E762Q). (A) C-terminal P1’ position preferences. (B) P1 position preferences. (C,D) Substrate preferences of CPD domain II (i.e., rhCPD E350Q). (C) C-terminal P1’ position preferences. (D) P1 position preferences. The number of times each amino acid was present in P1 or P1’ was determined for good substrates, weak substrates and non-substrates. For P1 analysis, only substrates with permissive P1’ residues (Lys, Arg) were considered.

Fig 8. Quantitative peptidomics scheme for the study of rhCPD substrate specificity using HEK293T derived peptide library and representative spectra.

(A) Quantitative peptidomics scheme. Peptides were extracted from HEK293T cell cultures treated for 1 h at 37°C with 0.5 μM bortezomib. The resultant peptide extract (i.e., HEK293T peptidome) was aliquoted and digested with no enzyme or different rhCPD concentrations (i.e., 0.1, 1, 10, and 100 nM at 37°C for 16 h). After incubation samples were labeled with one of five stable isotopic TMAB-NHS tags (D0 = No enzyme; D3 = 1 nM; D6 = 10 nM; D9 = 100 nM; and D12 = 0.1 nM). Then, samples were pooled and analyzed by LC-MS. Examples of representative data are shown for (B) non-substrates, (C) good substrates, (D) weak substrates and (E) products.

Fig 9. Analysis of the P1’ substrate preferences of rhCPD using a HEK293T derived peptidomics library.

The number of times each amino acid was present in P1’ was determined for good substrates, weak substrates and non-substrates.

Fig 10. Structural modeling of the active sites of CPD domains I, II and III.

Electrostatic potential molecular surfaces of the catalytic sites of CPD domains I (A), II (B) and III (C) in presence of a modeled peptide. Lower panels show the best docking pose based on GlideScore for the truncated peptide GQKR (green sticks) within the active sites of domains I (D), II (E), and III (F), with the two last N-terminal residues omitted for clarity. Zn coordinates (yellow sphere) are added from the template structure used for modeling (PDB 1H8L). Hydrogen bonds are depicted in black dashed lines, π-cation interactions in pink and metal coordination in orange dashed lines. Within lower panels, the side chains of residues directly involved in the zinc binding are shown (i.e., His69, Glu72 and His196 using bCPA1 numbering; corresponding to His139/564, Glu142/567, and His257/671 in CPD domains I/ II), as well as the catalytic residue Glu270 (Glu350/762 in CPD domains I/II). In addition, the side chains of some of the putative residues determining respectively the S1’ and S1 specificity pockets are shown for domains I/ I; i.e., Arg218/635 (corresponding to Arg145 of bCPA1 numbering); Tyr328/740 (Tyr248 of bCPA1); and Gln335/747 (Ile255 of bCPA1). See S3 Fig for other residue equivalences.