Impaired Prohormone Convertases in Cpe fat / Cpe fat Mice*

A spontaneous point mutation in the coding region of the carboxypeptidase E (CPE) gene results in a loss of CPE activity that correlates with the development of late onset obesity D., Varlamov, O., P. M., Rouille, Y., Steiner, D. F., Carroll, R. J., Paigen, B. J., and Leiter, E. H. (1995) Nat. Genet. 10, Examination of the level of neuropeptides in these mice showed a decrease in mature bioactive peptides as a result of a decrease in both carboxypeptidase and prohormone convertase activities. A defect in CPE is not expected to affect endoproteolytic processing. In this report we have addressed the mechanism of this unexpected finding by directly examining the expression of the major precursor processing endoproteases, prohormone convertases PC1 and PC2 in Cpe fat mice. We found that the levels of PC1 and PC2 are differentially altered in a number of brain regions and in the pituitary. Since these enzymes have been implicated in the generation of neuroendocrine peptides (dynorphin A-17, b -endorphin, and a - melanocyte-stimulating hormone) involved in the control of feeding behavior and body weight, we compared the levels of these peptides in Cpe fat and wild type animals. We found a marked increase in the level of dynorphin A-17, a decrease

Peptide hormones and neuropeptides are synthesized as propeptides that undergo a series of modifications before they are released as active peptides. A number of enzymes involved in the processing of neuropeptides have been identified and characterized (Ref. 2 and references therein). These include subtilisin/kexin-like pro-proteins such as furin, prohormone convertase 1 (PC1, 1 also known as PC3), PC2, PACE4, PC4, PC5/PC6, and PC7/PC8. Carboxypeptidases are required to remove cleavage site residues, usually Lys and Arg, from the C terminus of peptide hormone and neuropeptides after the action of prohormone convertases. Carboxypeptidase E (CPE, also known as CPH), a metallocarboxypeptidase with neuroendocrine distribution, is the major enzyme responsible for the C-terminal trimming of the majority of peptide hormones and neuropeptides (3). Support for this comes from a mouse model lacking active CPE (designated Cpe fat ) that shows accumulation of C-terminally extended peptides. Although a full-length messenger RNA transcript is produced in these mice, the translated mutant protein is degraded in the endoplasmic reticulum due to a point mutation (serine 202 to proline) in a highly conserved region of the enzyme (1). Thus, these mice show deficiency in C-terminal trimming of basic residues (1, 4 -8). In addition, these mice exhibit impaired endoproteolytic processing as characterized by precursor accumulation (4) and/or altered gene expression (6,7). These changes point to an alteration in the endoproteolytic processing of neuropeptides and hormone precursors in Cpe fat mice.
It is well established that feeding behavior is regulated by a number of peptides (Ref. 9 and references therein). Among them, dynorphin A-17 (Dyn A-17) (9) and ␤-endorphin 1-31 (Ref. 10 and references therein) have been shown to stimulate feeding, whereas acetylated forms of ␣-melanocyte-stimulating hormone (␣-MSH) exert a tonic inhibition on feeding behavior (Ref. 11 and references therein). On the other hand, postranslational processing of ␤-endorphin through N-terminal acetylation and C-terminal proteolysis eliminates effects on food intake (10). Dynorphins, ␤-endorphin, and ␣-MSH are peptides derived from prodynorphin (ProDyn) and proopiomelanocortin (POMC) by the action of prohormone convertases (PC1 and PC2) followed by the removal of C-terminal basic residues by CPE (3).
A number of studies (both in vitro as well as in vivo) implicate PC1 and PC2 in ProDyn processing (12)(13)(14). Studies using mice lacking PC2 activity (PC2 K/O) show a complete lack of Dyn A-8 and substantially reduced levels of Dyn A-17 and Dyn B-13 (13,14). Thus PC2 appears to be involved in the generation of Dyn A-17, Dyn A-8, and Dyn B-13 from ProDyn under physiological conditions (13,14). As in the case of ProDyn, PC1 and PC2 have been implicated in POMC processing (Ref. 15 and references therein). The cleavage profile of POMC by PC1 produces a pattern very similar to the one normally found in the anterior pituitary, high levels of adrenocorticotropic hormone (ACTH) and ␤-lipotropin (␤-LPH) and low levels of ␤-endorphin (15,16). PC2 on its own does not release ACTH from POMC but cleaves POMC into N-terminally extended cortico-tropin containing the joining peptide, ␣-MSH, and ␤-endorphin (15,16). A combination of PC1 and PC2 produces a pattern similar to that seen in the neurointermediate lobe, high levels of ␣-MSH, corticotropin-like intermediate lobe peptide (CLIP), ␤-endorphin, and low levels of ACTH and ␤-LPH (15). Thus, PC1 and PC2 are distinct proprotein convertases acting alone or together to produce a set of tissue-specific maturation products in the brain and pituitary. The mechanism by which the defect in CPE reduces the processing of these peptide hormones and neurotransmitters is not well understood.
In the present study we tried to determine if the absence of CPE leads to alterations in the level of PCs (protein activity and mRNA), which in turn could affect the endoproteolytic processing of ProDyn and POMC in the brain, and pituitary of Cpe fat mice. We found that the maturation of PC1 is decreased and the levels of PC2 are increased in Cpe fat mice as compared with wild type (WT) mice. We also found that the proteolytic processing of ProDyn and POMC are decreased in Cpe fat mice, resulting in reduced levels of Dyn A-8, Dyn B-13, and ␣-MSH among other bioactive peptides.

EXPERIMENTAL PROCEDURES
Animals-Mice were bred at The Jackson Laboratory as described previously (1). The identity of Cpe fat (Ϫ/Ϫ) animals was confirmed by genotyping using MIT primers (D8MT69 F and R; D8MIT131 F and R) from Research Genetics according to the protocol supplied by The Jackson Laboratory. Nonobese littermates (ϩ/Ϫ) or (ϩ/ϩ) referred to as WT were used as controls. The age of the animals ranged from 17 to 20 weeks.
Tissue Preparation-WT and Cpe fat mice were decapitated between 10:00 a.m. and 12:00 noon. Whole brains and pituitaries were collected and dissected into regions or immediately frozen on dry ice. For regional distribution studies brains were dissected into seven regions as described by Glowinski and Iversen (17). Frozen tissues were stored at Ϫ70°C until use.
PC2 Enzyme Assay-For PC2 activity determination, frozen tissues were extracted with 50 mM Tris-Cl, pH 7.5, containing 0.5% Triton X-100, 10% glycerol, 1 M E-64, 1 M pepstatin, 10 M leupeptin, 300 M phenylmethylsulfonyl fluoride, and 5.0 g/ml aprotinin (buffer A). After sonication, extracts were centrifuged at 14,000 ϫ g for 20 min at 4°C. Supernatants were collected, separated into aliquots, and stored at Ϫ70°C until further use. PC2 activity was measured using 200 M Pyr-Arg-Thr-Lys-Arg-aminomethylcoumarin (American Peptide Co.) as a substrate in 100 mM sodium acetate buffer, pH 5.0, containing 1 mM CaCl 2 and 0.1% Tx-100 in the presence of a protease inhibitor mixture (0.28 mM N-tosyl-L-phenylalanine chloromethyl ketone, 0.14 mM N-␣p-tosyl-L-lysine chloromethyl ketone, 1 M E-64, 1 M pepstatin A, and 10 M captopril). The inhibitor mixture 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 M human C-terminal (CT) peptide 1-31 (a gift from Dr. I. Lindberg, Louisiana State University Medical Center). The release of 7-amino-4-methylcoumarin was measured using a PerkinElmer Life Sciences spectrofluorimeter ( excitation ϭ 360 nm; emission ϭ 480 nm), and the amount of product formed was calculated using free 7-amino-4-methylcoumarin as a standard. The activity inhibited by the C-terminal peptide was taken as PC2 activity.
Inactivation of PC2 by C-terminally Extended Dyn-and Leu-Enkderived Peptides-Recombinant PC2 (10 -50 ng) (a gift from Dr. I. Lindberg, Louisiana State University Medical Center) was incubated with 100 M Dyn A-17 and 5-8 concentrations of C-terminally extended peptides (shown in Table II) for 30 min at 37°C in 0.1 M sodium acetate buffer, pH 5.0, containing 0.1% Triton X-100 and 1 mM CaCl 2. The reactions were terminated by boiling for 10 min. The pH of the reaction mixture was adjusted to 7.5 with 1 M Tris-Cl, pH 8.0, followed by incubation with 6 ng of CPB/100 l of reaction mixture for another 30 min at 37°C, and the assay was terminated by boiling. The Dyn A-8 released from Dyn A-17 was quantified by radioimmunoassay (RIA) as described previously (18,19).
Western Blot Analysis of CPE, CPD, ProDyn, POMC, PC1, and PC2-Before Western blot analysis buffer conditions were optimized to exclude the possibility of enzyme and precursor degradation during extraction. Two different buffers were tested and compared with the buffer used for PC2 activity determination (buffer A). These buffers contained different combinations of protease inhibitors and/or denaturing agents. The composition of the tested buffers were 50 mM Tris-HCl, pH 7.4 containing 10% glycerol, 2% SDS, 300 M phenylmethylsulfonyl fluoride, 10 M leupeptin, 1 M pepstatin A, 1 M E-64, 5 g/ml aprotinin (buffer B), and 50 mM Tris-HCl, pH 7.4, containing 1 mM leupeptin, 10 mM EDTA, 1% Triton X-100, 10% glycerol (buffer C). Tissue homogenates of various brain regions or of a whole brain or pituitary from WT and Cpe fat mice extracted with buffer A were subjected to SDS gel electrophoresis on 10% (CPE, CPD, PC1, PC2) or 16% gels (ProDyn, POMC) and analyzed by Western blotting. CPE and CPD antisera (a gift from Dr. L. Fricker, Albert Einstein College of Medicine) were directed against the C-terminal region of the enzymes (20,21). The 13 S antiserum, directed against mid-portion of Dyn B-13, was used to detect ProDyn (4). The anti-POMC antiserum (Phoenix Pharmaceuticals, Inc.), directed against the N-terminal amino acid sequence 27-52, was used at a dilution 1:1000. The anti-PC1 and anti-PC2 antisera, directed against the N-and C-terminal regions of the enzymes, respectively, were used at a dilution of 1:1000 (14). The blots were normalized by tubulin using anti-tubulin antiserum (Sigma) at a dilution of 1:2000. Western blots were visualized using an ECL kit (Pierce). Densitization of blots was done using NIH Image software.
RNA Analysis-Total cellular RNA was extracted from mouse brain and pituitary using a TRIzol reagent kit (Life Technologies, Inc.). Total RNA (30 g) was electrophoresed through 1.5% agarose gel containing 2.2 M formaldehyde using 0.04 M MOPS, pH 7.0, buffer containing 10 mM sodium acetate and 1 mM EDTA. RNA was transferred to nitrocellulose (Gene screen membrane, PerkinElmer Life Sciences) overnight, and the filter was air-dried and baked at 80°C for 2 h. After prehybridization the blots were hybridized with appropriately labeled probes as described previously (22). Antisense cRNA probes were synthesized using SP6 (POMC) or T7 (PC1, PC2, ProDyn) RNA polymerase in the presence of [␣-32 P]CTP using a Riboprobe ® in vitro transcription system (Promega). The fragments used to generate each of the riboprobes were as follows: POMC, a 923-base pair EcoRI/HindIII fragment from a full-length mouse POMC cDNA clone (23), and ProDyn, a 300-base pair HindIII/BamHI fragment of the main exon of the rat ProDyn gene (24,25). Rat (r) PC1 and rPC2 cDNAs were obtained using polymerase chain reaction as described previously (26). The rPC1 probe (492 base pairs) was 97% identical to mouse PC1, corresponding to nucleotides 715-1206. The rPC2 probe (450 base pair) was 95% identical to mouse PC2, corresponding to nucleotides 878 -1326. The blots were washed three times for 15 min at room temperature with 2ϫ SSC (1ϫ SSC ϭ 0.15 M NaCl and 0.015 M sodium citrate), 0.05% SDS followed by 2-3 washes for 15 min at 65-70°C with 0.2ϫ SSC, 0.1% SDS and exposed to a phospho screen for 3 h or overnight at room temperature. Autoradiograms were quantified by scanning densitometry and normalized to 18 S RNA visualized by ethidium bromide staining.
Peptide Extraction-For peptide analysis whole brains were homogenized in 0.1 M CH 3 COOH at 100°C (10 volumes for the brain) and incubated at this temperature for 15 min. After centrifugation (13,000 ϫ g for 30 min at 4°C) the supernatants were concentrated on Speed Vac (Savant) and stored at Ϫ20°C. Anterior lobe (AL) and neurointermediate lobe of the pituitary (NIL) were homogenized with buffer used for PC2 extraction (buffer A, glycerol was excluded from the buffer composition, 100 l/structure). Upon removal of a small aliquot for PC2 activity determination (to ensure the accuracy of dissection into AL and NIL), an equal volume of 2 M CH 3 COOH was added to the remainder. After centrifugation, the supernatants were concentrated on Speed Vac and stored at Ϫ20°C. Before RIA and/or gel exclusion chromatography, samples were resuspended in methanol:0.1 N HCl (v:v).
Size Exclusion Chromatography and RIAs-Gel exclusion chromatography was performed on Superdex ® peptide HR 10/30 column (Amersham Pharmacia Biotech). Samples in a total volume of 50 -100 l (usually 2-3 brains combined together) were applied to the column and separated with 1% formic acid containing 0.02% protease-free bovine serum albumin (Sigma). Individual extracts of AL and NIL of the pituitary were applied to the column in a total volume of 25 l and separated with 30% acetonitrile in 0.1% trifluoroacetic acid. The flow rate was 0.5 ml/min, and 0.5 ml fractions were collected. Fractions were dried, resuspended, and subjected to RIA to quantitate ProDyn/Dyn (14,18,19)-or POMC-derived peptides. For POMC-derived peptides (␤endorphin, ␣-MSH, and ACTH) RIAs were carried out as described for ProDyn/Dyn-derived peptides. ␤-Endorphin antiserum (National Institute of Drug of Abuse) was used at a titer of 1:90,000; this antiserum has an IC 50 of 30 fmol/tube and completely cross-reacts with ␤-LPH. ACTH antiserum (Peninsula Laboratories, Inc.) was used at 1:75,000 final dilution; this antiserum has an IC 50 of 100 fmol/tube and com-pletely cross-reacts with human ACTH 18 -39 but does not cross-react with ␤-LPH or ␤-endorphin. ␣-MSH antiserum (Peninsula Laboratories, Inc.) was used at a dilution of 1:30,000; this antiserum has an IC 50 of 20 fmol/tube and exhibits 67% cross-reactivity with (Nle4, D-Phe7)-␣-MSH but does not cross-react with ACTH or ␤-endorphin.
Characterization of POMC Peptides by Reverse Phase (RP) HPLC-For RP-HPLC, 200 -250 l of material containing the highest immunoreactivity after fractionation of the NIL of the pituitary on Superdex peptide column were injected into a peptide C18 column (Vydac) and eluted with a gradient of 80% acetonitrile in 0.1% trifluoroacetic acid (solvent B) against 0.1% trifluoroacetic acid (solvent A). For the fractionation of ␤-endorphin and ACTH, the gradient was 30 -35% B for the initial 5 min, 35-45% B for the next 45 min, and then to 50% B over the next 5 min (27). For the fractionation of ␣-MSH, the gradient was 15-45% B for the first 30 min and then to 65% B over the next 10 min. The flow rate was 1 ml/min, and 0.5-ml fractions were collected, concentrated in Speed Vac concentrator and dissolved in methanol/HCl for RIA. Synthetic ␤-endorphin, ACTH, CLIP, and ␣-MSH peptides (from Peninsula Laboratories) were used as standards to calibrate the column.

RESULTS
CPE activity is deficient in Cpe fat mice since a single mutation in the coding region results in an unstable enzyme that is quickly degraded before maturation and transport to the Golgi apparatus (20). To confirm the absence of mature CPE in Cpe fat mice, we carried out Western blot analysis with an antiserum that detects both pro-CPE as well as mature CPE. We found a 53-kDa band corresponding to mature CPE in the brain of WT mice (Fig. 1). This size is consistent with the form of mature CPE previously reported in various bovine and rat tissues (28). In contrast, we found an immunoreactive band of ϳ56 kDa that corresponds to the size of pro-CPE (i.e. the precursor form of CPE with 14 additional N-terminal amino acids) in Cpe fat mice. Mature CPE (53 kDa) is not seen in these animals, which is consistent with previous findings (4). The absence of mature CPE in Cpe fat mice could result in a compensatory up-regulation of other related carboxypeptidases such as CPD. By Western blot analysis we found a protein band of 180 kDa, representing the major form of CPD in all regions of both WT and Cpe fat animals (Fig. 1). This size is consistent with the predominant form of CPD previously reported in bovine pituitary and other tissues (21,29). Careful comparison of the relative levels of CPD between WT and Cpe fat mice show no significant differences, suggesting that the lack of active CPE does not result in an increase in CPD levels in Cpe fat mice.
Previous studies found an accumulation of C-terminally extended peptides in Cpe fat mice (1, 4 -8). In addition, these mice exhibit an accumulation of partially processed intermediates as well as the precursor, ProDyn (4). However, these studies did not examine the extent of ProDyn accumulation in various brain regions. We examined this as well as the relative level of another peptide hormone precursor, POMC, in Cpe fat mouse brain regions. We found a predominant band of 30 kDa, representing ProDyn in all brain structures (Fig. 2). The level of immunoreactive ProDyn is 1.5-4.0-fold higher in Cpe fat mouse brain regions as well as in pituitary as compared with WT mice. Western blot analysis of POMC in pituitaries shows a predominant band of 23 kDa and minor bands of 32 and 28 kDa (Fig.  2). The 30 -32-kDa form of POMC has been previously reported in various cell culture systems (30 -32). The other bands probably represent unglycosylated forms or processing intermediates of POMC. We found an increase in the level of all these forms in Cpe fat animals. These results indicate that there is an accumulation of both ProDyn and POMC precursors in Cpe fat mice.
To examine if the apparent increase in the precursor levels is due to enhanced mRNA levels, we carried out Northern blot analyses. The levels of ProDyn (seen as an ϳ2.4-kilobase band) or POMC (seen as an ϳ1.2-kilobase band) mRNAs in WT are comparable with those in Cpe fat mice, indicating that the changes in precursor levels are not accompanied by modifications in ProDyn or POMC gene expression (Fig. 3). It should be pointed out that Cpe fat mice with an average weight of 33.7 Ϯ 1.5 g exhibit no change in POMC mRNA levels compared with WT mice (27.6 Ϯ 0.6 g; p Ͻ 0.01). In contrast, Cpe fat mice with an average weight of 47.6 Ϯ 1.5 g exhibit a 2-3-fold increase in POMC mRNA level in the pituitary as compared with WT mice (data not shown). These data suggest that POMC mRNA changes became evident as mice grow and develop severe obesity.
The lack of a significant difference in ProDyn and POMC mRNA contents would suggest that the increase in precursor amounts could be due to changes in the level of the processing enzymes. We attempted to determine if the mRNA levels of the two PCs implicated in ProDyn and POMC processing (namely PC1 and PC2) are altered in Cpe fat mice. We found that PC1 and PC2 mRNA are detected as major bands of 3.0 and 2.8 kilobases, respectively, both in Cpe fat and WT mice (Fig. 3) and that there are no significant differences in PC1 or PC2 mRNA levels in these mice. Taken together, the results (showing no changes in gene expression) would suggest that the precursor accumulation could be mainly due to alterations in the level of active enzymes in Cpe fat mice.
We compared the levels of PC1 and PC2 in Cpe fat and WT mice by Western blot analysis. The PCs are synthesized as inactive zymogens, and their maturation involves a series of endoproteolytic steps such as the removal of the pro-domain and truncation of the C terminus presumably by autocatalysis (Refs. 2, 33-35, and 36 and references therein). To exclude the possibility of artifactual processing of PCs during extraction, three buffers with different combinations of protease inhibitors and/or denaturing agents (see "Experimental Procedures" for FIG. 1. Western blot analysis of CPE (total brain) and CPD in the brain regions of WT (؉/؉) and Cpe fat (؊/؊) mice. Approximately 15 g of protein from total brain or each region was subjected to gel electrophoresis and Western blotting using the polyclonal antiserum directed against CPE and CPD or tubulin antiserum and analyzed as described under "Experimental Procedures." The positions and molecular weights of prestained protein standards (Sigma) are indicated.

FIG. 2. ProDyn and POMC immunoreactivity in WT (؉/؉) and
Cpe fat (؊/؊) mouse brain regions and pituitary. Tissue samples from each brain region (30 g) and pituitary (10 g for POMC analysis) were subjected to immunoblot analysis with ProDyn (13 S) or POMC antisera as described under "Experimental Procedures." The blots were reprobed with a monoclonal anti-tubulin antiserum (Sigma) for normalization and analyzed as described. The position and the size of molecular weight markers are indicated on the left margin. details) were tested, and the buffer that preserved the enzymatic activity but blocked nonspecific hydrolysis during extraction was used for further analysis. We found a decrease in the level of 68-kDa PC1 and an increase in the level of 87-kDa in different brain regions and pituitary of Cpe fat mice (Fig. 4A). The 87-kDa form has been shown to be present in a partially active state before the secretory granules, where it is processed to a maximally active state (36). Quantitation of the enzyme levels shows that the 68-kDa form is 1.5-5 times lower in Cpe fat mice compared with WT mice in all regions examined (Fig. 4A,  lower panel). The largest decrease is detected in the pituitary, cerebellum, hippocampus, striatum, and cortex followed by midbrain, pons, and medula oblongata and hypothalamus. It should be noted that we found a concomitant increase in the 87-kDa form of the enzyme in all the regions examined (Fig. 4A,  upper panel). Nonetheless, the total level of PC1 (sum of the levels of 87 and 68 kDa) in all brain regions of Cpe fat mice is lower compared with WT animals. These data suggest that the apparent decrease in the 68-kDa form of PC1 could result not only from a decrease in enzyme maturation but also from other changes at the translational and/or post-translational level.
PC2 is seen as a major protein of 68 -71-kDa forms in all brain regions examined (Fig. 4B). When compared with WT mice, the level of immunoreactive PC2 is increased by ϳ1.3-2.0-fold in every brain region as well as in the pituitary of Cpe fat mice (Fig. 6). We further examined this unexpected finding by testing if the increase in ir-PC2 results in an increase in PC2 activity. For this we used a simple assay that takes advantages of the 31-amino acid C-terminal peptide derived from human 7B2 to selectively inhibit PC2 (37), thus allowing specific determination of the enzyme activity in tissue homogenates. We found that PC2 activity is also increased about 1.3-2-fold in all brain regions and pituitary of Cpe fat mice as compared with WT mice (Table I). Moreover, this is in good correlation with the increase in PC2 immunoreactivity seen in Cpe fat animals.
To better understand this paradoxical finding as to how decreased processing could result from an increased level of PC2, we attempted to determine if peptides that accumulate in Cpe fat mice, i.e. peptides with C-terminally basic residue extensions, would inhibit PC2. We found that Dyn B-14, Dyn A-7, and Dyn A-6 inhibited recombinant PC2 (Table II). Dyn A-7 with double Arg extensions is the most potent, and peptides with single Arg extensions are more potent than those with single Lys extensions. It is possible that a substantial accumulation of these peptides and/or other peptides with C-terminal basic residues within the secretory granules could lead to a significant inhibition of PC2 in Cpe fat mice. Indeed, previously it has been reported that processing of ProDyn and Dyn A-17 is less efficient in the absence of CPE (13). Another possibility is that a portion of the neuroendocrine protein 7B2, with the CT peptide 1-18 containing Lys-Lys at the C terminus, remains associated with the protein and, thus, inhibits the enzyme, whereas in WT mice this inhibition is removed by CPE. It has been previously shown that CT peptide 1-18 is a powerful inhibitor of PC2, whereas CT peptide 1-16 (lacking the Cterminal Lys-Lys) displays little inhibitory effect even at extremely high concentrations (38).
Next, we examined the effect of altered PC1 and PC2 levels on the extent of processing of two peptide precursors, namely ProDyn and POMC. The processing profile of ProDyn was characterized in the brain and the profile of POMC in the AL and NIL of the pituitary. Immunoreactive peptides were measured with specific RIAs (14,18,19) in fractions after size exclusion chromatography. We found that the levels of ir-Dyn A-8 and ir-Dyn B-13 (products of monobasic processing) are ϳ6and 2.0-fold lower, respectively, in Cpe fat as compared with WT mouse brains (Fig. 5, left top and bottom panels). At the same time the levels of ir-Dyn A-17 (the precursor of Dyn A-8) were increased by ϳ2-fold in Cpe fat as compared with WT mouse brain (Fig. 5, left middle panel). To determine if the reduction in the levels of Dyn A-8 and Dyn B-13 were due to C-terminal extension of basic residues (Lys and/or Arg), the fractions were treated with CPB (a treatment that removes C-terminal basic residue extensions) after separation on a gel filtration column. We found that this treatment does not cause a substantial alteration in the amount of ir-Dyn A-8 and ir-Dyn B-13 in fractions from WT brain but causes an increase in the levels of these peptides in Cpe fat brain (Fig. 5, right top and bottom  panels). A 5-fold increase in the level of ir-Dyn A-8 and ϳ2.5fold increase in the levels of ir-Dyn B-13 is seen that essentially restores the levels of ir-Dyn peptides in Cpe fat mouse brain to the levels in WT brain. In contrast, CPB treatment does not cause a substantial increase in the level of ir-Dyn A-17 in Cpe fat mouse brain (Fig. 5, right middle panel). We found that a decrease in the amount of fully processed peptides is accompanied by an increase in the higher molecular weight intermediates in Cpe fat mouse brain (Fig. 5). These results are consistent with the reduced endoproteolytic processing in Cpe fat animals due to alterations in the levels of PC1 and PC2 activity.
To better understand how differentially altered levels of PC1 (decreased) and PC2 (increased) affect neuropeptide and hormone processing, we examined the processing profile of POMC, a well studied peptide hormone with a distinct processing profile in the pituitary. AL corticotrophs express very high levels of PC1 mRNA but very little PC2 mRNA, whereas NIL mela-   (n ϭ 3). IC 50 is determined using 10 -50 ng of purified PC2, 100 M Dyn A-17, and different concentrations of C-terminally extended peptides as described under "Experimental Procedures." ir-Dyn A-8 was measured by RIA after sample treatment with CPB. notrophs express high levels of PC2 mRNA and low levels of PC1 mRNA (26). We examined the forms of POMC-derived peptides by size exclusion chromatography and RP-HPLC in WT and Cpe fat mice. The POMC-processing pattern is different in the AL from that seen in the NIL (Fig. 6). In the AL of WT animals, two ir-␤-endorphin peaks are seen, and their sizes of (ϳ 10 and 3.5 kDa) are consistent with ␤-LPH and ␤-endorphin (Fig. 6A). The levels of both these peptides are reduced in Cpe fat animals (Fig. 6A). In contrast, a single peak of ␤-endorphin is seen in NIL of WT animals, and its level does not change in Cpe fat mice (Fig. 6D). The ir-ACTH peptides appear as two peaks, and their sizes are consistent with ACTH (1-39) and CLIP (ACTH 18 -39). The relative levels of these peptides vary in the two lobes of WT mice pituitary (Fig. 6, B and E). Both peptides are substantially reduced in both lobes in Cpe fat mice (Fig. 6, B and E). ir-␣-MSH appears as a single peak in both Al and NIL of the pituitary of WT mice, and its levels are significantly reduced in both lobes of Cpe fat mice (Fig. 6, C and F). It seems likely that the differences in PCs leads to the differences in peptide levels observed between the pituitary lobes of WT and Cpe fat mice. Since size exclusion chromatography does not differentiate among different molecular forms of ␤-endorphin, ␣-MSH, and ACTH peak fractions from gel filtration chromatography of NIL were analyzed by RP-HPLC. By this analysis we found four peaks of ir-␤-endorphin corresponding to the desacetyl ␤-endorphin 1-31, ␣-N-acetyl ␤-endorphin 1-31, ␣-N-acetyl ␤-endorphin 1-27, and ␣-N-acetyl ␤-endorphin 1-26 in WT mice (Fig. 7). In contrast, a similar analysis of Cpe fat mice reveals an increase in ␣-N-acetyl ␤-endorphin 1-31, with a substantial decrease of ␣-N-acetyl ␤-endorphin 1-27 and a complete loss of ␣-N-acetyl ␤-endorphin 1-26. These data point to a reduction in the C-terminal processing of ␤-endorphin in Cpe fat mice. The major component of the ACTH immunoreactivity in the WT and Cpe fat NIL has the retention time of CLIP (Fig. 7). We found that the level of ir-CLIP is reduced enormously in Cpe fat mice as compared with WT mice (Fig. 7). Interestingly, the main form of ␣-MSH immunoreactivity in the WT mice elutes as diacetyl-␣-MSH; a low levels of desacetyl ␣-MSH and monoacetyl-␣-MSH are also seen (Fig. 7). We found that the relative proportions of nonacetylated and acetylated forms of ␣-MSH in WT mouse pituitary are similar to those reported previously (39). Although the same forms are detected in Cpe fat mice, the levels of all these forms were highly reduced (Fig. 7). These data demonstrate that POMC processing in both lobes of the pituitary of Cpe fat mice is greatly reduced at the level of endo-and exoproteolysis, and this significantly affects the level of differentially processed peptides. Taken together these results support the notion that changes in the level of PC1 and PC2 enzymes affect the processing of neuropeptides and peptide hormones, requiring the action of both convertases for complete cleavage of the precursors.

DISCUSSION
The primary finding of the present study is that the two major neuropeptide-processing endoproteases of the secretory pathway, namely PC1 and PC2, are altered in mice lacking active CPE. The level of the more active form of PC1 is decreased with a concomitant increase in the lesser active form of the enzyme in Cpe fat mice. Furthermore, the level and activity of PC2 are increased in these mice. These enzymes are regulated at multiple levels. PC1 and PC2 are synthesized as zymogens and require proteolytic release of the pro-region for activity. PC1 undergoes an autocatalytic intramolecular processing of its N-terminal pro-fragment in the endoplasmic reticulum, resulting in an 87-kDa active enzyme (33,40). This 87-kDa form is targeted to the regulated secretory pathway, where it is further shortened by removal of 135 amino acids from its C-terminal tail, leading to a 66-kDa form (Ref. 36 and references therein). The C-terminal tail of PC1 has been reported to act as an inhibitor of the enzyme (36). It is possible that in Cpe fat mice, accumulation of the C-terminal tail with an Arg-Arg at the C-terminal end inhibits PC1 activity, thus pre-venting further activation of the enzyme and leading to an increase in lesser active forms of the protein (Fig. 4A). It is possible that in WT mice CPE removes the C-terminal extension of the C-terminal tail of PC1, releasing the inhibition. Another possibility is that in Cpe fat mice PC1 is inhibited by endogenous inhibitors such as the newly discovered proSAAS (41). In vitro studies show that purified proSAAS inhibits PC1 activity, and overexpression of proSAAS in AtT-20 cells leads to a reduction in the extent of POMC processing (41).
PC2 is also synthesized as a zymogen of 75 kDa that undergoes proteolysis to yield 68 -71-kDa forms (33)(34)(35). The binding protein, 7B2, regulates the maturation of PC2. 7B2 contains two domains, a 21-kDa N-terminal domain required for PC2 maturation and a CT region that inhibits PC2 at nanomolar concentrations. There is evidence that the 7B2 CT peptide is cleaved at the internal paired basic site, most likely by PC2 itself (38). It is likely that the resultant peptide (CT peptide 1-18) remains associated with PC2, inhibiting the enzyme. We found that C-terminal basic residue-containing peptides inhibit PC2 activity. The chronic inhibition of PC2 in Cpe fat mice may result in a compensatory increase in the level of the enzyme ( Fig. 4B and Table I). Thus efficient removal of the pro-segment, C-terminal tail and endogenous inhibitors might represent multiple regulatory steps in the activation of PC1 and PC2 and hence in the processing of peptide hormones.
A considerable body of evidence implicates PC1 and PC2 in the endoproteolytic processing of POMC (Ref. 39 and references therein). PC1 is capable of cleaving POMC to ACTH and ␤-LPH in heterologous expression systems. Overexpression of PC1 in AtT20 cells speed up initial steps of POMC processing, which leads to a more extensive cleavage of the precursor to smaller products by PC2 (42). Antisense RNA to PC1 has also been shown to block the processing of POMC in AtT20 cells. The reduction in the level of ir-ACTH and ␤-LPH in AL of Cpe fat mice (Fig. 6, A and B) is in a good agreement with the decreased level of the active form of PC1. ACTH and ␤-LPH are derived from POMC by PC1 in the corticotropic cells of the anterior pituitary and are further processed by PC2 in NIL with the formation of ␣-MSH, CLIP (ACTH18 -39), ␥-LPH, and ␤-endorphin. A decrease in the level of ir-ACTH in AL of Cpe fat mice leads to a reduction in the amount of ␣-MSH and CLIP formed by PC2 in NIL (Figs. 6, E and F) despite increased levels of the enzyme. The level of ir-␤-endorphin in NIL (Fig. 6D) is virtually unchanged, indicating that the latter may be formed in vivo by the action of PC2 directly from the POMC precursor. This is consistent with results from heterologous expression systems, where it was shown that PC2 is able to process POMC, leading to the formation of ␤-endorphin (16). These data are consistent with the hypothesis that the production of peptides requiring the action of both PC1 and PC2 is reduced in Cpe fat mice pituitary. An impaired pattern of processing is also seen for ProDyn. PC2 is the major enzyme involved in the generation of small mature Dyn peptides (13,14). We found that the PC2 activity is increased about two times in Cpe fat mice as compared with WT; however, the same degree of increase in the processing of ProDyn is not seen. It should be pointed out that the PC2 activity was determined under conditions that are substantially different than the physiological environment where the concentration of C-terminally extended peptides required to inhibit PC2 activity could be relatively high. Although these concentrations can occur in an immature or mature secretory granule, which would result in partial inhibition of PC2 activity, tissue homogenization might result in a significant dilution of these peptides and, thus, release the enzyme from the inhibitory intracellular effect. Thus, the 2-fold increase in PC2 activity determined in vitro might not represent the actual activity in the context of the cellular environment.
We found an accumulation of the acetylated form of ␤-endorphin 1-31, a substantial decrease in the acetylated form of ␤-endorphin 1-27, and an absence of acetylated ␤-endorphin 1-26 in Cpe fat mice pituitary (Fig. 7). This finding is in a good agreement with a previous report showing that CPE catalyzes the conversion of ␤-endorphin 1-27 into ␤-endorphin 1-26 (43). It is possible that CPE also catalyzes the C-terminal proteolysis of acetylated ␤-endorphin 1-31, which terminates in Gln; in the absence of CPE, we found an accumulation of the latter in Cpe fat mice (Fig. 7). In the absence of CPE one would expect an increase in the ir amount of ␤-endorphin 1-31 in the hypothalamus of Cpe fat mice. ir-␤-endorphin 1-31 is a potent stimulator of feeding (Ref. 10 and references therein). In contrast, ␤-endorphin processing through C-terminal proteolysis or N-terminal acetylation eliminates the effect on food intake. Altogether, these data indicate that an accumulation of Dyn A-17 and ␤-endorphin 1-31 with a concomitant decrease in the amount of ir-acetylated ␣-MSH (potent inhibitor of feeding) may contribute to the development of obesity in Cpe fat mice.
We found that POMC mRNA levels were not increased in the brain and pituitary of Cpe fat mice as compared with WT. However, we did see an increase in POMC mRNA in Cpe fat as they develop severe obesity (data not shown). This finding is consistent with a previous report showing no significant difference in POMC mRNA levels between 5-week-old fatty and lean rats, even though significantly higher POMC mRNA levels were observed in 12-week-old fatty rats compared with lean littermates (44). Thus the difference in POMC mRNA content between lean and obese animals becomes apparent as they grow and develop severe obesity. Moreover, it has been demonstrated that leptin, usually rising with obesity, stimulates expression of anorexigenic peptides such as POMC (45,46). However, low concentrations of leptin (usually observed with overfeeding or moderate obesity) did not affect the expression of POMC mRNA level (Ref. 46 and references therein).
The distinguishing characteristics of mice homozygous for the Cpe fat mutation are early and severe hyperproinsulinemia, transient hyperglycemia, but no hypercorticosteronemia (1). Moderate obesity develops progressively, starting between 8 and 12 weeks of age. The molecular mechanism by which inactivation of CPE leads to the development of obesity in these animals is unclear. Our data are consistent with a deficit in CPE activity affecting maturation of major endoproteolytic enzymes (PC1 and PC2) as well as neuropeptides derived from ProDyn and POMC. An abnormal neuropeptide and hormone precursor processing is a general phenomenon in Cpe fat mice. Defects in the production of neuropeptides such as ␣-MSH, ␤-endorphin, dynorphins, melanin-concentrating hormone, neurotensin (5), gastrin (6,7), and cholecystokinin (8) suggest multiple roots for the development of obesity. Critical to understanding the etiology of obesity is detailing the processes that occur after initiation of the obesity-stimulating event. It remains to be elucidated if the obesity in Cpe fat mice is induced by the disruption of the melanocortin-and leptin-signaling pathway (47,48) or a new uncharacterized pathway. Five novel peptides derived from a common precursor proSAAS have been identified in the brain of mice lacking active CPE (41). It is possible that some of these or hitherto undiscovered peptides may promote obesity through a novel and/or already established pathway.