Induced and spontaneous mutations at Ser202 of carboxypeptidase E. Effect on enzyme expression, activity, and intracellular routing.

Carboxypeptidase E (CPE) is involved in peptide processing in the brain and various neuroendocrine tissues. In mice homozygous for the Cpefat mutation, the virtual absence of CPE activity in islets of Langerhans and pituitary was associated with a missense mutation effecting a Ser202 to Pro shift (Naggert, J. K., Fricker, L. D., Varlamov, O., Nishina, P. M., Rouille, Y., Steiner, D. F., Carroll, R. J., Paigen, B. J., and Leiter, E. H. (1995) Nat. Genet. 10, 135-142). To examine the importance of Ser202 in CPE function, several mutations in this position were generated (Pro202, Ala202, Gly202, and Phe202). When the mutant proteins were expressed in a Baculovirus system, both Phe202 and Pro202CPE were enzymatically inactive, were unable to bind to a substrate affinity column, and were not secreted from Sf9 cells. In contrast, Ala202CPE or Gly202CPE exhibited enzymatic properties similar to those of wild-type CPE and were secreted from Sf9 cells. When expressed in AtT-20 cells, a mouse pituitary-derived cell line, CPE with Pro202 and Phe202 were not secreted. Pulse-chase analysis with [35S]Met indicated that Pro202CPE was degraded in AtT-20 cells within several hours. This degradative process was blocked by incubation at 15 degrees C but not by brefeldin A or by lysosomotrophic drugs. Pulse-chase analysis using dispersed pituitary cells from C57BLKS/Lt-Cpefat/Cpefat mutant mice shows similar results; Pro202-CPE produced in these cells was not secreted but rather was degraded within 5 h. Immunofluorescence analysis of epitope-tagged CPE revealed Ser202CPE to be present primarily in secretory vesicles, whereas Pro202CPE was localized to the endoplasmic reticulum and not the secretory vesicle-like structures. These results support the previous finding that Cpefat/Cpefat mice are defective in CPE activity because of the point mutation producing the Ser202 to Pro substitution. Furthermore, these results are consistent with a model that Ser202 is important for the intracellular folding of CPE.

Most peptide hormones and neurotransmitters are produced from larger precursors by limited proteolysis. Initially, en-dopeptidases such as prohormone convertase 1, prohormone convertase 2, and furin cleave the prohormone precursor at multiple basic amino acid cleavage sites (1)(2)(3)(4)(5). Then, a carboxypeptidase removes the basic amino acids from the C terminus of the peptide to generate either the bioactive product or a precursor for the formation of the C-terminal amide group (6,7). The carboxypeptidase step is important for the production of numerous bioactive peptide in many tissues (6,7). Carboxypeptidase E (CPE, 1 also known as carboxypeptidase H and enkephalin convertase) is the major carboxypeptidase involved with the processing of many peptides (6,7). A second enzyme, carboxypeptidase D, also may participate in the intracellular processing of peptide hormones and neurotransmitters, although the exact role of carboxypeptidase D is not yet clear (8). Both CPE and carboxypeptidase D are members of the metallocarboxypeptidase gene family, which includes carboxypeptidases A, B, M, and N and AEBP-1, a recently described transcription repressor (9 -16).
The importance of CPE in peptide processing is evident from the finding that C57BLKS/Lt-fat/fat mice have a reduced ability to convert proinsulin into insulin (17). The mutation responsible for the defect in these mice has been localized to a single point mutation within the coding region of the Cpe gene on chromosome 8. The mutant allele, now designated Cpe fat , contains proline in place of serine at residue 202. Ser 202 is conserved in human, bovine, rat, mouse, fish, and Aplysia CPE (18 -21) 2 and is also conserved in human carboxypeptidase M, human carboxypeptidase N, and human, rat, duck, and Drosophila carboxypeptidase D (11,13,22,23). 3 However, in the digestive enzymes (CPA and carboxypeptidase B), the amino acid in the corresponding position is either Phe (CPA) or Tyr (carboxypeptidase B). Furthermore, in the bacterial metallocarboxypeptidases, the amino acid in a comparable position is Ala (Streptomyces griseus), His (Streptomyces capreolus), Leu (Thermoactinomyces vulgaris), and Val (Bacillus cereus). The crystal structures of carboxypeptidases A, B, and T (the T. vulgaris carboxypeptidase) have been determined (24 -26) and are generally similar despite the moderate amount of amino acid homology between the various sequences (approximately 50% amino acid identity between CPA and carboxypeptidase B and approximately 25% amino acid identity between CPA and carboxypeptidase T). In all structures, the residue in the position that corresponds to Ser 202 of CPE (Phe 192 of bovine CPA) is present in a ␤ sheet. Substitution of a Pro into this position of CPA is predicted to perturb the structure. 4 There is virtually no CPE activity in Cpe fat /Cpe fat mouse pituitary or pancreatic islets. Although a small amount of CPElike activity is detected in these tissues, this action is due entirely to other metallocarboxypeptidases, 5 such as carboxypeptidase D (8). The absence of CPE in Cpe fat /Cpe fat mouse tissues is consistent with our previous finding that CPE with a substitution of Pro 202 is not active when expressed in the Baculovirus system (17). In this study, we investigated whether other amino acids can substitute for Ser in position 202. The amino acids chosen for this substitution are those that are present in a comparable position in other metallocarboxypeptidases (Phe, Ala, Gly). In addition, we examined CPE with a Pro 202 in further detail. The results of this analysis indicate that Ser 202 is not essential for the production of enzymatically active CPE since Ala or Gly can replace the Ser. Analysis of the fate of the Pro 202 CPE in AtT-20 cells and in Cpe fat /Cpe fat mice is consistent with the hypothesis that this mutant form of CPE is misfolded and subsequently degraded within the cell.

MATERIALS AND METHODS
Generation of Constructs-Site-specific mutations of the Ser 202 in CPE were generated using the Altered Sites II in vitro mutagenesis system (Promega). Rat CPE cDNA was subcloned into the BamHI and EcoRI sites of the vector pALTER-1, and the mutagenic oligonucleotides with different codon changes in position 202 were used to generate CPE with the single amino acid substitutions. The constructs were subcloned into the BamHI and EcoRI sites of the Baculovirus transfer vector pVL1393 or into eukaryotic expression vector pcDNA-3. For eukaryotic expression, the epitope recognized by the 12CA5 monoclonal antisera (YPYDVPDYA) was inserted into a PstI site located immediately downstream of the pentabasic propeptide cleavage site, with the resulting amino acid sequence RRRR RLQSRYPYDVPDYALQQED . . . (the gap indicates the propeptide cleavage site). Dideoxynucleotide sequencing was performed to confirm the sequence of the mutated region of CPE.
Expression of Proteins in AtT-20 Cells-The expression plasmids containing the sequences encoding CPE Ser 202 mutants were transfected into AtT-20 cells using the standard calcium phosphate procedure (27). Stably transfected cells were selected using 0.6 mg/ml of geneticin (G418). Cell expression was detected by Western blot analysis and by immunoprecipitation using the 12CA5 monoclonal antiserum raised against HA epitope tag (a gift of Dr. Jonathan Backer, Molecular Pharmacology, Albert Einstein College of Medicine).
Expression of Mutant Proteins in Baculovirus-Recombinant Baculovirus expressing different constructs was generated using the Baculovirus system (Pharmingen) as described (17). The cells from 25 ml of the culture were recovered by centrifugation at 1000 ϫ g for 10 min. The cells were homogenized (Polytron, Brinkmann Instruments) in 5 ml of 100 mM sodium acetate buffer, pH 5.5. Aliquots of cell homogenate and media were assayed for carboxypeptidase activity using 0.2 mM dansyl-Phe-Ala-Arg in 100 mM sodium acetate buffer, pH 5.5, as described (28). Protein was determined using the Bradford reagent. Aliquots were also analyzed on a Western blot as described (17). To purify wild-type CPE and Ser 202 CPE mutants, the cell homogenates were adjusted to 10 ml with 1 M NaCl, 1% Triton X-100 in 100 mM sodium acetate buffer, pH 5.5; sonicated; and centrifuged at 30,000 ϫ g for 30 min. The supernatants containing cellular CPE were subjected to purification on a paminobenzoylarginine-Sepharose 6B affinity column as described (29). The CPE was recovered in 2 ml of 50 mM Tris buffer, pH 8, containing 150 mM NaCl and 0.01% Triton X-100. The ability of the various CPE constructs to bind the column was assessed by Western blot analysis using an antiserum against the CPE C-terminal sequence KMMSETLNF, as described (30).
Pulse-Chase Analysis of AtT-20 Cells at Different Conditions-AtT-20 cells expressing wild-type CPE and Pro 202 CPE were labeled (pulse) with [ 35 S]Met (60 Ci/ml) for 20 min and then washed twice with Dulbecco's modified Eagle's medium (DMEM). Some of the plates were incubated without radiolabel (chase) for 30 min in control media or for 3 h under different conditions: incubation at 15°C or at 37°C in media containing brefeldin A (5 g/ml), chloroquine (100 M), or NH 4 Cl (10 mM) or no addition. These conditions have been used previously to study the processing of proCPE in the AtT-20 cell line (38). Media were removed and the cells were washed with phosphate-buffered saline (PBS) and then frozen in 10 mM sodium acetate buffer, pH 5.5, with 1 mM phenylmethylsulfonyl fluoride. The cells and media were then subjected to immunoprecipitation as described (31) using the 12CA5 antiserum against the HA epitope.
Pulse-Chase Analysis of Freshly Dispersed Pituitary Cells-Pituitary glands were removed from five 20-week-old C57BLKS/Lt-Cpe fat /Cpe fat females and from five lean littermate controls (mixture of ϩ/ϩ and ϩ/Cpe fat genotypes). The pituitaries were rinsed in calcium-magnesium-free Hanks' solution, minced with fine scissors, and washed three times in calcium-magnesium-free Hanks' solution. The fragments were then digested for 7 min in a sterile Eppendorf plastic centrifuge tube at room temperature in 250 l of 0.25% trypsin (Life Technologies, Inc.) in calcium-magnesium-free Hanks' solution. Following this treatment, 500 l of DMEM (5.5 mM glucose) containing 10% heat-inactivated and dialyzed fetal bovine serum (Life Technologies, Inc.) were added, and the material was triturated using a siliconized Pasteur pipette. After three washes in this medium, cells were allowed to recover in the above medium plus serum for 45 min. Following this recovery period, the cell yield for each group (control versus mutant) was divided between two Eppendorf tubes, and each aliquot was incubated for 1 h at 37°C in 500 l of methionine-free DMEM plus 10% fetal bovine serum. This medium was then removed and replaced with 500 l of the methionine-free DMEM containing 300 Ci of [ 35 S]Met (ICN Trans-Label). The cells were labeled for 1.5 h. The radioactive medium was then removed, and cells in each tube were rapidly washed and gently pelleted three times in DMEM supplemented with 10 mM unlabeled methionine and 10% fetal bovine serum. One aliquot from control and mutant cell populations was then extracted in 75 l of 50 mM Tris (pH 8.0), 2% SDS, and the extract was boiled for 3 min and then frozen. The remaining pair of tubes was refed with 500 l of serum-containing DMEM and incubated for 5 h (chase). The chase medium was removed and saved, and the cells were extracted with 2% SDS in 50 mM Tris, pH 7.4, followed by heating at 95°C. Aliquots of the media and cells were subjected to immunoprecipitation as described, using an antiserum raised against the N-terminal region of CPE (29,31).
Immunofluorescence-Transfected and wild-type AtT-20 cells were cultured on 18-mm coverslips precoated with 1 mg/ml polylysine (Sigma). Cells were washed with PBS, fixed in 4% paraformaldehyde for 15 min, and then permeabilized for 15 min in 0.1% Triton X-100 in PBS. After 1 h of blocking in 3% bovine serum albumin, the cells were immunostained for 1 h with the primary antisera: mouse monoclonal antisera 12CA5 against the HA epitope (1:10,000 dilution); rabbit polyclonal antisera directed against the C-terminal region of CPE (1:5,000 dilution); or rabbit polyclonal antisera to calnexin (1:1000). Cells were washed three times with PBS containing 0.2% Tween 20 and then incubated with fluorescein-labeled anti-mouse or anti-rabbit secondary antibody (Vector Laboratories Inc., 1:200 dilution) for 1 h followed by extensive PBS washing. Immunofluorescence staining was examined using a Nikon confocal microscope.

RESULTS
On expression in Sf9 insect cells using the Baculovirus system, CPE with Ser 202 , Gly 202 , Ala 202 , Phe 202 , or Pro 202 was expressed at relatively high levels in the cells (Fig. 1, left). However, the amount of CPE protein detected by Western blot analysis does not correlate with the amount of CPE activity present in the same cell extracts; only CPE with Ser 202 , Gly 202 , and Ala 202 showed carboxypeptidase activity greater than the amount of activity detected in cells infected with control Baculovirus (Table I). Analysis of the media from infected cells revealed that only the forms of CPE that showed enzyme activity (Ser 202 , Gly 202 , and Ala 202 ) were secreted in detectable amounts (Fig. 1, middle). CPE protein could not be detected by Western blot analysis of media from cells expressing the Phe 202 or Pro 202 mutants (Fig. 1, middle), and the amount of CPE-like enzyme activity in these fractions was comparable to the level in media from cells infected with control virus (Table I). In two separate experiments, the level of CPE-like activity in the cells and media varied as much as 50% for the same construct, which is presumably due to variables associated with the viral infection. In both experiments, cells expressing Ser 202 CPE, Gly 202 CPE, or Ala 202 CPE showed high levels of CPE activity, and cells expressing Pro 202 or Phe 202 CPE showed background levels. To investigate which of the CPE variants in the cell extracts were capable of binding to a C-terminal Arg, extracts were purified on a p-aminobenzoylarginine-Sepharose affinity column. Only the enzymatically active forms of CPE (Ser 202 , Gly 202 , and Ala 202 ) were bound by the substrate affinity resin (Fig. 1, right).
The properties of CPE purified from the media of Baculovirus-infected Sf9 cells were examined to determine whether substitution of Ser 202 by Gly or Ala altered the enzymatic properties. All three forms of CPE had acidic pH optima in the 5-6 range, with essentially no activity detectable at pH 7.4 (not shown). CoCl 2 (1 mM) activated all three forms approximately 620 -690%, and NiCl 2 (1 mM) activated all forms by 65-83% (not shown). ZnCl 2 , CaCl 2 , and MgCl 2 (1 mM) had no substantial effect on enzyme activity. All three forms were similarly inhibited by 1 mM CdCl 2 and, more potently, by 1 M CuSO 4 and HgCl 2 (not shown). The metal-chelating agent 1,10-phenanthroline inhibited all three forms of CPE (not shown).
The stability of CPE to thermal denaturation was investigated by preincubating the purified CPE at various temperatures and then assaying for activity at 37°C. The mutation of Ser 202 to Ala or Gly did not substantially alter the stability of CPE (Fig. 2). All three forms of CPE are stable between 4°C and 42°C, and lose activity upon incubation at temperatures greater than 47°C (Fig. 2). The results shown in Fig. 2 were performed at pH 7.4; similar results were obtained at pH 5.5 (not shown).
To investigate whether Ser 202 is important for the expression of CPE in a mammalian cell line, several of the Ser 202 mutations were modified to contain the influenza hemagglutinin epitope recognized by the monoclonal antisera 12CA5 at the N terminus of the mature form of CPE. This epitope tag, which increases the molecular mass of the 56-kDa proCPE by approximately ϳ2 kDa, is necessary to distinguish mutant CPE from endogenous wild-type CPE present in AtT-20 cells. The epitope tag in the N-terminal position does not interfere with enzyme activity, thermostability, or sorting to the regulated secretory pathway. 6 As found for Baculovirus-mediated expression in Sf9 cells, epitope-tagged CPE with Ser 202 mutated to Pro, Gly, or Phe was expressed in the cells, and the Ser 202 and Gly 202 forms of CPE were secreted (Fig. 3). Similar results were found when the various forms of CPE were transiently expressed in COS cells and examined by immunoprecipitation and Western blot analysis (not shown). In addition to the 58-kDa epitope tagged CPE, the 12CA5 antisera detected several 40-to 50-kDa proteins in the AtT-20 cells; since these proteins were also present in the untransfected AtT-20 cells (Fig. 3, wt), they are not related to the epitope-tagged CPE. When AtT-20 cell extracts expressing the various mutations were purified on a substrate affinity resin, only the Ser 202 CPE and Gly 202 CPE were bound to the resin (data not shown).
AtT-20 cells expressing epitope-tagged Ser 202 CPE or Pro 202 CPE were examined using [ 35 S]Met pulse-chase analysis. As previously found for wild-type CPE, [ 35 S]Met-labeled epitope-tagged Ser 202 CPE shows a major band around 58 kDa after a 20-min pulse (Fig. 4, left). After a 3-h chase, this band is substantially reduced, which is due to secretion into the media and proteolytic processing which removes N-and C-terminal Met residues (33). When the chase is performed at 15°C, which blocks transport between the endoplasmic reticulum and Golgi, or performed in the presence of brefeldin A, the amount of epitope-tagged Ser 202 CPE is slightly reduced, compared with the level in the "no-chase" control ( Fig. 5, middle). Compounds that interfere with the acidification of granules and lysosomes (ammonium chloride, chloroquine) had little effect on the decrease in levels of cellular Ser 202 CPE (Fig. 5, middle).
The Pro 202 form of CPE also showed a major band of approximately 58 kDa after labeling for 20 min with [ 35 S]Met (Fig. 4,  right). Unlike Ser 202 CPE, the apparent size and the intensity of the labeled Pro 202 CPE did not change following 30 min of chase. After 3 h of chase, the intensity of this band was considerably reduced, although in contrast to Ser 202 CPE, no radiolabeled Pro 202 CPE could be detected in the media (Fig. 4,  right). Incubation of the cells at 15°C prevents the degradation of Pro 202 CPE (Fig. 5, right). Densitization of these data, and 6 L. Song and L. Fricker, unpublished observations.

FIG. 1. Expression of Ser 202 CPE mutants in Sf9 cells.
The cells were infected as described under "Materials and Methods," collected by centrifugation, then extracted with 10 ml of 1 M NaCl, 1% Triton X-100 at pH 5.5, and centrifuged at 30,000 ϫ g. Either 10 l of the cell extract (left) or 50 l of the media (middle) were analyzed on a Western blot, using an antiserum directed against the C-terminal region of CPE to detect the protein (30). Right, CPE was purified using p-aminobenzoylarginine columns, as described (29). CPE was eluted in 2 ml of 50 mM Tris containing 150 mM NaCl and 0.01% Triton X-100 at pH 8, and 20 l of this eluate were analyzed on a Western blot. The positions and molecular masses (in kDa) of prestained protein standards are indicated. bv, control Baculovirus.  2. Thermostability of CPE. Affinity-purified Ser 202 CPE (q), Gly 202 CPE (f), or Ala 202 CPE (å) were diluted with 5 mM Tris-Cl, pH 7.4, containing 0.05% Triton X-100 and incubated at the indicated temperatures for 1 h. The reaction mixture was then combined with dansyl-Phe-Ala-Arg and sodium acetate buffer (final pH, 5.5), and the samples were incubated at 37°C. Product was detected using the standard procedure (28). Error bars, standard error (n ϭ 3). data from two additional experiments showed the 15°C group to be approximately 90% of the 'no chase' control group. In contrast, brefeldin A, chloroquine, and ammonium chloride had no effect on the stabilization of Pro 202 CPE. Similar results were observed in two additional experiments. These results are consistent with previous studies on a variety of unfolded proteins that have been found to be degraded by a temperature-sensitive endoplasmic reticulum or Golgi process that is not sensitive to lysosomotrophic agents (34 -37).
Pulse-chase analyses of pituitaries from control and Cpe fat/ Cpe fat mice were performed to determine whether the Pro 202 CPE, which occurs naturally in the Cpe fat mice, was also degraded before secretion. After labeling for 1.5 h, strong bands of radiolabeled CPE (approximately 52-55 kDa) were detected after immunoprecipitation in both the genotype-normal control and Cpe fat pituitary cell suspensions (Fig. 6). Whereas the Cpe fat pituitary cells showed a single band of 55 kDa for CPE, the genotype-normal cells showed two bands of 52 and 55 kDa, consistent with previous reports of the size of CPE in cell lines and tissues (29,38). A 5-h chase was associated with a large reduction in the level of labeled CPE in pituitaries from both mice. However, CPE is detected in the media only for the genotype control cells and not from the Cpe fat pituitary cells (Fig. 6).
To examine the intracellular distribution of Ser 202 and Pro 202 CPE in the AtT-20 cells, the stably transfected cells were fixed, permeabilized, and probed with an antiserum to the HA epitope. Epitope-tagged Ser 202 CPE showed punctate staining throughout the cell body, with dense staining at the tips of the cell processes (Fig. 7). In addition, some staining was observed in a perinuclear region (Fig. 7). This distribution of epitopetagged Ser 202 CPE is similar to the distribution of endogenous CPE in wild-type AtT-20 cells, as visualized with an antiserum directed against the C-terminal region (Fig. 7). The endogenous CPE is concentrated in the tips of the cell processes and shows punctate staining throughout the cell body (Fig. 7). This distribution of CPE is consistent with the distribution of other secretory vesicle proteins in AtT-20 cells (39,40). In contrast, Pro 202 CPE shows a diffuse, reticular network-like distribution throughout the cell body (Fig. 7). This distribution is similar to that of calnexin (Fig. 7), a resident endoplasmic reticulum protein (41). Wild-type AtT-20 cells showed negligible staining when probed with the antisera directed against the HA epitope (Fig. 7). Also, approximately 80% of the cells expressing either Ser 202 or Pro 202 CPE were not labeled with the HA epitopedirected antisera (shown for Pro 202 CPE in Fig. 7), suggesting that the mixed population of stably transfected cells included many cells that did not express the CPE construct. In contrast, all of the cells were labeled with antisera directed against CPE or calnexin. DISCUSSION A major finding of this study is that Ser 202 is not essential in the function of CPE since this residue can be replaced with Gly or Ala. The enzymatic properties of these two mutant forms of CPE are comparable to the properties of wild-type (Ser 202 ) CPE. This finding suggests that Ser 202 is not an essential residue in the folding of CPE, despite the conservation of this residue in CPE from various species and in other related carboxypeptidases such as carboxypeptidase M and carboxypeptidase N. In contrast, replacement of Ser 202 with Pro or Phe causes the complete loss of CPE activity. This effect is presumably due to misfolding of the protein since incorrectly folded proteins are typically degraded in the cell and not secreted. The finding that CPE with a Pro 202 substitution is inactive is consistent with our previous finding that this mutation, which is found in the Cpe fat mouse, leads to a loss of enzyme activity (17). The inability of Phe to substitute for Ser 202 is not surprising since the Phe is much larger than the Ser. Although a Phe is found in the comparable position of CPA and B, there are numerous other differences between these enzymes and CPE; the amino acid identity is only 20% between CPE and either CPA or carboxypeptidase B.
Our finding that Ser 202 can be mutated to Gly or Ala without adversely affecting the enzyme activity implies that if this Ser is a site for post-translational modification (phosphorylation or glycosylation) these modifications are not essential. It is unlikely that Ser 202 is a phosphorylation site since the consensus sequence for phosphorylation of secretory pathway proteins is Ser-X-Asp/Glu (42), which is not present in this region of CPE. The region of CPA, carboxypeptidase B, and carboxypeptidase T that is comparable to Ser 202 of CPE is a ␤ sheet that is buried in the interior of the protein (24 -26). If CPE has a similar structure for this region, Ser 202 would be an unlikely site for phosphorylation or glycosylation.
The finding that Pro 202 CPE is inactive in both Sf9 cells and AtT-20 cells is consistent with the identification of a single Ser 202 to Pro point mutation within CPE as the defect associated with the Cpe fat mutation (17). These mice have essentially no CPE activity; the small amount of CPE-like activity detected in pituitary and pancreatic islets is entirely due to carboxypeptidase D and other enzymes. 5 However, the effect of the CPE mutation on prohormone processing is complicated. A defect in CPE would be predicted to interfere only with the step catalyzed by CPE, the removal of C-terminal Lys and Arg residues from intermediates formed by the action of prohormone convertases 1 and 2 on the prohormone (43). However, proinsulin accumulates in very high levels in Cpe fat mice, suggesting that the defect in CPE affects the previous enzymatic step (17). In considering the mechanism by which the defect in CPE inter-feres with the endopeptidase processing of proinsulin, there does not appear to be a dominant negative effect of mutant CPE. Animals that are heterozygous for the Cpe fat mutation and that presumably produce Pro 202 CPE in addition to wildtype CPE have levels of proinsulin in the normal range (17).
There are several possible mechanisms by which the defect in CPE could interfere with the previous endopeptidase step. The high intragranular levels of C-terminally extended forms of insulin conversion intermediates (resulting from the absence of CPE) could mediate feedback inhibition on the prohormone convertase activities. A related possibility is that an endogenous inhibitor of the endopeptidases is inactivated by CPE, and the absence of CPE then leads to a decrease in prohormone convertase activity. Evidence for this possibility has recently been reported, although the inhibitor (7B2) blocks only prohormone convertase-2 activity and not prohormone convertase 1 (44). Another possibility is that CPE is directly involved in the activation of the prohormone convertases, perhaps by cleaving the C-terminal basic residues from the propeptide of these endopeptidases. Alternatively, it is possible that CPE plays a role in the transport and/or sorting of the endopeptidases into the regulated secretory pathway. From this study, the finding that Pro 202 CPE appears to be degraded in an endoplasmic reticulum or Golgi compartment and not sorted into the regulated pathway is important for understanding the defect of the Cpe fat mice. Thus, the defect is not merely in the absence of CPE activity but also in the absence of CPE protein from the regulated secretory pathway. CPE has recently been found to aggregate at acidic pH in the presence of Ca 2ϩ , and this may contribute to the sorting of CPE into the regulated pathway (45). It is possible that aggregation of CPE is important for the sorting of other proteins, and the absence of CPE protein, rather than CPE activity, is responsible for the apparent defect in the endopeptidase cleavage step.
The finding of comparable translation of CPE in control and Cpe fat pituitary cells immediately after the [ 35 S]Met pulse is consistent with the results of Pro 202 CPE expressed in AtT-20 cells and with the previous report that levels of CPE mRNA are not reduced in endocrine tissues of Cpe fat mice (17). This result is also consistent with the results of Pro 202 CPE expressed in Sf9 cells; in all cases, there are large amounts of mutant CPE found in the cells. Thus, the defect appears to be one of protein stability, presumably due to an altered structure caused by the Ser 202 to Pro mutation.