Purification and sequence of rat extracellular superoxide dismutase B secreted by C6 glioma.

An enzyme which converts radical oxygen, produced by phorbol 12-myristate 13-acetate activated neutrophils, into nonluminescent products is secreted by rat C6 glioma. The enzyme was purified from chemically defined conditioned media and identified as an extracellular superoxide dismutase (EC-SOD). The purified enzyme is distinct from human EC-SOD C (Hjalmarsson, K., Marklund, S. L., Engström, A., and Edlund, T. (1987) Proc. Natl. Acad. Sci. U.S.A. 84, 6340-6344) by its elution from heparin-Sepharose at 300-400 mM NaCl, its pI of 6.1-7.2, and its native M(r) of 85,000 +/- 20,000. The rat EC-SOD is a dimer with a subunit M(r) of 34,000-36,000 and is extensively modified by post-translational processing. Although rat EC-SOD has a high sequence homology with the catalytic center and the polybasic heparin-binding site near the COOH terminus of human EC-SOD C, its NH2-terminal sequence and the sequences flanking the heparin-binding site differ substantially. The sequence of the isolated rat EC-SOD cDNA fully confirms the data obtained from amino acid sequence analysis. The amino acid sequence of the enzyme and its biochemical properties support its identification as the rat EC-SOD B.

enzymes are conserved throughout evolution. More recently Marklund and co-workers (3)(4)(5) have shown that several extracellular forms of the enzyme exist in the plasma of mammalian species. These extracellular forms can be divided into three classes namely A, B, and C according to their increasing affinity for heparin (6)(7)(8). Until now only the human extracellular SOD C has been purified, sequenced, cloned, and expressed.
EC-SOD C avidly binds to anchorage-dependent cell lines but has almost no affinity for blood cells, including neutrophils (9). The enzyme mostly binds to heparan sulfate proteoglycan in the glycocalyx of cell surfaces and in the connective tissue matrix. The EC-SOD A and B subtypes mainly exist in the extracellular fluid (8,10). The binding of EC-SOD C to sulfated proteoglycan and heparin is due to a hydrophilic COOH-terminal sequence containing 6 arginine and 3 lysine residues in the 21 carboxyl-terminal amino acids (4, 11). The C form is absent in rat and therefore, in this species the B form is the only EC-SOD that is able to bind to heparin (8). It has been suggested that the A and B forms are derived from the C form by post-translational proteolytic cleavage at the carboxyl terminus (11)(12)(13).
Whereas in other mammals EC-SODS have been characterized as tetrameric CuZn-binding glycoproteins with a subunit M, of 30,000-35,000 and a native M, of 150,000, rat EC-SOD has a native M, of 97,000 and seems to be a dimer (8).
EC-SOD is secreted by fibroblast and glial but not by epithelial and endothelial cells (14).
During our studies on rat Cs glioma, we observed that conditioned media of the latter cells contain a protein that strongly reduces the amount of superoxide released by activated neutrophils. In this article we describe the purification, the amino acid and nucleotide sequence determination, and the biochemical properties of the EC-SOD enzyme and identify it as the rat EC-SOD B form with properties distinct from the human EC-SOD C.

Materials
The rat C6 glioma cell line (ATCC CCL 107) was obtained from Flow Laboratories (Irvine, Scotland).

Cell Culture and Conditioned Medium
with 10% fetal calf serum, 2 mM L-glutamine, 1% (v/v) MEM vita-Rat c6 glioma cells were cultivated in Ham's F-10 supplemented B Secreted by Rat C6 Glioma 24615 mins 100 X, 1% (v/v) MEM nonessential amino acids 100 X, 100 units/ml penicillin, and 100 pg/ml streptomycin. Incubation was at 37 "C in a 5% CO2 humidified atmosphere. At confluence the cells were washed twice with PBS and harvested using 0.025% (w/v) trypsin and 0.01% (w/v) EDTA in PBS. Purification of EC-SOD All chromatographies were performed at room temperature. Hydroxyapatite Chromatography-Six harvests of 2 liters of CM (400 ml/doubletray) were diluted (2:1, v/v) with 50 mM Tris/HCl (pH 7.5) and the proteins concentrated in batch by binding onto Bio-Gel-HTP hydroxyapatite (4 "C). Hydroxyapatite was equilibrated with 10 mM Tris/HCl (pH 7.5), 10 p M CaClz (buffer A), and 10 g of resin added per 1.8 liter of diluted CM. After extensive washing with buffer A the bound material was discontinuously eluted with 500 mM sodium phosphate (pH 7.5).
Hydrophobic Interaction Chromatography-The active fractions eluted from the heparin-Sepharose CL-GB columns were pooled, dialyzed (m.c.0. 3, 500) against 500 mM (NH4)'S04, 50 mM sodium phosphate (pH 7.0), and applied onto a Bio-Rad phenyl-HPLC column (75 X 7.5 mm) on-line with a phenyl-Sepharose column (65 X 10 mm) equilibrated with the binding buffer. The bioactivity was not retained on the first phenyl-HPLC column, but was bound on the phenyl-Sepharose column. Bioactive material was eluted from the latter column with a 30-min linear gradient from binding buffer to 50 mM sodium phosphate (pH 7.0).
Reverse Phase Chromatography-The active fractions eluted from Luminescence Assay for Inhibition of the Respiratory Burst Neutrophils were purified (>97%) from blood taken from healthy human volunteers (15).
Neutrophils (3 X lo4 in 150 pl of RPMI) were incubated together with 150 pl of RPMI or test sample diluted in RPMI for 90 min at 37 "C. Subsequently 60 pl of a lucigenin solution M ) was added and the tubes placed in a Berthold 6-channel Biolumat LB 9505 apparatus (37 "C) to record the background luminescence. Finally 60 pl of phorbol 12-myristate 13-acetate (1 pg/ml) in RPMI was added and the luminescence recorded. The peak values were corrected for background and the inhibition calculated in relation to the value of the reference tubes without inhibitor. One unit of SOD is defined as the dilution which gives a 50% reduction of the luminescence of phorbol 12-myristate 13-acetate-activated neutrophils. One unit in our assay is equivalent to the activity of approximately 50 ng of intracellular SOD and 5.8 ng of the purified rat EC-SOD B ( Table  11).

Radioimmunoassay for EC-SOD
For the preparation of polyclonal anti-EC-SOD B, rabbits were primed and boosted twice with 15-20 pg of purified EC-SOD B together with complete Freund's adjuvant. An IgG fraction was obtained by Na2S04 precipitation (39.1 mg/ml).
The radioimmunoassay for EC-SOD B was performed as follows: cups of a microtiter plate were coated with 50 pl of SOD-containing fractions, postcoated with bovine serum albumin (2 mg/ml in PBS), incubated with 50 pl of the rabbit anti-EC-SOD B IgG (1:250 in PBS), washed three times (PBS containing 0.2% Tween 80), incubated with '?-labeled sheep anti-rabbit IgG antibodies, washed, and counted in a Packard y-counter.
Under these conditions less than 1 ng of rat c 6 EC-SOD can be detected. Cross-reactivity of our antiserum with commercial bovine and human erythrocyte CuZn-SOD is about 1.5%.

Protein Electrophoresis and Electroblotting
One-dimensional SDS-and two-dimensional PAGE were carried out as described (16)(17)(18). Proteins prepared for amino acid sequencing were carboxymethylated (19). Electroblotting onto ProBlott PVDF membranes was in the presence of 10 mM CAPS (20). Electrophoresis, blotting, and staining was in the presence of 1 mM thioglycolic acid.

Amino Acid Sequence Determination of EC-SOD B
Purified peptides were sequenced using a pulse-liquid Model 477A sequenator equipped with an on-line 120 phenylthiohydantoin analyzer from Applied Biosystems (Foster City, CA). In the case of NH,-

TABLE I1
Purification of rat c 6 EC-SOD B Samples were dialyzed against 50 mM Tris/HCl (pH 7.5) before measurement of the SOD activity and the protein concentration. The latter was determined using bovine serum albumin as a standard.   Table 111).
terminal sequencing of PVDF bound proteins, the excised bands or two-dimensional spots were sliced (2 X 2 mm) and placed in the reaction cartridge on top of a glass fiber filter to achieve an optimal flow of reagents.
Limited Acid Hydrolysis and Peptide Mapping-Partial acid hydrolysis of proteins immobilized in polyacrylamide gels was as described (21). The peptides were separated on a narrow-bore PTC C,. column (2.1 X 220 mm) by a gradient from (6 min to 15% B, 6-55 min to 60% R).
Cyanogen Bromide Cleavage-EC-SOD was deglycosylated with Nglycanase (22). CNBr-cleaved peptides were generated hy a modification of the method described by Gross (23). The reaction was performed in the dark in 70% (v/v) formic acid in water with a 50or 200-fold molar excess of CNBr/methionine residues. Incubation was a t 30 "C for 24 h. The reaction was stopped by dilution of the reaction sample with 15 volumes of water and subsequent freezedrying.

Cloning and Sequencing of EC-SOD
Preparation of mRNA from Rat C, Glioma-Total RNA was isolated by the acid-guanidinium-thiocyanate phenol method as described by Chomczynski and Saachi (25). Poly(A)+ RNA was purified concentrated fractions on a heparin-Sepharone column was as described. Activity was eluted using a NaCl gradient (-). Fractions of 3 ml were collected and assayed for SOD activity (block diagram). Optical density (-).
Fractions (horizontal bar) were pooled and subjected to hydrophobic interaction chromatography. R, hydrophobic interaction chromatography on phenyl-Sepharose 4R. The SOD activity was eluted from the column with a descending gradient of (NH,)ZSO, (-). Fractions of 3 ml were collected and tested for SOD activity (block diagram). Optical density (-).
The active pool is indicated by the horizontal bar. C, reverse phase chromatography on C,. The bound material was eluted with a gradient of acetonitrile (-). Fractions of 1 ml were collected. 10 pl of the fractions were freeze-dried and tested for SOD activity (block diagram). Optical density (-).
The indicated pool is used for amino acid sequence determination.

TABLE Ill
Determined amino acid sequences of rat EC-SOD R The peptides generated by formic acid, tryptic and hydroxylamine cleavage were separated by reverse phase chromatography. The peptide mixture obtained after CNBr cleavage was separated by one-dimensional PAGE and blotted onto a PVDF membrane (Fig. 3). The sequenced polypeptides were generated by the indicated hydrolysis method carboxyl-terminal of the amino acid between parentheses. The latter were deduced from the complete sequence of the rat EC-SOD B. The sequences were aligned with human EC-SOD C ( 4 ) (Table 111). E , peptides of a tryptic digest, carried out on the blot, were applied onto a C, reverse phase column and eluted with an acetonitrile gradient (-). Optical density (-).
Only the sequence data of peak T8 gave additional sequence information (Table 111). C, polypeptides of formic acid-hydrolyzed EC-SOD were separated on a PTC-CIR reverse phase column and the peptides eluted with an acetonitrile gradient (-). Optical density (-).
The indicated peaks have been sequenced. The data of the peaks AH7, AH17, AH24, AH29, and AH33 were partial sequences of peaks AH16, AH22, and AH23. D, peptides generated by hydroxylamine cleavage of rat EC-SOD were separated on a C, reverse phase column. The peptides were eluted with an acetonitrile gradient (-). The peptides H1 and H2 were sequenced. Optical density (-).
cDNA Synthesis and Cloning-The first cDNA strand was synthesized as descrihed (27) Fig. 4R) was synthesized and used to screen 7.5 X lo5 plaques of the cDNA library. cDNA was subcloned into pBluescript plasmid and sequenced (28.30).
The cDNA isolated from the library lacked 5"sequence information and ligase-anchored polymerase chain reaction was used to obtain the 5'-end (29). After ligation of an anchor oligo to the 3'ends of the first strand cDNA synthesized using olig~(dT)~S as a primer, EC-SOD specific sequences were amplified using an internal antisense primer complementary to the cDNA sequence 5'-GAC-CAAGCCTGTGATCTGCGG-3' (nucleotides 380-400, fig. 4B) and a primer corresponding to the anchor oligo S'-GC(;GCCGCTTAT-TAACCCTCACTAAA-3' (29). A 440-bp f r a m e n t was obtained and purified. Polymerase chain reaction with a nested antisense primer
Protein concentrations were determined using the microprotein assay (31). Residues are numbered according to rat mature EC-SOD. Amino acids determined by protein sequencing (summarized in Table 111) are indicated ( -1.

Secretion
The other residues are deduced from the nucleotide sequence of EC-SOD B presented in B. Residues identical to rat EC-luminescence assay. The extent of the decrease is maximal in the logarithmic phase of the cell growth (Table I). However, expressed per cell, the amount of secreted enzyme decreases as a function of the cultivation time. As the lactate dehydrogenase activity in the CM never exceeded 3-5% of the total amount of intracellular lactate dehydrogenase, we exclude the possibility that the SOD activity detected in the CM is derived from intracellular SOD by cell lysis.
Purification of EC-SOD from Rat Cs Conditioned Medium- The Cs secreted factor, identified as EC-SOD B (see below), was purified as summarized in Table 11. Hydroxyapatite chromatography was used to concentrate the proteins of approximately 12 liters of CM. Discontinuous elution with 500 mM sodium phosphate (pH 7.5) concentrated the proteins and increased the specific activity approximately 1.8-fold. The concentrated proteins were subsequently applied to a heparin-Sepharose column (Fig. 1A). EC-SOD eluted from the column between 300 and 400 mM NaCl. The specific activity increased more than 10-fold without loss in total activity. A further purification was achieved by chromatography on a Bio-Gel phenyl-5PW HPLC column on-line with a phenyl-Sepharose column. Dialysis of the active pool of heparin-Sepharose against the binding buffer for hydrophobic interaction chromatography strongly decreased the total activity. Of the applied protein nearly 90% was retained on the phenyl-HPLC column while the EC-SOD activity did not bind onto the first column but was retained on the phenyl-Sepharose column (Fig. 1B). The activity was eluted from the latter with a descending gradient of (NHJ2S04 in 50 mM sodium phosphate (pH 7.0). Only 15% of the total EC-SOD activity was recovered after this step. The decrease in activity is mainly due to a loss in EC-SOD protein as measured by radioimmunoassays using a rabbit polyclonal antibody against Cs EC-SOD (Table 11). Nevertheless the specific activity of the remaining EC-SOD increased 2-fold.
Reverse phase chromatography on a C4 column was used as a final purification step (Fig. 1C). The activity eluted at approximately 30% acetonitrile. However, the acidic conditions lowered the activity to 0.1% of the initial value and reduced the specific activity to 6,000 units/mg. Since no loss in EC-SOD protein was measured by radioimmunoassays we assume that the drop in activity is due to alterations in the metal binding sites below p H 5.0 as observed for bovine intracellular SOD (32). From our results we deduced that 1 unit of SOD corresponds to 5.8 ng of EC-SOD B.
Analysis by SDS-PAGE of the purified SOD indicates the presence of two polypeptides migrating with a M, of 34,000 and 36,000. The final preparation was more than 95% pure (Fig. 2 A ) . Two-dimensional gel electrophoresis under reducing conditions resolved the doublet protein band into multiple isoforms (Fig. 2B). The two arrays of protein spots varied in PI from 6.1 to 7.2 and in M , from 33,000 to 37,000, suggesting extensive post-translational processing of the polypeptides. The two-dimensional gel pattern is characteristic for highly glycosylated proteins. Identification of the Protein as EC-SOD by Amino Acid Sequencing-The purified protein preparation was separated by two-dimensional PAGE and blotted onto a PVDF membrane. NH2-terminal amino acid sequences were determined for different protein isoforms as indicated in Fig. 2C. All 4 spots proved to have the same NH,-terminal sequence (Table  111). These results indicate that the different protein clusters are derived from the same protein by post-translational modifications. The NH2-terminal sequence could not be aligned with a common alignment algoritm to any protein in the data bank (PIR, Swissprot, Genbank).
The same material was subjected to limited acidic hydrolysis, CNBr cleavage, tryptic digestion, and hydroxylamine cleavage (Fig. 3). This resulted in amino acid sequences, as shown in Table 111, which could be aligned with recombinant human EC-SOD C using the FASTA program (33) (Fig. 4A). The hydroxylamine cleavage product H1 contained the carboxyl-terminal sequence of the purified enzyme.
The CNBr-generated fragment of M , 10,000 could be aligned with a sequence located in the active site of SOD and was generated by a cleavage carboxyl-terminally of Met'4R in the rat EC-SOD sequence (Fig. 4A). The CNBr fragments of M , 20,800 and 13,000 were generated by cleavage carboxylterminally of Met57. These sequences have a low homology with the human EC-SOD C.
Except for the NH2-terminal sequence all sequence data could be aligned with recombinant human EC-SOD C from up to the carboxyl terminus. More than 75% of the EC-SOD sequence could be determined from peptide sequencing data (Table 111, Fig. 4A). Several other polypeptides, generated by acidic hydrolysis and separated by reverse phase, have been sequenced (Fig. 3). The latter are already included in the sequences of AH 16, AH 22, and AH 23 and support the existence of one protein with a complex two-dimensional pattern. The sequence data identify the enzyme as a rat EC-SOD. As the rat does not contain the C-form of EC-SOD the purified enzyme has to be the EC-SOD B form (8).
Complete Amino Acid Sequence of Rat EC-SOD B Deduced from the cDNA Nucleotide Sequence-Based on the determined amino acid sequences and their alignment to human EC-SOD C (Table 111, Fig. 4A), two degenerate primers were synthesized and used to isolate a rat EC-SOD cDNA. Based on the obtained cDNA sequence data a probe was synthesized and used to screen a Xgtll cDNA library of C, glioma. A clone contained the SOD sequence from nucleotide position 269 to position 1729 (Fig. 4B). Ligase-anchored polymerase chain reaction was used to obtain the full size nucleotide sequence. The latter is 1729 base pairs long and contains an open reading frame of 244 amino acids flanked by 133 base pairs in the 5'-and 864 bp in the 3"untranslated regions. 129CAGCCAUG'36 is the consensus sequence for translational initiation (34). The consensus sequence '706AUUAAA'711 is . located 19 bp upstream of the polyadenylation signal (35).
The mature rat EC-SOD contains 224 amino acids and is preceeded by a signal sequence of 20 amino acids. The human and rat EC-SODS are very homologous in regions which align with the intracellular SODs. These regions mainly comprise the active site of the enzymes (Fig. 4A). Except for 2 residues all invariant amino acids of the intracellular SODs can be identified in the primary structure of their extracellular counterparts (Table IV).
The hydrophilic COOH-terminal region, which has been proposed as the heparin-binding site of human EC-SOD C   and the consensus sequence for glycosylation at Amy4 are also conserved in the rat homologue (22). The main differences in the primary structure of the human and rat EC-SOD are observed in the NH2-terminal residues Met'-Ile'j5 and in regions flanking the heparin-binding site, comprising residues Thr'y9-Lys210 and Thrzz3-ThrZz4 (Fig. 4A). The sequence located NH2-terminally of the heparin-binding site lacks the Gly-Pro-Gly sequence present in human EC-SOD suggesting tertiary structure differences near the heparin-binding site. Using the FASTA program (33), the rat NH2-terminal sequence only aligns with human EC-SOD if residues carboxyl-terminal of are included in the alignment search. Biochemical Characteristics of Rat EC-SOD B-Although very homologous to the human EC-SOD C, the purified rat enzyme has properties distinct from its human counterpart. The rat enzyme has a slightly lower affinity for heparin-Sepharose and is eluted from the column at 300-400 mM NaCl as proposed for the B-form of human EC-SOD (6-8). In comparison, the human EC-SOD C eluted from the latter column above 500 mM NaC1.
Contrary to other SODs which are acid stable, the purified rat enzyme is acid labile. Its lability was demonstrated by incubation at pH 6.0, 4.0, and 2.0 at 4 "C and for 16 h. After readjustment to pH 7.4, the remaining bioactivity was deter-mined. The latter decreased to 55, 13, and 8%, respectively.
The PI of 6.1-7.2 measured for the rat enzyme (Fig. 2) is also significantly different from the PI of other SODs. Intracellular SODs have a PI close to 4.7 (36) and the human EC-SOD C has a PI of 4.5 ( 7 ) .
The native M , for the rat EC-SOD measured by gel filtration on Superose 12 was approximately 85,000 k 20,000 (Fig.  5) and is significantly less than the M, of 140,000 reported for the human EC-SOD C (3). The M , was confirmed by gel filtration on Toyopearl HW55 (data not shown). The estimated M , is indicative for a dimeric structure rather than a tetrameric structure as reported for the human enzyme.

DISCUSSION
Rat c.5, a glioma cell line with oligodendrocytic and astrocytic properties, secretes high amounts of EC-SOD in the logarithmic growth phase. An amount of 100-150 ng/ml is secreted in 72 h by c6 which is approximately 20-30-fold higher than the maximal amount of EC-SOD measured for several other cell lines by Marklund (14). Of the more than 15 rat and human cell lines tested' only rat B50 neuroblastoma secretes approximately the same amount of SOD. The secretion of SOD, which converts the superoxide produced by activated neutrophils, may be one of the mechanisms by which C6 survives in rodents and is able to form solid tumors (37,38).
Purification, amino acid sequence determination, cDNA cloning, and determination of the deduced amino acid sequence shows that the enzyme is 68% homologous to human EC-SOD C, the only member of the family of extracellular SODS from which the complete amino acid sequence has been described (4).
EC-SODS are ubiquitous CuZn-containing enzymes, present in the extracellular fluids of all mammals so far tested (6). The latter enzymes are heterogenous in their affinity for heparin-Sepharose and according to Marklunds (14) definition, three subtypes A, B, and C can be distinguished having no, moderate, or high affinity for heparin, respectively (3). Under appropriate conditions, all three isoforms can be detected in the extracellular fluids of humans and other mammalia (8). Rat seems to differ from the other mammalia by the absence of the C-isoform (8).
Recently, research on EC-SOD has been focused on the differences in heparin binding. It was shown that human rEC-SOD C has a carboxyl-terminal sequence containing a cluster of 6 arginine and 3 lysine residues in the last 21 amino acids which are involved in the binding to heparan sulfate proteoglycans and related molecules (10). To account for the observed decrease in heparin binding it has been proposed that the A and B form could be derived from the C form by modifications in the COOH-terminal sequence.
Two explanations, supported by experimental evidence, have been proposed. First, it was shown that EC-SOD A and especially EC-SOD B could be glycated in vitro and in uiuo, and that nonenzymatic glycation of some of the lysine residues at the COOH terminus could be responsible for the decrease in affinity for heparin (12). Second, the affinity for heparin is not influenced by truncations induced in the human rEC-SOD C downstream of G1u216. However, removal of the latter or its substitution by basic residues abolishes the heparinbinding properties. From these data Sandstrom et al. (11) concluded that the human B-form could be a heterotetramer composed of heparin binding C forms and truncated nonheparin binding A forms. Although each of the former explanations seems plausible, the amino acid sequence of rat EC-SOD indicates that the B form may exist as a distinct entity which is not derived from the C form.
Although important domains such as the catalytic site, Nglycosylation site, and heparin-binding domain are rather homologous, sequence comparison between human EC-SOD C and rat EC-SOD B indicates an overall sequence divergence of 32%. The observed sequence difference is larger than what is seen between rat and human intracellular CuZn-SOD and mitochondrial Mn-SOD, with 17 and 12% sequence divergence, respectively. Furthermore, interspecies divergence is manifested by sequence variations which are more or less evenly distributed along the molecule. In contrast, the divergence between rat EC-SOD B and human EC-SOD C is less random.
The alignment shows that the rat EC-SOD B has a NH2terminal sequence in which the nonidentity amounts to 55% over the first 70 residues. The COOH-terminal sequence, which lacks 3 amino acids, contains another stretch of nonidentical amino acids Thr'99-LysZ'o. An important difference between rat and human EC-SOD is the presence of a GlyZo'-Pro-Gly203 motif in the human enzyme. The latter sequence has a preference to form turns in protein structures (39) and probably favors the exposure of the heparin-binding site in EC-SOD C. The lack of a Gly-Pro-Gly motif in the rat enzyme may be responsible for its reduced heparin affinity. Based on these observations we conclude to the existence of an EC-SOD B isoform distinct from the human EC-SOD C.
SOD has been used in therapeutic approaches to reduce the noxious effects of reactive oxygen in pathophysiological conditions. Its main application is protection against postischaemic reperfusion damage in various organs (40). Intracellular CuZn-SODS, however, are rapidly cleared from the circulation and consequently they confer only limited protection. Therefore, in recent experiments human rEC-SOD was used (41), especially since it seems to bind in vivo to the proteoglycans of the endothelial cell surface. This could be due to the Gly-Pro-Gly sequence flanking the heparin-binding site which may give the enzyme more flexibility to interact with the proteoglycans.
Also, chimeric constructs of intracellular SOD and the heparin-binding site of a serpin protein C inhibitor coupled via the Gly-Pro-Gly motif have been used to improve the therapeutic effectiveness of SOD (39). Consequently, the lack of this Gly-Pro-Gly motif in the isolated rat EC-SOD B probably influences its cell binding and cell specificity. In conclusion, EC-SOD B which has biochemical properties distinct from EC-SOD c , is a new SOD to be tested in a number of therapeutic applications related to all types of tissue damage by superoxide.