Furin Proteolytically Processes the Heparin-binding Region of Extracellular Superoxide Dismutase*

Extracellular superoxide dismutase (EC-SOD) is an antioxidant enzyme that attenuates brain and lung injury from oxidative stress. A polybasic region in the carboxyl terminus distinguishes EC-SOD from other superoxide dismutases and determines EC-SOD's tissue half-life and affinity for heparin. There are two types of EC-SOD that differ based on the presence or absence of this heparin-binding region. It has recently been shown that proteolytic removal of the heparin-binding region is an intracellular event (Enghild, J. J., Thogersen, I. B., Oury, T. D., Valnickova, Z., Hojrup, P., and Crapo, J. D. (1999) J. Biol. Chem. 274, 14818–14822). By using mammalian cell lines, we have now determined that removal of the heparin-binding region occurs after passage through the Golgi network but before being secreted into the extracellular space. Specific protease inhibitors and overexpression of intracellular proteases implicate furin as a processing protease. In vitro experiments using furin and purified EC-SOD suggest that furin proteolytically cleaves EC-SOD in the middle of the polybasic region and then requires an additional carboxypeptidase to remove the remaining lysines and arginines. A mutation in Arg213renders EC-SOD resistant to furin processing. These results indicate that furin-dependent processing of EC-SOD is important for determining the tissue distribution and half-life of EC-SOD.

A distinguishing feature of EC-SOD is its high affinity for heparin and the extracellular matrix. Six basic amino acid residues (Arg-Lys-Lys-Arg-Arg-Arg) within the last 13 amino acid residues (8) are essential for this high affinity (9,10).
Mutations within this region lead to elevations of EC-SOD in plasma, presumably from decreased affinity toward the extracellular matrix (11)(12)(13). Removal of the polybasic region decreases the tissue half-life of EC-SOD by ϳ10-fold but does not affect the enzymatic activity of EC-SOD (14). Therefore, the most likely function of the polybasic region is to determine the bioavailability of EC-SOD within the extracellular matrix.
We have reported recently (15) that EC-SOD can be secreted in two forms that have a different carboxyl terminus. These different forms of EC-SOD are not because of mRNA differences but instead are due to intracellular proteolytic processing of the carboxyl terminus containing the polybasic residues. The cleaved EC-SOD has lower heparin affinity because the heparin-binding region is absent. The proteolytic removal of the carboxyl terminus could serve as a regulatory step by altering the affinity of EC-SOD for the extracellular matrix. Indeed, variability in the ratio of cleaved and intact EC-SOD has been observed among different mouse tissues (15) and during lung development (16), suggesting that there might be both spatial and temporal expression patterns of proteolytic processing of EC-SOD.
To explore fully the significance of the heparin-binding region of EC-SOD, it is necessary to determine how proteolytic processing occurs; however, neither the mechanism nor the responsible processing protease has been identified. In this current study we characterize the EC-SOD processing pathway, and we identify furin as a putative processing protease. The location of processing is shown to occur in a stretch of 6 consecutive basic amino acids within the last 13 amino acids and depends on the presence of an arginine residue at position 213.
Metabolic Labeling and Pulse Analysis-Cells were grown in tissue culture plates until 80% confluent according to the ATCC or supplier guidelines. The cells were then washed twice with Earles' balanced salt solution and incubated for 30 min in methionine-deficient RPMI media. The cells were then incubated for 30 min in media containing [ 35 S]methionine (ICN). Cells were then rinsed twice with serum-free media and incubated with unlabeled serum-free media for 3 h.
Immunoprecipitations-A polyclonal antiserum directed against EC-SOD and monoclonal antibody for nerve growth factor (provided by Dr. Rae Nishi, Oregon Health Sciences University) were used to recover [ 35 S]methionine-labeled proteins as described previously (19 -21). In brief, media from metabolically labeled cells were removed, and protease inhibitors were added (final concentration of 0.2 mM 3,4-dichloroisocoumarin, 0.04 mM E64, and 4 mM 1,10-phenanthroline). Cells were then washed twice with phosphate-buffered saline and scraped free in 1 ml of lysis buffer (25 mM Tris-HCl, 500 mM NaCl, 5 mM EDTA, pH 7.5, with 0.5% Triton X-100, and protease inhibitors). Cell debris was pelleted by centrifugation. Samples were cleared by adding 10 l of preimmune antiserum and 40 l of protein G-Sepharose FF (Amersham Biosciences), rotating at room temperature for 1 h, pelleting by centrifugation, and then removing the supernatant. Subsequently 10 l of primary antiserum and 40 l of protein G-Sepharose were added, and the mixture was rotated overnight at 4°C. The following day the protein G-Sepharose-primary antibody complex was pelleted by centrifugation and then washed five times with 500 mM NaCl, 25 mM Tris-HCl, 5 mM EDTA, pH 7.5, with 0.5% Triton X-100, spun through 1 M sucrose, 500 mM NaCl, 25 mM Tris-HCl, 5 mM EDTA, pH 7.5, and finally washed twice in 10 mM Tris-HCl, 1 mM EDTA. The radiolabeled proteins were recovered by boiling in reducing SDS sample buffer with 50 mM dithiothreitol and analyzed by PAGE.
PAGE-We used SDS-12% polyacrylamide gels. Proteins were visualized by immunoblotting with a polyclonal EC-SOD antibody or autoradiography using a PhosphorImaging system (Molecular Dynamics).
Recombinant Virus Infections-Vaccina viruses overexpressing furin, PC6, and nerve growth factor using a cytomegalovirus promoter were kindly provided by Dr. Gary Thomas, Vollum Institute. Adenovirus overexpressing human EC-SOD using a cytomegalovirus promoter was purchased from Gene Transfer Vector Core, University of Iowa College of Medicine, Iowa City. Optimal multiplicity of infection (m.o.i.) was determined empirically in each cell line in a 6-well plate. LoVo cells were infected at 80% confluence; C6 and L2 cells were infected at 50% confluence. Virus was added to 700 l of fresh media for each well and then incubated at 37°C. After 1 h, 3 ml of additional fresh media were added to each well, and the cells were kept overnight at 37°C. Because LoVo cells secreted only a very small amount of EC-SOD and the vaccinia virus was mildly toxic to these cells at our working m.o.i., we augmented the secretion of EC-SOD in LoVo cells by co-infecting them with an adenoviral vector that increased secretion of EC-SOD. Nerve growth factor was used as a positive control for furin activity and was completely processed at an m.o.i. of 20.
In Vitro Furin Activity Assays-One unit of human recombinant furin (ϳ0.19 pmol) (Alexis) is defined as that able to cleave 1 pmol of fluorescent substrate t-butoxycarbonyl-Arg-Val-Arg-Arg-7-amino-4methylcoumarin (Alexis) in 1 min. For both the mutant and wild type EC-SOD, 1 unit of furin was added per 100 pmol of EC-SOD in 25 l of buffer (100 mM HEPES, 1 mM CaCl 2 , 0.5% Triton X-100, 1 mM␤-mercaptoethanol, pH 7.5) and incubated at 37°C overnight. Reactions were stopped with 5 volumes of 1 mM ZnCl 2 and protease inhibitors (final concentration of 2 mM 3,4-dichloroisocoumarin, 0.4 mM E64, and 40 mM 1,10-phenanthroline). After reactions were stopped, the samples were immediately boiled for 5 min in SDS samples buffer with 50 mM dithiothreitol. Furin activity was confirmed using a Spectrofluorometer (380 nm excitation and 460 nm emission) and the fluorescent furin substrate. EC-SOD processing was determined by Western blotting and quantitated using ECL imaging system (Molecular Dynamics).
Protein Structure Determination-The amino terminus was sequenced using automated Edman degradation carried out in an Applied Biosystems 477A automated protein sequencer with on-line analysis of the phenylthiohydantoin using an Applied Biosystems 120A high pressure liquid chromatography. The carboxyl terminus was determined by electrospray mass spectrometry as follows. Four l of EC-SOD (0.9 g/l) was reduced to 0.5 l of EC-SOD with 1.55 M dithiothreitol for 30 min. One l of this solution was applied to a microseparation gel loader tip packed with Poros 1 and eluted with 5 l of 50% MeOH and 5% formic acid in water. The sample was then analyzed on a Micromass Q-time of flight mass spectrometer with a nanoflow source using the same solvent. Carbohydrate residues were determined by measuring mass before and after treatment with the exoglycosidase neuraminidase.

EC-SOD Processing Varies among Tissues and Cell Lines-
Comparative immunoblotting of different human organs ( Fig.  1A) revealed that the total amount of EC-SOD protein was most abundant in the lung and least abundant in the brain, similar to that reported previously (22) in tissue activity assays. Moreover, immunoblotting revealed two distinct isoforms of EC-SOD migrating as 34-and 32-kDa species. We have demonstrated recently (15) that the lower molecular weight band represents EC-SOD that is missing the 13 carboxyl-terminal amino acids that are essential for heparin binding. Processing of EC-SOD varied among tissues, being higher in lung, liver, and heart yet notably low in the kidney and brain. The differences persisted from experiment to experiment, suggesting that the presence of the heparin-binding region might be more important in the kidney and brain.
To study intermediates in the processing of EC-SOD, we screened 15 mammalian cell lines (see "Experimental Procedures") for secretion of EC-SOD. Cells were screened by pulse labeling with [ 35 S]methionine and then immunoprecipitating the media and lysate. In addition to the rat lung L2 cell line described previously, we identified rat glial C6 cells as capable of secreting EC-SOD (Fig. 1B). Both L2 and C6 lines secreted EC-SOD with two molecular weights corresponding to the intact and proteolytically processed EC-SOD. The difference in molecular weight persisted when immunoprecipitates were treated with peptide:N-glycosidase F, which removes carbohydrate chains from N-linked residues by cleaving the aspara-FIG. 1. The ratio of intact to cleaved EC-SOD varies among tissue types and cell types. A, 50 g of lung tissue and 100 g of homogenized kidney, liver, heart, and brain tissue were submitted to reducing SDS-PAGE followed by immunoblotting for EC-SOD. The lung had the highest amount of EC-SOD, followed by the kidney, liver, heart and brain (exposure time for lung 10 s, kidney 30 s, liver and heart 1 min, and brain 10 min) Processing of EC-SOD was lowest in brain and highest in the lung. B, pulse-chase labeling followed by immunoprecipitation with EC-SOD antibody also revealed different ratios of intact and cleaved EC-SOD among cells lines. The L2 cell line (rat lung epithelial derived) secreted more processed EC-SOD than the C6 cell line (rat glial cell).
gine-GlcNAc bond (23) (data not shown). This suggested that the lower molecular weight EC-SOD secreted into cell culture was not a glycosylation variant. Interestingly, the (L2) lung cells had a similar percentage of processing as lung tissue (42 Ϯ 3 versus 45 Ϯ 3%), and the brain cells had a similar level of processing as brain tissue (21 Ϯ 4 versus 19 Ϯ 7%).
Although other cell lines, such as human aortic smooth muscle cells, are known to make EC-SOD mRNA (24), we were unable to detect secretion of EC-SOD protein by immunoblotting or pulse-chase analysis in these cell lines. Because EC-SOD is a matrix protein, its secretion may dependent on the presence of an intact extracellular matrix. Therefore, the cell lines were grown on Matrigel and retested for secretion of EC-SOD. Neither additional EC-SOD secretion nor a change in the ratio of intact to cleaved EC-SOD was noted when cells were grown on Matrigel. Therefore, secretion and proteolytic processing of EC-SOD appear to be unaffected by the extracellular matrix.
Studies with Inhibitors of Secretory Vesicle Trafficking-Proteins that are destined for secretion pass from the endoplasmic reticulum, through the Golgi network (TGN), and then exit via the plasma membrane into the extracellular space. Although prior work (15) had suggested that EC-SOD was proteolytically processed before it reached the extracellular space, we did not know the step at which this processing occurred. To answer this question, we used several techniques in an attempt to trap proteolytically processed EC-SOD in cells. For instance, brefeldin A and nocodazole were used to block transit from the endoplasmic reticulum, and the cells were held at 20°C to accumulate cargo proteins in the TGN. These experiments neither diminished the percentage of cleaved EC-SOD in the media nor caused an intracellular accumulation of cleaved EC-SOD (Fig. 2). Thus, proteolytic processing of EC-SOD appeared to occur after passage through the TGN. To determine whether processing occurred in a pH-sensitive vesicle, similar experiments were performed with bafilomycin; however, bafilomycin treatment did not inhibit proteolytic processing. The sensitivity of EC-SOD processing to calcium concentrations remains unknown because we were unable to remove calcium completely from the media without causing severe cell toxicity (25). Thus, proteolytic processing of EC-SOD most likely occurs after transit through the Golgi and before reaching the extracellular space.
EC-SOD Proteolytic Processing by Proprotein Convertases-Many extracellular proteins undergo proteolytic processing during their transit through the secretory pathway. To facilitate characterization and identification of the class of EC-SOD processing protease, protease inhibitors were added during the pulse-chase experiments. At multiple concentrations, neither cysteine protease inhibitors (E64), serine protease inhibitors (phenylmethysulfonyl fluoride), nor matrix metallo-protease inhibitors (1,10-phenanthroline) were able to inhibit selectively secretion of cleaved EC-SOD without causing significant cell toxicity. Because a few extracellular proteins have been reported to be proteolytically processed by cytosolic proteasomes, the proteasome inhibitor lactacystin was added during the chase (26); however, there was no inhibition of proteolytic cleavage. The failure to prevent processing using broad inhibitors of vesicular trafficking and protease activity led us to consider more specific approaches to identify the processing protease.
Because proteolytic processing occurred near polybasic residues and after transit through the TGN, the role of the PC proteases was examined. We found that a previously described inhibitor of PC proteases, decanoyl-Arg-Val-Lys-Arg-chloromethyl ketone (Dec-RVKR-CMK), blocked proteolytic process-ing of EC-SOD at low micromolar concentrations (Fig. 3). At similar concentrations neither the related elastase inhibitor N-methoxysuccinyl-Ala-Ala-Pro-Val-chloromethyl ketone nor the trypsin inhibitor tosyl-L-lysine chloromethyl ketone had any effect on proteolytic processing of EC-SOD. These results suggested that the PC family of proteases is required for EC-SOD processing.
To study the PC family of proteases, we used multiple different approaches, all of which implicated furin as an EC-SOD processing protease. First, we examined whether a cell line that lacked PC activity would proteolytically process EC-SOD. LoVo cells, which are well known to be deficient in furin activity (27), secreted only intact EC-SOD (Fig. 4A). Second, overexpression of furin enhanced EC-SOD processing in C6 and L2 cells (Fig. 4B) and restored EC-SOD processing in the furindeficient LoVo cells (Fig. 4C). Overexpression of a related PC protease, PC6, did not restore EC-SOD processing in LoVo

FIG. 2. Inhibition of the secretory pathway does not lead to accumulation of cleaved EC-SOD.
If the secretory pathways were blocked after proteolytic processing of EC-SOD, we might expect an intracellular accumulation of cleaved EC-SOD. To test this hypothesis, various techniques were used to block the secretory pathway. First, L2 cells were grown to confluence and then pulsed for 15 min with [ 35 S]methionine. A, control pulse-chase at 37°C. There is only intact EC-SOD in the lysate, but both cleaved and intact are present in the media after 1 h. B, brefeldin A, which prevents transit from the endoplasmic reticulum to the Golgi complex, inhibits secretion of EC-SOD but does not lead to intracellular accumulation of EC-SOD. C, reduction of temperature to 20°C leads to accumulation of cargo proteins in the TGN, but does not lead to intracellular accumulation of cleaved EC-SOD. Secretion of EC-SOD is delayed but appears normal. These experiments suggest that proteolytic processing occurs after transit through the TGN.

FIG. 3. A furin convertase inhibitor specifically inhibits proteolytic processing of EC-SOD in cells. L2 cells were labeled with [ 35 S]methionine and then chased for 2 h in the presence of the protease inhibitors.
A furin-specific inhibitor (Dec-RVKR-CMK) prevented proteolytic processing at micromolar concentrations. Similar results were seen when the experiment was repeated with C6 cells.
cells. Third, we found that recombinant furin was capable of cleaving both purified and recombinant human EC-SOD in vitro (Fig. 5).
Determination of the Amino Acid Residues Involved in Proteolytic Processing of EC-SOD by Furin-Although these experiments strongly suggested that furin was involved in EC-SOD processing, several observations led us to examine further the proteolytic processing of EC-SOD by furin. First, although both recombinant furin and purified furin were capable of proteolytically processing intact EC-SOD in vitro, immunoblotting revealed that the molecular weight of the furin treated EC-SOD was higher than that of cleaved EC-SOD (Fig. 5). This suggested that furin cleaved EC-SOD after Glu 209 and that there might be an additional protease that trimmed the remaining amino acid residues to Glu 209 . To test this hypothesis we treated EC-SOD first with furin and then with carboxypeptidase B. After only 30 min, all of the furin-processed EC-SOD migrated the same distance as cleaved EC-SOD, suggesting that a carboxypeptidase was capable of rapidly trimming the remaining basic residues from furin-treated EC-SOD. Although the identity of the EC-SOD carboxypeptidase is unknown, it has been reported recently (28 -30) that intracellular carboxypeptidases D and E are capable of removing carboxylterminal basic residues of peptide hormones such as insulin. Furin processing was also inhibited by specific inhibitors such as Dec-RVKR-CMK (Fig. 5) and nonspecific inhibitors such as zinc chloride (Fig. 5) and EDTA (not shown).
The second line of evidence that suggested that furin was not the only processing protease was prior work in our laboratory that showed that the carboxyl-terminal sequence of cleaved EC-SOD is Glu 206 -His 207 -Ser 208 -Glu 209 (15). This cleavage site is not consistent with previously published data (31) that suggest that furin preferentially recognizes the cleavage site sequence Arg-Xaa-(Lys/Arg)-Arg. Based on this consensus site, one should expect furin to cleave EC-SOD after Arg 210 -Lys 211 -Lys 212 -Arg 213 . However, human pro-parathyroid hormone (32), proalbumin (33), and pro-protein C (34) also serve as furin substrates but do not have basic P 4 residues, although their P 6 residues are all either Arg or Lys. Thus, an alternative cleavage site for EC-SOD could be after Arg 210 -Lys 211 -Lys 212 -Arg 213 -Arg 214 -Arg 215 . In an attempt to resolve these discrepancies, purified human recombinant EC-SOD and furin were incubated at 37°C and then subjected to protein sequencing. Edman degradation revealed that the amino terminus was intact, suggesting that furin was indeed processing the carboxyl terminus of EC-SOD. Mass spectrometric analysis of the furincleaved EC-SOD revealed a number of peaks that could be related to carboxyl-terminal cleavage (Table I). Surprisingly, the mass spectrometry results suggested that the carboxylterminal amino acids were both Glu 206 -His 207 -Ser 208 -Glu 209 -Arg 210 -Lys 211 and His 207 -Ser 208 -Glu 209 -Arg 210 -Lys 211 -Lys 212 . These results suggest that furin cleaves in the heparin-binding region but left some ambiguity as to which residue was crucial for furin processing. Although a silver stain of the recombinant furin preparations revealed that the predominant protein was furin, we cannot exclude that there was a small amount of carboxypeptidase activity responsible for the loss of the carboxyl-terminal arginines.
Because furin typically processes proteins with consensus sites RX(K/R)R or (K/R)XXX(K/R)R (31), we suspected that the two potential cleavage sites would be at Arg 213 or Arg 215 . To help determine which of these two were most likely, we purified EC-SOD from the serum of a patient homozygous for a Arg 213 3 Gly mutation. Western blot of this protein showed predominantly unprocessed EC-SOD (Fig. 6). Furthermore, this mutation rendered the protein resistant to proteolytic processing by furin but not other proteases that can proteolytically cleave EC-SOD in vitro (e.g. trypsin). This suggested that furin processes EC-SOD at Arg 213 and then a carboxypeptidase trims the remaining basic residues back to Glu 209 (Fig. 7). DISCUSSION In this report we have identified furin as an EC-SOD processing candidate, at least for the initial step of proteolytic processing. Several lines of evidence strongly implicate furin as the processing protease in vivo. 1) LoVo cells, which are deficient in furin activity, do not secrete cleaved EC-SOD. 2) EC-SOD can be cleaved by recombinant purified furin in vitro. 3) Processing is inhibited by a specific furin inhibitor. 4) Processing occurs after passage through the Golgi, consistent with the subcellular localization of furin (31); mutation within the furin consensus sequence hinders processing both in vivo and in vitro. 5) Although the carboxyl-terminal amino acid in fully processed EC-SOD is Glu 209 , a carboxypeptidase is capable of trimming the remaining basic residues back to Glu 209 , which is the actual carboxyl-terminal amino acid in cleaved EC-SOD. Therefore, proteolytic processing of EC-SOD likely occurs via a multistep mechanism. In the first step, the endoprotease furin cleaves EC-SOD within its polybasic site at Arg 213 . Subsequently, a carboxypeptidase trims the remaining basic amino acid residues back to Glu209. Although the carboxypeptidase that completes these final step(s) has not yet been identified, potential candidates include carboxypeptidase D and E (28 -30).
Re-examining the results of several studies published previously strengthens the evidence for a multistep mechanism. First, ϳ2-4% of the population has a variant EC-SOD gene in which Arg 213 is substituted to Gly 213 (11)(12)(13)35). This mutation results in the absence of cleaved EC-SOD in the serum (12,18), and we found that it rendered EC-SOD resistant to furin processing. Thus, Arg 213 was likely to play a crucial role in the initial processing step. Second, studies of EC-SOD mutants revealed that when the carboxyl-terminal residue is arginine, all the carboxyl-terminal basic residues are rapidly removed by an unknown intracellular protease (10). This strongly suggests the existence of an in vivo carboxypeptidase that trims the remaining basic residues in a second processing step. The variability of EC-SOD processing among tissues might be explained by two mechanisms. First, within organs, some cell types could secrete only intact EC-SOD and some could secrete only cleaved EC-SOD. Second, different cell types could secrete different ratios of intact to cleaved EC-SOD depending on the efficiency of furin processing within these cells. We identified three cell lines that consistently secrete different ratios of cleaved and intact EC-SOD, suggesting that individual cells might regulate intracellular proteolytic processing EC-SOD. In LoVo cells, we were able to demonstrate that EC-SOD processing could go from 0 to 100% depending on the relative amounts of furin.
We propose that secretion of intact EC-SOD is useful for highly localizing EC-SOD activity, whereas secretion of cleaved EC-SOD is useful for a more generalized EC-SOD activity. For instance, organs that are composed of millions of small identically functioning units, such as the alveoli in the lung and the acinii in the liver, may not need EC-SOD to be confined to the small functioning unit that secreted it. In these organs, we found a mix of cleaved and intact EC-SOD. On the other hand, FIG. 5. Proteolytic processing of EC-SOD in vitro. A, 1 unit of furin was added to 150 pmol of human EC-SOD either purified from aortas or made recombinantly in Chinese hamster ovary cells. Immunoblotting revealed that the molecular weight of EC-SOD decreased after treatment with furin; however, the high molecular weight band did not migrate completely to the position of the fully processed EC-SOD band, suggesting that there was incomplete proteolytic processing. Subsequent treatment with 1 unit of carboxypeptidase B (CPB) resulted in complete processing after 30 min. Furin processing was inhibited in the presence of a furin specific inhibitor (Dec-RVKR-CMK) 1.3 M and ZnCl 2 1 mM. B, scanning densitometry of the 1st three lanes of the purified EC-SOD shows that the intermediate band of furin-treated EC-SOD migrates to position of the fully processed EC-SOD after treatment with carboxypeptidase B. organs such as the brain and kidney have millions of units that do not function identically. In these organs we found mostly intact EC-SOD. The diffuse distribution of EC-SOD in the lung (36) and the highly localized distribution in the brain regions (37) have been confirmed by immunolocalization, but these studies did not distinguish between the distribution of intact and cleaved EC-SOD. Thus the putative physiologic role of EC-SOD processing remains speculative, and we are currently attempting to develop techniques that will distinguish intact from cleaved EC-SOD in situ. Using intracellular endoproteo-lytic cleavage to localize a protein is not unique to EC-SOD. Bone morphogenetic protein 2/4 is hypothesized to have short or long range activity, depending on the order of processing by intracellular endoprotease (38 -40). Thus, we propose that tissues regulate how tightly localized their EC-SOD activity is by regulating the amount and percentage of EC-SOD that is proteolytically cleaved. The molecular mechanisms that affect tissue distribution likely involve regulation of the polybasic residues within the intact carboxyl terminus (41)(42)(43)(44). Two lines of evidence suggest that proteolytic processing of the heparin-binding region is paramount. First, EC-SOD variants that do not have the heparin-binding region have markedly shortened tissue half-lives (14,45). Second, mutations within the heparin-binding region increase the plasma levels of EC-SOD (12,46,47). Thus, proteolytic processing of EC-SOD by furin is likely to be important in determining both the location and half-life of EC-SOD in the extracellular matrix. The clinical significance of proteolytic processing remains unknown, but it has been shown recently (13) that it inversely correlates with multiple risk factors for cardiovascular disease.
Our results demonstrate that furin is capable of proteolytic processing of EC-SOD in a two-step process. In the first step, furin cleaves EC-SOD within the heparin-binding polybasic region. A second carboxypeptidase is then necessary to remove remaining basic residues so that the final carboxyl-terminal residue is glutamic acid. Other intracellular processing proteases related to furin, such as PC7, have similar consensus sites and therefore cannot be ruled out as the authentic proc-FIG. 6. The Arg 213 3 Gly mutation prevents furin processing. Western blotting of 10 ng of EC-SOD reveals that wild type EC-SOD exists as a mixture of cleaved and intact protein, but EDC-SOD from a subject homozygous for the Arg 213 3 Gly mutation is associated with predominantly unprocessed protein. Furthermore, this mutation renders EC-SOD resistant to proteolytic processing by furin but not other proteases such as trypsin.
FIG. 7. Suggested scheme for proteolytic processing of EC-SOD. Furin cleaves intact EC-SOD (A) within the polybasic region (cross-hatched) of the carboxyl terminus. The resulting EC-SOD protein is partially processed and has a carboxyl terminus of Arg 213 . B, an unknown carboxypeptidase(s) then trims the remaining carboxyl-terminal residues to glutamic acid to yield fully processed EC-SOD (C). The Arg 213 3 Gly mutation prevents furin from proteolytically processing EC-SOD. essing proteases in vivo. The consequence of processing is that it diminishes the affinity of EC-SOD toward heparin. This potential regulatory mechanism could be exploited to change the tissue mobility of EC-SOD; however, because furin is an essential, ubiquitous enzyme, manipulation of its activity would have to be localized to specific organs.