By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Address correspondence and reprint requests to Dr. L. A. Devi at Department of Pharmacology, New York University School of Medicine, MSB 411, 550 First Avenue, New York, NY 10016, U.S.A. E-mail: Lakshmi.Devi@med.nyu.edu
Abstract: Prodynorphin, a multifunctional precursor of several important opioid peptides, is expressed widely in the CNS. It is processed at specific single and paired basic sites to generate various biologically active products. Among the prohormone convertases (PCs), PC1 and PC2 are expressed widely in neuroendocrine tissues and have been proposed to be the major convertases involved in the biosynthesis of hormonal and neural peptides. In this study we have examined the physiological involvement of PC2 in the generation of dynorphin (Dyn) peptides in mice lacking active PC2 as a result of gene disruption. Enzymological and immunological assays were used to confirm the absence of active PC2 in these mice. The processing profiles of Dyn peptides extracted from brains of these mice reveal a complete lack of Dyn A-8 and a substantial reduction in the levels of Dyn A-17 and Dyn B-13. Thus, PC2 appears to be involved in monobasic processing, leading to the generation of Dyn A-8, Dyn A-17, and Dyn B-13 from prodynorphin under physiological conditions. Brains of heterozygous mice exhibit only half the PC2 activity of wild-type mice; however, the levels of Dyn peptides in these mice are similar to those of wild-type mice, suggesting that a 50% reduction in PC2 activity is not sufficient to significantly reduce prodynorphin processing. The disruption of the PC2 gene does not lead to compensatory up-regulation in the levels of other convertases with similar substrate specificity because we find no significant changes in the levels of PC1, PC5/PC6, or furin in these mice as compared with wild-type mice. Taken together, these results support a critical role for PC2 in the generation of Dyn peptides.
Most neuropeptides are synthesized as larger precursors that undergo endoproteolysis at specific sites (Docherty and Steiner, 1982). These sites are usually multiple basic amino acids (Hutton, 1990), although some cleavages can occur at the single basic residues that usually fit a consensus sequence (Devi, 1991). The dynorphin (Dyn) precursor, prodynorphin (Prodyn), is a useful model for studying enzymes involved in neuropeptide processing because it contains both paired basic and single basic cleavage sites; cleavage at these sites generates various potent opioid peptides with differential receptor selectivity (Shulz et al., 1982). In the brain and neurointermediate lobe, Prodyn is processed to α- and β-neoendorphins, Dyn A-17, Dyn A-8, Dyn B-13, and Leu5-enkephalin (Leu-Enk). Generation of α- and β-neoendorphins, Dyn A-17, Dyn B-29, and Leu-Enk requires cleavage at paired basic sites, and an additional cleavage at a single basic site is needed to release Dyn B-13 and Dyn A-8 (Fig. 1). In the anterior lobe of the pituitary, Prodyn is processed primarily to higher-molecular-weight Dyn-containing intermediates; the levels of fully processed Dyn peptides are substantially lower in this tissue (Cone et al., 1983; Day and Akil, 1989). Thus, the relative amounts of Dyn peptides in the brain vary from region to region. Because differential proteolysis of a single precursor can result in the generation of several neuropeptides that activate a diverse set of receptors, the relative levels of endoproteases can have a regulatory effect on the “bioactivity” of the neuropeptide. Therefore, processing enzymes play a key role in the modulation of neuropeptide levelsin vivo.
Several neuropeptide-processing enzymes have been identified in mammalian cells (Fricker and Devi, 1995). Following the precursor cleavage at single or multiple basic residues by endoproteases, carboxypeptidases remove the basic amino acids from the C termini of the resulting peptides (Fricker and Devi, 1995). Endoproteases of the subtilisin family of serine proteases, such as prohormone convertases (PCs), cleave peptide precursors at paired basic residue cleavage sites (Seidah and Chretien, 1992; Steiner et al., 1992; Rouille et al., 1995). These enzymes are homologous to KEX-2, a yeast peptide-processing enzyme (Seidah and Chretien, 1992). PC1 (also known as PC3) and PC2 exhibit a restricted neuroendocrine distribution and have been widely accepted to be responsible for the endoproteolytic processing of the majority of neuroendocrine peptides (Rouille et al., 1995). Their acidic pH optima, requirement for calcium ions, and ability to function in the regulated secretory pathway are consistent with the putative role for PCs as the major endoproteases responsible for the maturation of a large number of hormone and neuroendocrine precursors (Mains et al., 1990; Benjannet et al., 1991; Seidah et al., 1993; Fricker et al., 1996; Furuta et al., 1998).
The involvement of PC1 and PC2 in Prodyn processing has been examined mainly byin vitro techniques using either purified recombinant enzymes or vaccinia virus overexpression system (Dupuy et al., 1993; Day et al., 1998). PC1 processes Prodyn primarily to generate 8-10-kDa Dyn-containing high-molecular-weight intermediates, a C-terminal fragment, and a 16-kDa N-terminal fragment (Dupuy et al., 1993). A small amount of shorter peptides resulting from single basic processing is also observed. In contrast, PC2 processes Prodyn at both paired basic and single basic sites, generating significant amounts of shorter peptides such as Dyn A-9 and Dyn B-14 (Day et al., 1998). Also, purified PC2 selectively cleaves peptides that fit the consensus for single basic (monobasic) processing (Berman et al., 1997), suggesting that PC2 could be involved in the single basic processing of Prodyn in vivo. It is interesting to note that other Prodyn-expressing tissues such as the striatum (Dores and Akil, 1985), hypothalamus (Day and Akil, 1989), or spinal cord (Xie and Goldstein, 1987; Day and Akil, 1989) process the Prodyn precursor more completely than the anterior lobe of the pituitary; these tissues express high levels of PC2. Thus, regional variation in Prodyn processing correlates with the level of PC1 and PC2 enzymes.
Previous studies examining the involvement of processing enzymes in Dyn biosynthesis have used overexpression systems or purified recombinant enzymes (Dupuy et al., 1993; Day et al., 1998). An important test of whether PC2 is physiologically involved in Dyn processing is to disrupt the enzyme and examine if peptide processing is altered. Previous studies using PC2 knockout (PC2 K/O) mice as a model have revealed severely impaired processing of proglucagon, prosomatostatin, and proinsulin in the α, δ, and β pancreatic cells (Furuta et al., 1997) and pro-melanin-concentrating hormone (Viale et al., 1999) and proenkephalin in the brain (Johanning et al., 1998). These studies provided indirect evidence for the absence of PC2 activity in PC2 K/O mice.
In the present study, we have directly measured the PC2 activity in these mice and examined the possible compensatory changes of other PCs involved in neuropeptide processing in the absence of active PC2. Using a highly specific enzymatic assay, we find a complete lack of PC2 activity in all brain regions of PC2 K/O mice. In these animals, processing of Prodyn at single basic sites is severely altered. However, studies with heterozygous mice show that 50% enzyme activity is sufficient to process Prodyn efficiently at both single basic and paired basic sites. The levels of other PCs, including PC1, furin, and PC5/PC6, do not show any compensatory changes due to PC2 gene disruption.
Wild-type and PC2 K/O mice brains were dissected into seven regions as described (Glowinski and Iversen, 1966). Tissues were immediately frozen on dry ice and stored at -70°C until further use. For PC2 activity determination, frozen tissues were extracted with 50 mM Tris-Cl (pH 7.5) containing 1% Triton X-100, 10% glycerol, 1 μM E-64, 1 μM pepstatin, 10 μM leupeptin, 30 μM phenylmethylsulfonyl fluoride, and 5.0 μg/ml aprotinin. After sonication, extracts were kept on ice for 30 min, followed by centrifugation at 16,000 g for 20 min at 4°C. Supernatants were collected, and aliquots were stored at -70°C until further use.
PC2 enzyme assay
PC2 activity was measured using 200 μM L-pyroglutamyl-Arg-Thr-Lys-Arg-7-amino-4-methylcoumarin as substrate in 100 mM sodium acetate buffer (pH 5.0) containing 1 mM CaCl2 and 0.1% Triton X-100 in the presence of a protease inhibitor cocktail (0.28 mMN-tosyl-L-phenylalanine chloromethyl ketone, 0.14 mMNα-p-tosyl-L-lysine chloromethyl ketone, 1 μM E-64, 1 μM pepstatin A, and 10 μM captopril). The inhibitor cocktail was included to protect the substrate from nonspecific enzymatic degradation by other proteolytic enzymes. All incubations were at 37°C for 30 min to 4 h. Parallel samples were incubated with 1 μM CT peptide (SVNPYLQGKRLDNVVAKK). This peptide is derived from the C-terminal region of 7B2 and serves as a specific inhibitor of PC2 (Lindberg et al., 1995; Zhu et al., 1996). The release of 7-amino-4-methylcoumarin was measured using a Perkin-Elmer spectrofluorimeter (λ excitation = 360 nm; λ emission = 480 nm), and the amount of product formed was calculated using free 7-amino-4-methylcoumarin as a standard. The activity inhibited by CT peptide was taken as PC2 activity.
For peptide analysis whole brains were homogenized in 10 volumes of 0.1 M HCl at 100°C and incubated at this temperature for 15 min. Following centrifugation (13,000 g for 30 min at 4°C), supernatants were concentrated on a Speed Vac (Savant) and stored at -20°C. Before radioimmunoassays (RIAs) and/or gel-exclusion chromatography, samples were resuspended in methanol/0.1 M HCl (1:1 vol/vol).
Size-exclusion chromatography and RIAs
Gel-exclusion chromatography was performed on a Superdex Peptide HR 10/30 column (Pharmacia). The sample, in a total volume of 50-100 μl, was applied to the column and separated with 1% formic acid containing 0.02% protease-free bovine serum albumin (Sigma). The flow rate was 0.5 ml/min; 0.5-ml fractions were collected, dried, resuspended, and subjected to RIA. RIAs for Prodyn/Dyn-derived peptides were carried out essentially as previously described (Berman et al., 1994, 1995). The anti-Dyn A-17 antiserum (Peninsula Laboratories) does not cross-react with α-neo-endorphin, β-endorphin, Leu-Enk, Dyn A-8, and Dyn B. The antisera against Dyn A-8 and Dyn B-13 (13S) have been extensively characterized (Cone and Goldstein, 1982; Cone et al., 1983; Xie and Goldstein, 1987): Dyn A-8 antiserum is directed against the COOH-terminal portion of Dyn A-8 peptide and does not recognize COOH-terminal extensions, whereas Dyn B antiserum recognizes both the N- and C-terminally extended Dyn B-13.
Western blot analysis of Prodyn, PC1, PC2, furin, and PC5
The levels of Prodyn, PC1, PC2, furin, and PC5 proteins in various brain regions of wild-type and PC2 K/O mice were examined by western blotting. For the detection of Prodyn, 13S antiserum was used as described previously (Fricker et al., 1996). PC1/PC3 antiserum directed against the N-terminal region of the enzyme corresponding to amino acids 84-100 was used at a dilution of 1:1,000 (Qian et al., 2000). For the detection of PC2, anti-PC2 polyclonal antiserum [directed against the C-terminal amino acid sequence (592-608) of the mouse enzyme] was used at a dilution of 1:1,000. PC5 antiserum (a gift from Dr. N. G. Seidah) raised against the N-terminal region of the enzyme corresponding to amino acids 116-132 (De Bie et al., 1996) was used at a dilution of 1:1,000. Furin antibody (a gift from Dr. R. Angeletti) directed against the N-terminal amino acid sequence 153-167 was used at a dilution of 1:1,000. The blots were normalized using tubulin as a marker of total protein detected with monoclonal anti-tubulin antiserum (Sigma) at a dilution of 1:2,000. The blots were visualized using the ECL kit (Pierce).
To examine directly the extent of involvement of PC2 in Prodyn processing in vivo, we took advantage of the availability of mice lacking enzymatically active PC2 (PC2 K/O). To confirm the lack of active PC2 in PC2 K/O animals, we measured the level of PC2 protein and activity. PC2 protein was detected by western blot analysis, and the activity was determined by an assay that specifically measures PC2 activity in tissue homogenates (Fig. 2 and Table 1). We detected 75- and 68-kDa proteins in wild-type mice; the size of 75 kDa is consistent with that of the proenzyme, and the size of the 68-kDa protein is consistent with the mature enzyme (Rouille et al., 1995). In PC2 K/O mice, a band of 72 kDa was observed; this is consistent with the expected size of PC2 lacking the 38 residues (amino acids 94-131) encoded by exon 3. Exon 3 contains the activation site at the junction between the propeptide and the mature enzyme (Furuta et al., 1997). The heterozygous animals showed three immunoreactive bands: 75- and 68-kDa forms (due to the presence of one wild-type copy of PC2) and a 72-kDa band (due to the presence of one mutant copy of PC2) (data not shown).
Table 1. PC2 activity levels in brain regions of PC2 K/O mice
PC2 activity (nmol/min/mg of protein)
The isolation of various brain regions was according to the procedure of Glowinski and Iversen (1966). Enzyme extraction and assay were carried out as described in Experimental Procedures. Enzyme activity represents the activity inhibited by 1 μM CT peptide. Data are mean ± SEM values of triplicate determinations from three animals.
To confirm the absence of PC2 activity in PC2 K/O mice, we developed a rapid fluorimetric assay. This assay takes advantage of the ability of CT peptide (an 18-amino acid peptide derived from the C terminus of 7B2) to selectively inhibit PC2 activity (Lindberg et al., 1995). We find no significant activity in PC2 K/O animals in any of the regions examined (Table 1). In contrast, there is substantial PC2 activity in all brain regions of wild-type animals, with highest activity in the hippocampus and hypothalamus and lowest activity in cerebellum and pons/medulla oblongata (Table 1). In the heterozygous mice the PC2 activity is ∼50% of that in wild-type animals (Table 1). This is consistent with the presence of a single copy of the wild-type gene in these animals. Taken together, these data indicate that the PC2 K/O animals lack the 68-kDa form of the mature enzyme and PC2 enzymatic activity.
We examined the Prodyn processing profile in PC2 K/O mouse brains and compared it with the profile in wild-type mice. These profiles were obtained by separation of peptides by size-exclusion chromatography followed by detection of Dyn peptides by RIA. We find that immunoreactive (ir)-Dyn A-8 is not detectable in PC2 K/O mouse brains (Fig. 3A), implying that PC2 is required for the generation of Dyn A-8. In contrast, a small amount of ir-Dyn A-17 and ir-Dyn B-13 can be detected in PC2 K/O mouse brains (Fig. 3B and C). The fact that PC2 K/O mice do not contain detectable amounts of Dyn A-8 supports a role for PC2 in single basic processing at this site. The presence of some fully processed Dyn A-17 and Dyn B-13 in these mice suggests that their generation is not absolutely dependent on PC2 activity and that other endoproteases, such as PC1, could be involved. It is possible that a decrease in the amount of fully processed opioid peptides would lead to an increase in the higher-molecular-mass intermediates. Consistent with this, we find an increase in the level of the 8-kDa form containing both ir-Dyn A and ir-Dyn B (Fig. 3B and C). Furthermore, we find a significant increase in the unprocessed precursor, Prodyn, in the majority of the brain regions of PC2 K/O mice (Fig. 4). This is in contrast to the wild-type or heterozygous mice, where the precursor was undetectable in the majority of the regions studied (Fig. 4). These results suggest that the impaired Prodyn processing in PC2 K/O mice is accompanied by the accumulation of the precursor and partially processed intermediates.
Next we compared the level of Dyn peptides in heterozygous mice (that contain 50% of active PC2) with PC2 K/O mice (that do not contain active PC2) or with wild-type mice. The levels of ir-Dyn A-8, ir-Dyn B-13, and ir-Dyn A-17 in heterozygous mice were comparable to levels in wild-type mice (Table 2). In contrast, the levels of ir-Dyn A-8 and ir-Dyn B-13 were substantially reduced in PC2 K/O mice; the levels of ir-Dyn A remained the same in these animals (Table 2). Taken together, these results suggest that PC2 is involved in the generation of Dyn A-8 and Dyn B-13 and that 50% enzyme activity in heterozygous animals is sufficient to process Prodyn efficiently at both single basic and paired basic sites. It is interesting to note that in PC2 K/O mice we find a substantial reduction in the fully processed Dyn peptides following size fractionation (Fig. 3), and yet we do not find the same extent of reduction in peptide levels in total homogenate without size fractionation (Table 1). This could be because the levels measured in tissue homogenates represent a combination of fully processed peptides and the partially processed intermediates, whereas the levels measured following size fractionation represent individual fully processed or partially processed peptides. One needs to consider this important issue while interpreting data from studies measuring peptide levels in tissue homogenates (without fractionation).
Table 2. Dyn peptide levels in PC2 K/O mouse brains
ir-peptide concentration (pmol/g of tissue) before extraction
The peptide extraction and RIAs were carried out as described in Experimental Procedures. Data are mean ± SEM values of triplicate determinations from three individual animals except for Dyn A-8, which are from six individual animals.
There is a small but significant amount of Dyn B-13 and Dyn A-17 in PC2 K/O animals, suggesting that PCs other than PC2 could be involved in processing at these sites. PC1 is a functional and structural homologue of PC2, andin vitro studies have implicated a role for PC1 in compensating for PC2 activity due to its overlapping specificity and localization (Dupuy et al., 1993; Seidah et al., 1998). To examine whether PC1 levels are up-regulated in the absence of PC2, we examined the levels of PC1 in various brain regions of PC2 K/O animals and compared these with the levels in wild-type animals. We find that the level of PC1 in PC2 K/O mice is comparable to that in wild-type animals in the majority of the regions examined (Fig. 5). The level of PC1 protein was some-what reduced in the cerebellum of PC2 K/O mice; the basis for this is not clear. PC1 is seen as 87- and 68-71-kDa forms (Fig. 5). The 87-kDa form corresponds to PC1 after N-terminal fragment removal, and the 68-71-kDa doublet represents PC1 truncated at the C terminus (Vindrola and Lindberg, 1992; Benjannet et al., 1993; Scougall et al., 1998).
The levels of other PCs such as PC5/PC6 and furin, which are also known to cleave at multiple basic sites, were found to be comparable to those in wild-type animals in the majority of the regions examined (Fig. 5). PC5 is found as a major form of 65 kDa and a minor form of 40 kDa (Fig. 5); because the PC5 antibody used for western blot analysis is N-terminally directed, both forms represent C-terminal truncations of the enzyme. The 40-kDa form may represent a truncation of the 65-kDa band; a C-terminally truncated 65-kDa form has been previously reported in AtT20 cells (De Bie et al., 1996). Furin is found as a major 98-kDa form and a minor 60-kDa band (Fig. 5). The 98-kDa form has been reported to be the major form for bovine furin (Nakayama, 1997). The minor protein species of ∼60 kDa probably arises from additional cleavages in the cysteine-rich region (Bravo et al., 1994). Taken together, these results indicate that the lack of PC2 does not lead to compensatory changes in the level of PC1, PC5, or furin and that PC2 plays a major role in the biosynthesis of mature Dyn peptides.
There is a large body of evidence implicating PC1 and PC2 in neuropeptide processing (Mains et al., 1990; Benjannet et al., 1991; Dupuy et al., 1993; Rouille et al., 1995; Day et al., 1998). These studies have used recombinant enzymes and precursors. This in all likelihood does not represent physiological conditions in tissues expressing both enzymes. Furthermore, both PC1 and PC2 are required for efficient processing of neuropeptide precursors, and PC1 is thought to function before PC2 within the secretory pathway. In the present study we have addressed the contribution of PC2 to Prodyn processing using mice lacking the active enzyme. A major finding is that under physiological conditions PC2 is involved in the generation of products of monobasic processing, Dyn A-8 and Dyn B-13. The generation of these peptides requires processing at single basic as well as paired basic residues. The fact that in the absence of PC2 the levels of these bioactive Dyns are reduced is consistent with the proposal that PC2 is able to process at both single basic and paired basic sites.
In PC2 K/O mice we find larger (∼8-kDa) Dyn-containing peptides. Its size and the fact that this form is recognized by both the Dyn A and Dyn B antisera suggest that it is composed of α-neo-endorphin, bridge peptide, Dyn A, and Dyn B. This form is presumably generated by the action of PC1 becausein vitro studies have shown that PC1 is able to process Prodyn primarily to higher-molecular-mass Dyn-containing peptides such as 10- and 8-kDa forms (Dupuy et al., 1993). It is not known if PC1 is able to generate Dyn AB-32 from Prodyn. The fact that small but significant amounts of Dyn AB-32, Dyn A-17, and Dyn B-13 are found in mice lacking active PC2 indicates that other enzyme(s) may be involved, albeit with lower efficiency.
We find an increase in Prodyn precursor in PC2 K/O mice. This was confirmed by measuring ir-Leu-Enk levels following trypsin/carboxypeptidase B treatment of gel-filtrate fractions from PC2 K/O mouse brain (data not shown); this treatment releases ir-Leu-Enk from larger fragments (Fricker et al., 1996). At the elution position of Prodyn we noted a two- to threefold increase in the amount of ir-Leu-Enk in PC2 K/O mice versus control ones. This is consistent with the level of increase seen by western blot analysis (Fig. 4). The level of Prodyn protein precursor in wild-type mice correlates well with Prodyn mRNA levels reported previously (Pittius et al., 1987). In the mouse brain Prodyn is unevenly distributed, with highest levels in striatum, hippocampus, and hypothalamus, and lowest levels in midbrain, pons/medulla oblongata, and cortex; it is not detectable in cerebellum (Devi et al., 1987; Pittius et al., 1987). PC2 is found in high levels in some of the same regions (Schafer et al., 1993; Day et al., 1998). PC2 mRNA is highest in hippocampus, followed by striatum, hypothalamus, cortex, and midbrain, and lowest in cerebellum and pons/medulla oblongata (Schafer et al., 1993). This is comparable with the PC2 activity we find in various regions. Tissues expressing high levels of PC2 process Prodyn much more completely than those expressing high levels of PC1 (Dores and Akil, 1985; Day and Akil, 1989). Taken together, these observations substantiate our data suggesting that PC2 is the major convertase in the processing of Prodyn into mature opioid peptides such as Dyn B-13, Dyn A-17, and Dyn A-8.
It appears that PC2 is involved in the generation of mature opioid peptides from other precursors, as well. A study examining proenkephalin processing in PC2 K/O mice has reported a severe depletion of mature enkephalin with an increase in the levels of Met-enkephalin-containing peptides (Johanning et al., 1998). We find a drastic reduction in Dyns with a significant accumulation of higher-molecular-weight Dyn-containing peptides and Prodyn precursor. These results suggest that modulation of PC2 activity could have significant effects on the levels of opioid peptides. The maturation of PC2 and its activity are tightly controlled by various factors such as the local concentration of Ca2+ (Shennan et al., 1995), pH (Lamango et al., 1999), and helper proteins such as 7B2 (Lindberg et al., 1995; Zhu et al., 1996). Thus, factors governing the activity of neuroendocrine processing enzymes may affect the levels of opioid peptides and other neuropeptidesin vivo.