Purification and characterization of pig kidney aminopeptidase P. A glycosyl-phosphatidylinositol-anchored ectoenzyme

Sequence and structure comparison suggest that methionine aminopeptidase, prolidase, aminopeptidase P, and creatinase share a common fold

Amino acid sequence comparison suggests that the structure of Escherichia coli methionine aminopeptidase (EC 3.4.11.18) and the C-terminal domain of Pseudomonas putida creatinase (EC 3.5.3.3) are related. A detailed comparison of the three-dimensional folds of the two enzymes confirms this homology: with an approximately 260-residue chain segment, 218 C alpha atoms of the structures superimpose within 2.5 A; only 41 of these overlapping positions (i.e., 19%) feature identical amino acids in the two protein chains. Notwithstanding this striking correspondence in structure, methionine aminopeptidase binds and is stimulated by Co2+, while creatinase is not a metal-dependent enzyme. Searches of protein data banks using sequence and structure-based profiles reveal other enzymes, including aminopeptidase P (EC 3.4.11.9), prolidase (EC 3.4.13.9), and agropine synthase, that likely share the same "pita-bread" fold common to creatinase and methionine aminopeptidase.

Characterization of an antibody to the cross-reacting determinant of the glycosyl-phosphatidylinositol anchor of human membrane dipeptidase

A polyclonal antiserum raised to the phospholipase C-solubilized form of membrane dipeptidase (EC 3.4.13.11) purified from human kidney was found to cross-react with unrelated trypanosomal and porcine glycosyl-phosphatidylinositol anchored proteins. Those antibodies recognising the cross-reacting determinant (CRD) were isolated by chromatography on a column of immobilized phospholipase C-solubilized porcine aminopeptidase P (EC 3.4.11.9), and the epitopes involved in the recognition were then characterized by immunoelectrophoretic blot analysis and by a competitive ELISA. The phospholipase C-solubilized forms of human and porcine membrane dipeptidase, porcine aminopeptidase P and trypanosome variant surface glycoprotein were recognised by the anti-CRD antiserum, and this recognition was abolished by prior treatment of the proteins with either mild acid or nitrous acid. In contrast, the detergent-solubilized, membrane-forms of human and porcine membrane dipeptidase were not recognised. Of a range of components of the glycosyl-phosphatidylinositol anchor, only inositol 1,2-cyclic monophosphate and the insulin-mimetic disaccharide, glucosaminyl-1,6-inositol 1,2-cyclic monophosphate, inhibited in the micromolar range the binding of the anti-CRD antiserum to immobilized porcine aminopeptidase P. These results indicate that the major epitope recognised by this anti-CRD antiserum is the inositol 1,2-cyclic monophosphate formed on phospholipase C cleavage of the glycosyl-phosphatidylinositol anchor.

Inhibition of aminopeptidases N, A and W. A re-evaluation of the actions of bestatin and inhibitors of angiotensin converting enzyme

The effects of a range of metallopeptidase inhibitors on the activities of the porcine kidney cell surface zinc aminopeptidases, aminopeptidase A (AP-A; EC 3.4.11.2), aminopeptidase N (AP-N; EC 3.4.11.7) and aminopeptidase W (AP-W; EC 3.4.11.16), have been directly compared. Amastatin and probestin were effective against all three aminopeptidases, with the concentration of inhibitor required to cause 50% inhibition (I₅₀) in the low micromolar range (I₅₀ = 1.5–20 μM), except for probestin with AP-N which displayed an I₅₀ of 50 nM. Actinonin failed to inhibit significantly either AP-A or AP-W, and thus can be considered a relatively selective inhibitor (I₅₀ = 2.0 μM of AP-N. In contrast, bestatin was a relatively poor inhibitor of AP-N (I₅₀ = 89 μm) and failed to inhibit AP-A, but was more potent towards AP-W (I₅₀ = 7.9 μM). Thus, some of the observed chemotherapeutic actions of bestatin may be due to inhibition of cell-surface AP-W. A number of other metallopeptidase inhibitors, including inhibitors of endopeptidase-24.11 (EC 3.4.24.11) and membrane dipeptidase (EC 3.4.13.11), and the carboxylalkyl and phosphoryl inhibitors of angiotensin converting enzyme (EC 3.4.15.1) failed to inhibit significantly AP-A, AP-N or AP-W. However, AP-W was inhibited with I₅₀ values in the micromolar range by the sulphydryl converting enzyme inhibitors rentiapril (I₅₀ = 1.6 μM), zofenoprilat (I₅₀ = 7.0 μM) and YS 980 (I₅₀ = 17.7 μM). Neither AP-A nor AP-N were affected by these sulphydryl compounds. Inhibition of AP-W may account for some of the side effects noted with the clinical use of the sulphydryl converting enzyme inhibitors. The availability of compounds which are totally selective for AP-W over any of the other mammalian cell surface zinc aminopeptidases may aid in identifying endogenous substrates, and thus physiological or pathophysiological role(s) of AP-W.

Aminopeptidase P from human leukocytes

Cytosolic aminopeptidase P was obtained in highly purified form from human leukocytes by a four-step procedure. Buffy coats were the starting material. A M(r) of 140,000 was obtained by size-exclusion HPLC for the native enzyme. As shown by SDS/PAGE under reducing and denaturing conditions, the enzyme consisted of likely identical subunits with M(r) of 71,000. Purified aminopeptidase P cleaved off, specifically and efficiently, the N-terminal residues from peptides with N-terminal Xaa-Pro sequences. The penultimate proline was not replaceable by hydroxyproline, alanine and glycine in di-, tri- and tetrapeptides. Polyproline was not hydrolyzed. Dipeptides were cleaved (Arg-Pro, Phe-Pro > Trp-Pro > Pro-Pro) although slower than longer peptides. Cleavage was observed of several biologically active peptides; C-terminal fragment (residues 201-206) of C-reactive protein, oxytocin fragment Tyr-Pro-Leu-Gly, morphiceptin, peptide Gly-Pro-Arg-Pro (inhibitor of fibrin polymerization) and kentsin. In addition, cleavage of a protein, interleukin-6, was also demonstrated. Aminopeptidase P was maximally activated by Mn2+, and to a lesser extent by Co2+. The activity was optimal at pH 8. Ni2+, Zn2+ and especially Cd2+ caused marked inhibition. EDTA, 1,10-phenantroline and dithiothreitol were also inhibitory. Carbobenzoxy-phenylalanine, as well as several N-carbobenzoxy-proline-containing peptides, caused partial inhibition. The observed resistance of Gly-Pro, Pro-Gly, Pro-Phe and Pro-Ile to hydrolysis by the purified enzyme strongly indicates absence of known proline-specific dipeptidases in the aminopeptidase-P preparation.

Kininase activity in human platelets: cleavage of the Arg1-Pro2 bond of bradykinin by aminopeptidase P

A proline-specific peptidase aminopeptidase P (APP, EC 3.4.11.9) that cleaves the Arg1-Pro2 bond of bradykinin was isolated from human platelets by liquid chromatography. The enzyme was purified 557 times. The native molecule has a M(r) of 223,000. Human platelet APP exists as a trimer with a subunit M(r) of 71,000. The apparent Km of platelet APP is 66 mumol/L for bradykinin and 47 mumol/L for the internally quenched fluorogenic substrate Lys (2,4-dinitrophenyl)-Pro-Pro-NH-CH2-CH2-NH-2-aminobenzoyl. 2HCl which is used for the routine determination of the enzyme activity. The optimum pH for hydrolysis of the fluorogenic substrate is 8.0, and the optimum temperature is 43 degrees. Platelet APP is inhibited by 1,10-phenanthroline and activated by Mn2+, thus confirming its metalloprotease nature. Cu2+, Zn2+ and Hg2+ are strongly inhibitory. Inhibition by cysteine protease inhibitors suggests the presence of a thiol group essential for enzymatic activity. Serine protease inhibitors do not affect the enzyme activity.

A survey of membrane peptidases in two human colonic cell lines, Caco-2 and HT-29

The expression of cell-surface peptidases was examined in two human colon carcinoma cell lines, Caco-2 and HT-29. Enzymic assays revealed the presence of eight cell-surface peptidases on a Caco-2 cell line (passage number 82-88), namely aminopeptidase N, dipeptidyl peptidase IV, peptidyl dipeptidase A (angiotension-converting enzyme), aminopeptidase P, aminopeptidase W, endopeptidase-24.11, gamma-glutamyl transpeptidase and membrane dipeptidase. The presence of dipeptidyl peptidase IV and endopeptidase-24.11 was also confirmed immunochemically. After 15 days culture, the activities of aminopeptidase P, peptidyl dipeptidase A and alkaline phosphatase activities on Caco-2 cells reached a plateau, and that of membrane dipeptidase began to decline. In contrast, aminopeptidase N, dipeptidyl peptidase IV and endopeptidase-24.11 activities were still rising after 26 days in culture. Caco-2 cells of passage number 181-183 were found to lack endopeptidase-24.11, but maintained dipeptidyl peptidase IV expression. Two populations of HT-29 cells were surveyed. Both the standard, undifferentiated population and a differentiated population expressed only three peptidases: dipeptidyl peptidase IV, aminopeptidase W and carboxypeptidase M. In the differentiated HT-29 cells the activity of dipeptidyl peptidase IV after 14-21 days was beginning to plateau whereas aminopeptidase W activity was still rising and that of carboxypeptidase M had begun to decline. These differences in activity profiles observed among this group of cell-surface peptidases indicate that these cell lines, especially Caco-2, are useful models to study the regulation of their expression.

Inhibition by converting enzyme inhibitors of pig kidney aminopeptidase P

Several inhibitors of angiotensin converting enzyme were also found to inhibit aminopeptidase P, whereas inhibitors of other mammalian aminopeptidases were ineffective. Aminopeptidase P purified from pig kidney cortex was found to contain one atom of zinc per polypeptide chain, confirming its metalloenzyme nature. The concentrations of converting enzyme inhibitors required to cause 50% inhibition (I50) of aminopeptidase P were in the low micromolar range. The most potent converting enzyme inhibitors toward aminopeptidase P were the carboxylalkyl compounds, cilazaprilat, enalaprilat, and ramiprilat (I50 values of 3-12 microM). The sulfhydryl compounds captopril (I50 110 microM) and YS980 (I50 20 microM) were slightly less potent at inhibiting aminopeptidase P. In contrast, the carboxylalkyl compounds benazeprilat, lisinopril, and pentoprilat; the sulfhydryl compound rentiapril; and the phosphoryl compounds ceranopril and fosinoprilat had no inhibitory effect against aminopeptidase P. This compares with I50 values in the 1-6 nM range for these inhibitors with angiotensin converting enzyme. Inhibition of aminopeptidase P may account for some of the effects or side effects noted with the clinical use of converting enzyme inhibitors. These results may provide the basis for the design of more selective inhibitors of angiotensin converting enzyme or mixed inhibitors of aminopeptidase P and angiotensin converting enzyme, or both.

Purification and characterization of guinea pig serum aminoacylproline hydrolase (aminopeptidase P)

Aminoacylproline hydrolase (EC 3.4.11.9) of guinea pig serum has been obtained as two apparently homogeneous isoforms. Dialyzed serum was chromatographed successively on Affi-gel blue, hydroxyapatite, DE-cellulose, phenyl-Sepharose, an affinity matrix for angiotensin converting enzyme and concanavalin-Sepharose. On the latter matrix, 68% of the enzyme activity was eluted with alpha-methyl mannoside at 10 and 100 mM, and 29% was eluted with alpha-methyl glucoside, 500 mM, at 56 degrees C. The two fractions ('biantennary' and 'high mannose' fractions, respectively) were concentrated and then chromatographed separately on Sephacryl S-200HR. Both fractions were eluted as expected for a globular protein of Mr 217,000. On SDS-PAGE, under reducing and non-reducing conditions, each of the concanavalin-Sepharose fractions was separated into two protein bands, Mr 89,000 and Mr 81,500. Each of the bands was found to be N-blocked when N-terminal amino acid sequencing was attempted. The reaction of the 'biantennary' fraction with the synthetic substrate Arg-Pro-Pro-[3H]benzylamide was characterized in part: Km 0.7 microM, kcat 124.6 min-1, kcat/Km 1.78.10(8) M-1 min-1. Hydrolysis of the substrate was strongly inhibited by bradykinin and those of its lower homologs that contain two adjacent proline residues. Cu2+ was strongly inhibitory. Co2+ at 30 microM activated the enzyme, as did Mn2+, Mg2+ and Ca2+ at higher concentrations. Sulfhydryl compounds, including captopril, inhibited the enzyme as did 1,10-phenanthroline. Iodoacetamide and N-ethylmaleimide had no effects, but 4-hydroxymercuribenzoate conferred a partial inhibition over a remarkably wide concentration range: 0.34-1400 microM. Amastatin and bestatin did not inhibit the enzyme. Aminoacylproline hydrolase of guinea pig serum appears to be a heterogeneous, glycosylated metallo-enzyme with a high affinity for bradykinin and related peptides in which the sequence Pro-Pro, Xaa-Pro-Pro or Xaa-Pro-Hyp is N-terminal.

Aminopeptidase P: purification of a membrane-bound bradykininase from rat lung

Aminopeptidase P that hydrolyzes the Arg1-Pro2-bond of bradykinin was solubilized from rat lung microsomes using phosphatidylinositol-specific phospholipase C. The enzyme was purified 420-fold by chromatography on decylagarose (two steps), omega-aminodecyl-agarose and DEAE-Sephacel. A single stained band was observed following native gradient (4-15%) polyacrylamide gel electrophoresis. Dipeptidylaminopeptidase IV-like activity was also present in the final preparation and co-migrated with aminopeptidase P in the above gel system.

A survey of vasoactive peptide metabolizing enzymes in the rat mesenteric arterial bed perfusate

We have demonstrated that the isolated perfused rat mesenteric arterial bed (MAB) secretes peptidases capable of metabolizing bradykinin and angiotensin I. The major degradative pathway of bradykinin by enzymes found in the rat MAB perfusate was mediated by carboxypeptidase A-like activity, whereas angiotensin 1 degradation followed two main routes, one attributable to a carboxypeptidase A-like enzyme and the other to an endopeptidase. This latter enzyme seems to be a novel serine peptidase capable of releasing angiotensin II directly from both angiotensin I and renin substrate tetradecapeptide. The rat MAB perfusate was also shown to contain additional endo- and exopeptidases that might play a role in the metabolism of other vasoactive peptides. Our finding that isolated rat MAB secretes peptidases into the perfusion medium indicates that peptide processing within the microvasculature environment may be effected by enzymes besides those normally found in plasma or associated with cell membranes.

Metabolism of bradykinin agonists and antagonists by plasma aminopeptidase P

In addition to angiotensin I converting enzyme (ACE; EC 3.4.15.1) and carboxypeptidase N (CPN; EC 3.4.17.3), other peptidases contribute to bradykinin (BK) degradation in plasma. Rat plasma degraded BK by hydrolysis of the N-terminal Arg1-Pro2 bond, and the characteristics of hydrolysis are consistent with identification of aminopeptidase P (APP; EC 3.4.11.9) as the responsible enzyme. BK and BK[1-5] N-terminal hydrolysis was optimal at neutral pH, was inhibited by 2-mercaptoethanol, dithiothreitol, o-phenanthroline and EDTA, but was unaffected by the aminopeptidase inhibitors amastatin, puromycin and diprotin A, the endopeptidase-24.11 inhibitors phosphoramidon and ZINCOV, and the ACE and CPN inhibitors captopril and D,L-mercapto-methyl-3-guanidinoethylthiopropanoic acid (MERGETPA), respectively. Although kallidin (Lys-BK) was not metabolized directly by APP, conversion to BK by plasma aminopeptidase M (EC 3.4.11.2) resulted in subsequent degradation by APP. BK analogs containing N-terminal Arg1-Pro2 bonds, including [Tyr8-(OMe)] BK and [Phe8 psi(CH2NH)Arg9]BK (B2 agonists), des-Arg9-BK and [D-Phe8]des-Arg9-BK (B1 agonists), and [Leu8]des-Arg9-BK (B1 antagonist), were degraded by APP with Km and Vmax values comparable to those found for BK (Km = 19.7 +/- 2.6 microM; Vmax = 12.1 +/- 1.2 nmol/min/mL). In contrast, B2 antagonists containing D-Arg0 N-termini, including D-Arg[Hyp3,Thi5.8,D-Phe7]BK and D-Arg[Hyp3,D-Phe7,Phe8 psi(CH2NH)Arg9]BK, were resistant to APP-mediated hydrolysis. These data support a role for plasma aminopeptidase P in the degradation of circulating kinins, and a variety of B2 and B1 kinin agonists and antagonists. However, APP does not participate in the degradation of D-Arg0-containing antagonists.

Aminopeptidase P from rat brain. Purification and action on bioactive peptides

Aminopeptidase P (EC 3.4.11.9) was purified from rat brain cytosol. A subunit Mr of 71,000 was determined for the reduced, denaturated protein whereas an Mr of 143,000 was determined for the native enzyme. The purified aminopeptidase P selectively liberated all unblocked, preferentially basic or hydrophobic ultimate amino acids from di-, tri- and oligopeptides with N-terminal Xaa-Pro- sequences. Corresponding peptides with penultimate Ala instead of Pro were cleaved with much lower rates; oligopeptides with residues other than Pro or Ala in the penultimate position appeared not to be substrates for the enzyme. Several bioactive peptides with Xaa-Pro sequences, especially bradykinin, substance P, corticortropin-like intermediate lobe peptide, casomorphin and [Tyr]melanostatin were shortened by the N-terminal amino acid by aminopeptidase P action. Rat brain aminopeptidase P was optimally active at pH 7.6-8.0 in the presence of Mn2+. Chelating agents and SH-reacting reagents inhibited the enzyme, but common inhibitors of aminopeptidases, like amastatin or bestatin, of prolidase or of dipeptidyl peptidases II and IV, like N-benzoyloxycarbonyl-proline or epsilon-benzyl-oxycarbonyl-lysyl-proline, as well as antibiotics like beta-lactam ones, bacitracin or puromycin, had little or no effect.

Aminopeptidase P and dipeptidyl peptidase IV activity in human leukocytes and in stimulated lymphocytes

Human white blood cells were shown to contain high aminopeptidase P activity. The specific activities found in the high-speed supernatant of the extracts of granulocytes, lymphocytes and monocytes ranged from 30 to 70 units per mg protein. Culturing lymphocytes during 7 days in the presence of phytohaemagglutinin resulted in a 70-200% increase in the specific aminopeptidase P activity and a 200% increase in the specific activity of dipeptidyl peptidase IV. The time-course of the activity of both aminopeptidase P and dipeptidyl peptidase IV during the stimulation of human T-lymphocytes by phytohaemagglutinin indicates an involvement of these two enzymes in the proliferative process of these immunocompetent cells. Due to their substrate specificity their potential substrates must have the N-terminal Xaa-Pro sequence known to be present in several immunologically important polypeptides.

Characterization of antibodies to the glycosyl-phosphatidylinositol membrane anchors of mammalian proteins

Two polyclonal antisera were raised in rabbits to the phospholipase C-solubilized forms of pig renal dipeptidase (EC 3.4.13.11) and pig aminopeptidase P (EC 3.4.11.9). These antisera were purified and shown to cross-react with other glycosyl-phosphatidylinositol (G-PI)-anchored proteins isolated from pig, human and trypanosomes. The epitopes involved in this cross-reactivity were characterized by Western-blot analysis after mild acid or nitrous acid treatment of the G-PI-anchored proteins and by a competitive e.l.i.s.a. with other G-PI-anchored proteins and individual components of the anchor structure. These studies revealed that the primary epitope for both antisera is the inositol 1.2-(cyclic)monophosphate that is formed on phospholipase C cleavage of the intact G-PI anchor. Other minor epitopes, such as phosphoethanolamine, probably involve side-chain modifications to the core anchor structure that may be species-specific.

cDNA cloning and expression in Xenopus laevis oocytes of pig renal dipeptidase, a glycosyl-phosphatidylinositol-anchored ectoenzyme

Clones expressing renal dipeptidase (EC 3.4.13.11) have been isolated from a pig kidney cortex cDNA library after employing the polymerase chain reaction technique to amplify a region of the dipeptidase cDNA. The complete primary sequence of the enzyme has been deduced from a full length cDNA clone. This predicts a protein of 409 amino acids, a cleavable N-terminal signal sequence of 16 residues and two N-linked glycosylation sites. At the C-terminus of the predicted sequence is a stretch of mainly hydrophobic amino acids which is presumed to direct the attachment of the glycosyl-phosphatidylinositol membrane anchor. Expression of the mRNA for pig renal dipeptidase in Xenopus laevis oocytes led to the production of a disulphide-linked dimeric protein of subunit Mr 48,600 which was recognized by a polyclonal antiserum raised to renal dipeptidase purified from pig kidney cortex. Bacterial phosphatidylinositol-specific phospholipase C released renal dipeptidase from the surface of the oocytes and converted the amphipathic detergent-solubilized form of the dipeptidase to a hydrophilic form, indicating that Xenopus laevis oocytes can process expressed proteins to their glycosyl-phosphatidylinositol anchored form.