Processing site blockade results in more efficient conversion of proenkephalin to active opioid peptides.

Prohormones are known to be processed at various cleavage sites in a defined temporal order, suggesting the possibility of sequential unfolding of processing sites. In order to investigate whether sequential processing at predefined sites is in fact required for proper processing, site-directed mutagenesis was performed to block known initial cleavage sites within proenkephalin. Pulse-chase/immunoprecipitation experiments were employed to analyze the fate of mutant and native proenkephalins in stably transfected AtT-20 cells. While processing did not occur at blockaded sites, surprisingly, overall processing of mutant proenkephalins proceeded efficiently, and alternative sites were chosen. When compared with native proenkephalin, processing of mutant proenkephalins occurred more slowly at early stages and more quickly at later stages. Experiments employing endoglycosidase H indicated that the early slow processing of mutant proenkephalins may be due to delays in intracellular transport. Metabolic labeling studies showed that more efficient production of bioactive opioids occurred in all processing site blockade mutants examined; these results were confirmed using several different radioimmunoassays of stored peptide products. We conclude that efficient processing of prohormone precursors does not require a specific temporal order of processing events. The fact that mutant proenkephalins were more fully processed than native proenkephalin may provide a route for more efficient production of opioid peptides in applications for chronic pain treatment.

In endocrine and neuronal cells, peptide hormones and proneuropeptides are synthesized as large precursor proteins that are then endoproteolytically processed during intracellular transport to produce the biologically active molecules. After posttranslational modification, these precursors are sorted in the trans-Golgi network and packed into dense core secretory vesicles for storage until release upon stimulation (1,2). Common recognition sites present in these proproteins are pairs of basic amino acid residues, such as the preferentially cleaved Lys-Arg and Arg-Arg sites, and less frequently Lys-Lys and Arg-Lys pairs (reviewed in Refs. 1 and 2). Since not all dibasic pairs are actually cleaved, structural features surrounding these sites are probably important for enzyme recognition and processing (1)(2)(3). Prohormone convertase 1 (PC1, 1 also known as PC3) (4 -6) and prohormone convertase 2 (PC2) (4,7) are thought to be the primary endoproteolytic enzymes responsible for prohormone cleavage in neuroendocrine tissues (2). The exposed C-terminal basic residues are then removed by specific carboxypeptidases.
Proenkephalin (PE), which in AtT-20 cells is synthesized in both N-glycosylated and unglycosylated forms (8), is endoproteolytically cleaved at 12 pairs of basic residues into enkephalin-containing peptides and mature enkephalins. Posttranslational processing of PE in the brain results mainly in the production of penta-to octapeptide opioids, while adrenal medullary chromaffin cells predominantly process this precursor to intermediate sized enkephalin-containing peptides (9,10). AtT-20 cells, a mouse anterior pituitary tumor cell line, represent a convenient model system in which to study prohormone processing. In this cell line, cleavage of PE occurs at Lys-Lys sites to produce intermediate sized peptides (8); similar Lys-Lys cleavage within other precursors such as proopiomelanocortin (11) or proneuropeptide Y (12) occurs inefficiently, suggesting that structural features within proenkephalin are permissive for cleavage at this site. Site-directed mutagenesis of prohormones at cleavage sites has been used to investigate the importance of particular types of sites in order to potentially relate site usage to the action of the two prohormone convertases. Many prohormones have been studied in this manner, among others proinsulin (13), prosomatostatin (14), proopiomelanocortin (15), proneuropeptide Y (12), egg-laying hormone (16), and insulin-like growth factor 1 (17).
Processing of precursor molecules including PE is known to occur in a sequential order, i.e. a predefined order exists with regard to the various cleavage events (1,8,18). The elements controlling this sequence are unclear. A possible mechanism underlying this phenomenon is that cleavage at one site causes conformational rearrangements in the molecule, which are required to expose subsequent processing sites. Indeed, evidence that initial cleavage of proinsulin by PC1 is required prior to PC2 cleavage has been presented (19). PE is initially cleaved at a site that results in the production of Peptide B; shortly thereafter, another cleavage at Lys 196 -Arg 197 occurs, resulting in intermediate sized enkephalin-containing peptides that can ultimately be processed to enkephalins (8). If the orderly nature of prohormone processing is due to sequential unfolding and conformational rearrangements following each cleavage, then blockade of processing at initial sites would be expected to block further processing events. The purpose of these experi-ments was to examine the effect of blockade of such initial processing in the model precursor PE. Stably transfected AtT-20 cells were used in this study, since it has been amply demonstrated that these cells have a high rate of biosynthesis of peptide products and are capable of processing exogenous prohormones, including proenkephalin. This cell line was transfected with either native or mutant forms of PE containing blockade(s) at known cleavage sites. Surprisingly, we found that processing of PEs with processing site blockade mutations proceeded more slowly during initial stages but more quickly at later stages, ultimately resulting in higher levels of stored opioid peptides.

Mutagenesis of Proenkephalin and Construction of the Expression
Vectors-Rat preproenkephalin cDNA (20) was excised from the plasmid pEV/rENK (21) and ligated into pRcCMV (ϩ) (Invitrogen) as described previously (8). Three individual mutations and one double mutation of paired basic residues of PE were performed as shown in Fig. 1 (22) and were confirmed by dideoxy DNA method using standard methods. The oligonucleotides employed as mutagenesis primers were as follows: 5Ј-TAGAGACTCAGCAAATTTGTGCAGGAAGCCTCC-3Ј for K1 mutant PE; 5Ј-TTCCAGCTGGGGGCTTTTGTGGAGGCCTCTCAT-3Ј for the K2 mutant PE; 5Ј-TCCCTCATCTGCATCTTTGTGCATGAAAC-CGCC-3Ј for the K3 mutant PE; and 5Ј-CAGGAAGCCTCCGTATTTGT-GCTGATAGTCCAT-3Ј for the K4 mutant PE. Plasmids were produced in Escherichia coli and isolated by standard procedures (23). Before use in transfection procedures, plasmids were purified twice by centrifugation through CsCl 2 gradients.
Cell Culture and DNA Transfection-AtT-20/dv16 cells (obtained from R. E. Mains and B. A. Eipper) were grown in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) containing 2.5% fetal bovine serum (Irvine Scientific, CA) and 10% NuSerum (Irvine Scientific) at 37°C in an atmosphere containing 5% CO 2 . The cells were stably transfected with the following expression plasmids containing the mutations: His 237 -Lys 238 (K1), His 196 -Lys 197 (K2), His 141 -Lys 142 (K3), and His 210 -Lys 211 -His 237 -Lys 238 (K4). The Lipofectin (Life Technologies, Gaithersburg, MD) method was used for transfection. Following removal of transfection medium at 5 h, cells were trypsinized and diluted into a peel-top T-162 flask. Twenty-four stable transfectants resistant to the neomycin analog G418 (0.6 mg/ml, ϳ50% active, Life Technologies, Inc.) were subcloned by the soft agar method as described previously (24). RIA screening was performed to select proenkephalin-expressing clones. Cells were grown in 35-mm wells for 2 days before use to approximately 80% confluence. Cells were scraped into 1 ml of ice-cold 1 M acetic acid, 20 mM HCl, and 0.1% (v/v) ␤-mercaptoethanol (solution A). Cell extracts were frozen and thawed, and insoluble material was removed by centrifugation. Samples were then subjected to a Met-enk-Arg-Phe radioimmunoassay as described previously (25). The highest expressing clones of each cell line were selected for further study.
Pulse-Chase Experiments and Immunoprecipitations-For the pulsechase experiments, subconfluent 35-mm wells (1-2 ϫ 10 6 cells) of native PE-expressing and PE blockade mutant cell lines were subjected to the labeling procedure essentially as described (25). In brief, for the 20-min pulse labeling, 1 ml of methionine-free medium containing 1 mCi of [ 35 S]methionine (average specific activity Ͼ1190 Ci/mmol, Amersham Corp.) was added to each well. Cells were then chased for 0, 0.5, 1, 2, or 4 h. After each chase incubation, cells were scraped into 1 ml of ice-cold solution A, and extraction of PE-derived peptides was performed as described earlier (25). Immunoprecipitation was also performed as described previously (8, 25) using 0.1 ml of the cell extracts and two different enkephalin antisera added to different sets of samples. The antisera used were raised against Peptide B/Met-enk-Arg-Phe (designated Xandra) (8) and the 8.6-kDa peptide, Peptide F/Met-enk (designated Betty). Antiserum Betty was raised to a conjugate of Met-enkephalin coupled to keyhole limpet hemocyanin using glutaraldehyde, following a procedure described previously (26). Immunoglobulins were purified from this antiserum using ammonium sulfate fractionation prior to use for immunoprecipitation.
Immunoprecipitated proteins were extracted from the protein A-Sepharose beads using 130 l of 32% (v/v) acetonitrile containing 0.1% trifluoroacetic acid, 1 N acetic acid, and 6 M urea prior to high pressure gel permeation chromatography (HPGPC). Alternatively, beads were resuspended in a solution of 50 mM Tris-HCl, pH 6.8, 2% SDS, and 1% ␤-mercaptoethanol (solution B), boiled for 3 min, and spun, and the supernatant was collected for endoglycosidase H digestion (see below). Pulse-chase experiments were repeated three times on separate preparations of cells with qualitatively similar results. Pulse-chase experiments were performed using two or three different clones of each mutant proenkephalin, with comparable results.
HPGPC-HPGPC was used to size-fractionate labeled immunoprecipitated peptides. A Beckman high pressure liquid chromatography system was used to pump the eluant through the column array, which consisted of a TSK gel precolumn (7.5 ϫ 75 mm, Toso Hass, Montgomeryville, PA), two high pressure gel permeation columns connected in series, a Protein Pak SW 300 column (300 ϫ 7.8 mm, Waters, Milford, MA), and a Bio-Sil TSK-125 column (300 ϫ 7.5 mm, Bio-Rad). Fractions were collected with a flow rate of 0.5 ml/min, and radioactivity in each was determined by liquid scintillation spectroscopy following the addition of Ready Safe scintillation mixture (Beckman, Fullerton, CA). The recovery of radioactivity from the column array was quantitative. Standardization of the HPGPC system was performed as described elsewhere (25) using [ 35 S]proenkephalin, Peptide B, Met-enk-Arg-Phe, Peptide F, and Met-enk.
Total cellular content of various enkephalins was determined by RIA in the native and mutant cell lines using the same amounts of total protein (for sample normalization, to account for differences in cell growth). Protein content was determined using a 20-l sample of total cell extract after scraping the cells into cold solution A (Bio-Rad; Coomassie Brilliant Blue assay). The remainder of the samples were centrifuged for 10 min in a microcentrifuge at 4°C. Supernatants were collected and stored at Ϫ20°C until further analysis. Aliquots (77-100 l) of the clarified cell extracts were injected onto the HPGPC together with 90 g of gel filtration standards (Bio-Rad). Fractions were collected into polypropylene tubes to which 5 g of bovine serum albumin was added as carrier. Duplicate aliquots of each cell line extract were subjected to radioimmunoassay using specific enkephalin antisera (8,25).
Endoglycosidase H Treatment of PE-In order to determine whether or not there was any difference in the time that labeled precursor spent in transit from the endoplasmic reticulum to the later compartments of the trans-Golgi network, PE processing was compared between the native PE-expressing cells and the double-mutant PE-expressing cell line, K4. Native and K4 cells were metabolically labeled as described above. However, cells were chased for 0, 20, 40, 60, and 90 min. After resuspension of the immune complexes in 130 l of solution B, samples were boiled for 3 min to denature proteins and centrifuged to collect the supernatants. Duplicate samples (30 l) of each supernatant were diluted 1:2 with 100 mM sodium acetate, pH 5.5, containing 1 mM phenylmethylsulfonyl fluoride. Samples corresponding to each cell line were divided into two sets, a control and an endoglycosidase H (endo H) treatment group, and the procedure that followed was as described previously (27) except that samples were subjected to SDS-polyacrylamide gel electrophoresis in 20% acrylamide gels (16 ϫ 16 cm). Gels were fixed, fluorimpregnated, dried, and exposed to Hyperfilm (Amersham Corp.) using intensifying screens for 1 week. Gels were then quantitated by phosphoimaging analysis. Experiments were performed two times on separate preparations of cells with similar results.
Radiosequencing-In order to identify the metabolic intermediate eluting at fraction 35 in the K1 cells, a subconfluent 35-mm well of this cell line (approximately 1-2 ϫ 10 6 cells) was used for labeling and immunoprecipitation experiments. One ml of methionine-deficient medium (Amersham) containing 1 mCi of [ 35 S]methionine was added, and cells were incubated at 37°C in an atmosphere containing 5% CO 2 for 6 h. The immunoprecipitation procedure followed was as described above, using a combination of Xandra antiserum (8) and JAS antiserum (Met-enk-Arg-Phe) (28). Immunoprecipitated peptides were separated by HPGPC. A major peak of radioactivity, which had an apparent molecular size of 5 kDa, was pooled and subjected to automated Edman degradation (performed by the San Diego State University Microchemical Core Facility). Each cleaved residue was collected for liquid scintillation counting. Radiosequencing was performed only once.
RIA-Cells from each native and mutant PE-expressing clone were counted and 2.6 ϫ 10 6 cells grown for 2 days in 10-cm Petri dishes. Duplicate 5-l diluted (1:10) samples of the clarified cell extracts were subjected to RIA analyses to determine the overall proenkephalin ex-pression level in each clone. In addition, duplicate aliquots of each fraction obtained from the HPGPC (as described above), were vacuumdried in polypropylene tubes in the presence of bovine serum albumin as carrier protein. Fractions were resuspended in 100 l of RIA buffer (0.1 M sodium phosphate, pH 7.4, containing 0.1% heat-treated bovine serum albumin, 50 mM sodium chloride, 0.1% sodium azide, and 0.1% ␤-mercaptoethanol) and subjected to RIA as described previously (26). Briefly, specific antiserum, sample or standard, and ϳ10,000 cpm of 125 I-labeled peptide (Amersham) were incubated overnight at 4°C. The antisera used for RIA were raised against Met-enk-Arg-Phe (JAS antiserum) as described in Ref. 28, against Met-enk-Arg-Gly-Leu (29), or against Met-enk (RB4 antiserum; Ref. 30). Carrier ␥-globulin and 25% polyethylene glycol were added to precipitate the bound label, which was separated by centrifugation. Radioactivity in pellets was determined using an LKB ␥-counter. RIAs of all cell lines were carried out at least four times on independent preparations of cells with similar results.

RESULTS
In order to examine the effect of site blockade on the processing of PE, four mutant cell lines were constructed. Fig. 1 depicts a diagrammatic representation of PE showing the different sites of mutation and the peptides known to be generated from the precursor molecule.
The  1). Fig. 2 shows the fate of the Met-enk-Arg-Phe portion of native and mutant PEs. For these experiments, immunoprecipitation was performed using antiserum directed against Met-enk-Arg-Phe, the carboxyl-terminal peptide of PE (this antiserum recognizes all peptides terminating in this heptapeptide). A single peak, corresponding to the approximate elution position of PE, was observed in both cell lines after a 20-min pulse. Native proenkephalin was initially processed more quickly than the K1 mutant (Fig. 2, compare panels A and B at 0.5 h of chase), as shown by a stronger decrease in the peak size of PE relative to 0 h as well as the presence of a processing product in native PE, which was not observed in the K1 mutant at this time. The K1 mutant instead exhibited an intermediate peak of unknown identity (designated as I 1 ; fraction 35, approximately 5 kDa) and a small peak that corresponded to the elution position of the Met-enk-Arg-Phe standard (fraction 45).  (Fig. 3). Taken together with the ability of the peptide to be immunoprecipitated with Metenk-Arg-Phe antiserum and the molecular size of the peptide, these data indicate that the peptide sequenced was most likely generated by cleavage on the carboxyl side of Lys 210 -Arg 211 . This extended form of Peptide B is indicated by an asterisk in the PE diagram (Fig. 1). The radiosequencing results thus suggest that an alternative site, but in fact not the nearest paired basic site, was used for initial cleavage when the naturally preferred site was unavailable.

The Double Blockade Mutant PE (K4) Is Also Processed at a Faster Rate during Initial Cleavage Stage and Gives Rise to an
Alternative Intermediate Peptide-Since the processing of PE was apparently not hampered by a mutation at the initial processing site, we decided to blockade the newly chosen site (Lys 210 -Arg 211 ) as well in the K1 mutant (Fig. 1). This new PE mutant, termed K4, was subjected to a similar pulse-chase analysis as described above.
Results from the metabolic labeling of PE in native PE and double mutation PE are shown in Fig. 4. At the 0.5-h chase time, more than two-thirds of total PE in native, but only one-half of total PE in the K4 mutant, was processed (compare panels A and B). Also at 0.5 h, a second peak was observed in both cases. For native PE this peak corresponded to Peptide B, while in the K4 mutant the second peak corresponded to another intermediate (designated as I 2 ). Given the molecular mass of this peptide and its ability to be immunoprecipitated with Met-enk-Arg-Phe antiserum, we speculate that this peak is a product of alternative cleavage at the carboxyl side of residues Arg 217 -Arg 218 . Following 1 h of chase, the molar ratio of Peptide B to Met-enk-Arg-Phe was 2.7:1 in native PE. In contrast, the molar ratio of I 2 :Met-enk-Arg-Phe was 0.3:1 in K4 at the chase time. These results indicate that processing of I 2 to Met-enk-Arg-Phe was more extensive in K4 PE than the analogous processing event in native PE (Fig. 4). At 2 and 4 h of chase, this same trend continued, with higher Met-enk-Arg-Phe production in the K4 mutant PE as compared with native PE. At 4 h of chase, the molar ratio of Peptide B to Met-enk-Arg-Phe in native PE was 0.8:1; however, in the K4 mutant the intermediate peptide had been completely converted to Metenk-Arg-Phe. In summary, these results demonstrate that like the K1 mutant, K4 mutant PE was processed more slowly at initial steps, but more rapidly and more completely at later steps, than was native PE. Thus blockade of processing at two of the preferred initial processing sites also did not halt proteolytic cleavage but instead resulted in a shift of cleavage to yet another site. Similar kinetics of processing were observed in pulse-chase and immunoprecipitation experiments using two additional clones from the K4 mutant expressing either comparable or lower levels of proenkephalin.
Endoglycosidase H Sensitivity in the K4 Mutant PE Indicates Longer Retention in the Endoplasmic Reticulum-The meta- bolic labeling studies above indicated that the initial stages of PE processing were slower in cells containing mutant forms of PE. In order to examine if this could be a result of slower arrival at processing compartments, i.e. longer retention in the endoplasmic reticulum, we subjected double mutant K4 and native PEs to endo H digestion in order to estimate the appearance of endo H-resistant forms. Fig. 5 shows the profile of endo H sensitivity of radiolabeled PE at different time points after synthesis. At 0 h of chase, native PE and K4 mutant PE showed two bands, corresponding to comparable amounts in the two cell lines of glycosylated and unglycosylated PE forms. At this time point, glycosylated PEs in both mutant and native cell lines were completely endo H-sensitive. After 20 min of chase, the glycosylated native PE began to exhibit some endo H resistance, while the glycosylated mutant PE was still almost completely endo H-sensitive. In addition, both PE bands in native PE were less intense when compared with the double mutant, indicating a loss of native PE by further processing. This provided confirmation of the pulse-chase analyses in which native PE was processed at a higher rate at initial stages (Fig. 4A). Following 40 min of chase, endo H-treated native PE showed an approximately equal ratio of glycosylated to unglycosylated PE (seen best in a longer exposure of the autoradiograph), indicating that the glycosylated form had acquired complete endo H resistance. At this time, the K4 mutant PE still showed some endo H sensitivity (Fig. 5). In addition, a lower molecular weight band had appeared in native PE, corresponding to Peptide B, and in the double mutant, representing I 2 . At 60 min of chase, about half of the glycosylated form of the mutant PE still remained endo H-sensitive; intact native PEs were almost completely processed. These results provide evidence that the mutant PE takes longer to arrive at the medial Golgi compartments (where resistance to endo H is achieved) than native PE, most probably due to longer residence in the endoplasmic reticulum.
Processing of PE in Initial K1 Blockade Mutants Results in More Efficient Production of Active Opioid Peptides-To confirm the efficient processing of mutant PEs into active opioids, additional pulse-chase labeling and immunoprecipitation experiments were performed using native PE and the PE and mutant K1 cell lines employing antiserum against another part of the proenkephalin precursor, Met-enkephalin (Met-enk). The antiserum recognizes peptides terminating in this pentapeptide but recognizes the PE precursor very poorly; therefore, only 2-and 4-h chase point periods are presented. present in all cell lines (Figs. 7 and 8). However, at this time, only about one-third of the total sample radioactivity was present in intact native PE, while about one-half or more was present in intact mutant PEs (Figs. 7 and 8, compare panels A and B), indicating slower initial processing of mutant PEs. At 1 h of chase, some mutant PE still remained unprocessed, unlike native PE. At 2 h of chase, the amount of radioactivity in Peptide B compared with Met-enk-Arg-Phe was higher in native PE as compared with the mutant K2 and K3 PEs, indicating faster processing into Met-enk-Arg-Phe in the latter. These results were confirmed at the 4-h chase time. Thus, these PE mutants exhibited the same profile of processing as the previous two tested, i.e. depressed initial rates of processing and enhanced later rates. Comparable results were obtained during two different sets of experiments using both K mutants and another K2 clone expressing a similar level of PE.
RIA of Stored Peptides in Mutant Cell Lines Supports the Finding of More Extensive Processing of PE into Active Opioids-To determine the profile of stored immunoreactive enkephalins in mutant and native cell lines, RIAs were performed on gel filtration fractions of cellular extracts. The Met-enk-Arg-Phe RIA revealed two peaks of immunoreactivity, Peptide B and Met-enk-Arg-Phe, in extracts from cell lines expressing native PE; however, only one peak, Met-enk-Arg-Phe, was present in all four mutant cell lines (Fig. 9A). Fig. 9B shows the results from the RIAs using antiserum against Met-enk-Arg-Gly-Leu as described previously (8). Native PE was processed into one significant Met-enk-Arg-Gly-Leu-ir peak. However, the K1, K3, and K4 blockade mutants showed two peaks, corresponding to the 5.3-kDa PE-derived peptide and Met-enk-Arg-Gly-Leu. K3 mutants are not shown, since blockade at this site resulted in no Met-enk-Arg-Gly-Leu-ir peptide production. These RIA results support the pulse-chase analyses of mutant cell lines in demonstrating more efficient processing of the PE blockade mutants relative to native PE. The overall PE expression levels were determined in each clone and expressed as total Met-enk-Arg-Phe immunoreactivity per 10-cm plate, as follows: native PE (73 pmol/plate), K1 (440 pmol/plate), K2 (623 pmol/plate), K3 (258 pmol/plate), K4 (250 pmol/plate). Because of the more extensive processing in mutant PE clones as compared with native PE, and because free Met-enk-Arg-Phe reacts better in the Met-enk-Arg-Phe assay than does Peptide B, these values should be treated with caution. Western blot analyses using antibodies against PC1 indicated that the expression level of this enzyme was comparable in all four mutants and in the native PE-expressing clones. 2

DISCUSSION
The posttranslational processing of proenkephalin has been examined in various types of cells, such as bovine adrenal medullary chromaffin cells (29 -31) and PE-transfected AtT-20 (8,25) and rat insulinoma Rin5f cell lines (25). Results from these studies indicate that processing of PE begins at the 2 K. Johanning and I. Lindberg, unpublished data. carboxyl-terminal side of the molecule, producing Peptide B. Subsequently, other cleavages such as the cleavage at Lys 196 -Arg 197 results in intermediate sized peptides that will become processed at later stages, eventually producing the bioactive penta-to octa peptides. Very little is known about the biochemical basis for the observation of temporally ordered cleavage of prohormone precursors at specific sites (reviewed in Ref. 1). Rhodes et al. (19) have provided evidence for sequential unfolding of the proinsulin molecule such that cleavage at one site provides a more favorable substrate for the action at another site.
In the work reported here, we have employed site blockade of the initial cleavage sites of PE in order to examine whether or not processing of this precursor to the mature form must occur in a predefined order. Contrary to our expectations, processing profiles of these mutant PEs indicated that cleavage need not occur in a specific order. Our data demonstrated that all mutant forms of PE were efficiently expressed and processed and that alternative initial cleavage sites were utilized when the preferred sites were unavailable. These results are reminiscent of those obtained with Aplysia egg-laying hormone in AtT-20 cells (16). In this study, removal of an initial tetrabasic processing site by deletion resulted in alternative cleavage at a tribasic site on eventual efficient production of egg-laying hormone (16). Wilson and co-workers (32) have recently presented data on the processing of human PE to larger intermediates in bovine chromaffin cells using Western blotting and monoclonal antibodies. In this study, human PE was mutated at 12 dibasic processing sites by converting Lys-Arg sequences to Lys-Gln, and Arg-Arg sites to Arg-Gln. At four of these cleavage sites, blockade of processing was achieved, and while none of the usual processing products were produced, proenkephalin did not remain intact but was processed into unnatural intermediates. Thus, for at least two different precursors, a major alteration of cleavage site usage appears to result in efficient enzyme processing of alternative sites and processing via unnatural intermediates.
While processing of proenkephalin to end products proceeded without hindrance in mutant PEs, our results clearly demonstrate differences in the rate and extent of processing of mutant PEs as compared with native PE. Initial processing was slowed in all mutant PEs examined, as judged from the disappearance of intact PE, the production of initial intermediates, and the acquisition of endo H resistance. Since the introduction of mutations may be expected to alter the normal conformation of the precursor, we speculate that these altered conformations are recognized as unnatural by quality control mechanisms in the endoplasmic reticulum, potentially resulting in a delay in export to the trans-Golgi network. However, once mutant PEs reached the Golgi and underwent initial cleavage at alternative sites, processing of aberrant intermediates appeared to occur at an enhanced rate compared with native PE.
Perhaps the most interesting aspect of the processing of these mutant PEs is the extent of processing of the aberrant intermediates to the penta-and heptapeptide bioactive enkephalins. Both metabolic labeling and RIAs indicated a much greater generation and storage of mature enkephalins in all four mutant PE cell lines as compared with the native PEexpressing cell line. Overexpression of PC1 relative to proopiomelanocortin has been shown to result in enhanced processing of this precursor (33). However, since mutant PE lines expressed PE at levels higher than or comparable with that of the native PE line (see Figs. 2, 4, 7, and 8) and since PC1 expression was comparable in all cell lines, substrate:enzyme ratio considerations probably do not play a role in the enhanced processing observed in mutant PE cell lines. One possible explanation for the faster processing rate in mutant PE lines may be related to an increase in enzyme affinity for the aberrant intermediates, which might be expected to exhibit less structure, or to bind to other granule components with a different affinity, than native intermediates. Alternatively, earlier exposure of the intermediates to 66-kDa PC1, which is the predominant form of PC1 in regulated secretory vesicles (34 -36) may somehow trigger more complete proteolysis, since this form of PC1 is known to be more active than the 87-kDa form (37). PC2 could potentially also play a role in the more complete proteolysis of mutant intermediates, although the presence of this enzyme in AtT-20 cells is controversial (18,38). Alternatively, other proteases might be involved in the processing of mutant proenkephalins (39).
In summary, we conclude that 1) processing of proenkephalin does not require a predefined order, since alternative sites, which lead to the production of unnatural intermediates, are selected when preferred sites are unavailable; 2) initial processing of site-blockaded proenkephalin is slowed, potentially by longer residence in the endoplasmic reticulum and by the necessity to use alternative sites; 3) unnatural intermediates are processed more quickly into the final end product enkephalins; and 4) mutant PEs are ultimately more completely cleaved into enkephalins than native PE.
Several studies have demonstrated that adrenal medullary cell transplantation into rat spinal cord can be used as a method to produce and release opioid peptides for the relief of chronic intractable pain (40 -42). More recently, researchers have attempted to use cell therapy with genetically engineered cell lines to produce bioactive opioids to treat pain (43,44). Implantation of genetically modified AtT-20 cells expressing human PE in mice has been demonstrated to produce analgesia and offers a potential therapy to relieve chronic pain (45). Our data suggest that it might be possible to improve upon nature by engineering a PE molecule that can deliver more opioid peptide units per molecule of precursor. The mutant PEs described in this report may represent the first step in this direction.