help button home button Endocrine Society Endocrine Reviews
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Shrimpton, C. N.
Right arrow Articles by Lew, R. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Shrimpton, C. N.
Right arrow Articles by Lew, R. A.
Endocrine Reviews 23 (5): 647-664
Copyright © 2002 by The Endocrine Society

Soluble Metalloendopeptidases and Neuroendocrine Signaling

Corie N. Shrimpton, A. Ian Smith and Rebecca A. Lew

Baker Medical Research Institute (C.N.S., A.I.S., R.A.L.), Melbourne, Australia 8008; and Thrombosis Research Section (C.N.S.), Department of Medicine, Baylor College of Medicine, Houston, Texas 77030

Correspondence: Address all correspondence and requests for reprints to: Rebecca A. Lew and Ian Smith, Baker Medical Research Institute, PO Box 6492, St. Kilda Road Central, Melbourne, Australia 8008. E-mail: rebecca.lew{at}baker.edu.au and ian.smith{at}baker.edu.au


   Abstract
Top
 Abstract
I. Introduction
 II. Endopeptidase EC 3.4.24.15
 III. Endopeptidase EC 3.4.24.16
 IV. Endooligopeptidase A
 V. Functions of EP24.15...
 VI. Conclusions
 References
 
Peptidases play a vital and often highly specific role in the physiological and pathological generation and termination of peptide hormone signals. The thermolysin-like family of metalloendopeptidases involved in the extracellular processing of neuroendocrine and cardiovascular peptides are of particular significance, reflecting both their specificity for particular peptide substrates and their utility as therapeutic targets. Although the functions of the membrane-bound members of this family, such as angiotensin-converting enzyme and neutral endopeptidase, are well established, a role for the predominantly soluble family members in peptide metabolism is only just emerging. This review will focus on the biochemistry, cell biology, and physiology of the soluble metalloendopeptidases EC 3.4.24.15 (thimet oligopeptidase) and EC 3.4.24.16 (neurolysin), as well as presenting evidence that both peptidases play an important role in such diverse functions as reproduction, nociception, and cardiovascular homeostasis.

I. Introduction

II. Endopeptidase EC 3.4.24.15

A. Biochemical properties

B. Substrate specificity

C. Distribution and subcellular localization

III. Endopeptidase EC 3.4.24.16

A. Biochemical properties

B. Distribution and subcellular localization

IV. Endooligopeptidase A

V. Functions of EP24.15 and EP24.16

A. Reproductive axis

B. Pain perception—opioid processing

C. NT inactivation

D. Cardiovascular/renal homeostasis

E. Alzheimer’s disease

F. Antigen presentation

VI. Conclusions


   I. Introduction
Top
 Abstract
 I. Introduction
II. Endopeptidase EC 3.4.24.15
 III. Endopeptidase EC 3.4.24.16
 IV. Endooligopeptidase A
 V. Functions of EP24.15...
 VI. Conclusions
 References
 
BIOACTIVE PEPTIDES FUNCTION as chemical messengers acting via specific receptors on target cells to mediate cell-to-cell communication and thus integrate life processes in multicellular organisms. Many hormones, neurotransmitters, and growth factors are peptides, acting in autocrine, paracrine, or endocrine fashion. The production of bioactive peptides can be regulated at the transcriptional, translational, or posttranslational levels, leading to increased or decreased production and release from the cell. In addition, as there are no known reuptake systems for peptides, their levels in the extracellular space are also determined by the extent of degradation by specific proteolytic enzymes. Thus, peptidase-mediated proteolysis is critical for both the generation and termination of peptide signals.

Peptides are initially synthesized as inactive precursors (preprohormones) that require proteolytic processing to generate smaller active peptides (1) (Fig. 1Go). The peptides are first excised from the precursor by specific endoproteolysis directed toward basic residues (Arg, Lys) that typically flank the peptide in pairs (1, 2, 3). To date, eight distinct mammalian Kex2/subtilisin-related prohormone convertases [furin, PC 1/3, PC2, PC4, PACE4, PC5/6, PC7/8 (LPC), and SKI-1] have been identified (3, 4). Among these, PC1 and PC2 are the most relevant to neuroendocrine function as their expression is typically restricted to neuroendocrine cells (5). They are appropriately located in secretory granules, are optimally active at acidic pH, and cleave prohormones targeted to the constitutive and regulated secretory pathways. After endoproteolysis, the basic residue extensions at the carboxy terminus are removed by carboxypeptidase H (6). Regulatory peptides also undergo other posttranslational modifications including disulfide bond formation, glycosylation, sulfation, phosphorylation, N-acetylation, and C-terminal {alpha}-amidation (7, 8). Amidation is mandatory for the bioactivity of many peptides and, along with other posttranslational modifications, prolongs the half-life of many peptide messengers, protecting them from exopeptidase action in the extracellular space.



View larger version (21K):
[in this window]
[in a new window]
  		Figure 1. Schematic of intracellular and extracellular peptide processing events. Peptide hormones are initially synthesized as preprohormones, which are first processed within the secretory vesicle by signal peptidase, prohormone convertases, and carboxypeptidase E. Several alternative pathways exist for further processing; three such examples are depicted. In pathway 1, a glycine-extended precursor is amidated by peptidylglycine {alpha}-amidating monooxygenase (PAM) to form an active moiety that can then be secreted upon stimulation (examples include GnRH, vasopressin, gastrin, neuropeptide Y, among others). Most nonamidated peptides are also fully processed intracellularly and stored before secretion. In a second, less common scenario, an inactive precursor is stored, and excision of the active moiety is a cosecretory event (atrial natriuretic peptide, endothelin). A third possibility is that the peptide is secreted in a precursor form, and activating cleavages occur extracellularly, at a site remote from the site of secretion (e.g., angiotensin II). Inactivation of peptides occurs almost exclusively outside the cell (by membrane-bound ectoenzymes or by circulating or interstitial soluble enzymes), although the possibility exists for intracellular degradation after ligand-receptor internalization.

 
Peptide processing is not uniquely intracellular, but may continue in the extracellular environment. The enzymes involved in extracellular processing events are present as ectoenzymes on the cell surface or as soluble enzymes, with both playing an important role in regulating peptide-mediated cell-to-cell communication. Such enzymes can convert an inert precursor to an active moiety or cleave peptides to generate products unable to activate receptors. In some cases, the same peptidase may carry out both actions. For example, angiotensin-converting enzyme (ACE) converts angiotensin I to the active angiotensin II and contributes to the inactivation of bradykinin. In addition, such enzymes may also modulate the peptidergic signal by generating smaller biologically active fragments of the parent peptide with either a different receptor binding selectivity or a distinct biological activity. For example, Peptide E loses µ-opioid receptor affinity after extracellular proteolytic cleavage and loss of the amino-terminal tyrosine but maintains high {delta}-opioid receptor affinity (9).

Although extracellular examples of each of the four major classes of protease (aspartic, serine, cysteine, and metal- dependent) can be found, the majority of enzymes involved in extracellular peptide metabolism are metallopeptidases. Of particular significance are the thermolysin-like mammalian zinc metalloendopeptidases, which hydrolyze peptide bonds in substrates of less than 40 amino acids (the length of most bioactive peptides) and are maximally active at neutral pH. These metalloendopeptidases have been classified by Rawlings and Barrett (10) as belonging to Clan MA, and include ACE (EC 3.4.15.1), endothelin-converting enzyme (ECE; EC 3.4.24.71), neutral endopeptidase (NEP; EC 3.4.24.11), endopeptidase EC 3.4.24.15 (thimet oligopeptidase; EP24.15), and endopeptidase EC 3.4.24.16 (neurolysin; EP24.16). These enzymes require a zinc atom in the active site and their sequences contain the classic HEXXH motif (where X = any amino acid residue) typical of zinc metallopeptidases.

The physiological significance of the membrane-bound members of this family (ACE, NEP, and ECE) is well established, and the functions of these enzymes have been extensively reviewed (11, 12, 13, 14). In contrast, the role of metalloendopeptidases that do not contain a transmembrane region and are thus predominantly soluble is less clear. These peptidases are generally referred to as "soluble metalloendopeptidases," despite evidence that membrane-associated forms exist, as we shall discuss. Although both EP24.15 and EP24.16 are well characterized biochemically, their contribution to peptide metabolism in vivo has yet to be thoroughly established. In the following pages, we will review the current understanding of EP24.15 and EP24.16 and discuss the evidence supporting their involvement in neuroendocrine signaling. As the biochemistry, cell biology, and physiology of the two peptidases are so similar, they will be discussed at length for EP24.15, while the section on EP24.16 will highlight the unique features of this particular enzyme.


   II. Endopeptidase EC 3.4.24.15
Top
 Abstract
 I. Introduction
 II. Endopeptidase EC 3.4.24.15
III. Endopeptidase EC 3.4.24.16
 IV. Endooligopeptidase A
 V. Functions of EP24.15...
 VI. Conclusions
 References
 
A. Biochemical properties
EP24.15 was initially isolated from the soluble fractions of bovine pituitary (15) and rat brain (16) and shown to hydrolyze a number of biologically active peptides with a general specificity for peptide bonds on the carboxyl side of aromatic or basic residues (17). High levels of EP24.15 activity have been localized, both catalytically and immunohistochemically, to the brain, pituitary, and testis with lower levels in tissues such as the lung, liver, kidney, and spleen (18). Although primarily associated with the soluble fraction of tissue homogenates, subcellular fractionation of EP24.15 in rat brain shows that 20–25% of the total enzyme activity is associated with membrane fractions, including synaptosomes (19). Given the distribution of enzyme activity and its ability to degrade several bioactive peptides, Orlowski and colleagues (16) proposed a possible function in neuropeptide metabolism.

Isolated brain EP24.15 was described as a neutral metalloendopeptidase, as it is optimally active at neutral pH, is inhibited by metal ion chelators, and can be reactivated by divalent cations (16). Reactivation of the enzyme with Zn2+ occurs at quite low concentrations (0.025 mM), and the presence of a zinc atom in the EP24.15 catalytic core was confirmed by atomic absorption spectrometry (20). Although different degrees of inhibition of EP24.15 by EDTA have been reported, this may largely reflect different preincubation times; as the metal ion is tightly bound, its dissociation is a slow time-dependent process. As with thermolysin, the zinc atom is coordinated by the side chains of the two active-site histidines, plus a glutamate residue located 25 residues carboxy terminal to the second His (21). The active-site Glu also participates in the coordination of the zinc, via an activated water molecule (Fig. 2Go).



View larger version (37K):
[in this window]
[in a new window]
  		Figure 2. Model of zinc-coordinating residues within the rat EP24.15 active site. Represented are the active site histidines (H473, H477) and the remote glutamate (E502), which coordinate with the catalytic zinc ion. A second glutamate residing within the active site (E474) also participates in the coordination of the zinc, via an activated water molecule. [Reprinted from P. M. Cummins et al.: J Biol Chem 274:16003–16009, 1999 (21 ) with permission from The American Society for Biochemistry and Molecular Biology.]

 
A rather weak and variable degree of inhibition was observed with relatively high concentrations of thiol blocking agents, which initially confounded the classification of the enzyme. Dithiothreitol was shown to have a dual effect, in that low levels (0.5 M) activated the enzyme, while higher levels (5 mM) resulted in inhibition (16, 20, 22, 23). Inhibition has been attributed to the disruption of disulfide bridges within the enzyme and to the thiophilicity of the catalytic zinc ion (24). The unique mechanism by which EP24.15 is activated by thiols has been shown to involve the conversion of an inactive multimer (where substrate access to the catalytic site is restricted) to an active monomer via disruption of intermolecular disulfide bridges (25). Given that no natural inhibitors of EP24.15 have been identified to date, this conversion may represent a mechanism by which EP24.15 activity is regulated through changes in intra- and extracellular redox potential.

In 1989, EP24.15 was purified to homogeneity from rat testis (22); in addition to being rich in EP24.15, this source was chosen since a greater majority of the enzyme activity (92%) is found in the soluble fraction. The testicular enzyme, with a molecular mass of about 70 kDa, was shown to be immunologically and catalytically related, if not identical, to the brain enzyme. Polyclonal antibodies raised against the soluble form of rat testis EP24.15 were used for expression cloning of the gene from a rat testis cDNA library (26). The cloned enzyme was reported to comprise 645 amino acids with a molecular mass of 73 kDa, as found for the isolated brain and testis enzymes. The enzyme sequence was novel, with no significant homology with any known protein beyond the putative active-site sequence (HEXXH). Southern blots with EP24.15 cDNA from rat testis indicated that the enzyme is the product of a single gene (26). In 1993, it was shown that a sequencing error had resulted in two segments of the rat protein being read out of frame: the revised deduced amino acid sequence contains 687 residues (Fig. 3Go), representing a protein of 78 kDa (27). The pig (28), mouse (29), and human (30) homologs of EP24.15 have since been cloned. The cDNA encoding the rat EP24.15 has been used to determine the chromosomal localization of the human gene to chromosome 19 (31, 32), and the promoter region of the rat enzyme has recently been reported (33).



View larger version (48K):
[in this window]
[in a new window]
  		Figure 3. Sequence alignment of rat EP24.15 and EP24.16. Residues that are identical between the two enzymes are indicated by the vertical lines. The active site motif is in boldface type.

 
Examination of the deduced sequence reveals that EP24.15 lacks a signal sequence and membrane-spanning domain, so that the enzyme is unlikely to act as a membrane-bound ectozyme. Although the primary sequence of the enzyme contains one potential site for N glycosylation, there is no evidence to date that EP24.15 undergoes glycosylation. The sequence also contains several putative sites for phosphorylation that may alter and/or regulate EP24.15 activity. Indeed, Tullai et al. (29) have recently demonstrated that EP24.15 can be phosphorylated by protein kinase A at Ser644 in vitro, and that kinetic parameters for certain substrates, particularly GnRH, are altered by such phosphorylation. Furthermore, phosphorylated forms of native EP24.15 were detected in rat PC12 and mouse AtT-20 cells, both basally and in response to forskolin stimulation of protein kinase A. Moreover, cleavage of an EP24.15 fluorescent substrate by PC12 cell extracts is decreased after treatment by cAMP analogs (34). Thus, EP24.15 activity may be subject to regulation by kinases activated by extracellular receptors, although no specific examples of such a mechanism have yet been documented.

B. Substrate specificity
Upon isolation of EP24.15 from rat brain, Orlowski and colleagues (16) used a series of model synthetic substrates to study the specificity of the enzyme. They showed that EP24.15 preferentially cleaves bonds on the carboxyl side of hydrophobic amino acids, with an additional preference for an aromatic or basic residue in the P2 position (nomenclature of Schechter and Berger, Ref. 34A ). An additional aromatic residue (Phe, Tyr) at P3', some distance from the hydrolyzed bond, greatly increases the binding affinity of the substrate and rate of reaction (18). Work by Dando et al. (35), using purified human EP24.15 from erythrocytes, suggested that the favorable location (P2') of a proline residue and a free C terminus (favorable in position P3', P4', or P5') may be as important as the hydrophobic residues in the P1, P2, and P3' positions. Furthermore, they found no cleavage of substrates greater than 17 amino acid residues, contrary to earlier reports of larger substrates, which may be attributed to trace contamination by other peptidases. Recent studies of the recombinant enzyme support these findings; the largest substrates cleaved are 17 residues long, while the minimum length required is 6 residues (36, 37). In addition, the P1 position can be occupied by any amino acid, although cleavage is favored with Phe, Ala, or Arg, at least in the bradykinin-based fluorescent substrates analyzed. Interestingly, the cleavage site can shift, depending on residues some distance away. This implies a degree of flexibility in the binding of substrates to the active site.

With respect to natural substrates (Fig. 4Go), purified EP24.15 was shown to hydrolyze bradykinin at the Phe5-Ser6 bond (16, 22), GnRH at the Tyr5-Gly6 bond, the resulting N- terminal fragment at the His2-Trp3 bond (15, 16, 18), and neurotensin (NT) between two arginine residues (Arg8-Arg9) (16, 18, 22). Other natural substrates include somatostatin, substance P, and angiotensin I and II (18, 22). However, EP24.15 does not cleave the peptide bond required to convert angiotensin I to angiotensin II; rather, the primary cleavage occurs at the Pro7-Phe8 bond, producing angiotensin1–7, which has bioactivity distinct from angiotensin II (38). Some larger opioid peptides such as dynorphin A1–8, ß-neoendorphin, metorphamide, and Met-enkephalin-Arg-Gly-Leu are also rapidly converted by the enzyme to their corresponding enkephalins.



View larger version (20K):
[in this window]
[in a new window]
  		Figure 4. Cleavage sites of natural substrates of EP24.15 and EP24.16. The peptide sequences of substrates are given in the three-letter amino acid code. Sites of cleavage by both peptidases are shown as arrows; the different cleavage sites in NT are marked for each enzyme.

 
In 1994, our own group (39) provided evidence for a two-step mechanism of GnRH metabolism in the hypothalamus by prolyl endopeptidase and EP24.15. Our results demonstrated that GnRH, an {alpha}-amidated decapeptide, is itself a poor substrate for EP24.15; cleavage by brain extracts reflects the initial removal of the N-terminal glycinamide (Gly10-NH2) by prolyl endopeptidase, followed by rapid cleavage of GnRH1–9 at the Tyr5-Gly6 bond by EP24.15. This finding highlighted the need for pure preparations of EP24.15 for accurate characterization. Indeed, more recent results using recombinant EP24.15 support the view that cleavage of intact GnRH by the peptidase is slow (36, 40) and may be hindered by the C-terminal amide (23). For the most part, these studies have confirmed the substrate specificity seen with the purified enzyme, although some discrepancies remain, which probably reflect different assay conditions (e.g., salt concentration, presence of thiols, etc.) or even contamination by bacterial protease activity.

C. Distribution and subcellular localization
EP24.15 is ubiquitously distributed across mammalian species and throughout many tissue and cell types. The highest levels of EP24.15 activity have been found in brain, testis, and anterior and posterior pituitary, in keeping with its proposed neuroendocrine function. Peripheral tissues such as spleen, liver, kidney, lung, adrenals, and thyroid contain EP24.15 activity at levels approximately 10–20% of those found in the brain. Within rat brain, the highest EP24.15 activity is present in the cerebellum, with intermediate activities in the hippocampus, substantia nigra, cortex, striatum, and hypothalamus and low levels in midbrain, thalamus, and medulla/pons (18). EP24.15 has been localized immunohistochemically to the rat trachea, lung tissue, and alveolar macrophages, suggesting it may modulate the activities of bioactive peptides within the lung (41). Additional EP24.15 staining has been detected in mast cell granules of spleen lymphocytes (42), while a particulate distribution consistent with an endosomal localization is seen in chicken embryonic fibroblasts (43). EP24.15 activity has also been detected in semen (44) and in plasma (45).

The immunocytochemical localization of EP24.15 in brain with a polyclonal antibody raised against the rat testis enzyme shows EP24.15 staining in both glial and neuronal cells, concentrated in the cell nuclei (46). This finding, that EP24.15 staining was predominantly nuclear rather than cytoplasmic, was surprising given the biochemical association of EP24.15 with the soluble fraction. This result also contrasts with previous subcellular fractionation experiments showing very low endooligopeptidase A (Endo-A) activity (an enzyme similar or identical to EP24.15) in the nuclear fraction of rat neural tissue (47). The majority of the Endo-A activity appears to be associated with the cytosol, whereas around 20% of the activity is associated with the P2 mitochondrial particulate fraction, distributed between mitochondrial and synaptosomal membranes. The observed discrepancy may be attributed to a possible perinuclear localization of soluble EP24.15, which in conventional immunocytochemistry may appear to be within the nucleus, or due to cross-contamination during the subcellular fractionation procedure. Much of the confusion surrounding the exact subcellular localization of EP24.15 may also arise from the presence of one or more very closely related enzymes, such as Endo-A and EP24.16. Both activity assays and antisera thought to be specific for EP24.15 may have also detected either or both of these enzymes.

More recent studies using antisera that have been demonstrated to distinguish between EP24.15 and EP24.16 suggest that rat brain EP24.15 is primarily nuclear, whereas EP24.16 is mainly extranuclear, and particularly found in cell processes (Fig. 5Go and Ref. 48). These findings were recently extended to the electron microscopic level by the same research group (49), who report that both EP24.15 and EP24.16 are present in the cytoplasm of both glia and neurons, often associated in the latter with elements of the neurosecretory system, such as the Golgi apparatus, synaptic vesicles, and endosomes, as well as cytoskeletal structures. Strong nuclear staining for EP24.15 was prevalent, whereas that of EP24.16 was infrequent and much less intense; interestingly, it appeared that the level of nuclear EP24.15 was inversely correlated with its abundance in the cytoplasm, suggesting that entry of this enzyme into the nucleus may be a regulated event.



View larger version (109K):
[in this window]
[in a new window]
  		Figure 5. Immunohistochemical distribution of EP24.15 (a, c, e, and g) and EP24.16 (b, d, f, and h) in the rat brain. Note that the immunoreactivity for the EP24.15 appears mainly located in the nucleus, while EP24.16 is present in fibers and cell bodies’ cytoplasm (arrows). a and b, Secondary somatosensory cortex; c and d, cerebellar cortex; e and f, lateral vestibular nucleus; g and h, hypoglossal nucleus (XII) and dorsal motor nucleus of the vagus (X). CC, Central canal; LVe, lateral vestibular nucleus; S2, secondary somatosensory cortex. [Reprinted from E. E. Massarelli et al.: Brain Res 851:261–265, 1999 (48 ) with permission from Excerpta Medica Inc. © 1999.]

 
Several recent studies using the mouse corticotrope cell line AtT-20 have helped to clarify the subcellular distribution of EP24.15, at least in this cell type. Garrido et al. (50) have demonstrated that immunoreactive EP24.15 stains in a punctate fashion throughout the soma and cell processes and, in some cells, intense staining was seen within the nucleus, as confirmed by confocal microscopy (Fig. 6Go). This contrasted with staining for EP24.16, which was confined to the perinuclear area. Interestingly, Western analysis revealed a smaller EP24.15-immunoreactive band (69 kDa), which was predominant in the nuclear-enriched fraction, compared with the major 77-kDa species (identical with the recombinant enzyme) found in the postnuclear fraction; the basis for this difference is unknown but may reflect either an alternative splice variant or selective posttranslational processing of the nuclear form. The sequence of EP24.15 includes several putative nuclear localization sequences; however, similar sequences can also be found in EP24.16, and site-directed mutagenesis studies are needed to determine the exact motifs necessary for EP24.15 to reach the nucleus. Furthermore, any possible function of nuclear EP24.15 has yet to be explored.



View larger version (77K):
[in this window]
[in a new window]
  		Figure 6. Immunocytochemistry of EP24.15 in AtT-20 cells. One of the four EP24.15-immunoreactive cells seen in this field exhibits intense nuclear labeling (arrow) in addition to cytoplasmic labeling. Scale bar, 10 µm. [Reprinted from P. A. G. Garrido et al.: DNA Cell Biol 18:323–331, 1999 (50 ) with permission from Mary Ann Liebert Inc.]

 
In the AtT-20 cell, both EP24.15 and EP24.16 colocalized with syntaxin-6, a marker for the trans-Golgi network, and with ACTH, contained within secretory granules (50). However, both enzymes showed a more extensive distribution than these markers, suggesting that only a portion of these intracellular peptidases is present in the secretory pathway. A similar conclusion was reached by Ferro et al. (51), who found 15% of the total EP24.15 activity and immunoreactivity in AtT-20 cell homogenates sedimented in the denser fractions of a Ficoll gradient. These fractions overlapped significantly with those containing ß-endorphin immunoreactivity, again suggesting the localization of the peptidase within the secretory pathway.

The presence of EP24.15 within the classical secretory pathway is an unexpected finding, given the lack of any obvious signal sequence. However, several studies suggest that the enzyme may be secreted, although the exact mode of secretion remains undetermined. The first report of EP24.15 secretion was by Ferro et al. (52), who observed the time-dependent appearance of Endo-A activity in the medium of cultured glioma C6 cells. The relative lack of cellular glucose-6-phosphate dehydrogenase activity released during incubation supported the concept of secretion of EP24.15-like activity rather than nonspecific release as a result of cell lysis. More recent work (51) has demonstrated secretion of EP24.15 from AtT-20 cells. Again, the appearance of activity and immunoreactivity was time dependent, and unlikely to be due to cell death. Furthermore, release into the medium was approximately doubled by both the specific secretagog CRH, and by the calcium ionophore A23187, which induces exocytosis of secretory vesicles. This stimulated release of EP24.15 was sensitive to both brefeldin A and nocodazole, agents that disrupt the progression of proteins through the secretory pathway. However, basal EP24.15 release was unaffected by these agents; taken together, these results suggest that EP24.15 may be released via both the classical regulated secretory pathway, as well as by an alternate mechanism.

Clearly, if EP24.15 is to participate in the degradation of peptide signals, either the peptides must be internalized before cleavage, for which there is little supporting evidence, or the peptidase must attain an extracellular location. Secretion is one means by which EP24.15 could reach the extracellular environment; another possibility is that a portion of the enzyme is targeted to, and remains associated with, the plasma membrane. Recent evidence from Crack et al. (53) suggests the latter may indeed occur. The bulk of the EP24.15 activity in the postnuclear membrane fraction of AtT-20 cells could be found in the same Percoll gradient fractions as a specific plasma membrane marker, fluorescein isothiocyanate-concanavalin A. Interestingly, the size of the EP24.15 found in the enriched plasma membrane fraction (75 kDa) was about 2 kDa smaller than that found in the cytoplasm (77 kDa), again suggesting a membrane-specific form of the peptidase. In addition, a thiol-cleavable, membrane-impermeable cross-linker was used to positively identify the presence of immunoreactive EP24.15 on the extracellular side of the membrane. Immunocytochemistry of nonpermeabilized cells (i.e., primary antibody is bound before fixation) revealed distinct plasmalemmal patches of EP24.15 staining, in contrast to the additional staining of cytoplasmic elements in cells exposed to primary antibody after fixation. It should be noted that this group reports detecting nuclear staining only in dividing cells; this observation may provide a clue as to the function of nuclear EP24.15. Finally, membrane-associated EP24.15 accounted for approximately one-third of the total extracellular activity, the remainder being due to secreted enzyme. Thus, despite the apparent lack of both a signal sequence and any membrane-binding motifs, a portion of cellular EP24.15, albeit minor, appears to attain an extracellular location, both via secretion of the soluble enzyme and by attachment to the plasma membrane, at least in AtT-20 cells.


   III. Endopeptidase EC 3.4.24.16
Top
 Abstract
 I. Introduction
 II. Endopeptidase EC 3.4.24.15
 III. Endopeptidase EC 3.4.24.16
IV. Endooligopeptidase A
 V. Functions of EP24.15...
 VI. Conclusions
 References
 
A. Biochemical properties
EP24.16 (neurolysin) was originally purified and characterized from rat synaptic membranes (54), and later from tissues such as rat ileum and kidney (55, 56), on the basis of its ability to cleave NT at Pro10-Tyr11, generating the biologically inactive fragments NT1–10 and NT11–13. EP24.16 was shown to be distinct from other enzymes capable of cleaving NT, such as ACE, NEP and prolyl endopeptidase, as it was not inhibited by captopril, thiorphan, or serine protease inhibitors, respectively. Furthermore, EP24.16 could be separated from EP24.15, which hydrolyses NT exclusively at the Arg8-Arg9 bond, by chromatography on hydroxyapatite (54, 57). Indeed, NT represents one of the few differences in substrate specificity between the two enzymes (Fig. 4Go and Table 1Go); most other natural substrates are cleaved at identical sites (40). Recent studies of recombinant or highly purified enzymes and synthetic substrate/inhibitor libraries reveal only very subtle differences in the specificity of EP24.16 relative to EP24.15 (37, 58, 59, 60).


View this table:
[in this window]
[in a new window]
  		Table 1. Differences between the closely related metalloendopeptidases EP24.15 and EP24.16

 
Some distinction between EP24.15 and EP24.16 can also be made by the use of the inhibitors N-[1-(R,S)-carboxy- 3-phenylpropyl]-Ala-Ala-Phe-p-aminobenzoate (cFP-AAF-pAB) and N-[1-(R,S)-carboxy-3-phenylpropyl]-Ala-Ala-Tyr-p-aminobenzoate (cFP-AAY-pAB), and the EP24.16-specific dipeptide inhibitor Pro-Ile (61) (Table 1Go). However, Pro-Ile is not very potent and cFP-AAF-pAB/cFP-AAY-pAB can also block EP24.16 activity, although less potently than they inhibit EP24.15 (62, 63). The recent development of potent and highly specific phosphinic peptide inhibitors against EP24.15 and EP24.16 (59, 60) should aid further assessment of their functional roles. Another important difference between these enzymes is the unique activation of EP24.15 by low levels of thiol reducing agents. As mentioned previously, EP24.15 forms inactive multimers in the absence of thiols, while EP24.16 is strictly monomeric (25). Like other metallopeptidases, both EP24.15 and EP24.16 are inhibited by higher concentrations of thiol reductant, either by disruption of intramolecular disulfide bonds and/or by binding to the catalytic zinc atom (Ref. 24 and Table 1Go).

The cloning of rat brain EP24.16 (64) revealed a 704-amino-acid protein that is approximately 60% identical to EP24.15 (Fig. 3Go) and is considered to be the rat homolog of the pig liver soluble angiotensin-binding protein (65) and rabbit liver microsomal endopeptidase (66). The structure of EP24.16 has recently been solved by x-ray crystallography (67) and suggests that a narrow channel restricts substrate access to the active site, thereby limiting the length of peptides able to be cleaved (Fig. 7Go). Modeling of the binding of NT to EP24.16 suggests that the N terminus of the peptide slides into the channel, aligning the cleavage site (Pro10-Tyr11) with the catalytic zinc ion. Furthermore, flexible loop elements lining the channel may account for the ability of EP24.16 to cleave at a variety of residues. A comparison of the tertiary structures of EP24.16 and EP24.15 by computer modeling may reveal the structural basis for the differences in their substrate and inhibitor specificities.



View larger version (55K):
[in this window]
[in a new window]
  		Figure 7. Model of NT bound to the deduced structure of EP24.16. A molecular surface representation of neurolysin sectioned to show the large cavity at the bottom of the active site channel is shown with the 13-residue substrate NT, with the N terminus of the peptide at the top. [Reprinted with permission from C. K. Brown et al.: Proc Natl Acad Sci USA 98:3127–3132, 2001 (67 ). © National Academy of Sciences, U.S.A.]

 
B. Distribution and subcellular localization
Like EP24.15, EP24.16 exists in both soluble and membrane-associated forms. Most reports suggest that the soluble form is predominant: 10–20% of EP24.16 activity has been found to be membrane associated in pig (57) and rat (68) brain, rat kidney (56), and in whole neuron and astrocyte homogenates (69). In other studies, however, both pig and rat brain EP24.16 activity was 80% particulate (62). Again, like EP24.15, there is evidence to suggest that a portion of EP24.16 may, in fact, be genuinely membrane bound. First, neuronal membrane-associated EP24.16 has been shown to resist treatment with digitonin and Na2CO3, indicating the enzyme was neither inside vesicles nor weakly membrane adsorbed (70). Second, cryoprotection experiments and substrate degradation studies, using stably transfected human kidney cells, indicated that these transfects expressed the membrane- associated form of EP24.16 with its active site facing the extracellular medium (69).

In addition to a cytosolic and plasma membrane location, EP24.16 exists in the mitochondrial compartment, shown by biochemical characterization (20) and partial sequencing (71) to be distinct from EP24.15. This new activity was referred to as mitochondrial oligopeptidase and was later shown to be identical to the recently cloned EP24.16 (64). The fact that EP24.16 is the product of a single gene raises the question as to how it can be localized to more than one compartment. The presence of a putative mitochondrial targeting sequence at the N terminus of the precursor to EP24.16 (mitochondrial oligopeptidase) was noted by Serizawa et al. (71). More recently, Kato et al. (72) demonstrated the existence of alternate initiation sites for transcription of EP24.16. The longer transcript contains a cleavable mitochondrial target sequence, which directs the enzyme to the mitochondria, whereas the shorter form lacking such a signal remains in the cytosol.

As discussed previously, confocal microscopy has revealed EP24.16 in the classical secretory pathway in AtT20 cells, and its colocalization with ACTH further suggests the enzyme may be targeted to the regulated secretory pathway (50). Indeed, Vincent et al. (70) have demonstrated that EP24.16 activity is secreted from astrocytes, but not neurons, in culture; however, its release is insensitive to a range of agents that affect classical secretion. Furthermore, secretion was temperature dependent, being significantly curtailed at 4 C and accelerated at 42 C, as has been observed for IL-1ß secretion (73). Thus, EP24.16 secretion may occur via a nonclassical mechanism, despite its apparent presence within the secretory pathway.

Immunocytochemical studies of EP24.16 at both the light and electron microscopic levels in rat brain have shown that the enzyme is both cytoplasmic and membrane associated in neurons (49, 74). Neuronal staining was found primarily in dendrites, often in dense patches immediately adjacent to intensely immunoreactive regions of plasma membrane (74). Staining was also observed over perikarya (only occasionally intranuclear), as well as in axons and axon terminals, where it was associated with the neurosecretory system. Interestingly, immunoreactive EP24.16 could be found on both the cytoplasmic and luminal sides of vesicular membranes (49); this contrasts with vesicle-associated EP24.15, which was exclusively cytoplasmic, and suggests that EP24.16 may traverse intracellular membranes by an as-yet-unknown mechanism. If so, EP24.16 may be better placed for the degradation of neuropeptides than the predominantly nuclear EP24.15.

The membrane-associated form of neuronal EP24.16 was shown to be marginal in the early stages of the neuronal differentiation process, but markedly increased during maturation in vitro (70), thus supporting a role for EP24.16 in maturation and development of brain tissue. EP24.16 appears ubiquitously and heterogeneously distributed in the brain, with high concentrations of EP24.16 localized to the olfactory bulb and tubercle, cingulate cortex, medium striatum, and globus palidus (63). This distribution parallels that reported for NT receptors (75) and supports a physiological role for EP24.16 in NT inactivation, as discussed below. As with EP24.15, EP24.16 activity, immunoreactivity, and mRNA are also found in a wide range of peripheral tissues, including kidney, ileum, and testis (64, 76, 77).

A putative variant of EP24.16 (EP24.16B) has been purified from rat testis (78). While similar in biochemical characteristics, this variant reportedly exhibits different substrate specificity toward NT, cleaving at both the Pro10-Tyr11 and Arg8-Arg9 bonds, and different sensitivity to Pro-Ile. However, the possibility of contamination with endopeptidase EP24.15 has not been sufficiently excluded.


   IV. Endooligopeptidase A
Top
 Abstract
 I. Introduction
 II. Endopeptidase EC 3.4.24.15
 III. Endopeptidase EC 3.4.24.16
 IV. Endooligopeptidase A
V. Functions of EP24.15...
 VI. Conclusions
 References
 
A bradykinin-hydrolyzing activity present in the cytosol of rabbit brain was described in 1973 by Camargo et al. (79) and was subsequently reported to cleave NT and enkephalin precursors in a manner identical to EP24.15 (80, 81, 82). This activity was named Endo-A (EC 3.4.22.19) and has been the subject of protracted debate as to whether it is identical to or distinct from EP24.15 (83, 84, 85). The consensus in the field in recent years has been that the two activities are attributable to the same enzyme. However, more than 25 yr after its initial description, Hayashi et al. (86) have reported the cloning of a truncated form of Endo-A from rabbit brain cDNA. The deduced sequence shows no significant homology to any metallopeptidase, including EP24.15, except for the presence of the active site motif near the C terminus. The sequence also contained a high number of cysteine residues (18 of 512), which may account for the reported thiol dependence of Endo-A. Northern blot analysis indicated a high level of expression in the brain relative to peripheral tissues. The expressed protein possessed enzymatic activity against peptide substrates, which was the same as that reported for purified Endo-A. Antisera raised against the recombinant enzyme reduced the cleavage of a quenched fluorescent enkephalin-related substrate by brain cytosol by more than 70% but was without effect on recombinant EP24.15 activity. Thus, not only are studies of EP24.15-like peptidases confounded by the existence of homologous enzymes such as EP24.16, but it appears that peptidases belonging to entirely different families may also display similar characteristics.


   V. Functions of EP24.15 and EP24.16
Top
 Abstract
 I. Introduction
 II. Endopeptidase EC 3.4.24.15
 III. Endopeptidase EC 3.4.24.16
 IV. Endooligopeptidase A
 V. Functions of EP24.15...
VI. Conclusions
 References
 
As stated before, the precise physiological roles of EP24.15 and EP24.16 are unclear. The fact that both enzymes are widely distributed and are primarily cytoplasmic suggests their participation in general intracellular degradative processes, although evidence for such a role is only beginning to emerge. However, given their prevalence in the nervous system and their ability to cleave bioactive peptides in vitro, these peptidases have also been hypothesized to play a part in the metabolism of neuropeptides. In the following sections, we will discuss the evidence supporting a role for EP24.15 and/or EP24.16 in the regulation of the reproductive axis, pain nociception, NT inactivation, and the regulation of cardiovascular and renal function. Furthermore, more recent and controversial evidence regarding their possible role in antigen presentation and the metabolism of amyloid peptides will also be discussed. In many studies, distinction between EP24.15 and EP24.16 was not possible or was not attempted; thus, activities attributed to EP24.15 may also be due to EP24.16.

A. Reproductive axis
As stated, the highest levels of EP24.15 activity are found in the testis, implying a role for the enzyme in reproduction. Testicular EP24.15 enzyme activity, immunoreactivity, and expression levels increase linearly with age during maturation in rats, with the protein being localized primarily in elongating spermatids and residual bodies (87, 88). A similar cellular distribution was also observed in human testes (87), suggesting the peptidase may have a role in spermatogenesis. In the female rat, ovarian EP24.15 activity is relatively high (similar to brain levels), and increases sharply at puberty, but returns to prepubertal levels in adulthood (88). In addition, EP24.15 activity increases somewhat on the afternoon of proestrus (88). The exact ovarian cell type that expresses EP24.15 is not known, nor is its function in this organ.

The production of gonadal steroids is regulated by LH and FSH secreted from the anterior pituitary. These gonadotropins are in turn controlled by the pulsatile release of GnRH by neurons of the hypothalamus. The concentration of GnRH that reaches the anterior pituitary appears to be controlled, in part, by the rate of GnRH degradation within the hypothalamus and/or pituitary (89). The presence of an N-terminal pyroglutamate residue and a C-terminal amide bond renders GnRH resistant to most exopeptidases and, as such, enzyme degradation must be controlled by endopeptidases. A number of GnRH-degrading activities, particularly EP24.15, have been demonstrated in the soluble and particulate fractions of the hypothalamus, median eminence, and pituitary, where they may contribute to the regulation of GnRH release and action (Fig. 8Go).



View larger version (21K):
[in this window]
[in a new window]
  		Figure 8. Schematic of the possible role of EP24.15 in gonadatropin secretion. Secretion of LH and FSH from the anterior pituitary is primarily regulated by the pulsatile release of GnRH. GnRH is synthesized by hypothalamic neurons, released in the median eminence, and reaches the anterior pituitary via the hypophysial portal vessels. Together with prolyl endopeptidase (PEP), EP24.15 is postulated to degrade GnRH at any or all of these sites (i.e., within the median eminence, the portal plasma, or the anterior pituitary). EP24.15 may also participate in the conversion of dynorphin A (DynA) into enkephalin (Enk), which can then inhibit GnRH synthesis and secretion at the hypothalamic level. Furthermore, the product of EP24.15 cleavage, GnRH1–5, may interfere with GnRH secretion via blockade of the NMDA glutamate (Glu) receptor.

 
As discussed above, EP24.15 cleaves the Tyr5-Gly6 bond of GnRH (16, 89), albeit slowly. Studies in vitro have shown that the EP24.15 inhibitor cFP-AAY-pAB prevents cleavage of GnRH by hypothalamus and pituitary extracts (18, 90). High levels of EP24.15 activity have been shown in the rat median eminence, anterior pituitary, and preoptic area of the hypothalamus (88), the latter being the major site of GnRH synthesis (91). The same regions of the sheep hypothalamo-pituitary axis also exhibited high levels of EP24.15 activity, in particular the median eminence (92). Early immunocytochemical studies detected EP24.15 in the inner, but not the outer, lamina of the rat median eminence (46). Given that the outer lamina of the median eminence is the site at which hypothalamic regulatory hormones are released into the hypophysial portal circulation, Healy and Orlowski (46) questioned the involvement of EP24.15 in processing or degradation of peptides released at this site. More recently, however, EP24.15 immunoreactivity has been demonstrated in the perivascular space of the median eminence, in the tanacytic fibers, and in regions of the median eminence where GnRH axon terminals and hypophysial portal vessels are present (45). In addition, the enzyme has been detected in hypophysial portal blood, suggesting the enzyme is secreted within the median eminence (45). As mentioned previously, EP24.15 has been identified both at the cell surface (53) and in the culture medium (51) of AtT-20 cells derived from a mouse corticotrophic tumor; whether pituitary gonadotrophs, the target cells for GnRH, also express extracellular EP24.15 is still unknown.

The use of specific inhibitors suggests that EP24.15 may play a role in the in vivo metabolism of GnRH. A number of studies have demonstrated that cFP-AAF-pAB administration by either the intracerebroventricular or iv route results in approximately an 8-fold increase in the half-life of exogenously administered GnRH (89, 93, 94). The increased half-life is typical of that of a GnRH agonist, in which the Tyr5-Gly6 bond (the site of EP24.15 cleavage) has been modified to resist peptidase degradation. It should be noted that due to the polar nature of cFP-AAY-pAB, it is unlikely to cross cellular membranes, and thus any effects observed after its administration in vivo probably reflect inhibition of extracellular rather than intracellular enzyme. Further, there is evidence to suggest that EP24.15 is under the control of gonadal steroids and hence that the changes in GnRH portal plasma levels that occur at the onset of puberty or with each estrous cycle are due, in part, to alterations in the degradation of GnRH. Advis et al. (95) showed that an activity directed against the GnRH Tyr5-Gly6 bond from the soluble fraction of rat hypothalamus and anterior pituitary decreased before the first estrous cycle at puberty. In addition, the GnRH-degrading activity is lower in median eminence homogenates of estradiol-primed ovariectomized rats given progesterone to induce an LH surge (96). The activity of EP24.15 is stimulated in the anterior pituitary and several regions of the hypothalamus after ovariectomy and has been shown to increase in the preoptic area and decrease in the anterior pituitary before puberty in rats (88). In addition, the presence of EP24.15 immunoreactivity in the median eminence appears to fluctuate on the proestrous day of the rat estrous cycle (45), although in the ewe, EP24.15 activity remained constant in the preoptic area and median eminence throughout the estrous cycle (97).

N-Methyl-D-aspartate (NMDA) glutamate receptors are located in the hypothalamus and median eminence and are involved in the pulsatile secretion of GnRH. The product of EP24.15 action on GnRH, GnRH1–5, has been shown to function as a NMDA receptor antagonist, selectively inhibiting GnRH secretion evoked by NMDA (98). Recent evidence suggests that such an ultra-short negative feedback mechanism may be operative in the prepubertal rat and may be regulated through changes in prolyl endopeptidase activity (99). Finally, the feedback regulation of LH release appears mediated, in part, by endogenous opioid peptides that suppress hypothalamic GnRH release (88). Given that EP24.15 has been shown to be involved in opioid catabolism (see below), the enzyme may participate in the release of gonadotropins in a number of ways, either directly via GnRH degradation, indirectly via the generation of active opioid peptides or, alternatively, via the generation of an NMDA receptor antagonist (Fig. 8Go).

B. Pain perception—opioid processing
Opioids exert their action by binding to specific membrane receptors distributed throughout the central nervous system within structures involved in transmission, modulation, and sensation of pain. Endogenous opioid peptides (enkephalins, dynorphins, and endorphins) are released into the brain and blood after stress and/or pain (100). The ability of EP24.15 to process opioid precursors (Fig. 4Go) suggests a putative role for the enzyme in the modulation of nociception. Studies by Kest and colleagues (101) utilizing the tail flick and jump tests demonstrated that intracerebroventricular administration of cFP-AAF-pAB into rats produced significant antinociceptive effects, in both a dose- and time-dependent manner. The degree of antinociception was comparable to that obtained after central administration of NEP inhibitors, which act by preventing enkephalin degradation. Naloxone, an opioid receptor antagonist, significantly reduced the antinociceptive effects of EP24.15 inhibitors, suggesting that the actions of EP24.15 are largely mediated through the opioid peptide system. The authors did acknowledge, however, that EP24.15 inhibitors may prevent the degradation of other peptides that facilitate opioid-mediated nociception (101). In a subsequent study, the inhibitor potentiated the antinociceptive effects of coadministered Met-enkephalin-Arg-Gly-Leu and dynorphin A1–8 (102); this is consistent with the protection of exogenous dynorphin A1–8 by cFP-AAF-pAB in cerebrospinal fluid demonstrated by Molineaux and Ayala (103). Although EP24.15 activity is not detectable in CSF, the enzyme may be located in structures such as the choroid plexus, which are exposed to CSF (103).

Acute exposure to environmental stressors can also produce antinociception that is opioid mediated. The opioid mediation of intermittent cold water swims (ICWS) antinociception in rats has been confirmed by its sensitivity to naloxone antagonism. Central pretreatment with cFP-AAF-pAB significantly, and in a dose-dependent manner, increased ICSW nociception, without affecting basal nociceptive thresholds or latencies. Given that EP24.15 acts on longer-chain endogenous opioids such as dynorphin A1–8, the increase in ICWS antinociception in response to the EP24.15 inhibitor indicates that these peptides may participate in opioid forms of environmental antinociception (104).

The opioid receptor-like (ORL)1 orphan receptor is closely related to opioid receptors both in primary sequence and in function, although it does not bind any known opiate ligands with high affinity (105). Recently, two groups identified a potent endogenous ligand of ORL1, a heptadecapeptide (FGGFTGARKSARKLANQ) known as nociceptin/orphanin FQ, which structurally resembles dynorphin A (106, 107). Although the functional role of this peptide is unclear, the wide distribution of ORL1 mRNA and nociceptin/orphanin FQ precursor in the CNS and several areas known to be involved in pain perception suggests a nociceptive role (108). EP24.15 and aminopeptidase N (APN), a zinc-containing proteolytic ectoenzyme (109), have been shown to be the two main enzymes involved in the metabolism of nociceptin/orphanin FQ in mouse brain cortical slices. APN cleaves the Phe1- Gly2 bond while EP24.15 hydrolyses the Ala7-Arg8, Ala (11)-Arg (12) and Arg (12)-Lys13 bonds (105). The combined use of inhibitors against APN (bestatin) and EP24.15 [Z-(L, D)Phe{psi}(PO2CH2)(L, D)Ala-Arg-Phe; Ref. 58)] potentiated the marked reduction in motor activity induced by nociceptin/orphanin FQ (108), indicating a role for these peptidases in the degradation of this novel neuropeptide in vivo.

C. NT inactivation
NT is a tridecapeptide (pGlu-Leu-Tyr-Glu-Asn-Lys-Pro-Arg-Arg-Pro-Tyr-Ile-Leu-OH) originally isolated from bovine hypothalamus and widely distributed throughout the mammalian central nervous system and periphery (110). When administered iv, NT exerts a broad range of cardiovascular and endocrine effects including hypotension, analgesia, and hypothermia. Several studies have established that EP24.15 and EP24.16 are the main contributors to the inactivation of NT in vitro. The early investigation of NT metabolism in various membrane preparations and cell lines of central and peripheral origin showed that EP24.16 was the only peptidase ubiquitously contributing to NT inactivation, being present in all tissue where NT receptors are located (68). Indeed, the enzyme, which cleaves at the Pro10-Tyr11 bond (Fig. 4Go), was initially referred to as NT-degrading enzyme, as several lines of evidence suggested it may participate in the physiological inactivation of NT. EP24.16 appears to colocalize with NT receptors in pure, differentiated, cultured neurons from mouse embryos (111), whereas NT analogs that are substituted at position 11 by a D-amino acid were shown to be totally resistant to degradation in vitro and in vivo by brain tissue (112). EP24.15, on the other hand, is responsible for the cleavage of the Arg8-Arg9 bond of NT (16) (Fig. 4Go), and cleavage at this bond can be reduced in rat brain synaptosomal membranes by a specific inhibitor of EP24.15 (113). The examination of NT catabolism in rat hypothalamic slices has suggested that EP24.15 plays a predominant role in NT degradation (114, 115). Both peptidases contribute to the degradation of NT by rat astrocytes in culture (116).

More direct evidence for the involvement of EP24.15 and EP24.16 in the physiological degradation of NT came with the development of specific inhibitors. The dipeptide inhibitor of EP24.16, Pro-Ile, significantly enhanced the recovery of iv infused NT, accompanied by a concomitant decrease in the generation of NT1–10 in the ileum of the anesthetized dog (117). More recently, the newly designed EP24.16-specific inhibitor, P33, was shown to completely prevent the formation of the NT1–10 fragment by primary cultured neurons from mouse embryos (118). NT has been shown to elicit naloxone-resistant analgesia after its central administration in mice (119, 120) and, as such, NT may participate in non-opioid-mediated nociception. Intracerebroventricular administration of P33 dramatically potentiated the NT-induced analgesia of mice in the hot plate test (118). Similarly, in the same model of antinociception, administration of the EP24.15-specific inhibitor phosphodiepryl 21 also augmented the analgesic effects of submaximal doses of NT (121). These studies support a significant role of both EP24.15 and EP24.16 in the inactivation of NT, both centrally and in the periphery.

Interestingly, the neuroleptic haloperidol has been shown to decrease NT metabolism in intact rat brain slices by significantly reducing EP24.15 activity in the treated rats (122). NT has been shown to colocalize with dopamine (123, 124) and NT receptors with dopaminergic neurons (125) in many brain regions. There are indications that alterations in NT concentration may account for some of the pathophysiology associated with schizophrenia. Decreased NT immunoreactivity in CSF has been observed in certain subpopulations of schizophrenics (126, 127), and the subnormal NT concentrations in CSF have been shown to return to normal after treatment with neuroleptics (128). Thus, EP24.15 may play a role in the pathogenesis of schizophrenia via the inactivation of NT.

D. Cardiovascular/renal homeostasis
In addition to possible functions within the brain, EP24.15 may also play a role in the periphery. Endopeptidase 24.15 has been shown to efficiently degrade the potent vasodilatory peptide bradykinin in vitro at the Phe5-Ser6 bond (16, 22) (Fig. 4Go). Bradykinin (Arg-Pro-Pro-Gly-Phe-Ser-Pro-Phe-Arg) is one of the most important kinins, derived by cleavage of the plasma precursor kininogen through the kallikrein/kinin system. It is a powerful influence in stimulating extravascular smooth muscle (i.e., bronchial, uterine) contraction, inducing hypotension, dilating blood vessels and increasing vascular permeability. In vivo, bradykinin is widely considered to be cleaved primarily by carboxypeptidase N and ACE, although several other enzymes have been implicated in the inactivation of bradykinin (129).

In 1991, Genden and Molineaux (130) showed that the iv infusion of the EP24.15 inhibitor cFP-AAF-pAB produced an immediate drop in mean arterial pressure in normotensive rats, and that this decrease could be blocked by B2 bradykinin receptor antagonism. Furthermore, cFP-AAF-pAB was shown to potentiate the hypotensive effect of iv bradykinin infusion. These data suggested a role for EP24.15 in the inactivation of exogenous and endogenous bradykinin and led Genden and Molineaux to postulate a direct involvement of EP24.15 in blood pressure regulation. However, the EP24.15 inhibitor was subsequently shown to be rapidly degraded in the circulation to form a potent ACE-inhibiting metabolite (cFP-AA), suggesting the actions of cFP-AAF-pAB in vivo were attributable to the inhibition of ACE rather than EP24.15 (Fig. 9Go) (131, 132, 133). The Genden and Molineaux interpretation of their data is thus contestable, and the role of EP24.15 in the regulation of blood pressure unclear. Recently, however, a new EP24.15 inhibitor, equipotent to cFP-AAY-pAB, but biologically stable, has been reported (134) (Fig. 9Go). This inhibitor (JA-2) was shown to potentiate bradykinin-induced hypotension, without affecting the hypertensive effects of angiotensins I and II, thus suggesting a role for EP24.15 in the metabolism of circulating bradykinin and blood pressure control (135). However, administration of JA-2 to rabbits rendered hypertensive by renal wrapping did not affect mean arterial pressure or a number of other cardiovascular and renal parameters (136); whether EP24.15 participates in bradykinin metabolism in other models of hypertension is yet to be explored.



View larger version (17K):
[in this window]
[in a new window]
  		Figure 9. Structures of the EP24.15 and EP24.16 inhibitors, cFP- AAY-pAB and JA-2. As shown in panel A, cFP-AAY-pAB is subject to proteolytic cleavage by EP24.11, generating a potent ACE inhibitor, cFP-AA. Substitution of the second alanine residue in the inhibitor with an {alpha}-aminoisobutyric acid (Aib) results in an inhibitor (JA-2) equipotent to cFP-AAY-pAB, but completely resistant to EP24.11 cleavage (panel B).

 
Although bradykinin is generally considered a vasodilator, in the hepatic portal circulation it is a potent hypertensive agent. As such, the liver exhibits a great capacity to inactivate the peptide. Molina et al. (137) have recently shown that EP24.15 is the major kininase present in rat liver homogenates, as cleavage of fluorogenic bradykinin-based substrates was inhibited more than 90% by cFP-AAF-pAB, whereas captopril had no effect. Thus, hepatic EP24.15 may contribute significantly to the clearance of exogenous as well as endogenous bradykinin.

The kidney is another potential site of bradykinin generation, action, and clearance. Although the bulk of renal kininase activity can be attributed to NEP (138), a role for EP24.15 has also been postulated. Indeed, EP24.15 activity has recently been demonstrated in Madin-Darby canine kidney cells (139). In 1994, Yang et al. (140) reported that cFP-AAY-pAB administration to rats decreases mean arterial pressure, while increasing renal blood flow, glomerular filtration rate, urine volume, and urinary sodium and potassium excretion. The systemic blood pressure and renal blood flow effects could be blocked by pretreatment with the ACE inhibitor enalaprilat, suggesting that they were mediated by the conversion of cFP-AAY-pAB to cFP-AA. The other effects were unchanged by ACE inhibition and presumed to be due to inhibition of EP24.15 or a related peptidase. However, subsequent studies in the rabbit suggest that the dose of cFP-AAY-pAB used in the rat may have been high enough to partially block NEP. Tomoda et al. (141) found that coadministration of captopril and the NEP inhibitor SCH39370 to conscious rabbits reduced arterial pressure and increased heart rate, renal blood flow, and sodium excretion. No further changes in these hemodynamic and renal parameters was observed when cFP-AAY-pAB was added to the other peptidase inhibitors, suggesting that EP24.15 does not play a significant role in regulating kidney function. Interestingly, the renal effects of combined ACE and NEP inhibition were not blocked by B2 receptor antagonism, suggesting that these enzymes affect renal function via the metabolism of other peptides.

In addition to degrading bradykinin, EP24.15 also converts angiotensin I in vitro to the biologically active peptide angiotensin1–7 (18). The reported physiological actions of this peptide fragment favor a blood pressure-lowering action, potentially synergizing with the effects of bradykinin (142). The metabolism of angiotensin I to angiotensin1–7 by both vascular smooth muscle cells in culture (143) and by rat hind limb perfusate ex vivo (144) was reduced by more than 85% in the presence of an EP24.15 inhibitor. Thus, EP24.15 may participate in cardiovascular homeostasis via both the activation (angiotensin1–7) and inactivation (bradykinin) of vasodilatory peptides.

E. Alzheimer’s disease
More controversially, EP24.15 has also been implicated in the metabolism of the ß-amyloid peptide (Aß) associated with Alzheimer’s disease. ß-Amyloid is produced by the processing of the ß-amyloid precursor protein, a type I membrane-spanning glycoprotein ubiquitously expressed in mammalian cells (145). The production of Aß is thought to derive from proteolytic attacks at its N and C termini by ß- and {gamma}-secretases, respectively (146). Several early studies suggested that EP24.15 could be responsible for the ß-secretase activity, based either on its ability to cleave synthetic substrates containing the ß-secretase cleavage site (147), or on the degradation of intact ß-amyloid precursor protein by recombinant or coexpressed EP24.15 (148, 149). Other studies, however, contradicted these findings (150, 151), and the recent cloning of a membrane-bound aspartic protease with undisputed ß-secretase activity (152, 153) has dismissed the possibility that EP24.15 is the ß-secretase.

Although EP24.15 does not produce Aß, a recent study suggests the enzyme may actually participate in its degradation. Yamin et al. (154) observed that the amount of Aß present in conditioned medium from neuroblastoma cells was inversely correlated with the level of EP24.15 expression in mock, sense-, and antisense-EP24.15-transfected cells. This is clearly contrary to expectations for a ß-secretase and implies the involvement of EP24.15 in Aß clearance. Several other metalloendopeptidases have recently been implicated in Aß degradation, including NEP (155), ECE (156), ACE (157), and insulin-degrading enzyme (158, 159). However, unlike these peptidases, recombinant EP24.15 could not cleave synthetic Aß (154); thus, it is still unclear what role the enzyme may play in the regulation of Aß metabolism in the cell. The authors suggest that EP24.15 may increase the activity of a serine protease, perhaps by conversion of a zymogen form of the Aß-cleaving protease, or by degradation of an endogenous inhibitor, but this mechanism is difficult to envison, given the general inability of EP24.15 to cleave large proteins. Thus, given the evidence to date, a role for EP4.15 in Aß processing or metabolism remains speculatory and largely unsupported.

F. Antigen presentation
Recently, a role for EP24.15 in the processing of major histocompatibility (MHC) class I antigenic peptides has been postulated. In antigen-presenting cells, cytosolic proteins destined for degradation are targeted to the proteasome, in which the action of a number of proteolytic activities results in the generation of short peptides approximately 10 residues in length. These peptides are then transported to the cell surface and bound to the MHC class I molecule for presentation at the plasma membrane. Portaro et al. (160) observed that a range of such peptides resisted cleavage by both recombinant EP24.15 and crude macrophage cytosol. Furthermore, most of the peptides tested could inhibit EP24.15 activity, many with high affinity (Ki = 0.2–10 µM). This suggests that cytosolic EP24.15 may actively bind the products of proteasome activity and thus protect the peptides from other intracellular peptidases while en route to the cell surface. This hypothesis is supported by the observation that the proliferation and cytotoxicity of CD8 T cells can be increased by loading of antigen-presenting cells with recombinant EP24.15, and decreased by loading with cFP-AAY-pAB (161).

In contrast, Saric et al. (162) have recently published evidence that antigenic peptides are efficiently cleaved by EP24.15. Six antigenic peptides were rapidly degraded by soluble extracts from HeLa cells; addition of cFP-AAF-pAB or Z-(L, D)Phe{psi}(PO2CH2)(L, D)Ala-Arg-Phe significantly inhibited the cleavage of most of the peptides, suggesting a major role for EP24.15 in cytosolic peptide degradation. Furthermore, immunodepletion of EP24.15 from HeLa cell extracts using a specific antiserum that does not recognize EP24.16 also markedly reduced peptide cleavage. Finally, in contrast to the work of Portaro et al., this group found that these antigenic peptides were also readily cleaved by recombinant EP24.15. The reasons for the stark differences between these two studies are unknown, and further work is clearly necessary to determine the exact role of EP24.15 in the destruction or presentation of MHC class I antigens. In addition, the contribution of this endopeptidase to general cellular peptide degradation has yet to be explored, although Saric et al. allude to work in progress which specifically addresses this issue.


   VI. Conclusions
Top
 Abstract
 I. Introduction
 II. Endopeptidase EC 3.4.24.15
 III. Endopeptidase EC 3.4.24.16
 IV. Endooligopeptidase A
 V. Functions of EP24.15...
 VI. Conclusions
References
 
The pivotal role of the membrane-bound members of the zinc metalloendopeptidase family, particularly ACE, NEP, and ECE, in peptide metabolism is well documented and undisputed; indeed, the development of specific inhibitors of these enzymes as therapeutics has been, and continues to be, a significant focus of the pharmaceutical industry. The physiology of their soluble counterparts, EP24.15 and EP24.16, is less well understood, yet the evidence presented in this review is suggestive of a considerable contribution to peptide metabolism within specific peptide signaling pathways. Clearly, a number of outstanding issues remain to be resolved; for example, how are these soluble enzymes associated with both intra- and extracellular membranes? By what mechanism do they reach the extracellular space? Is there any role for cytoplasmic peptidases in the degradation of extracellular signaling peptides? Do they participate in an entirely separate aspect of cellular function, such as degradation of antigen peptides? The recent advances in our knowledge of the molecular and cellular biology of EP24.15 and EP24.16, coupled with both the development of inhibitors with greater specificity, should facilitate further analysis of the contribution of the soluble metalloendopeptidases to neuroendocrine signaling.


   Acknowledgments
 
The authors thank Professor John W. Funder for his encouragement and his critical appraisal of the manuscript.


   Footnotes
 
Abbreviations: Aß, ß-Amyloid peptide; ACE, angiotensin-converting enzyme; APN, aminopeptidase N; cFP-AAF-pAB, N-[1-(R,S)-carboxy-3-phenylpropyl]-Ala-Ala-Phe-p-aminobenzoate; cFP-AAY-pAB, N-[1-(R,S)-carboxy-3-phenylpropyl]-Ala-Ala-Tyr-p-aminobenzoate; CSF, cerebrospinal fluid; ECE, endothelin converting enzyme; Endo-A, endooligopeptidase A; EP24.15, EC 3.4.24.15 (thimet oligopeptidase); EP24.16, EC 3.4.24.16 (neurolysin); ICWS, intermittent cold-water swims; MHC, major histocompatibility; NEP, neutral endopeptidase; NMDA, N-methyl-D-aspartate; NT, neurotensin; ORL, opioid receptor-like.


   References
Top
 Abstract
 I. Introduction
 II. Endopeptidase EC 3.4.24.15
 III. Endopeptidase EC 3.4.24.16
 IV. Endooligopeptidase A
 V. Functions of EP24.15...
 VI. Conclusions
 References
 

  1. Docherty K, Steiner DF 1982 Post-translational proteolysis in polypeptide hormone biosynthesis. Annu Rev Physiol 44:625–638[CrossRef][Medline]
  2. Gainer H, Russell JT, Loh YP 1985 The enzymology and intracellular organization of peptide precursor processing: the secretory vesicle hypothesis. Neuroendocrinology 40:171–184[Medline]
  3. Steiner DF 1998 The proprotein convertases. Curr Opin Chem Biol 2:31–39[CrossRef][Medline]
  4. Seidah NG, Mowla SJ, Hamelin J, Mamarbachi AM, Benjannet S, Toure BB, Basak A, Munzer JS, Marcinkiewicz J, Zhong M, Barale JC, Lazure C, Murphy RA, Chretien M, Marcinkiewicz M 1999 Mammalian subtilisin/kexin isozyme SKI-1: a widely expressed proprotein convertase with a unique cleavage specificity and cellular localization. Proc Natl Acad Sci USA 96:1321–1326[Abstract/Free Full Text]
  5. Hook VYH, Azaryan AV, Hwang SR, Tezapsidis N 1994 Proteases and emerging role of protease inhibitors in prohormone processing. FASEB J 8:1269–1278[Abstract]
  6. Fricker LD, Snyder SH 1982 Enkephalin convertase: purification