ARCHIVES
OF BIOCHEMISTRY
AND
BIOPHYSICS
Vol. 297, No. 1, August 15, pp. 155-161, 1992
The Heparin Binding Site of Human ExtracellularSuperoxide Dismutase’ Tetsuo Adachi,” Tsutomu Kodera, Hideki Ohta, Kyozo Hayashi, and Kazuyuki Hirano Department
of Pharmaceutics,
Received February
Gifu Pharmaceutical
University,
11, 1992, and in revised form May 1, 1992
Extracellular-superoxide dismutase (EC-SOD) is a secretory glycoprotein that is major SOD isozyme in extracellular fluids. We revealed the possible structure of the carbohydrate chain of serum EC-SOD with the serial lectin affinity technique. The structure is a biantennary complex type with an internal fucose residue attached to asparagine-linked N-acetyl-o-glucosamine and with terminal sialic acid linked to iV-acetyllactosamine. ECSOD in plasma is heterogeneous with regard to heparin affinity and can be divided into three fractions: A, without affinity; B, with intermediate affinity; and C, with high affinity. It appeared that this heterogeneity is not dependent on the carbohydrate structure upon comparison of EC-SOD A, B, and C. No effect of the glycopeptidase F treatment of EC-SOD C on its heparin affinity supported the above results. A previous report showed that both lysine and arginine residues, probably at the C-terminal end, contribute to heparin binding. Recombinant EC-SOD C treated with trypsin or endoproteinase Lys C, which lost three lysine residues (Lys-2 11, Lys-2 12, and Lys-220) or one lysine residue (Lys-220) at the C-terminal end, had no or weak affinity for the heparin HPLC column, respectively. The proteinase-treated r-EC-SOD C also lost triple arginine residues which are adjacent to double lysine residues. These results suggest that the heparin-binding site may occur on a “cluster” of basic amino acids at the C-terminal end of EC-SOD C. EC-SOD is speculated to be primarily synthesized as type C, and types A and B are probably the result of secondary modifications. It appeared that the proteolytic cleavage of the exteriorized lysine- and arginine-rich C-terminal end in vivo is a more important contributory factor to the formation of ECSOD B and/or EC-SOD A. o 1992 Academic PWS. IN.
r This work was supported in part by Grants-in-Aid for the Encouragement of Young Scientists from the Ministry of Education, Science, and Culture of Japan (02771711) to T.A. * To whom correspondence should be addressed. 0003.9861/92 $5.00 Copyright All rights
Gifu 502, Japan
0 1992 by Academic Press, of reproduction in any form
Extracellular-superoxide dismutase (EC-SOD)3 (EC 1.15.1.1) is a secretory, tetrameric copper- and zinc-containing glycoprotein with a subunit molecular weight of about 30 kDa (1,2). EC-SOD is the major SOD isozyme in extracellular fluids, such as plasma, lymph (3), and synovial fluids (4), whereas this isozyme is the least predominant in tissues compared to copper,zinc-SOD (Cu,Zn-SOD) and manganese-SOD (Mn-SOD) (5, 6). The specific activity (l), the ligands to Cu and Zn, the cysteins forming the intrasubunit disulfide bridge, and the arginine residue in the entrance to the active site of EC-SOD (7) can be identical to those of Cu,Zn-SOD. The conformation of EC-SOD around the active site is therefore expected to be very similar to that of Cu,Zn-SOD. A prominent feature of EC-SOD is its affinity for heparin. EC-SOD in plasma in man (8) and other mammalian species (9) is heterogeneous with regard to affinity for heparin-Sepharose and can be divided into three fractions: A, without affinity; B, with weak affinity; and C, with relatively strong affinity. The background to the heterogeneity in heparin affinity is still unresolved. Studies with amino acid-specific reagents indicated that both lysine and arginine residues seemed to contribute to heparin binding (10). Compared to other SODS, another special feature of EC-SOD is that it is a glycoprotein. A possible N-glycosylation site was speculated to be Asn-89 (7). It is very important to know that the carbohydrate moiety of recombinant EC-SOD C (r-EC-SOD C) is indistinguishable 3 Abbreviations used: EC-SOD, extracellular-superoxide dismutase; Cu,Zn-SOD, copper,zinc-superoxide dismutase; Mn-SOD, manganesesuperoxide dismutase; r-, recombinant; n-, native; Con A, concanavalin A; LCA, lentil lectin; RCA120, Ricinus communis agglutinin 120; WGA, wheat germ agglutinin; PHA-E4, phytohemagglutinin E4; PNA, peanut lectin; a-MM, cu-methyl-D-mannoside; GlcNAc, N-acetyl-D-ghrcosamine; Gal, galactose; Galpl-4GlcNAc, N-acetyllactosamine; TNBS, trinitrobenzene sulfonic acid; PBS, phosphate-buffered saline; BSA, bovine serum albumin; TPCK, L-l-p-tosylamino-2-phenylethyl chloromethyl ketone; TLCK, N-p-tosyl-L-lysinechloromethyl ketone; IgG, immunoglobulin G. 155
Inc. reserved.
156
ADACHI
from that of native EC-SOD for r-EC-SOD C is valuable for biological as well as clinical use. However, the structure of the carbohydrate chain and its contribution to the heterogeneity of plasma EC-SOD are still unresolved. In the present investigation, the effect of the proteolytic cleavage of the C-terminal region of r-EC-SOD C on affinity for heparin-Sepharose is explored. Moreover, the carbohydrate chains of serum EC-SOD A, B, and C were compared with each other and with that of r-EC-SOD C by using lectin affinity techniques. EXPERIMENTAL
PROCEDURES
Materials. Human r-EC-SOD C prepared as described previously (2) was kindly provided by SYMBICOM AB, UmeH, Sweden. Endoproteinase Lys-C (lysylendoproteinase) was purchased from BoehringerMannheim GmbH, Mannheim, Germany. TPCK-trypsin and TLCKchymotrypsin were purchased from Sigma, St. Louis, Missouri, U.S.A. Neuraminidase (sialidase) was purchased from Nacalai Tesque, Kyoto, Japan. Glycopeptidase F (N-glycanase) was purchased from Takara Shuzo, Kyoto, Japan. Alkaline phosphatase-conjugated rabbit anti[mouse IgG (H + L)] antibody was purchased from Zymed Laboratories, San Francisco, California, U.S.A. Peroxidase-conjugated goat anti[rabbit IgG (H f L)] antibody was purchased from Organon Teknika Co., Durham, North Carolina, USA. Heparin-Sepharose, concanavalin A (Con A)-Sepharose, and CNBr-activated Sepharose 4B were products of Pharmacia LKB Biotechnology, Uppsala, Sweden. Con A-, lentil lectin (LCA)-, Ricinus communis agglutinin 120 (RCAlPO)-, wheat germ agglutinin (WGA)-, phytohemagglutinin E4 (PHA-E4)-, and peanut lectin (PNA)-agarose columns were purchased from Seikagaku Co., Tokyo, Japan. Immunoplate was purchased from Nunc, Roskilde, Denmark. EZJSA of EC-SOD concentration. ELISA of EC-SOD proceeded by the method described previously (11). S-Carboxymethylated r-EC-SOD was assayed by the method as follows, because the above method could not detect S-carboxymethylated r-EC-SOD. Portions (70 ~1) of S-carboxymethylated r-EC-SOD (standard) or sample diluted with 50 mM sodium carbonate, pH 9.5, containing 0.02% NaN, were added into each well of an immunoplate and left to stand overnight at 4°C. Each well was washed with phosphate-buffered saline (PBS, 10 mM sodium phosphate, pH 7.4, containing 0.15 M NaCl) containing 0.05% Tween 20 and 0.05% merthiolate (washing buffer). The remaining protein-binding sites were blocked with 300 ~1 of PBS containing 1% bovine serum albumin (BSA), 0.05% Tween 20, and 0.05% merthiolate (blocking buffer). The plate was then left to stand overnight at 4°C and washed three times with the washing buffer. Then 80 al of 9.1 pg/ml rabbit anti-EC-SOD antibody diluted with blocking buffer was added to each well, and the plate was incubated for 3 h at room temperature and washed with the washing buffer. A 90+1 portion of peroxidase-conjugated goat anti-[rabbit IgG (H + L)] antibody at a 1:lOOO dilution in blocking buffer was added to each well and the plate was incubated again for 3 h at room temperature, followed by washing three times with washing buffer. Substrate solution (150 ~1 of 0.1 M McIlvain buffer, pH 6.5, containing 16.5 mM o-phenylenediamine and 0.02% H,O,) was then added to each well and the plate was incubated for 30 min at room temperature. The enzyme reaction was stopped by the addition of 50 ~1 of 4.5 N H$O,, and the absorbance at 492 nm was measured with a Corona MTP-120 microplate reader. Heparin affinity chromatography. Heparin affinity chromatographies were performed on a heparin-Sepharose column (volume, 2 ml) or a HPLC column of TSK-gel heparin 5PW (Tosoh, Tokyo, Japan). The column chromatography was operated at a flow rate of 0.7 ml/min. A gradient system formed between buffer A (25 mM sodium phosphate, pH 6.5) and buffer B (25 mM sodium phosphate, pH 6.5, containing 1 M NaCl) was used.
ET AL. Purification of EC-SOD A, B, and C with anti-(r-EC-SOD C) monoC) monoclonal antibody-conjugated Sepharose 4B. Anti-(r-EC-SOD clonal antibody-coupled gel as previously prepared (11) was used. The samples (12 ml of EC-SOD A, 24 ml of EC-SOD B, and 20 ml of ECSOD C fractions in serum separated by heparin-Sepharose column chromatography) were applied to this column by the method previously described (11). Samples (150 pl of serum, Lectin ufinity column chromatography. 100 ~1 of 1 rg/ml r-EC-SOD C, or 100 ~1 of 300-500 rig/ml of EC-SOD A, B, and C fractions) were applied to Con A (volume, 1 ml)-, LCA (volume, 1 ml)-, RCA120 (volume, 1 ml)-, WGA (volume, 1 ml)-, and PNA (volume, 0.5 ml)-agarose columns equilibrated with PBS and washed with the same buffer. The bound fractions were then eluted with above buffer containing 0.25 M a-methyl-D-mannoside (a-MM) (for Con A and LCA columns), 0.25 M galactose (Gal) (for RCA120 and PNA columns), and 0.25 M N-acetyl-D-glucosamine (GlcNAc) (for the WGA column), respectively. Serial &tin affinity technique. Structure studies of the carbohydrate chain were performed with the method established by Cummings and Kornfeld (12). Sample (300 ~1) was applied to a Con A-Sepharose column (volume, 2 ml) equilibrated with PBS containing 10 pg/ml BSA and washed with the same buffer. The bound fraction was then eluted by stepwise addition of two different concentrations (0.01 and 0.5 M) of aMM. Three fractions, an unbound fraction (Fraction I), a weakly bound fraction (Fraction II), and a strongly bound fraction (Fraction III), were dialyzed against PBS and followed by concentration with YM-10. These fraction were further applied to PHA-E4 (volume, 0.5 ml), LCA (volume, 1 ml), and WGA (volume, 1 ml) columns, respectively, equilibrated with above buffer. The unbound and the bound fractions on the respective columns were separated by using 0.1 M GlcNAc, 0.2 M a-MM, and 0.1 M GlcNAc, respectively, as elution buffers. S-Curboxymethylution of r-EC-SOD C. Thirty microliters of 15 mg/ ml r-EC-SOD C treated with or without proteinases was reduced with 100 ~1 of 0.2 M dithiothreitol dissolved in 0.5 M Tris-HCl, pH 8.5, containing 7 M guanidine-HCl and 10 mM EDTA, flushed with nitrogen, and incubated for 2 h at room temperature. After the addition of I2 ~1 of 2.7 M iodoacetic acid, the reaction mixture was incubated in the dark for 3 h. Hydroxylnmine cleavage of reduced and S-curboxymethykzted r-ECSOD C. S-Carboxymethylated r-EC-SOD C’s prepared as above (450 rg) were dialyzed against 15 mM sodium acetate, pH 6.0, at 4°C overnight, followed by rotary evaporation to dryness. The residues were dissolved with 200 ~1 of 2 M hydroxylamine-HCl, 0.2 M K&O,, and 6 M guanidineHCl (pH 10.5). After incubation at 45’C for 4 h, the mixture were acidified by the addition of 30 ~1 of formic acid (13). The Asn-Gly peptide bond (Asn-180-Gly-181 in EC-SOD) was cleaved by this treatment. The products were separated by HPLC on a Cl8 reverse-phase column (PBondasphere, 5 pm, Nihon Waters Ltd., Tokyo, Japan). The column chromatography was operated at a flow rate of 0.7 ml/min. A gradient system formed between solvent A (0.1% trifluoroacetic acid in Milli-Q . water) and solvent B (0.1% trifluoroacetic acid in acetonitrile) was used. Peptides were eluted with a linear gradient up to 60% of solvent B for 60 min and continuously up to 100% of solvent B for 10 min. The peptides were detected on the basis of the absorbance at 210 nm. C-terminal peptides (Gly-181, C-terminal, retention time 33-34 min; amino acid sequences of peptides were determined; data not shown) were isolated and followed by rotary evaporation to dryness. The residues of a CChymotryptic digestion of C-terminal peptides. terminal fragment isolated as described above were dissolved with 100 ~1 of 10 mM Tris-HCl, pH 8.6, followed by the addition of 10 pl of 10 rig/ml TLCK-chymotrypsin. After incubation at 25°C for 1 h, the products were separated with Cl8 reverse-phase HPLC under the conditions described above. Amino acid sequence determination. termined with an Applied Biosystems
Amino acid sequences were deprotein sequencer, Model 473A.
THE
HEPARIN
BINDING
SITE
OF HUMAN
EXTRACELLULAR-SUPEROXIDE
Sialidase Treatment
(4 CQQA
-0
10
20
30
5. w L!a
OLPHAD r;;40 k n 30 3 20 A cj w
i
r
"
I
$ I 2
10
n -0
10
20
"00
10
20
0
10
20
Fraction number (1 mlhl mlhube) Jbe)
FIG. 1.
Serial lectin affinity chromatographies of serum EC-SOD. Three hundred microliters of normal serum was applied to a Con ASepharose column (a). Three fractions, unbound fraction (Fraction I), a weakly bound fraction eluted with 0.01 M a-MM (Fraction II), and a strongly bound fraction eluted with 0.5 M u-MM (Fraction III) were dialyzed against PBS followed by concentration with a YM-10 membrane filter. These fractions were further applied to PHA-E4-agarose (b), LCAagarose (c), and WGA-agarose (d) columns, respectively.
RESULTS
Lectin Affinity
Chromatography
157
of Serum EC-SOD
For no affinity to the RCA120 column, the presence of sialic acid residues attached to N-acetyllactosamine (Galpl-4GlcNAc) in the N-linked carbohydrate chain is possible. To confirm this notion, the effect of sialidase treatment on serum EC-SOD was studied. Normal serum (150 ~1) was incubated with 50 mU of sialidase in the presence of 150 ~1 of 25 mM potassium phosphate buffer, pH 6.5, at 37°C for 24 h (14), followed by application on the RCA120 column. Sialidase-treated n-EC-SOD substantially bound on the RCA120 column, whereas nontreated enzyme failed to bind to this column (Fig. 3a). nEC-SOD in serum treated with or without sialidase under the same conditions as those in the above experiments did not bind to the PNA-agarose column which required O-liked mucin-type sugar chains for binding (Fig. 3b).
Fraction number (2 mlhube)
50
DISMUTASE
of EC-SOD in Serum
The affinity for lectins was investigated to clarify the carbohydrate chain of n-EC-SOD. Normal serum (150 ~1) was applied to four kinds of lectin columns and EC-SOD in each fractions was detected by ELISA. n-EC-SOD in serum was found to bind to Con A- and LCA-agarose columns, whereas it did not bind to RCAlZO- and WGAagarose columns (data not shown). With the serial lectin affinity technique, substantially all of n-EC-SOD (85%) was found in the weakly bound fraction (Fraction II) for the Con A-Sepharose column (Fig. 1). Little of n-ECSOD was found in the unbound fraction (Fraction I, 5%) and in the strongly bound fraction (Fraction III, 10%). In order to further evaluate the lectin-binding affinity, Fractions I, II, and III obtained on the Con A-Sepharose column were further applied on PHA-E4, LCA, and WGA columns, respectively. Fraction I, a small proportion of Con A column chromatography, did not bind to the PHAE4 column. It was observed that Fraction II bound substantially to the LCA column, whereas Fraction III did not bind to the WGA column (Fig. 1). r-EC-SOD C expressed in the CHO cell (2) also bound to Con A- and LCA-agarose columns, whereas it did not bind to RCA120- and WGA-agarose columns. r-EC-SOD showed similar behavior with the serial lectin affinity technique (Fig. 2).
Lectin Column Chromatography
of EC-SOD A, B, and C
Serum EC-SOD could be separated into three fractions: A, without affinity; B, with weak affinity; and C, with relatively strong affinity for heparin-Sepharose, as shown previously (9, 11, 15). Pooled A, B, and C fractions were then isolated on anti-(r-EC-SOD C) monoclonal antibodyconjugated Sepharose 4B columns as described under Experimental Procedures. Subsequently the isolated fractions were chromatographed with Con A-, LCA-, RCAlBO-, and WGA-agarose columns. All three fractions showed similar chromatographs for lectins, bound to Con A and LCA columns, whereas they did not bind to RCA120
(4 Con
Fraction number (2 ml/tube) (b) PI-IA-E4
w !&A
(4
WGA
Fraction number (1 mlhube)
FIG. 2.
Serial lectin affinity chromatographies of r-EC-SOD C. Three hundred microliters of r-EC-SOD (900 rig/ml) was applied to a Con ASepharose column and chromatographed as described in the legend to Fig. 1.
158
ADACHI
(a) RCA1 20-aaarose
100
100 (ii)
(0 80
80
-2 @ n 60
60 z
8 Y
I &I
40
kt
6
20
$ I
40
d
20
0 l!!LLl_a. 0 2 4
0 6810
0
2
4
6
8
10
Fraction number (2 ml/tube)
(W PNA-aoarose (ii)
6)
8 W
20 ~1 of 10 mM Tris-HCl buffer, pH 8.6, and incubated at 30°C for 12 h. Portions (1 ~1) of r-EC-SOD treated with proteinases were mixed with 20 ~1 of 25 mM sodium phosphate buffer, pH 6.5, followed by heparin affinity chromatographies with the TSK-gel heparin 5PW HPLC column as described under Experimental Procedures. The enzyme treated with endoproteinase Lys-C had only very weak affinity for heparin and eluted at 0.05 M NaCI, whereas nontreated r-EC-SOD had strong affinity for heparin and eluted mainly at 0.6 M NaCl. Trypsin-treated r-EC-SOD had lost most of its heparin affinity (72% of the enzyme was observed in the unbound fraction and the remainder was eluted at 0.05 M NaCl) as shown in Fig. 6.
100
100
';i @ 8
ET AL.
80
80
60
60
40
40
85 t4
20
' o0
2
4
A 6610
o-
024
of C-Terminal Amino Acid Sequences of C’s Treated with Proteinases
C-terminal peptides (Gly-181-C-terminal) of r-ECSOD C’s treated with or without endoproteinase Lys-C or TPCK-trypsin were isolated and then digested with TLCK-chymotrypsin as described under Experimental
z (I) 5 t? d
20
d
Identification r-EC-SOD
6
8
10
Fraction number (2 ml/tube) FIG. 3. RCAlBO-agarose and PNA-agarose column chromatographies of desialylated n-EC-SOD. Normal serum (150 ~1) was incubated with (ii) or without (i) 50 mU sialidase in the presence of 150 rl of 25 mM potassium phosphate buffer, pH 6.5, at 37°C for 24 h, followed by applications on an RCA120-agarose column (volume, 1 ml) (a) and a PNAagarose column (volume, 0.5 ml) (b), and chromatographed as described under Experimental Procedures. The EC-SOD concentration in each fraction was determined by ELISA.
(a) EC-SOD A loo [ Con A
-0
2
4
0.N EC-SOD
and WGA columns as shown in Fig. 4. These behaviors were also the same as that of nonseparated serum ECSOD as described above. Heparin-Sepharose Column Chromatography of r-ECSOD C Treated with Glycopeptidase F It was suggested that the heterogeneity in heparin affinity of serum EC-SOD was not dependent on the carbohydrate chain (Fig. 4). To confirm this notion, the affinity of deglycosylated r-EC-SOD C for heparinSepharose was investigated. Portions (400 ~1) of S-carboxymethylated r-EC-SOD C (40 fig/ml) which was dialyzed against 10 mM Tris-HCl buffer, pH 8.6, were incubated with or without 5 mU glycopeptidase F at 37°C for 1 h. Complete deglycosylation of glycopeptidase Ftreated r-EC-SOD C was certified by the loss of its affinity for Con A-Sepharose. Deglycosylated r-EC-SOD C kept the high affinity for heparin-Sepharose as did nontreated r-EC-SOD C (Fig. 5). Heparin HPLC of r-EC-SOD C Treated with Proteinases Ten microliters of r-EC-SOD C (45 mg/ml) was mixed with 1 pg of endoproteinase Lys-C or TPCK-trypsin in
6
c LCA
610
0
2
4
r RCA120
6
610
0
4
6
610
0
WGA
2
4
6
610
B
loo [ Con A
F LCA
g!J!JpJk 0
2
r
2
4
6
F RCA120
WGA
1
610
0
2
4
6
6100
2
4
6
610
0
2
4
6
610
610
0
2
4
6
6100
2
4
6
610
0
2
4
6
610
(c) EC-SOD C
0
2
4
6
Fraction number (2 ml/tube)
FIG. 4. Lectin affinity chromatographies of EC-SOD A, B, and C obtained from normal serum. EC-SOD A (a), B (b), and C (c) in normal serum were separated with heparin-Sepharose column (volume, 2 ml) chromatography and then purified with an anti-r-EC-SOD C monoclonal antibody-conjugated Sepharose 4B column (volume, 2 ml) as described under Experimental Procedures. Isolated fractions were applied on each Con A-, LCA-, RCAlZO-, and WGA-agarose column (volume, 1 ml).
THE
HEPARIN
BINDING
SITE
OF HUMAN
(4 Con A-SeDharose 60 50 40 30 20 10 -0
2
4
6
8
" -0
10
2
4
6
6
10
Fraction number (1 ml/tube)
30
0) 1.0
f"" 7 e
20
/
0.6
0.6 Q g
i
10
0
M 0
I
5
0.4 0.2
/
0.0 10
15
20
25
0 0
5
10
15
20
25
Fraction number (1 ml/tube) FIG. 5. Con A-Sepharose and heparin-Sepharose column chromatographies of deglycosylated r-EC-SOD C. Four hundred microliters of reduced and S-carboxymethylated r-EC-SOD (40 pg/ml) was incubated with (ii) or without (i) 5 mU of glycopeptidase F at 37°C for 1 h. Aliquots (100 ~1) were applied to a Con A-Sepharose column (volume, 2 ml) (a) and a heparin-Sepharose column (volume, 2 ml) (b), and chromatographed as described under Experimental Procedures. The EC-SOD concentration in each fraction was determined by ELISA (0). NaCl concentration in buffer (. . . ).
Procedures. Figure 7 shows the Cl8 reverse-phase HPLC chromatographs. Peptides A-l and A-2 in nontreated rEC-SOD C; B-l and B-2 in endoproteinase Lys-C-treated r-EC-SOD C; and C-l, C-2, and C-3 in TPCK-trypsintreated r-EC-SOD C were isolated, followed by the determination of amino acid sequences. Amino acid sequence of the B-l peptide showed that r-EC-SOD C treated with endoproteinase Lys-C had lost the Arg-213Ala-222 peptide which contained one lysyl residue (Lys220) and three arginine residues (Arg-213, -214, and 215), compared to nontreated enzyme. It appeared that the TPCK-trypsin-treated enzyme had lost another two lysyl residues, Lys-211 and Lys-212, by the sequence analysis of C-l and C-2 peptides in Fig. 7. DISCUSSION
In the present study, a possible structure of the carbohydrate chain of EC-SOD is revealed by the serial lectin
EXTRACELLULAR-SUPEROXIDE
159
DISMUTASE
affinity technique. The core structure is shown to be a biantennary complex type without bisecting GlcNAc. The high-affinity binding to LCA-agarose shows the presence of an internal fucose residue attached to asparagine-linked GlcNAc. The appearance of affinity to RCA120-agarose after desialylation shows the presence of terminal sialic acid linked to Galpl-4GlcNAc. The structure of the carbohydrate chain of r-EC-SOD C produced in the CHO cell is indistinguishable from that of n-EC-SOD judging from lectin affinity column chromatographies. Our proposed carbohydrate structure of EC-SOD is consistent with that of r-EC-SOD, a complex biantennary type having a core fucose, which recently appeared (16). The carbohydrate structure of r-erythropoietin was also reported to be common to that of urinary n-erythropoietin (17, 18). However, the contribution of the carbohydrate chain of EC-SOD to biochemical and therapeutic properties such as enzymic activity, turnover rate in the body, and antigenicity is unclear; we could provide evidence that rEC-SOD C is valuable for experiments in biological and clinical research. Based on the amino acid sequence deduced from the cDNA sequence, EC-SOD appeared to have only one possible N-glycosylation site, Asn-89, and no 0-glycosylation site (7). This notion was confirmed by investigations using r-EC-SOD (16). An immobilized lectin, PNA, is known to preferentially bind to a typical
‘:.,o.o
;;;
0
10
Fraction
20
30
40
number
(0.7 ml/tube)
50
FIG. 6. Heparin HPLC of r-EC-SOD C treated with proteinases. rEC-SOD C (450 pg) was treated without (a) or with 1 pg of endoproteinase Lys-C (b) or TPCK-trypsin (c) in 20 ~1 of 10 mM Tris-HCl, pH 8.6, at 30°C for 12 h. Aliquots (1 hl) were mixed with 20 ~1 of 25 mM sodium phosphate, pH 6.5, and loaded on a column of TSK-gel heparin 5PW. The column was washed with the above buffer, and then bound components were eluted with a linear gradient of NaCl in the buffer in each fraction was assayed by (* * * ). The EC-SOD concentration ELISA (0).
160
ADACHI
I C-l I IJ
ET AL.
(c)
(b)
c-3
3-2
B-l
-L I
c-2
1 lJ
1'0
210
3b
I 40
I 5c
Retention time (min) Hydroxylamine I
180
Chymotrypsin 1
190
200
210
NGNAGRRLACCWGVCGPGLWERQAREESECK (a) r-EC-SOD
e
A-2
(b) Endoprotelnrse Lys-Ctreated r-ECSOD
jM
B _2
(c) TPCK-tryprintreated r-ECSOD
F *
c-3
220222
I :.
, II
A-l
h ’
-B-l -c-1 c-2
:
-’ _I e
FIG. 7. C-terminal amino acid sequences of the hydroxylamine-cleaved peptides of r-EC-SOD C treated with proteinases. C-terminal peptides (Gly-181~C-terminal) of r-EC-SOD C treated without (a) or with endoproteinase Lys-C (b) or TPCK-trypsin (c) were isolated and then digested with TLCK-chymotrypsin as described under Experimental Procedures. TLCK-chymotrypsin-digested peptides were separated again with Cl8 reverse-phase HPLC chromatographies. Amino acid sequences of isolated peptides were determined with an Applied Biosystems protein sequencer, Model 473A.
mucin-type glycochain, GalPl + 3GalNAccul + Ser/Thr, but not to this sialylated glycopeptide (19). We now show that desialylated n-EC-SOD in serum also has no affinity to the PNA column (Fig. 3), suggesting the absence of an O-linked mucin-type sugar chain in n-EC-SOD. A prominent feature of EC-SOD is its affinity for heparin. EC-SOD in plasma is heterogeneous with regard to affinity for heparin-Sepharose (8,9). The background to the heterogeneity in heparin affinity is still unresolved. The binding of EC-SOD to heparin is of an electrostatic nature (20). It is known that sterical hindrance due to carbohydrate chains and/or their electric charge contributes to some of the biochemical properties of proteins. The indistinctness of lectin affinities of EC-SOD A, B, and C, as showed in Fig. 4, suggested that heterogeneity in heparin affinity was not dependent on its carbohydrate moiety. No effect of the glycopeptidase F treatment of rEC-SOD C on its heparin affinity supports the above suggestion. Another possible heparin-binding site is a cluster of positively charged amino acid residues in the enzyme. Such a cluster occurs at the very hydrophilic C-terminal end of EC-SOD C, which contains three lysine and six arginine residues among the last 21 amino acids (7). We
have shown before that the subtle modification of only a few lysine residues by trinitrobenzene sulfonic acid (TNBS) caused a loss of heparin affinity (10). However, we could not discern which of the five lysines in the subunit was modified early. It was speculated that three lysines localized at the C-terminal end should be easily accessible to the reagents and were likely candidates for early modification (10). The environments of the others, Lys-23 and Lys-74, appear to be less distinctly exteriorized according to computer prediction (S. L. Marklund, unpublished data). In this report, we can suggest directly that lysine and arginine residues at the C-terminal end contribute to heparin binding. r-EC-SOD C digested with trypsin or endoproteinase Lys-C, which lost three lysine residues (Lys-211, Lys-212, and Lys-220) or one lysine residue (Lys-220) at the C-terminal end, had no or weak heparin affinity, respectively. This result is consistent with the previous data that TNBS modification of one or three lysines in subunits resulted in a loss of roughly half or most of its affinity for heparin-Sepharose (10). Treatment with endoproteinase Lys-C also resulted in loss of triple arginine residues, Arg-213, Arg-214, and Arg-215. In studies with an arginine-specific reagent, phenylglyoxal, it also appeared that arginine residues were in-
THE
HEPARIN
BINDING
SITE
OF HUMAN
volved in the binding to heparin (10). The two lysine residues, Lys-211 and Lys 212, are enclosed with four arginine residues, Arg-210-Arg-215, constitute a positively charged cluster, and may draw negatively charged residues in heparin. The in viva correlate of the heparin affinity is apparently binding to the heparan sulfate proteoglycan in the glycocalyx of the cell surface. Whereas subfractions A and B mainly reside in the plasma phase, the C type apparently forms an equilibrium between the plasma phase and the heparan sulfate proteoglycan of the glycocalyx of endothelial cell surfaces (20-22). EC-SOD is speculated to be primarily synthesized as type C in the body, because all of the investigated human cell lines express and secrete EC-SOD of the C type (23), and human tissues contain almost exclusively EC-SOD of the C type (S. L. Marklund, unpublished data). A-type and B-type EC-SODS are consequently probably the result of secondary modifications. Previous studies in our laboratory (11) indicated that nonenzymatic glycation was one such modification, especially contributing to formation of EC-SOD B. Another likely contributing factor is proteolytic cleavage of the exteriorized lysine- and arginine-rich C-terminal end. The fact that serum EC-SOD A was less glycated in uiuo than EC-SOD B may be due to more extensive proteolytic cleavage in this fraction (11). The results shown in Fig. 6 also support this possibility. Phagocytic leukocytes contain large numbers and amounts of proteinases. These cells release not only proteinases but also active oxygen in the inflammatory response. It is possible that the ECSOD in the vascular system suffers proteolytic cleavage under these conditions (24). The loss of heparin affinity by proteolytic cleavage of C-terminal peptides indicated that in uiuo the binding of EC-SOD to heparan sulfate in the glycocalyx of endothelial cell surfaces should be weakened under pathological conditions. This phenomenon may result in diminished ability to protect the endothelium from the active oxygen released from adhered leukocytes, perpetuates the local inflammatory process, and destroys the surrounding tissues.
EXTRACELLULAR-SUPEROXIDE
161
DISMUTASE
REFERENCES 1. Marklund,
S. L. (1982) Proc. Natl. Acad. Sci. USA 79, 7634-7638.
2. Tibell, L., Hjalmarsson, A., and Marklund,
K., Edlund, T., Skogman, G., Engstrom, S. L. (1987) Proc. Natl. Acad. Sci. USA 84,
6634-6638. 3. Marklund, S. L., Holme, E., and Hellner, L., (1982) Clin. Chim. Acta 126,41-51. 4. Marklund, S. L., Bjelle, A., and Elmqvist, L. G. (1986) Ann. Rheum. Dis.45,847-851.
5. Marklund, S. L. (1984) J. Clin. Invest. 74, 1398-1403. 6. Marklund, S. L. (1984) Biochem. J. 222,649-655. 7. Hjalmarsson, K., Marklund, S. L., Engstrom, A, and Edlund, T. (1987) Proc. Natl. Acad. Sci. USA 84, 6340-6344.
8. Karlsson, K., and Marklund, S. L. (1987) Biochem. J. 242,55-59. 9. Karlsson, K., and Marklund, S. L. (1988) Biochem. J. 255, 223228. 10. Adachi, T., and Marklund,
S. L. (1989) J. Biol. Chem. 264, 8537-
8541. 11. Adachi, T., Ohta, H., Hirano, K., Hayashi, K., and Marklund,
S. L.
(1991) Biochem. J. 279, 263-267. 12. Cummings, R. D., andKornfeld, S. (1982) J. Biol. Chem. 257,1123511240. 13. Hermodson, M. A., Ericsson, L. H., Neurath, H., and Walsh, K. A. (1973) Biochemistry 12, 3146-3153.
14. Uchida, Y., Tsukada, Y., and Sugimori, T. (1979) J. Biochem. (Tokyo) 86,1573-1585. 15. Karlsson, K., and Marklund, S. L. (1989) Lab. Znuest. 60,659-666. 16. Stromqvist, M., Holgersson, J., and Samuelsson, B. (1991) J. Chromatgr. 548, 293-301.
17. Sasaki, H., Bothner, Chem. 262,
B., Dell, A., and Fukuda, 12059-12076.
M. (1987) J. Biol.
18. Takeuchi,
M., Takasaki, S., Miyazaki, H., Kato, T., Hoshi, S., Kochibe, N., and Kobata, A. (1988) J. Biol. Chem. 263,3657-3663. 19. Osawa, T., and Tsuji, T. (1987) Annu. Reu. Biochem. 56, 21-42.
20. Karlsson, K., Lindahl, U., and Marklund, S. L. (1988) Biochem. J. 256,29-33. 21. Karlsson, K., and Marklund, S. L. (1988) J. Clin. Znuest. 82, 762766. 22. Karlsson, K., and Marklund, S. L. (1988) Biochim. Biophys. Acta 967,110-114. 23. Marklund, S. L. (1990) Biochem. J. 266, 213-219. 24. Reilly, P. M., Schiller, H. J., and Bulkley, G. B. (1991) Am. J. Surgery 161,488-503.
OF BIOCHEMISTRY
AND
BIOPHYSICS
Vol. 297, No. 1, August 15, pp. 155-161, 1992
The Heparin Binding Site of Human ExtracellularSuperoxide Dismutase’ Tetsuo Adachi,” Tsutomu Kodera, Hideki Ohta, Kyozo Hayashi, and Kazuyuki Hirano Department
of Pharmaceutics,
Received February
Gifu Pharmaceutical
University,
11, 1992, and in revised form May 1, 1992
Extracellular-superoxide dismutase (EC-SOD) is a secretory glycoprotein that is major SOD isozyme in extracellular fluids. We revealed the possible structure of the carbohydrate chain of serum EC-SOD with the serial lectin affinity technique. The structure is a biantennary complex type with an internal fucose residue attached to asparagine-linked N-acetyl-o-glucosamine and with terminal sialic acid linked to iV-acetyllactosamine. ECSOD in plasma is heterogeneous with regard to heparin affinity and can be divided into three fractions: A, without affinity; B, with intermediate affinity; and C, with high affinity. It appeared that this heterogeneity is not dependent on the carbohydrate structure upon comparison of EC-SOD A, B, and C. No effect of the glycopeptidase F treatment of EC-SOD C on its heparin affinity supported the above results. A previous report showed that both lysine and arginine residues, probably at the C-terminal end, contribute to heparin binding. Recombinant EC-SOD C treated with trypsin or endoproteinase Lys C, which lost three lysine residues (Lys-2 11, Lys-2 12, and Lys-220) or one lysine residue (Lys-220) at the C-terminal end, had no or weak affinity for the heparin HPLC column, respectively. The proteinase-treated r-EC-SOD C also lost triple arginine residues which are adjacent to double lysine residues. These results suggest that the heparin-binding site may occur on a “cluster” of basic amino acids at the C-terminal end of EC-SOD C. EC-SOD is speculated to be primarily synthesized as type C, and types A and B are probably the result of secondary modifications. It appeared that the proteolytic cleavage of the exteriorized lysine- and arginine-rich C-terminal end in vivo is a more important contributory factor to the formation of ECSOD B and/or EC-SOD A. o 1992 Academic PWS. IN.
r This work was supported in part by Grants-in-Aid for the Encouragement of Young Scientists from the Ministry of Education, Science, and Culture of Japan (02771711) to T.A. * To whom correspondence should be addressed. 0003.9861/92 $5.00 Copyright All rights
Gifu 502, Japan
0 1992 by Academic Press, of reproduction in any form
Extracellular-superoxide dismutase (EC-SOD)3 (EC 1.15.1.1) is a secretory, tetrameric copper- and zinc-containing glycoprotein with a subunit molecular weight of about 30 kDa (1,2). EC-SOD is the major SOD isozyme in extracellular fluids, such as plasma, lymph (3), and synovial fluids (4), whereas this isozyme is the least predominant in tissues compared to copper,zinc-SOD (Cu,Zn-SOD) and manganese-SOD (Mn-SOD) (5, 6). The specific activity (l), the ligands to Cu and Zn, the cysteins forming the intrasubunit disulfide bridge, and the arginine residue in the entrance to the active site of EC-SOD (7) can be identical to those of Cu,Zn-SOD. The conformation of EC-SOD around the active site is therefore expected to be very similar to that of Cu,Zn-SOD. A prominent feature of EC-SOD is its affinity for heparin. EC-SOD in plasma in man (8) and other mammalian species (9) is heterogeneous with regard to affinity for heparin-Sepharose and can be divided into three fractions: A, without affinity; B, with weak affinity; and C, with relatively strong affinity. The background to the heterogeneity in heparin affinity is still unresolved. Studies with amino acid-specific reagents indicated that both lysine and arginine residues seemed to contribute to heparin binding (10). Compared to other SODS, another special feature of EC-SOD is that it is a glycoprotein. A possible N-glycosylation site was speculated to be Asn-89 (7). It is very important to know that the carbohydrate moiety of recombinant EC-SOD C (r-EC-SOD C) is indistinguishable 3 Abbreviations used: EC-SOD, extracellular-superoxide dismutase; Cu,Zn-SOD, copper,zinc-superoxide dismutase; Mn-SOD, manganesesuperoxide dismutase; r-, recombinant; n-, native; Con A, concanavalin A; LCA, lentil lectin; RCA120, Ricinus communis agglutinin 120; WGA, wheat germ agglutinin; PHA-E4, phytohemagglutinin E4; PNA, peanut lectin; a-MM, cu-methyl-D-mannoside; GlcNAc, N-acetyl-D-ghrcosamine; Gal, galactose; Galpl-4GlcNAc, N-acetyllactosamine; TNBS, trinitrobenzene sulfonic acid; PBS, phosphate-buffered saline; BSA, bovine serum albumin; TPCK, L-l-p-tosylamino-2-phenylethyl chloromethyl ketone; TLCK, N-p-tosyl-L-lysinechloromethyl ketone; IgG, immunoglobulin G. 155
Inc. reserved.
156
ADACHI
from that of native EC-SOD for r-EC-SOD C is valuable for biological as well as clinical use. However, the structure of the carbohydrate chain and its contribution to the heterogeneity of plasma EC-SOD are still unresolved. In the present investigation, the effect of the proteolytic cleavage of the C-terminal region of r-EC-SOD C on affinity for heparin-Sepharose is explored. Moreover, the carbohydrate chains of serum EC-SOD A, B, and C were compared with each other and with that of r-EC-SOD C by using lectin affinity techniques. EXPERIMENTAL
PROCEDURES
Materials. Human r-EC-SOD C prepared as described previously (2) was kindly provided by SYMBICOM AB, UmeH, Sweden. Endoproteinase Lys-C (lysylendoproteinase) was purchased from BoehringerMannheim GmbH, Mannheim, Germany. TPCK-trypsin and TLCKchymotrypsin were purchased from Sigma, St. Louis, Missouri, U.S.A. Neuraminidase (sialidase) was purchased from Nacalai Tesque, Kyoto, Japan. Glycopeptidase F (N-glycanase) was purchased from Takara Shuzo, Kyoto, Japan. Alkaline phosphatase-conjugated rabbit anti[mouse IgG (H + L)] antibody was purchased from Zymed Laboratories, San Francisco, California, U.S.A. Peroxidase-conjugated goat anti[rabbit IgG (H f L)] antibody was purchased from Organon Teknika Co., Durham, North Carolina, USA. Heparin-Sepharose, concanavalin A (Con A)-Sepharose, and CNBr-activated Sepharose 4B were products of Pharmacia LKB Biotechnology, Uppsala, Sweden. Con A-, lentil lectin (LCA)-, Ricinus communis agglutinin 120 (RCAlPO)-, wheat germ agglutinin (WGA)-, phytohemagglutinin E4 (PHA-E4)-, and peanut lectin (PNA)-agarose columns were purchased from Seikagaku Co., Tokyo, Japan. Immunoplate was purchased from Nunc, Roskilde, Denmark. EZJSA of EC-SOD concentration. ELISA of EC-SOD proceeded by the method described previously (11). S-Carboxymethylated r-EC-SOD was assayed by the method as follows, because the above method could not detect S-carboxymethylated r-EC-SOD. Portions (70 ~1) of S-carboxymethylated r-EC-SOD (standard) or sample diluted with 50 mM sodium carbonate, pH 9.5, containing 0.02% NaN, were added into each well of an immunoplate and left to stand overnight at 4°C. Each well was washed with phosphate-buffered saline (PBS, 10 mM sodium phosphate, pH 7.4, containing 0.15 M NaCl) containing 0.05% Tween 20 and 0.05% merthiolate (washing buffer). The remaining protein-binding sites were blocked with 300 ~1 of PBS containing 1% bovine serum albumin (BSA), 0.05% Tween 20, and 0.05% merthiolate (blocking buffer). The plate was then left to stand overnight at 4°C and washed three times with the washing buffer. Then 80 al of 9.1 pg/ml rabbit anti-EC-SOD antibody diluted with blocking buffer was added to each well, and the plate was incubated for 3 h at room temperature and washed with the washing buffer. A 90+1 portion of peroxidase-conjugated goat anti-[rabbit IgG (H + L)] antibody at a 1:lOOO dilution in blocking buffer was added to each well and the plate was incubated again for 3 h at room temperature, followed by washing three times with washing buffer. Substrate solution (150 ~1 of 0.1 M McIlvain buffer, pH 6.5, containing 16.5 mM o-phenylenediamine and 0.02% H,O,) was then added to each well and the plate was incubated for 30 min at room temperature. The enzyme reaction was stopped by the addition of 50 ~1 of 4.5 N H$O,, and the absorbance at 492 nm was measured with a Corona MTP-120 microplate reader. Heparin affinity chromatography. Heparin affinity chromatographies were performed on a heparin-Sepharose column (volume, 2 ml) or a HPLC column of TSK-gel heparin 5PW (Tosoh, Tokyo, Japan). The column chromatography was operated at a flow rate of 0.7 ml/min. A gradient system formed between buffer A (25 mM sodium phosphate, pH 6.5) and buffer B (25 mM sodium phosphate, pH 6.5, containing 1 M NaCl) was used.
ET AL. Purification of EC-SOD A, B, and C with anti-(r-EC-SOD C) monoC) monoclonal antibody-conjugated Sepharose 4B. Anti-(r-EC-SOD clonal antibody-coupled gel as previously prepared (11) was used. The samples (12 ml of EC-SOD A, 24 ml of EC-SOD B, and 20 ml of ECSOD C fractions in serum separated by heparin-Sepharose column chromatography) were applied to this column by the method previously described (11). Samples (150 pl of serum, Lectin ufinity column chromatography. 100 ~1 of 1 rg/ml r-EC-SOD C, or 100 ~1 of 300-500 rig/ml of EC-SOD A, B, and C fractions) were applied to Con A (volume, 1 ml)-, LCA (volume, 1 ml)-, RCA120 (volume, 1 ml)-, WGA (volume, 1 ml)-, and PNA (volume, 0.5 ml)-agarose columns equilibrated with PBS and washed with the same buffer. The bound fractions were then eluted with above buffer containing 0.25 M a-methyl-D-mannoside (a-MM) (for Con A and LCA columns), 0.25 M galactose (Gal) (for RCA120 and PNA columns), and 0.25 M N-acetyl-D-glucosamine (GlcNAc) (for the WGA column), respectively. Serial &tin affinity technique. Structure studies of the carbohydrate chain were performed with the method established by Cummings and Kornfeld (12). Sample (300 ~1) was applied to a Con A-Sepharose column (volume, 2 ml) equilibrated with PBS containing 10 pg/ml BSA and washed with the same buffer. The bound fraction was then eluted by stepwise addition of two different concentrations (0.01 and 0.5 M) of aMM. Three fractions, an unbound fraction (Fraction I), a weakly bound fraction (Fraction II), and a strongly bound fraction (Fraction III), were dialyzed against PBS and followed by concentration with YM-10. These fraction were further applied to PHA-E4 (volume, 0.5 ml), LCA (volume, 1 ml), and WGA (volume, 1 ml) columns, respectively, equilibrated with above buffer. The unbound and the bound fractions on the respective columns were separated by using 0.1 M GlcNAc, 0.2 M a-MM, and 0.1 M GlcNAc, respectively, as elution buffers. S-Curboxymethylution of r-EC-SOD C. Thirty microliters of 15 mg/ ml r-EC-SOD C treated with or without proteinases was reduced with 100 ~1 of 0.2 M dithiothreitol dissolved in 0.5 M Tris-HCl, pH 8.5, containing 7 M guanidine-HCl and 10 mM EDTA, flushed with nitrogen, and incubated for 2 h at room temperature. After the addition of I2 ~1 of 2.7 M iodoacetic acid, the reaction mixture was incubated in the dark for 3 h. Hydroxylnmine cleavage of reduced and S-curboxymethykzted r-ECSOD C. S-Carboxymethylated r-EC-SOD C’s prepared as above (450 rg) were dialyzed against 15 mM sodium acetate, pH 6.0, at 4°C overnight, followed by rotary evaporation to dryness. The residues were dissolved with 200 ~1 of 2 M hydroxylamine-HCl, 0.2 M K&O,, and 6 M guanidineHCl (pH 10.5). After incubation at 45’C for 4 h, the mixture were acidified by the addition of 30 ~1 of formic acid (13). The Asn-Gly peptide bond (Asn-180-Gly-181 in EC-SOD) was cleaved by this treatment. The products were separated by HPLC on a Cl8 reverse-phase column (PBondasphere, 5 pm, Nihon Waters Ltd., Tokyo, Japan). The column chromatography was operated at a flow rate of 0.7 ml/min. A gradient system formed between solvent A (0.1% trifluoroacetic acid in Milli-Q . water) and solvent B (0.1% trifluoroacetic acid in acetonitrile) was used. Peptides were eluted with a linear gradient up to 60% of solvent B for 60 min and continuously up to 100% of solvent B for 10 min. The peptides were detected on the basis of the absorbance at 210 nm. C-terminal peptides (Gly-181, C-terminal, retention time 33-34 min; amino acid sequences of peptides were determined; data not shown) were isolated and followed by rotary evaporation to dryness. The residues of a CChymotryptic digestion of C-terminal peptides. terminal fragment isolated as described above were dissolved with 100 ~1 of 10 mM Tris-HCl, pH 8.6, followed by the addition of 10 pl of 10 rig/ml TLCK-chymotrypsin. After incubation at 25°C for 1 h, the products were separated with Cl8 reverse-phase HPLC under the conditions described above. Amino acid sequence determination. termined with an Applied Biosystems
Amino acid sequences were deprotein sequencer, Model 473A.
THE
HEPARIN
BINDING
SITE
OF HUMAN
EXTRACELLULAR-SUPEROXIDE
Sialidase Treatment
(4 CQQA
-0
10
20
30
5. w L!a
OLPHAD r;;40 k n 30 3 20 A cj w
i
r
"
I
$ I 2
10
n -0
10
20
"00
10
20
0
10
20
Fraction number (1 mlhl mlhube) Jbe)
FIG. 1.
Serial lectin affinity chromatographies of serum EC-SOD. Three hundred microliters of normal serum was applied to a Con ASepharose column (a). Three fractions, unbound fraction (Fraction I), a weakly bound fraction eluted with 0.01 M a-MM (Fraction II), and a strongly bound fraction eluted with 0.5 M u-MM (Fraction III) were dialyzed against PBS followed by concentration with a YM-10 membrane filter. These fractions were further applied to PHA-E4-agarose (b), LCAagarose (c), and WGA-agarose (d) columns, respectively.
RESULTS
Lectin Affinity
Chromatography
157
of Serum EC-SOD
For no affinity to the RCA120 column, the presence of sialic acid residues attached to N-acetyllactosamine (Galpl-4GlcNAc) in the N-linked carbohydrate chain is possible. To confirm this notion, the effect of sialidase treatment on serum EC-SOD was studied. Normal serum (150 ~1) was incubated with 50 mU of sialidase in the presence of 150 ~1 of 25 mM potassium phosphate buffer, pH 6.5, at 37°C for 24 h (14), followed by application on the RCA120 column. Sialidase-treated n-EC-SOD substantially bound on the RCA120 column, whereas nontreated enzyme failed to bind to this column (Fig. 3a). nEC-SOD in serum treated with or without sialidase under the same conditions as those in the above experiments did not bind to the PNA-agarose column which required O-liked mucin-type sugar chains for binding (Fig. 3b).
Fraction number (2 mlhube)
50
DISMUTASE
of EC-SOD in Serum
The affinity for lectins was investigated to clarify the carbohydrate chain of n-EC-SOD. Normal serum (150 ~1) was applied to four kinds of lectin columns and EC-SOD in each fractions was detected by ELISA. n-EC-SOD in serum was found to bind to Con A- and LCA-agarose columns, whereas it did not bind to RCAlZO- and WGAagarose columns (data not shown). With the serial lectin affinity technique, substantially all of n-EC-SOD (85%) was found in the weakly bound fraction (Fraction II) for the Con A-Sepharose column (Fig. 1). Little of n-ECSOD was found in the unbound fraction (Fraction I, 5%) and in the strongly bound fraction (Fraction III, 10%). In order to further evaluate the lectin-binding affinity, Fractions I, II, and III obtained on the Con A-Sepharose column were further applied on PHA-E4, LCA, and WGA columns, respectively. Fraction I, a small proportion of Con A column chromatography, did not bind to the PHAE4 column. It was observed that Fraction II bound substantially to the LCA column, whereas Fraction III did not bind to the WGA column (Fig. 1). r-EC-SOD C expressed in the CHO cell (2) also bound to Con A- and LCA-agarose columns, whereas it did not bind to RCA120- and WGA-agarose columns. r-EC-SOD showed similar behavior with the serial lectin affinity technique (Fig. 2).
Lectin Column Chromatography
of EC-SOD A, B, and C
Serum EC-SOD could be separated into three fractions: A, without affinity; B, with weak affinity; and C, with relatively strong affinity for heparin-Sepharose, as shown previously (9, 11, 15). Pooled A, B, and C fractions were then isolated on anti-(r-EC-SOD C) monoclonal antibodyconjugated Sepharose 4B columns as described under Experimental Procedures. Subsequently the isolated fractions were chromatographed with Con A-, LCA-, RCAlBO-, and WGA-agarose columns. All three fractions showed similar chromatographs for lectins, bound to Con A and LCA columns, whereas they did not bind to RCA120
(4 Con
Fraction number (2 ml/tube) (b) PI-IA-E4
w !&A
(4
WGA
Fraction number (1 mlhube)
FIG. 2.
Serial lectin affinity chromatographies of r-EC-SOD C. Three hundred microliters of r-EC-SOD (900 rig/ml) was applied to a Con ASepharose column and chromatographed as described in the legend to Fig. 1.
158
ADACHI
(a) RCA1 20-aaarose
100
100 (ii)
(0 80
80
-2 @ n 60
60 z
8 Y
I &I
40
kt
6
20
$ I
40
d
20
0 l!!LLl_a. 0 2 4
0 6810
0
2
4
6
8
10
Fraction number (2 ml/tube)
(W PNA-aoarose (ii)
6)
8 W
20 ~1 of 10 mM Tris-HCl buffer, pH 8.6, and incubated at 30°C for 12 h. Portions (1 ~1) of r-EC-SOD treated with proteinases were mixed with 20 ~1 of 25 mM sodium phosphate buffer, pH 6.5, followed by heparin affinity chromatographies with the TSK-gel heparin 5PW HPLC column as described under Experimental Procedures. The enzyme treated with endoproteinase Lys-C had only very weak affinity for heparin and eluted at 0.05 M NaCI, whereas nontreated r-EC-SOD had strong affinity for heparin and eluted mainly at 0.6 M NaCl. Trypsin-treated r-EC-SOD had lost most of its heparin affinity (72% of the enzyme was observed in the unbound fraction and the remainder was eluted at 0.05 M NaCl) as shown in Fig. 6.
100
100
';i @ 8
ET AL.
80
80
60
60
40
40
85 t4
20
' o0
2
4
A 6610
o-
024
of C-Terminal Amino Acid Sequences of C’s Treated with Proteinases
C-terminal peptides (Gly-181-C-terminal) of r-ECSOD C’s treated with or without endoproteinase Lys-C or TPCK-trypsin were isolated and then digested with TLCK-chymotrypsin as described under Experimental
z (I) 5 t? d
20
d
Identification r-EC-SOD
6
8
10
Fraction number (2 ml/tube) FIG. 3. RCAlBO-agarose and PNA-agarose column chromatographies of desialylated n-EC-SOD. Normal serum (150 ~1) was incubated with (ii) or without (i) 50 mU sialidase in the presence of 150 rl of 25 mM potassium phosphate buffer, pH 6.5, at 37°C for 24 h, followed by applications on an RCA120-agarose column (volume, 1 ml) (a) and a PNAagarose column (volume, 0.5 ml) (b), and chromatographed as described under Experimental Procedures. The EC-SOD concentration in each fraction was determined by ELISA.
(a) EC-SOD A loo [ Con A
-0
2
4
0.N EC-SOD
and WGA columns as shown in Fig. 4. These behaviors were also the same as that of nonseparated serum ECSOD as described above. Heparin-Sepharose Column Chromatography of r-ECSOD C Treated with Glycopeptidase F It was suggested that the heterogeneity in heparin affinity of serum EC-SOD was not dependent on the carbohydrate chain (Fig. 4). To confirm this notion, the affinity of deglycosylated r-EC-SOD C for heparinSepharose was investigated. Portions (400 ~1) of S-carboxymethylated r-EC-SOD C (40 fig/ml) which was dialyzed against 10 mM Tris-HCl buffer, pH 8.6, were incubated with or without 5 mU glycopeptidase F at 37°C for 1 h. Complete deglycosylation of glycopeptidase Ftreated r-EC-SOD C was certified by the loss of its affinity for Con A-Sepharose. Deglycosylated r-EC-SOD C kept the high affinity for heparin-Sepharose as did nontreated r-EC-SOD C (Fig. 5). Heparin HPLC of r-EC-SOD C Treated with Proteinases Ten microliters of r-EC-SOD C (45 mg/ml) was mixed with 1 pg of endoproteinase Lys-C or TPCK-trypsin in
6
c LCA
610
0
2
4
r RCA120
6
610
0
4
6
610
0
WGA
2
4
6
610
B
loo [ Con A
F LCA
g!J!JpJk 0
2
r
2
4
6
F RCA120
WGA
1
610
0
2
4
6
6100
2
4
6
610
0
2
4
6
610
610
0
2
4
6
6100
2
4
6
610
0
2
4
6
610
(c) EC-SOD C
0
2
4
6
Fraction number (2 ml/tube)
FIG. 4. Lectin affinity chromatographies of EC-SOD A, B, and C obtained from normal serum. EC-SOD A (a), B (b), and C (c) in normal serum were separated with heparin-Sepharose column (volume, 2 ml) chromatography and then purified with an anti-r-EC-SOD C monoclonal antibody-conjugated Sepharose 4B column (volume, 2 ml) as described under Experimental Procedures. Isolated fractions were applied on each Con A-, LCA-, RCAlZO-, and WGA-agarose column (volume, 1 ml).
THE
HEPARIN
BINDING
SITE
OF HUMAN
(4 Con A-SeDharose 60 50 40 30 20 10 -0
2
4
6
8
" -0
10
2
4
6
6
10
Fraction number (1 ml/tube)
30
0) 1.0
f"" 7 e
20
/
0.6
0.6 Q g
i
10
0
M 0
I
5
0.4 0.2
/
0.0 10
15
20
25
0 0
5
10
15
20
25
Fraction number (1 ml/tube) FIG. 5. Con A-Sepharose and heparin-Sepharose column chromatographies of deglycosylated r-EC-SOD C. Four hundred microliters of reduced and S-carboxymethylated r-EC-SOD (40 pg/ml) was incubated with (ii) or without (i) 5 mU of glycopeptidase F at 37°C for 1 h. Aliquots (100 ~1) were applied to a Con A-Sepharose column (volume, 2 ml) (a) and a heparin-Sepharose column (volume, 2 ml) (b), and chromatographed as described under Experimental Procedures. The EC-SOD concentration in each fraction was determined by ELISA (0). NaCl concentration in buffer (. . . ).
Procedures. Figure 7 shows the Cl8 reverse-phase HPLC chromatographs. Peptides A-l and A-2 in nontreated rEC-SOD C; B-l and B-2 in endoproteinase Lys-C-treated r-EC-SOD C; and C-l, C-2, and C-3 in TPCK-trypsintreated r-EC-SOD C were isolated, followed by the determination of amino acid sequences. Amino acid sequence of the B-l peptide showed that r-EC-SOD C treated with endoproteinase Lys-C had lost the Arg-213Ala-222 peptide which contained one lysyl residue (Lys220) and three arginine residues (Arg-213, -214, and 215), compared to nontreated enzyme. It appeared that the TPCK-trypsin-treated enzyme had lost another two lysyl residues, Lys-211 and Lys-212, by the sequence analysis of C-l and C-2 peptides in Fig. 7. DISCUSSION
In the present study, a possible structure of the carbohydrate chain of EC-SOD is revealed by the serial lectin
EXTRACELLULAR-SUPEROXIDE
159
DISMUTASE
affinity technique. The core structure is shown to be a biantennary complex type without bisecting GlcNAc. The high-affinity binding to LCA-agarose shows the presence of an internal fucose residue attached to asparagine-linked GlcNAc. The appearance of affinity to RCA120-agarose after desialylation shows the presence of terminal sialic acid linked to Galpl-4GlcNAc. The structure of the carbohydrate chain of r-EC-SOD C produced in the CHO cell is indistinguishable from that of n-EC-SOD judging from lectin affinity column chromatographies. Our proposed carbohydrate structure of EC-SOD is consistent with that of r-EC-SOD, a complex biantennary type having a core fucose, which recently appeared (16). The carbohydrate structure of r-erythropoietin was also reported to be common to that of urinary n-erythropoietin (17, 18). However, the contribution of the carbohydrate chain of EC-SOD to biochemical and therapeutic properties such as enzymic activity, turnover rate in the body, and antigenicity is unclear; we could provide evidence that rEC-SOD C is valuable for experiments in biological and clinical research. Based on the amino acid sequence deduced from the cDNA sequence, EC-SOD appeared to have only one possible N-glycosylation site, Asn-89, and no 0-glycosylation site (7). This notion was confirmed by investigations using r-EC-SOD (16). An immobilized lectin, PNA, is known to preferentially bind to a typical
‘:.,o.o
;;;
0
10
Fraction
20
30
40
number
(0.7 ml/tube)
50
FIG. 6. Heparin HPLC of r-EC-SOD C treated with proteinases. rEC-SOD C (450 pg) was treated without (a) or with 1 pg of endoproteinase Lys-C (b) or TPCK-trypsin (c) in 20 ~1 of 10 mM Tris-HCl, pH 8.6, at 30°C for 12 h. Aliquots (1 hl) were mixed with 20 ~1 of 25 mM sodium phosphate, pH 6.5, and loaded on a column of TSK-gel heparin 5PW. The column was washed with the above buffer, and then bound components were eluted with a linear gradient of NaCl in the buffer in each fraction was assayed by (* * * ). The EC-SOD concentration ELISA (0).
160
ADACHI
I C-l I IJ
ET AL.
(c)
(b)
c-3
3-2
B-l
-L I
c-2
1 lJ
1'0
210
3b
I 40
I 5c
Retention time (min) Hydroxylamine I
180
Chymotrypsin 1
190
200
210
NGNAGRRLACCWGVCGPGLWERQAREESECK (a) r-EC-SOD
e
A-2
(b) Endoprotelnrse Lys-Ctreated r-ECSOD
jM
B _2
(c) TPCK-tryprintreated r-ECSOD
F *
c-3
220222
I :.
, II
A-l
h ’
-B-l -c-1 c-2
:
-’ _I e
FIG. 7. C-terminal amino acid sequences of the hydroxylamine-cleaved peptides of r-EC-SOD C treated with proteinases. C-terminal peptides (Gly-181~C-terminal) of r-EC-SOD C treated without (a) or with endoproteinase Lys-C (b) or TPCK-trypsin (c) were isolated and then digested with TLCK-chymotrypsin as described under Experimental Procedures. TLCK-chymotrypsin-digested peptides were separated again with Cl8 reverse-phase HPLC chromatographies. Amino acid sequences of isolated peptides were determined with an Applied Biosystems protein sequencer, Model 473A.
mucin-type glycochain, GalPl + 3GalNAccul + Ser/Thr, but not to this sialylated glycopeptide (19). We now show that desialylated n-EC-SOD in serum also has no affinity to the PNA column (Fig. 3), suggesting the absence of an O-linked mucin-type sugar chain in n-EC-SOD. A prominent feature of EC-SOD is its affinity for heparin. EC-SOD in plasma is heterogeneous with regard to affinity for heparin-Sepharose (8,9). The background to the heterogeneity in heparin affinity is still unresolved. The binding of EC-SOD to heparin is of an electrostatic nature (20). It is known that sterical hindrance due to carbohydrate chains and/or their electric charge contributes to some of the biochemical properties of proteins. The indistinctness of lectin affinities of EC-SOD A, B, and C, as showed in Fig. 4, suggested that heterogeneity in heparin affinity was not dependent on its carbohydrate moiety. No effect of the glycopeptidase F treatment of rEC-SOD C on its heparin affinity supports the above suggestion. Another possible heparin-binding site is a cluster of positively charged amino acid residues in the enzyme. Such a cluster occurs at the very hydrophilic C-terminal end of EC-SOD C, which contains three lysine and six arginine residues among the last 21 amino acids (7). We
have shown before that the subtle modification of only a few lysine residues by trinitrobenzene sulfonic acid (TNBS) caused a loss of heparin affinity (10). However, we could not discern which of the five lysines in the subunit was modified early. It was speculated that three lysines localized at the C-terminal end should be easily accessible to the reagents and were likely candidates for early modification (10). The environments of the others, Lys-23 and Lys-74, appear to be less distinctly exteriorized according to computer prediction (S. L. Marklund, unpublished data). In this report, we can suggest directly that lysine and arginine residues at the C-terminal end contribute to heparin binding. r-EC-SOD C digested with trypsin or endoproteinase Lys-C, which lost three lysine residues (Lys-211, Lys-212, and Lys-220) or one lysine residue (Lys-220) at the C-terminal end, had no or weak heparin affinity, respectively. This result is consistent with the previous data that TNBS modification of one or three lysines in subunits resulted in a loss of roughly half or most of its affinity for heparin-Sepharose (10). Treatment with endoproteinase Lys-C also resulted in loss of triple arginine residues, Arg-213, Arg-214, and Arg-215. In studies with an arginine-specific reagent, phenylglyoxal, it also appeared that arginine residues were in-
THE
HEPARIN
BINDING
SITE
OF HUMAN
volved in the binding to heparin (10). The two lysine residues, Lys-211 and Lys 212, are enclosed with four arginine residues, Arg-210-Arg-215, constitute a positively charged cluster, and may draw negatively charged residues in heparin. The in viva correlate of the heparin affinity is apparently binding to the heparan sulfate proteoglycan in the glycocalyx of the cell surface. Whereas subfractions A and B mainly reside in the plasma phase, the C type apparently forms an equilibrium between the plasma phase and the heparan sulfate proteoglycan of the glycocalyx of endothelial cell surfaces (20-22). EC-SOD is speculated to be primarily synthesized as type C in the body, because all of the investigated human cell lines express and secrete EC-SOD of the C type (23), and human tissues contain almost exclusively EC-SOD of the C type (S. L. Marklund, unpublished data). A-type and B-type EC-SODS are consequently probably the result of secondary modifications. Previous studies in our laboratory (11) indicated that nonenzymatic glycation was one such modification, especially contributing to formation of EC-SOD B. Another likely contributing factor is proteolytic cleavage of the exteriorized lysine- and arginine-rich C-terminal end. The fact that serum EC-SOD A was less glycated in uiuo than EC-SOD B may be due to more extensive proteolytic cleavage in this fraction (11). The results shown in Fig. 6 also support this possibility. Phagocytic leukocytes contain large numbers and amounts of proteinases. These cells release not only proteinases but also active oxygen in the inflammatory response. It is possible that the ECSOD in the vascular system suffers proteolytic cleavage under these conditions (24). The loss of heparin affinity by proteolytic cleavage of C-terminal peptides indicated that in uiuo the binding of EC-SOD to heparan sulfate in the glycocalyx of endothelial cell surfaces should be weakened under pathological conditions. This phenomenon may result in diminished ability to protect the endothelium from the active oxygen released from adhered leukocytes, perpetuates the local inflammatory process, and destroys the surrounding tissues.
EXTRACELLULAR-SUPEROXIDE
161
DISMUTASE
REFERENCES 1. Marklund,
S. L. (1982) Proc. Natl. Acad. Sci. USA 79, 7634-7638.
2. Tibell, L., Hjalmarsson, A., and Marklund,
K., Edlund, T., Skogman, G., Engstrom, S. L. (1987) Proc. Natl. Acad. Sci. USA 84,
6634-6638. 3. Marklund, S. L., Holme, E., and Hellner, L., (1982) Clin. Chim. Acta 126,41-51. 4. Marklund, S. L., Bjelle, A., and Elmqvist, L. G. (1986) Ann. Rheum. Dis.45,847-851.
5. Marklund, S. L. (1984) J. Clin. Invest. 74, 1398-1403. 6. Marklund, S. L. (1984) Biochem. J. 222,649-655. 7. Hjalmarsson, K., Marklund, S. L., Engstrom, A, and Edlund, T. (1987) Proc. Natl. Acad. Sci. USA 84, 6340-6344.
8. Karlsson, K., and Marklund, S. L. (1987) Biochem. J. 242,55-59. 9. Karlsson, K., and Marklund, S. L. (1988) Biochem. J. 255, 223228. 10. Adachi, T., and Marklund,
S. L. (1989) J. Biol. Chem. 264, 8537-
8541. 11. Adachi, T., Ohta, H., Hirano, K., Hayashi, K., and Marklund,
S. L.
(1991) Biochem. J. 279, 263-267. 12. Cummings, R. D., andKornfeld, S. (1982) J. Biol. Chem. 257,1123511240. 13. Hermodson, M. A., Ericsson, L. H., Neurath, H., and Walsh, K. A. (1973) Biochemistry 12, 3146-3153.
14. Uchida, Y., Tsukada, Y., and Sugimori, T. (1979) J. Biochem. (Tokyo) 86,1573-1585. 15. Karlsson, K., and Marklund, S. L. (1989) Lab. Znuest. 60,659-666. 16. Stromqvist, M., Holgersson, J., and Samuelsson, B. (1991) J. Chromatgr. 548, 293-301.
17. Sasaki, H., Bothner, Chem. 262,
B., Dell, A., and Fukuda, 12059-12076.
M. (1987) J. Biol.
18. Takeuchi,
M., Takasaki, S., Miyazaki, H., Kato, T., Hoshi, S., Kochibe, N., and Kobata, A. (1988) J. Biol. Chem. 263,3657-3663. 19. Osawa, T., and Tsuji, T. (1987) Annu. Reu. Biochem. 56, 21-42.
20. Karlsson, K., Lindahl, U., and Marklund, S. L. (1988) Biochem. J. 256,29-33. 21. Karlsson, K., and Marklund, S. L. (1988) J. Clin. Znuest. 82, 762766. 22. Karlsson, K., and Marklund, S. L. (1988) Biochim. Biophys. Acta 967,110-114. 23. Marklund, S. L. (1990) Biochem. J. 266, 213-219. 24. Reilly, P. M., Schiller, H. J., and Bulkley, G. B. (1991) Am. J. Surgery 161,488-503.
