Mutational analysis of PC1 (SPC3) in PC12 cells. 66-kDa PC1 is fully functional.

The proteinase mPC1, a neuroendocrine member of the mammalian family of subtilisin-like enzymes, has previously been shown to be converted to a carboxyl-terminally truncated 66-kDa form during transport through the secretory pathway. The cleavage site and the function of this carboxyl-terminal truncation event are unknown. We have performed site-directed mutagenesis of two paried basic sites in the mPC1 carboxyl-terminal tail and expressed these constructs in PC12 cells, a rat pheochromocytoma known to lack endogenous PC1. We found that the most likely site for the truncation event was at Arg590-Arg591 since mutation of this site to Lys-His prevented processing of 87-kDa PC1. A PC1 mutant carboxyl-terminally truncated at this site and expressed in PC12 cells was efficiently routed to the secretory pathway and stored in secretory granules, indicating that the carboxyl-terminal extension is not required for sorting of this enzyme. The function of the various PC1 constructs was assessed by analyzing proneurotensin cleavage to various forms. The carboxyl-terminally truncated PC1 mutant was found to perform most of the cleavages of this precursor as well as wild-type PC1; however, the blockade mutant processed proneurotensin much less efficiently. Differences between the site preferences of the various enzymes were noted. Our results support the notion that carboxyl-terminal processing of PC1 serves to regulate PC1 activity.

BSTXI-5 and 591ST-3 (or 628 BL-3) were used as PCR primers in the first reaction, while  and XBAI-3 were used in the second reaction. After 10 cycles of amplification, the products were separated from template and primers on 1.5% agarose gel; the major DNA product of each PCR reaction was recovered and extracted from an agarose gel using a Geneclean II kit. In the third PCR reaction, the major products of the first two reactions were mixed and annealed to generate a template, and BSTXI-5 and XBAI-3 were used as PCR primers. The amplification was performed for 15 cycles, and the product (1005 bp) was separated from template and primers on a 1% agarose gel and DNA recovered using a Geneclean II kit. The product of the third PCR was digested using BstXI and XbaI as described above; the largest product was separated from small DNA segments using the Geneclean II kit. In order to ligate the PCR products into the digested vector (6800 bp), 60 ng of vector DNA was mixed with half of the PCR product and 1 l (400 units) T 4 DNA ligase (New England Biolabs) in a 20-l final volume for 45 min at room temperature. The ligation mixture was then transformed into Ultracomp INV␣FЈ cells (Invitrogen). To confirm the mutations and accuracy of PCR amplification, the 1005-bp PCR product region was sequenced using a DNA Sequenase version 2.0 kit with three different sequencing primers. Plasmids of correctly mutated mPC1SB/CMV (Arg 627 -Arg 628 3 Lys-Ala), mPC1DB/CMV (Arg 590 -Arg 591 3 Lys-His and Arg 627 -Arg 628 3 Lys-Ala), and mPC1ST/CMV (truncation at Gly 592 ) were then amplified in E. coli (XL1 Blue; Stratagene) and purified using a Qiagen plasmid kit.
Transfection of PC12 Cells with Mutated mPC1 Forms-PC12 cells, obtained from Drs. L. Elferink and R. Scheller (Stanford University), were grown in DMEM-high glucose (4.5 g/L) medium containing 10% fetal bovine serum and 5% heat-inactivated horse serum on collagencoated plates. PC12 cells expressing wild-type mPC1 have been described previously . Mutated PC1 forms were transfected into PC12 cells using Lipofectin (Life Technologies, Inc.). The day prior to transfection, 1-2 million PC12 cells were subcultured on 10-cm plates. The next day, after the cells were well attached, the plates were rinsed twice with prewarmed serum-free DMEM-high glucose medium. A 30-l aliquot of Lipofectin was placed into 3 ml of serum-free medium in a sterile 15-ml tube with 30 g of DNA from mPC1SB/CMV, mPC1DB/CMV, or mPC1ST/CMV, and kept for 10 min at room temperature. Experimental plates were incubated with 3 ml of medium containing Lipofectin/DNA, while the control plate was incubated with Lipofectin only. The medium was replaced with regular medium containing 300 g/ml G418 after 5 h of incubation at 37°C. Transfected cells produced colonies after approximately 3 weeks. The colonies were then subcloned using the agarose method described by Lindberg and Zhou (1995), and the cell lines producing the highest quantities of PC1 proteins were selected by immunoblotting 12-24 cell extracts with PC1 amino-terminal antiserum (Vindrola and Lindberg, 1992).
Collection of PC1 from PC12 Cells-PC12 cells stably transfected with mPC1s were cultured in medium containing 300 g/ml G418 in collagen-coated 35-mm plates. After the cells reached about 70 -80% confluence, the synthesis of proneurotensin was induced using medium containing 1 M forskolin, 1 M dexamethasone, 100 ng/ml nerve growth factor, and 10 mM LiCl . The plates were maintained in inducing medium in an atmosphere of 5% CO 2 at 37°C for 48 h. In stimulation experiments, two 35-mm wells of induced PC12 cells were each washed twice with 5 ml of prewarmed PBS. To one plate was added 1 ml of Earle's balanced salt solution (prepared as described by Lindberg et al. (1994)), while to the other plate was added 1 ml of Earle's balanced salt solution containing 50 mM KCl (and lacking a corresponding concentration of NaCl). The plates were incubated for 40 min at 37°C and the conditioned media were collected on ice, centrifuged at low speed to remove any floating cells, and then concentrated to 80 l using a Microcon 10 device (Amicon) prior to the addition of 10 l of 10 ϫ Laemmli sample buffer (Laemmli, 1970). For cell extraction, the cells in each 35-mm well were washed twice in 5 ml of prewarmed PBS and homogenized in 300 l of ice-cold Laemmli sample buffer. All samples were boiled for 5 min. Aliquots (70 l) of each sample were subjected to Western blotting for PC1 as described previously.
Pulse-chase Experiments of PC12 Cells-PC12 cells expressing wildtype and mutated forms of PC1 were grown in collagen-coated 35-mm wells in six-well plates to 80% confluence. The cells were rinsed and incubated in 1 ml of Met/Cys-free DMEM at for 30 min at 37°C, then labeled for 20 min at 37°C in 1 ml of Met/Cys-free DMEM containing 0.5 mCi of [ 35 S]Pro-Mix (Amersham Corp.). They were then chased in 1 ml of regular DMEM with 2% fetal bovine serum for various periods of time as indicated. The cells and the conditioned media were transferred into ice-cold Eppendorf tubes and centrifuged at 2,000 rpm for 5 min at 4°C. The supernatants were saved, and the pellets were resuspended in 100 l of boiling buffer containing 50 mM sodium phosphate, pH 7.4, 1% SDS, and 50 mM ␤-mercaptoethanol (Milgram and Mains, 1994). Both conditioned media and cell extracts were boiled for 5 min. One ml of AG buffer (Vindrola and Lindberg, 1992) was added to cell extracts (in 100 l of boiling buffer), and the solution was centrifuged to remove any pellet. Aliquots (1 ml) of each sample were subjected to immunoprecipitation, SDS-PAGE analysis, and autoradiography as described previously (Vindrola and Lindberg, 1992). Pulse-chase experiments were repeated three times.
Analysis of Neurotensin Produced in PC12 Cells-The ability of transfected mPC1s to process endogenous prohormone was assessed by measuring the various forms of mature and unprocessed neurotensinderived peptides in PC12 cell extracts. PC12 cells were cultured and induced in 10-cm plates as described above. After 48 h of induction, the cells were washed three times with 10 ml of Dulbecco's PBS, placed on ice, scraped into 1 ml of ice-cold 0.1 N HCl, and stored frozen at Ϫ70°C. Upon thawing, the samples were vortexed to produce a homogeneous suspension. A 100-l aliquot of cell extract was removed for protein determination and Western blotting of mPC1 (Fig. 2). The remainder was centrifuged on a microcentrifuge for 5 min, and the supernatant was removed and lyophilized. For the experiment shown in Fig. 5, the cell pellet obtained after removal of the HCl was solubilized in various amounts (between 0.3 and 0.6 ml) of Laemmli sample buffer containing 5 M urea to obtain a constant protein concentration of 3 mg/ml. Fifty l of each sample were subjected to Western blotting using the aminoterminal PC1 antiserum and the blot subjected to video densitometry for quantitation of the amount of immunoreactive PC1 in each dish. Proteins remaining in the gel following blotting were stained with Coomassie Blue; this procedure verified that the lanes had indeed been equally loaded with protein.
The rat proneurotensin precursor (Kislaukis et al., 1988) is depicted in Fig. 5a. It consists of a 169-residue polypeptide, which begins with a NH 2 -terminal signal peptide (1-22). Neurotensin is located near the COOH terminus of the precursor and is flanked by two Lys-Arg sequences at positions 148 -149 and 163-164. Neurotensin is preceded by a neuromedin N sequence located between Lys 140 -Arg 141 and Lys 148 -Arg 149 . A fourth Lys-Arg sequence occurs near the middle of the precursor at position 85-86. This doublet and the one at the NH 2 terminus of neuromedin N delimit a 53-residue peptide, which starts with a neuromedin N-like sequence (Lys-Leu-Pro-Leu-Val-Leu, designated K6L) and ends with an acidic sequence (Glu-Lys-Glu-Glu-Val-Ile, designated E6I).
The specificities of the neurotensin, neuromedin N, E6I, and K6L antisera used here have been described previously in detail Rovere et al., 1993). Briefly, the neurotensin and E6I antisera react with the free COOH termini, while the neuromedin N and K6L antisera recognize the free NH 2 termini, of their respective antigens. These antisera cross-react poorly (Ͻ1%) with antigenic sequences that are internal to proneurotensin or proneurotensin fragments. Thus, the neurotensin antiserum will detect all precursor products with a COOHterminal neurotensin sequence (including authentic neurotensin). Similarly, the E6I antiserum will measure all the precursor forms ending with the E6I sequence, while the neuromedin N and K6L antiserum will assay the precursor products bearing NH 2 -terminal neuromedin N and K6L sequences, respectively. The radioimmunoassay and reverse-phase HPLC procedures employed here to quantitate the various proneurotensin-derived peptides have been fully described elsewhere Rovere et al., 1993).
All cell extracts were directly assayed for their content in immunoreactive neurotensin (iNT), E6I (iE6I), and K6L (iK6L). Because of the above-described antisera specificity, the iNT, iE6I, and iK6L assays measure the amounts of precursor products that are processed at the Lys 163 -Arg 164 , Lys 140 -Arg 141 , and Lys 84 -Arg 85 sequences, respectively. Portions of the cell extracts were submitted to Arg-directed tryptic digestion Rovere et al., 1993) and then assayed for immunoreactive K6L. The value of CTiK6L thus obtained provides an index of the total amount of proneurotensin (either processed or unprocessed) that was synthesized and stored in the cells during the induction period. The remainder of the trypsin-treated samples was applied to reverse-phase HPLC, and the fractions were assayed for their immunoreactive neuromedin N (iNN) content. Previous studies have shown that trypsin-generated iNN can be resolved by HPLC into two peaks, one comigrating with synthetic neuromedin N and the other with neuromedin N bearing a COOH-terminal Lys-Arg extension Rovere et al., 1993). The latter peptide is produced by Arg-directed tryptic digestion of precursor forms in which the neuromedin N sequence is internal, whereas the former is generated by cleavage of peptides that end with a COOH-terminal neuromedin N sequence. Thus, the post-HPLC assay of trypsin-generated neuromedin N provides a measurement of all the precursor products that are processed at the Lys 148 -Arg 149 sequence (including authentic neuromedin N). The results were normalized for the amount of protein in each extract. The percentages of cleavage at the Lys 163 -Arg 164 , Lys 148 -Arg 149 , Lys 140 -Arg 141 , and Lys 84 -Arg 85 sequences were calculated by dividing, respectively, iNT, trypsin-generated neuromedin N, iE6I, and iK6L by CTiK6L and by multiplying these ratio values by 100. Duplicate independent samples were analyzed on two separate occasions.

RESULTS
Site-directed Mutagenesis of PC1-It has been documented that four paired basic residues exist within the mPC1 carboxylterminal region, however, only Arg 590 -Arg 591 and Arg 627 -Arg 628 are conserved across human, mouse, rat, and anglerfish PC1 (Seidah et al., 1991Smeekens et al., 1991;Bloomquist et al., 1991;Roth et al., 1993). If 87-kDa PC1 is indeed cleaved at these two sites, the larger products should approximate 62 and 70 kDa, respectively; these molecular masses are close to the sizes of the PC1 forms obtained by spontaneous cleavage of 87-kDa PC1 (66 and 74 kDa; Zhou and Lindberg (1994)). Therefore, we theorized that these two sites could represent the cleavage sites for PC1 carboxyl-terminal processing. To test this hypothesis, site-directed mutagenesis reactions were carried out to convert Arg 627 -Arg 628 to Lys-Ala or Arg 590 -Arg 591 to Lys-His and Arg 627 -Arg 628 to Lys-Ala, generating mPC1SB (single blockade at Arg 627 -Arg 628 ) and mPC1DB (double blockade at both sites). A site-directed mutation to replace Gly 592 with a stop codon was also performed to generate mPC1ST, the carboxyl-terminally truncated form of PC1. These mutations are diagrammed in Fig. 1.
Biosynthesis of the Mutated PC1 Forms-In order to determine the effect of the mutations at the mPC1 carboxyl-terminal region, the carboxyl-terminal processing of wild-type mPC1 and the mutated forms (mPC1SB, mPC1DB, and mPC1ST) were studied in PC12 cells stably transfected with the various constructs. Western blotting using PC1 amino-terminal antiserum revealed that wild-type mPC1 and mPC1SB were largely converted to a 66-kDa form, while the majority of mPC1DB still remained as 87-kDa PC1 (Fig. 2), suggesting that mutation at Arg 590 -Arg 591 (but not mutation at Arg 627 -Arg 628 ) could effectively block PC1 carboxyl-terminal conversion. Western blotting also showed that mPC1ST (truncated at residues 592) exhibited a molecular mass identical to that of the 66-kDa PC1, and this form did not undergo any further carboxyl-terminal cleavage. Following stimulation of PC12 cells with 50 mM KCl, wild-type mPC1 as well as mPC1DB and mPC1ST could be released from regulated secretory pathway (Fig. 3). These data indicate that neither deletion of the carboxyl-terminal region (residues 592 to 726) nor mutations at Arg 590 -Arg 591 and Arg 627 -Arg 628 could block PC1 targeting into the regulated secretory pathway. It was noted that a small portion of mPC1DB was still cleaved, but that the cleavage product possessed a molecular mass slightly larger than that of cleaved wild-type mPC1, indicative of the involvement of an alternative cleavage site.
In pulse-chase labeling experiments, we found that aminoterminal conversions of pro-mPC1DB and pro-mPC1ST were completed within the first 20 min of synthesis, similar to wildtype pro-mPC1 (data not shown). This result indicates that substitution of Arg 590 -Arg 591 and Arg 627 -Arg 628 or deletion of the PC1 carboxyl-terminal region (residues 592-726) apparently had little effect on proPC1 conversion. During the later stages of biosynthesis, intracellular 87-kDa mPC1DB remained intact after a 4-h chase period, while wild-type PC1 was converted to the 66-kDa form (Fig. 4). Constitutive secretion of the 87-kDa form of both mPC1DB and wild-type PC1 occurred within 1 h after the pulse period (Fig. 4). Similar studies of the mPC1ST mutant demonstrated efficient synthesis of a 66-kDa form of PC1 and constitutive secretion into the medium over the same time period (results not shown).
Physiological Function of the Mutated PC1 Forms-PC12 cells are known to greatly increase their synthesis of proneurotensin under inducing conditions . We have shown previously that transfection of mPC1 into PC12 cells can promote proneurotensin processing . To determine the physiological function of the mutated mPC1DB and mPC1ST, we compared the maturation of proneurotensin synthesized in PC12 cells transfected with the mutated mPC1DB and mPC1ST with that in PC12 cells expressing wild-type mPC1 and untransfected PC12 cells. It should be noted that the various cell lines expressed varying amounts of mPC1, with the double blockade mutant (mPC1DB) exhibiting the highest expression, and mPC1ST exhibiting the lowest. The relative amounts of PC1 in the cells used for this experiment were estimated by Western blotting a constant amount of protein from each dish used for neurotensin analysis and performing video densitometry of the blot. The ratios of the amount of mutant mPC1 (all forms) to wild-type mPC1 were approximately 0.3 (mPC1ST) and 9 (mPC1DB). As expected, untransfected PC12 cells exhibited no PC1 immunoreactivity.
The amount of proneurotensin (processed and unprocessed) stored in the various cell lines during the induction period ranged between approximately 10 and 30 pmol/mg protein (CTiK6L values are given in the legend to Fig. 5). Similarly to wild-type mPC1, mPC1DB and mPC1ST both possessed the ability to process proneurotensin (Fig. 5).
All forms of PC1 markedly increased proneurotensin cleavage at the Lys 140 -Arg 141 and Lys 148 -Arg 149 dibasic sites as compared to the control (Fig. 5b). They also cleaved, though less efficiently, the Lys 84 -Arg 85 site. This cleavage event was not observed in control PC12 cells. Only mPC1 and mPC1ST were able to increase processing at the Lys 163 -Arg 164 site above the level seen in the control, whereas mPC1DB appeared inactive in that respect. In general, and especially given the fact that it had the highest level of expression, mPC1DB was much less efficient in processing proneurotensin than mPC1 and mPC1ST. Interestingly, mPC1ST, the enzyme expressed at the lowest level, was the most active in processing the Lys 163 -Arg 164 site, in contrast to mPC1DB which apparently did not cleave this dibasic despite its high level of expression. Thus, there appear to be certain differences in proneurotensin processing efficiency and site usage between the 66-and 87-kDa PC1 forms. DISCUSSION PC1 is known to be cleaved within its carboxyl-terminal region at a late stage of its biosynthesis (Vindrola and Lindberg, 1992). Previous work has suggested that an autocatalytic mechanism may be involved in this process (Zhou and Lindberg, 1994); however, the site of this carboxyl-terminal cleavage event has not yet been identified. In this work, we have assumed that this cleavage occurs at a paired basic site, since these are known to represent consensus sequences for PC1 cleavage. Four paired basic residues are located in the mPC1 carboxyl-terminal region; cleavage at these sites can generate products with estimated molecular masses between 62 and 73 kDa. However, among these sites, only Arg 590 -Arg 591 and Arg 627 -Arg 628 are conserved among the PC1 sequences of human, rat, mouse, and anglerfish (Seidah et al., 1991Smeekens et al., 1991;Bloomquist et al., 1991;Roth et al., 1993); thus, these two sites were thought to represent likely candidate sites for PC1 carboxyl-terminal cleavage. By performing site-directed mutagenesis at these two sites, we found that mutation of Arg 627 -Arg 628 alone had little effect on the generation of the 66-kDa form, while mutations of both Arg 590 -Arg 591 and Arg 627 -Arg 628 were able to block the conversion of 87-kDa PC1 to the 66-kDa form. Furthermore, mPC1ST (truncation at Gly 592 ) exhibited a molecular mass on SDS-PAGE identical to that of endogenous 66-kDa PC1 converted from the 87-kDa wild-type mPC1. Subsequent to the generation of our mutants, the sequence of Aplysia PC1 was published (Chun et al., 1994); PC1a from this species contains the first of these dibasics, but not the second. Taken together, these data strongly suggest that Arg 590 -Arg 591 is the major cleavage site for the generation of 66-kDa PC1 in vivo. Since wild-type PC12 cells possess a regulated secretory pathway, but lack the ability to process prohormones at paired basic residues, PC1 cleavage at Arg 590 -Arg 591 within PC12 cells is likely to be attributable to an autocatalytic mechanism. This interpretation is supported by our in vitro work (Zhou and Lindberg, 1994) and in vivo results obtained in AtT-20 cells, which indicate that overexpression of PC1 results in increased COOH-terminal proteolytic processing (Zhou and Mains, 1994a).
The presence of an intermediate form of PC1 of approximately 74 kDa has been observed in in vitro studies (Zhou and Lindberg, 1994); however, little 74-kDa mPC1 is found in PC12 cells (this study) or in AtT-20 cells (Vindrola and Lindberg, 1992;Milgram and Mains, 1994). Taken together with the finding of lesser effects of the mutation of Arg 627 -Arg 628 in PC12 cells, these results suggest that 74-kDa PC1 may represent a minor product during the carboxyl-terminal processing of PC1 in vivo. A possible explanation for these differences is that the cleavage site generating the 74-kDa form is blocked in vivo, possibly due to an association of PC1 carboxyl-terminal region with membrane or with other proteins. The association of PC1 with membranes and association with other granule proteins have both been reported (Vindrola and Lindberg, 1992;Palmer and Christie, 1992).
Although mutation at Arg 590 -Arg 591 and Arg 627 -Arg 628 substantially blocked PC1 carboxyl-terminal conversion, a small portion of 87-kDa mPC1DB was still cleaved. The product was slightly larger than wild-type 66-kDa PC1 on SDS-PAGE, sug- FIG. 4. Biosynthesis of wild-type mPC1and mPC1DB in PC12 cells. PC12 cells transfected with either wild-type mPC1 or mPC1DB were labeled with [ 35 S]Met for 20 min, then chased in methioninecontaining medium with 2% dialyzed fetal bovine serum for various periods of time as indicated. The conditioned media and cell extracts were immunoprecipitated using PC1 amino-terminal antiserum. The immunoprecipitates were separated using SDS-PAGE and subjected to fluorography.

FIG. 5. Analysis of proneurotensin processing in PC12 cells expressing various forms of PC1.
Products of proneurotensin processing were detected by radioimmunoassay using specific radioimmunoassays directed against the epitopes indicated. A schematic diagram of proneurotensin is shown in panel a. Within each cell line, the percentage of each particular processing product is shown in panel b as a percentage of total proneurotensin. Numbers in parentheses refer to dibasic sites detected by each assay. Quantitation of total proneurotensin in each cell extract was determined by assaying iK6L following tryptic digestion (CTiK6L); the amounts of CTiK6L in PC12 Ctr (control), mPC1DB/PC12, mPC1/PC12, and mPC1ST/PC12 cells were 37.4 Ϯ 4.9, 10.1 Ϯ 0.9, 11.6 Ϯ 2.5, and 34.5 Ϯ 1.9 pmol/mg protein, respectively (mean Ϯ S.E., n ϭ 4).
gesting that an alternative cleavage site is involved. Through limited digestion using chymotrypsin, trypsin, and subtilisin, which possess different substrate specificities, we found that all three proteinases were able to convert 87-kDa recombinant PC1 to 66-and 74-kDa-like products in its carboxyl-terminal region (Zhou and Lindberg, 1994). These results suggest that the cleavage site in the PC1 carboxyl terminus is located in an exposed region which can readily be attacked. Therefore, the alternative cleavage site usage in PC12 cells may be due to the action of other proteinases located in the regulated secretory pathway. Alternatively, PC1 itself may also act at alternative cleavage sites. This idea is supported by our in vitro studies that demonstrate spontaneous carboxyl-terminal cleavage of 87-kDa mPC1DB purified from Chinese hamster ovary cells amplified for the production of this protein (results not shown). A possible alternative cleavage site for transfected mPC1 may be Lys 602 -Arg 603 , although this site is present only in mouse PC1. Cleavage at the Lys 602 -Arg 603 site can generate a product 12 amino acids longer than the product cleaved at Arg 590 -Arg 591 (66-kDa PC1); this molecular mass is also consistent with our observed molecular masses on SDS-PAGE. Construction of a PC1 vector encoding a further mutation at Lys 602 -Arg 603 will be required to investigate this possibility.
Carboxyl-terminal conversion of PC1 occurs mainly in regulated secretory granules, as evidenced by previous studies in AtT-20 and PC12 cells (Vindrola and Lindberg, 1992;Benjannet et al., 1992;Lindberg et al., 1994;Lindberg, 1994;Milgram and Mains, 1994;Zhou and Mains, 1994a). However, the functional significance of this conversion event is not clear. The activation of proPC1 occurs within the endoplasmic reticulum (Lindberg, 1994;Milgram and Mains, 1994;Goodman and Gorman, 1994), while peptide hormone precursors are thought to be cleaved within the later stages of the secretory pathway (reviewed by Loh et al. (1992)). It may thus be necessary for the cell to regulate PC1 function during intracellular transport. The decreasing pH gradient from the endoplasmic reticulum to the secretory granules may represent one important aspect of this regulation. The pH within the trans-Golgi network and regulated secretory granules corresponds well with the optimal pH of PC1 activity, between 5.0 and 6.5 (Zhou and Lindberg, 1993;Jean et al., 1993;Rufaut et al., 1993). On the other hand, since the timing and location of PC1 carboxylterminal processing coincide with the timing and location of prohormone processing, truncation of PC1 may also play a role in the regulation of enzyme activity. In vitro studies have shown that that carboxyl-terminal cleavage of PC1 dramatically increases PC1 activity against peptide and prohormone substrates (Zhou and Lindberg, 1994), suggesting that carboxyl-terminal cleavage of PC1 could potentially possess physiological significance.
In order to determine the function of the PC1 carboxylterminal region in vivo, we compared the processing of proneurotensin in PC12 cells stably transfected with wild-type mPC1, mutated mPC1DB, or mPC1ST. In contrast to AtT-20 cells, PC12 cells do not express prohormone convertases; thus, neurotensin is stored mainly in precursor form (Carraway et al., 1993;Rovere et al., 1993). When the varying expression levels are taken into account, the 66-kDa form of PC1 was found to be the most active against proneurotensin, especially relative to the 87-kDa blockade mutant, which was expressed at much higher levels. The comparatively low activity of mPC1DB against proneurotensin confirms our previous in vitro results, which indicate that the 87-kDa PC1 represents only a partially active PC1 form (Zhou and Lindberg, 1994); removal of the carboxyl-terminal region appears to be required to fully activate PC1. The finding that the various forms of PC1 are differentially active against proneurotensin supports our in vitro results showing that the 87-and 74/66-kDa recombinant PC1s exhibit differing specific activities (Zhou and Lindberg, 1994).
A recent study has also demonstrated that expression of carboxyl-terminally truncated PC1 (Stop D616 ) in AtT-20 cells increases the rate of conversion of proopiomelanocortin (Zhou and Mains, 1994b). Based on these in vivo and in vitro observations, we speculate that carboxyl-terminal processing of PC1 during transport through the secretory pathway may control the amount of PC1 activity available for the processing of prohormones.
In conclusion, we have demonstrated that PC1 carboxylterminal conversion largely occurs at Arg 590 -Arg 591 site through a possible autocatalytic mechanism; the PC1 carboxylterminal domain (Gly 592 to Asn 726 ) is required neither for activation nor for intracellular transport of PC1 to secretory granules. However, removal of this domain appears to increase total PC1 activity and alters cleavage site preference. In line with our previous in vitro data, these results support the idea that carboxyl-terminal processing is important for the regulation of PC1 function.