Eur. J. Biochem. 194,713-720 (1990) FEBS 1990

e

Human liver manganese superoxide dismutase Purification and crystallization, subunit association and sulfhydryl reactivity

',

Yukihiko MATSUDA '%', Shigeki HIGASHIYAMA', Yoshiyuki KIJIMA Keiichiro SUZUKI ', Kiyoshi KAWAN04, Morio AKIYAMA5. Sumio KAWATA', Seiichiro TARUI *, Harold F. DEUTSCH ' and Naoyuki TANIGUCHI' Department of Biochemistry, Osaka University Medical School, Japan Second Department and First Department of Internal Medicine, Osaka University Medical School, Japan Pathology Section, Osaka Rosai Hospital, Japan Department of Physics, Sapporo Medical College, Japan (Received July 27, 1990) - EJB 90 0906

Manganese superoxide dismutase (Mn-SOD) has been purified with a high yield (320 mg) from human liver (2 kg) and crystallized. Low-angle laser light scattering of the enzyme has shown that native enzyme is a tetrametic form. Four of the eight cysteine residues in the tetramer reacted with 5,5'dithiobis(2-nitrobenzoic acid) or with iodoacetamide. The others were only reactive in protein heated with SDS or urea after reduction with dithiothreitol or 2-mercaptoethanol. The reactive sulfhydryl group was found to be located at Cys196 by amino acid sequence analysis of Nbs,-reactive peptides isolated by activated thiol-Sepharose covalent chromatography. Incubation of Mn-SOD in 1O h SDS for 2 or 3 days at 25 "C or 5 min at 100"C gave material showing two prominent components on polyacrylamide gel electrophoresis in the presence of 0.1% SDS. The major component had a molecular mass of 23 kDa; the other, 25 kDa. Reduction of the protein by dithiothreitol or 2-mercaptoethanol heated in SDS produced only the 25-kDa monomer species. Essentially, no thiol groups were detected in the 23-kDa form, in which two cysteine residues appear to have been oxidized to form an intrasubunit disulfide. This indicates that Cys196 has a reactive sulfhydryl and appears to be a likely candidate for a mixed disulfide formation in vivo. Superoxide dismutases (SOD) are found in all aerobic organisms but not in obligate anaerobes [1, 21. The enzyme catalyzes the dismutation of superoxide anion into molecular oxygen and hydrogen peroxide [3]. The physiological function of the enzyme appears to protect cells against reactive free radicals by scavenging superoxide anions produced under oxidative conditions [I -31. In mammalian tissue, three SOD isozymes are found: Cu, Zn-SOD is predominantly located in the cytosol[4 - 71, extracellular SOD exists in the extracellular fluids [S] and Mn-SOD appears to be localized mostly in the mitochondria1 matrix [9]. Cytosolic Cu,Zn-SOD is a well-characterized protein containing two identical subunits, each having an intrachain disulfide and two sulfhydryls [lo, 111. The complete amino acids sequence of human Mn-SOD has been determined by Barra et al. [12] and, except for minor differences, is in good agreement with the structure deduced from its cDNA [13, 141. All MnSOD appear to be composed of four identical subunits [6,15, 161, although a recent study of a recombinant form suggests a dimeric enzyme [17]. The gross structural properties of such

Correspondence to N. Taniguchi, Department of Biochemistry, Osaka University Medical School, 2-2 Yamadaoka, Suita, Japan 565 Abbreviutions. SOD, superoxide dismutase; Nbs2, 5,5'-dithiobis(2-nitrobenzoic acid); LALLS, low-angle laser light scattering; HPGC, high-performance gel chromatography. Enzymes. Superoxide dismutase (EC 1.15.1.1); xanthine oxidase (EC 1.2.3.22); carboxypeptidase A (EC 3.4.12.2); carboxypeptidase B (EC 3.4.12.3); carboxypeptidase Y (EC 3.4.16.1); pepsin (EC 3.4.23.1); phosphorylase (EC 2.4.1.1); carbonic anhydrase (EC 4.2.1.1); ribonuclcase (EC 3.1.27.5).

proteins have received little study to date, although preliminary crystallographic study was performed using recombinant Mn-SOD [18]. The present investigation details sulfhydryl reactivities and the subunit association of human Mn-SOD. EXPERIMENTAL PROCEDURES Chemicals

Nitrobenzene tetrazolium, 4-amidinophenylmethylsulfonyl fluoride, guanidinium chloride, SDS and iodoacetamide were purchased from Wako Chemicals, dithiothreitol, Nethylmaleimide 5,5'-dithiobis(2-nitrobenzoic acid) (Nbs2; Ellman's reagent), urea and xanthine from Nacalai Tesque, xanthine oxidase and carboxypeptidases A, B and Y from Boehringer Mannheim. Pepsin and oxidized glutathione was purchased from Sigma. DEAE-cellulose (DE-52) was obtained from Whatman Ltd, hydroxylapatite from Wako Chemicals; Sephadex G-25 (PD-10 column), Sephacryl S-300, Red Sepharose, Polybuffer Exchanger 94, activated thiolSepharose 4B and Polybuffer 96 from Pharmacia LKB. Enzyme activity assay

Enzyme assays were carried out according to Beauchamp and Fridovich [19]. Activity of Mn-SOD incubated with or without denaturing agent and/or reducing agent was assayed by a xanthinelxanthine oxidase/nitrobenzene tetrazolium system in sodium carbonate buffer at pH 10.2. Protein concentrations of samples were measured by the method of Bradford [20] using bovine serum albumin as the standard.

714 Metal and EPR spectrometer analysis

The purified enzyme was extensively dialysed against metal-free water (Milli Q"' quality) and subjected to the analyzer. In order to determine manganese contents of the protein, 10 pl of the enzyme (36 pg/ml) was subjected to a polarized Zeeman atomic absorption spectrophotometer (Hitachi, Z-8000). To identify Mn(I1) or Mn(II1) in the native form, 1 ml enzyme (27 mg/ml) was employed in the EPR spectrometer system (Varian, Type E-12). Crystallization of Mn-SOD Purified Mn-SOD (10 mg/ml in pH 8.2, 50 mM Tris/HCl buffer) was crystallized by dialysis against 2.8 M ammonium sulfate solution containing pH 8.2, ?ris/HCl buffer at 4°C. A preliminary crystallographic study On Mn-SoD be published elsewhere. Stability of Mn-SOD

To investigate the pH stability of Mn-SOD, the purified enzyme (0.24 mg/ml) was incubated for up to 3 days at 4°C in various buffers at pH 5.5-9.3. Buffers used were 0.1 M sodium acetate, pH 5.5, sodium phosphate, pH 7.2, Tris/HCl, pH 8.0, or monoethanolamine acetate, pH 9.3. Aliquots of the mixture at appropriate intervals were removed and subjected to enzyme assay. This experiment was carried out in duplicate. To clarify stability of activity and subunit association of Mn-SOD in denaturing solutions, the enzyme (0.24 mg/ml) was incubated at 25°C in pH 8.0, 0.1 M Tris/HCl and 1 mM EDTA in the presence of 1% SDS, 8 M urea or 6 M guanidinium chloride. The mixtures were adjusted to pH 8.0, if necessary, with 0.1 M HC1. At appropriate time intervals, aliquots were removed for enzyme assay and subjected to SDS/PAGE. SDS/ PAGE

Discontinuous polyacrylamide gel electrophoresis employing a SDS/Tris/glycine buffer was carried out according to Laemmli et al. [21]. An aliquot of the non-heated protein was mixed with an equal volume of the sample buffer containing 1% SDS and 50% glycerol, and left at room temperature for 20 min prior to electrophoresis. The heated samples with or without 5 mM 2-mercaptoethanol were cooled, mixed with the sample buffer then electrophoresed. The Pharmacia molecular mass calibration kit was employed. Low-angle laser light scattering (LALLS)

Molecular masses of the purified Mn-SOD and various fractions were also determined by a LALLS technique in combination with high-performance gel chromatography (HPGC/LALLS) [22]. The 0.05 M sodium phosphate/O.l M NaCl buffer, pH 7.5, employed was first passed through a 0.2-pm pore ultrafilter then through a degasser (ERC-3312, Erma Optical Works). The buffer solution was supplied to the 7.8 mm x 300 mm TSK-GEL G3000 SW-XL column (TOSOH) at a constant flow rate of 0.5 ml/min. The column temperature was maintained at 20°C by a thermostated water supply (BL-51, Yamato Science Co. Ltd). Elution from the column was monitored by three types of detectors connected in series: an ultraviolet spectrophotometer (TOSOH, UV-

8000), a precision differential refractometer (TOSOH, RI8000) and a LALLS photometer (TOSOH, LS-8000). Bovine serum albumin (66267 Da, E&' = 0.670 mgiml), ovalbumin (42750 Da, 0.735 mg/ml) and ribonuclease A (13700 Da, 0.706 mg/ml) standards were assayed three times. Native Mn-SOD was diluted in the elution buffer to give concentrations of 0.243,0.486,1.217 and 2.434 mg/ml. 100 pl of each enzyme sample or of the standard sample were employed. The peak heights of the elution profile of the chromatogram were used in the calculation of molecular masses and molecular absorption coefficients of Mn-SOD by a standard method 1221. ThioI grouv titrations -

&

Thiol titrations with Nbs2 were carried out essentially by the method ofEllman [23]. Mn-SOD (1 mg/ml) in 0.1 M Tris/ HCl buffer, pH 8.0, containing 1 mM EDTA was incubated for 1 h at 25% or 5 min at 106°C in the presence or absence of 1% SDS or 8 M urea. The samples were then mixed with an equal volume of 10 mM Nbs2, kept at 25°C for 20 min, and the increase in A412 was compared with that of a control lacking Mn-SOD. A molar absorption coefficient of 14200 M- cm- for the 5-mercapto-2-nitrobenzoate anion was used. To block free thiol groups of Mn-SOD, 0.75 mg/ml native enzyme was incubated for 12 h in the dark at 37°C in pH 8.0, 0.1 M Tris/HCl containing 50 mM iodoacetamide. After desalting on a column of Sephadex G-25 (Pharmacia/ LKB, PD-10 column), the alkylated Mn-SOD was subjected to thiol titration and SDS/PAGE. The modified enzyme was heated in the presence of 8 M urea prior to electrophoresis.

'

Identification

of

a reactive sulfiydryl group

In order to identify the reactive site to sulfhydryl groups, human liver Mn-SOD (100 nmol) was dissolved in buffer A (0.2 M N-ethylmorpholine, pH 8.0 and 1 mM EDTA) containing 5 mM Nbs2 and incubated for 30 min at 25°C. The mixture was then loaded on to a commercial column of Sephadex G-25 (Pharmacia/LKB, PD-10) to remove low-molecular-mass reactants from the protein. The enzyme solution obtained was then mixed with a 2-ml slurry of activated thiolSepharose 4B. After incubation at 37 "C for 30 min, the mixture was poured into a small column. Approximately 75% of the Mn-SOD added was bound. After extensive washing of the column with buffer B (formic acid/acetic acid/water, 1 :4: 45, pH 1.9), the column contents were removed, suspended in the 2 ml of the above buffer B and digested with 80 pg pepsin for 12 h at 37°C. The level of pepsin used is approximately 0.8% of the amount of Mn-SOD bound to the resin. The digest was then poured into the original column, and washed successively with 10 ml each of buffer B, buffer A containing 1 M NaC1, and buffer A. The peptides bound to the column were then eluted with buffer A containing 20 mM 2-mercaptoethanol. The eluted peptide material was subjected to gas-phase amino acid sequence analysis and the resulting amino acids labeled with phenylhydantoin were analyzed by HPLC (Applied Biosystems, Model 120 A). RESULTS Isolation of Mn-SOD

Since liver and other tissues contain various forms of SOD, the precise calculation of yields of Mn-SOD from the starting

715 Tdblc 1. Purificution table o f Mn-SOD

Step

Supernatant

DEAE-Cellulose H ydroxylapatite Sephacryl S-300

Chromatofocusing 2nd hydroxylapatite Red Sepharose

Volume

Total protein

ml

9 000 20000 21 0 240 165 182 6.6

Total activity

Specific activity

Increase in specific activity

mg

lo-" x units

units/mg

-fold

289000 67 700

27.2 8.93

-

-

-

-

5230 3 290 1280 454 320

1.91 2.49 2.10 2.33 1.96

369 757 1640 51 33 6100

1 2.1 4.5 14.0 16.6

extracts based on activity is not possible, even though the Cu, Zn-SOD is largely inhibited by cyanide. For this reason, the fractionation steps will be described without presentation of activity/protein relationships for the crude extract and the DEAE-cellulose chromatography eluent. After the hydroxylapatite chromatographic step, fractions were found to be free from cyanide-sensitive proteins that affect the enzyme assay, therefore the calculations were possible. The following fractionation procedure was employed. Human normal liver was obtained at autopsy. Autopsy and research procedures were in accord with the ethical standards of the Helsinki Declaration of 1975. The frozen liver (2 kg) was thawed and homogenized with 4 vol. cold 10 mM sodium phosphate buffer, pH 7.8, containing 1 mM benzamidine and 10 pM 4-amidinophenylmethylsulfonyl fluoride to inhibit proteases. All purification procedures were performed at about 4°C unless otherwise indicated. The homogenate was centrifuged at 55000 x g for 60 min and 9 1 supernatant were chromatographed by DE-52. 20 1 flow-through active fractions were applied directly on to a hydroxylapatite column. Following washing of the column with 100 mM sodium phosphate, pH 7.2, the enzyme was eluted at about 25°C with 300 mM sodiumphosphate, pH 7.2 (Fig. IA). The active fractions were pooled and dialyzed against 25 mM monoethanolamine acetate, pH 9.4, then concentrated to 50 ml with an Amicon PM-10 membrane. This material was subjected to gel-exclusion chromatography on a Sephacryl S-300 column which had been equilibrated with the pH-9.4 buffer (Fig. 1B). The enzymatically active fractions were pooled and the resulting 240 ml applied to a Polybuffer Exchanger 94 chromatofocusing column. Following washing, elution with tenfold diluted Polybuffer 96 acetate (pH 6.0) was carried out. Mn-SOD was eluted at about pH 8.0 (Fig. 1C). The result indicates that while relatively small amount of enzymatically inactive material were separated in the hydroxylapatite and S-300 gel-filtration step, the chromatofocusing experiment led to a significant loss of protein. The pooled fractions (165 ml) were applied to a second hydroxylapatite column. Following washing, the protein was eluted with a linear gradient made by adding 500 ml400 mM sodium phosphate, pH 7.2, to the same volume of 100 mM sodium phosphate, pH 7.2. Inspection of Fig. 1D indicates that this second hydroxylapatite chromatographic step led to the loss of a considerable amount of inert protein. The 182-ml volume of the active fractions was extensively dialyzed against 20 mM sodium phosphate buffer, pH 7.2, and appled to a Red Sepharose column. The enzyme was eluted with a linear salt gradient made by adding 300 ml 20 mM phosphate buffer, pH 7.2, containing 1 M NaC1, to 300 ml of the same buffer

without NaCI. It can be seen from Fig. 1 E that the enzymatic activity elutes as a single symmetrical component. The specific activity of this purified Mn-SOD showed an increase of 17-fold from the protein eluted from the first hydroxylapatite column (Fig. 1A). The cyanide-resistant SOD activity in the starting homogenate supernatant comprises about 40% of the total reported in Table 1. If one assumes that this represents only Mn-SOD the overall recovery at the final stage of the Red Sepharose chromatography (Fig. 1E) is about 18%. This may be compared with the recovery of protein. The homogenate supernatant from 2 kg of liver showed the presence of 1.73 g Mn-SOD when assayed by an ELISA method employing a monoclonal antibody [9, 241. When this was compared with 320 mg purified enzyme recovered, there was an overall yield of 18.5%. Thus, the recoveries based on enzymatic and immunochemical methods are in good agreement. The isolated Mn-SOD was homogeneous on SDS/PAGE. Its specific activity was 6100 units/mg which is a higher value than that reported previously [16]. The enzyme contained 0.314% (by mass) manganese, as judged by atomic absorption spectrophotometry. This corresponds to 1.04 Mn/subunit, equivalent to 4 Mn/mol native tetramer. In order to identify Mn(I1) or Mn(II1) in the native enzyme, EPR spectrum was recorded. However, no signals were obtained, indicating that the native Mn-SOD has Mn(II1) because in the native Mn containing enzymes, only Mn(I1) can be detected by EPR as reported by Sugiura et al. [25]. The purified Mn-SOD was concentrated with an Amicon PM-10 membrane to 45 mg/ml then stored at -35°C in the presence of 50% glycerol until use. A photomicrograph of the crystallized Mn-SOD is shown in Fig. 2.

Determination of molecular mass by HPGCiLALLS The elution pattern of Mn-SOD (48.6 pg) when subjected to the HPGCiLALLS method is shown in Fig. 3. The MnSOD was eluted after 18 min as a single peak. The glycerol is particularly prominent in the refractometer assay and is seen to elute after 26 min. The apparent molecular mass of native Mn-SOD was calculated to be 88.6 f 2 kDa (mean f SD). Its molecular absorption coefficient at 280 nm was determined as 1.926 x lo5 M-' cm-'. A monomeric form of the enzyme prepared by treatment with 1 % SDS, as will be described later, exhibited a single component with an apparent molecular mass of 21.3 f 4 kDa. This indicates that Mn-SOD is a noncovalently linked tetrameric enzyme. Amino acid composition and partial sequence data have indicated that its subunits are identical (data not shown), as reported in [13, 141.

716 I

1

A A

Fraction

number

Fraction

number

.

Fraction

Fraction

number

to

l5Ic

-0

number

50

100

150 Fraction

200

250

300

350

400

number

Fig. 1 . Chromatographic patternsfor purification of human Mn-SOD. The size of each column and volume of each fraction collected are given in parentheses. (A) Hydroxylapatite column (7 cm x 42 cm) chromatography( I8 mi). (B) Sephacryl S-300column ( 5 cm x 125 cm) chromatography (10 ml). (C) Chromatofocusing with Polybuffer Exchanger 94 column (1.8 cm x 55 cm) chromatography ( 5 ml). (D) The second hydroxylapatite column (2.5 cm x 30 cm) chromatography ( 5 ml). (E) Red Sepharose column (2.2 cm x 23 cm) chromatography (3.6 ml). Each eluate fraction in A -E was assayed for cyanide-insensitive SOD activity (0)and protein (--- -). For experimental details, see Results

p H stubility

Stability in denaturing solutions

As shown in Fig. 4, Mn-SOD was most stable in the buffer at pH 7.2 where 98% of the original activity was retained for 1 day at 4 C. At pH 8.0 or 9.3, 80-90Y0, and at pH 5.5,65% activity was retained. The enzyme lost up to 30% of its activity after 3 days incubation at pH 7.2, 8.0 or 9.3, but about 50% of initial activity at pH 5.5.

Mn-SOD loses about 20% of its activity in pH 8.0, 0.1 M Tris/HCl after 3 days at 25"C, retaining its tetrameric structure as shown in Fig. 5 . In the presence of 1% SDS, all activity is lost within 6 h. The enzyme, however, retained its molecular mass of 88.6 kDa within a day, then gradually dissociated to a monomeric form; finally, most of the enzyme appeared to be

717

m 2

m 0

1

r h l

3

Day Fig. 2. Crystals of human liver Mn-SOD. Magnification, x 400

A B C

n

D

E

F

G H I 4

88kDa

25kDa *23kDa

I /

0

/

/

Day 1

/

20 10 Retention t i m e (rnin)

Fig. 3. The elution patterns ofhuman Mn-SODfrom a TSK-GEL G3000 SW-YL ( TOSOH, 7.8 mm x 300 mm) f o r molecular mass determination by HPGCILA LLS. The concentration gradients were obtained photometrically (LS), refractometrically (RI) and spectrophotometrically (UV). The elution patterns in the figure are for 48.6 wg enzyme. The gain settings of the detectors were 32 for the photometric, 64 for the refractometric, and A = 0.32 for the spectrophotometric measurements

0

1

Day 2

30

2

3

Day 3

B

Fig. 5 . The labile nature of Mn-SOD at 25°C in 0.1 M TrislHCI, p H 8.0, under varying conditions ( A ) and SDSjPAGE ( B ) . (A) ( 0 ) Control; (0)8 M urea; ( A ) 6 M guanidinium chloride; (U ) 1YoSDS. An aliquot of each sample was removed at the indicated timcs for activity measurements. Each point presents the mean of a duplicate experiment. (B) Sequential analysis was carried out for changes of molecular structure of the enzyme on SDSjPAGE as described in the text. Lanes A - C, D - F and G -I represent the samples incubated for 1,2 and 3 days, respectively. Three samples under different conditions were electrophoresed and demonstrated as follows: lanes A, D and G, control; lanes B, E and H, 8 M urea; lanes C, F and I, 1% SDS

in the monomeric form after 3 days. This indicates that MnSOD is a non-covalently linked tetrameric enzyme. Two monomeric components corresponding to molecular masses of 23 kDa and 25 kDa were distinguished. After 2 days, the amount of the 23-kDa and 25-kDa species appeared to be equivalent, but after 3 days, the amount of the 23-kDa form seemed to be greater than that the 25-kDa form. This indicates that the 25-kDa species appeared to convert to the 23-kDa species. In 6 M guanidinium chloride, about 50% of the activity was lost in 3 days. In 8 M urea, the enzyme remains active and a tetrameric form was observed which was essentially similar to the control form, even though minor monomeric components were observed.

Incubationtime ( day )

Fig. 4. The stability of Mn-SOD at 4°C in the following 0.1-M buffers ut various pH. ( 0 )Sodium acetate buffer, pH 5.5; (0)phosphate buffer, pH 7.2; ( A ) Tris/HCl buffer, pH 8.0; (El) monoethanolamine acetate buffer, pH 9.3. At the denoted time, an aliquot of the solution was removed for enzyme assays and calculated in duplicate

Thiol titration of Mn-SOD Under non-denaturing conditions, 4 thiols/mol(88.6 kDa) enzyme (Table 2) were derivatized. In the presence of 1YOSDS or 8 M urea solution for 1 h at 25°C a similar value was

71 8 -Fable 2. ThioI t i r i a t i o n of liiiniun Mn-SOD under various conditions Thiol titrations mere curried out with 5 mM Nbsz at pH 8.0, as described in the text. All mixtures contained 0.1 M Tris/HCl and 1 inM EDTA, pH 8.0. Thiol titration was carried out after removing dithiothreitol by gel filtration. Alkylation of Mn-SOD is described in the text. Each d u e reprcsenls the mean of a duplicate or a triplicate experiment Incubation conditions

Sulfhydryl group/Mn-SOD 25 c for 1 h

100 c for 5 min

niol/mol Natlvc Mn-SOD Alkylated Mn-SOI> Native Mn-SOD 1 % S D S N a t ~ v cMil-SOLI 1 % SDS I 0 m M dithiothreitol Ndtive Mil-SOD 8M urea Native Mil-SOD XM uiea 10 mM dithiothieilol

kDa

A

B

C

D

-

41 03 39

40 24

40 41

19 33

38

82

E

F

sulfhydryl residues reactive. This indicates that the unreactive thiol groups are exposed by these treatments. However, in the absence of dithiothreitol, these residues apparently combine with the reactive sulfhydryls to form oxidative intrachain disulfides, even in the presence of 5 mM EDTA. Amino acid analyses of the carboxylamidomethylated native protein showed that almost 4 mol sulfhydryl groups were alkylated (data not shown). Ellman titration of the alkylated protein gave a value of only 0.3 thiols/mol. However, in 8 M urea, followed by heating at 100°C after carboxylamidomethylation, the titration showed 4.0 sulfhydryls. Thus 4 out of the 8 cysteine residues in the tetramers are reactive with the Ellman's reagent as well as to alkylation by iodoacetamide, whereas the other residues are not, unless the protein is completely denatured in the presence of reducing reagent. Effect

of thiol seagents on the enzjwze nctivitj

The enzymatic activity was not changed by treatment with iodoacetamide, iodoacetic acid, Nbs2, N-ethylmaleimide, oxidized glutathione, 2-mercaptoethanol or dithiothreitol under conditions where the 4 reactive thiols noted could be derivatized. This indicates that sulfhydryl modification of MnSOD did not affect the enzyme activity.

C

94 67

SDSIPAGE

43 30 4

25kDa

20.1 14.4 Fig. 6. Rrsu1t.s of .SDS,:PAGE ufter incubation of M n - S O D under wrying condition.c in 0.1 M Tris!HCI, p H 8.0. The samples, which had all been heated at 100 C for 5 niin in 1 % SDS, were treated as follows: (A) standards; (B) control (10 wg); (C) reduced with 5 mM 2-mercaptoethanol (20 pg): (D) control (2 pg); (E) reduced with 5 mM 2-mercaptoethanol (2 pg); (F) alkylated with iodoacetamide (2 pg); ( G ) alkylatcd with iodoacetamide and reduced with 5 m M 2-mercaptoethanol ( 2 pg)

obtained. This probably suggests incomplete denaturation, since the tetraineric molecule has eight non-covalently crosslinked cysteine residues. However, heat treatment (1OO'C for 5 min) in the presence of SDS or urea gave relatively low values, suggesting that some of the thiol groups were oxidized to form disulfides. Incubation of the enzyme with SDS for 3 days at 25 C also gave low values. The conversion of the tetramer to monomers accompanied by the loss of reactive sulfhydryls means that some intrasubunit disulfide bonds are readily formed rather than the oxidative formation of disulfide-bridged oligomers. Following reduction of Mn-SOD with 10 mM dithiothreitol in 8 M urea or in lo/" SDS at 25°C for 1 h, only 4 sulfhydryl groups were found. This indicates that the other 4 unreactive cysteine residues have no mixed disulfide bonds on the outer sphere and are probably buried in the molecule. Mn-SOD treated with 10 mM dithiothreitol in 1% SDS or in 8 M urea and heated at 100 'C for 5 min showed almost 8

As shown in Fig.6, after heating at 100 C for 5 min in pH 8.0, 0.1 M TrisjHCl containing 1% SDS or 1% SDS and 5 mM 2-mercaptoethanol, the Mn-SOD was subjected to SDS/PAGE. When heated in SDS alone, a major component corresponding to 23 kDa and a minor one of 25 kDa were observed. Another two faint bands were observed, with molecular masses of 43 kDa and 54 kDa. However, in the presence of 5 mM 2-mercaptoethanol, a single band corresponding to a monomer of 25 kDa was observed. This indicates that the 43-kDa and 54-kDa molecules are probably proteins oligomerized by oxidative disulfide formation which convert to 25-kDa species by 2-mercaptoethanol. The enzyme alkylated with iodoacetamide followed by heating in the presence of SDS or SDS plus 2-mercaptoethanol, revealed the same result on SDS/PAGE, where a major molecule of 25 kDa and a minor one of 27 kDa were observed, but neither 43-kDa nor 54-kDa molecules. A minor band of 27 kDa is probably a byproduct of the carboxyamidomethylation reaction to certain amino acid residues due to long incubation time. We conclude that when the heat treatment of Mn-SOD in SDS is carried out in the absence of 2-mercaptoethanol or dithiothreitol, the likely formation of intrasubunit disulfide bonds occurred, and this is associated with conformational changes that give an apparent molecular mass of 23 kDa in SDSjPAGE experiments. These results are consistent with those obtained by thiol titrations. One might speculate that a proteolytic enzyme contaminant was activated by heat treatment and converted the 25-kDa species to one of 23 kDa. However, this was unlikely since alkylation or treatment with 2-mercaptoethanol after the heat treatment produced the 25-kDa species. This indicates a reversible conversion from the 23-kDa to the 25-kDa species. Identification of the reactive suif;hq'dsyl group of Mn-SOD The two cysteine residues known to be present in each subunit are Cys140 and Cys196. As described above, only

719 one cysteine residue is alkylated or titrated at 25°C under denaturing conditions. In order to identify this residue, the Mn-SOD was incubated with Nbs2, then subjected to thiolSepharose chromatography. The Nbs2-reactive cysteine was expected to bind to the thiol-Sepharose column through a SH/ SS exchange reaction. After peptic digestion of the Mn-SOD material bound to the column, it was eluted with 2-mercaptoethanol. Amino acid sequence analysis indicated the presence of two peptides in this eluate. Both were found to correspond to the C-terminal region, one comprising residues 182- 198 and the other 194 - 198 (Met-Ala-Xaa-Lys-Lys). The cysteine residues were not detected. These peptides are formed by cleavages on the carboxyl side of Trpl81 and Tyr193, respectively, and are in keeping with the high specificity of pepsin for cleavage at such residues. The binding of these peptides to the thiol-Sepharose column clearly indicate that Cys196 was the reactive sulfhydryl and is most probably located on the outer surface of the enzyme. In a separate experiment, the Mn-SOD which had been treated with iodoacetic acid or iodoacetamide was digested with a mixture of carboxypeptidases Y or A and carboxypeptidase B, and the amino acids released were analyzed. However, 2 mol lysine (Lys197 and Lys198) were quantitatively released from the C-terminus of Mn-SOD, but none of the cysteine that has been alkylated with iodoacetic acid or iodoacetamide was detected. The alkylated cysteine residue apparently was not susceptible to cleavage by either carboxypeptidase A or Y.

calculated from the amino acid composition data [13,14]. The subunit molecular mass on SDSjPAGE are somewhat higher than the value reported previously for other Mn-SOD isolates [6,15, 161. However, the N-terminal analysis of the 30 residues was identical with those reported by Barra et al. [12]. This indicates that the our isolated human Mn-SOD does not occur in a precursor form [17, 321. The Stokes' radius of the native enzyme, obtained as subdata from the HPGCILALLS system, was estimated to be 3.8 nm which corresponds to a molecular mass of 75 kDa for a globular protein. This indicates that native Mn-SOD is a non-globular tetrameric enzyme. As anticipated from the sequence data [12] Mn-SOD was found to contain 8 mol cysteine. In chicken liver Mn-SOD, the 8 cysteines appear to be involved in disulfide bond formation within each subunit [6]. In the present study, we found that human liver Mn-SOD contains four readily reactive cysteine residues/mol tetramer under both denaturing and nondenaturing conditions. Heating in the presence of 1% SDS showed no reactive sulfhydryls and predominantly converted the native tetramer (88.6 kDa) into monomers (23 kDa) containing an intimubunit disulfide bond. The presence of a Mn-SOD monomer of 23 kDa was noted in the SDSjPAGE experiments following heat treatment in urea or SDS solution. It seems reasonable to assume that formation of this material involves intrasubunit disulfide bond formation with an associated conformational change to a more compact type of molecule. On the other hand, we have no conclusive data to explain the appearance of 43 kDa and 54 kDa species. These molecules are probably oligomers of the 25 kDa monomer with interchain disulfide, such as DISCUSSION dimer or trimer. Such an assumption appears to be reasonable, Our isolation method did not employ carboxymethyl- since reduction with 2-mercaptoethanol diminished the cellulose chromatography at pH 5.5 since its use appeared to amount of these species and generated the 25 kDa form. As inactivate the enzyme as described [15]. As shown in Fig. 4, shown in Fig. 5, the structural changes responsible for generpH stability test of purified human Mn-SOD also supported ating the 23-kDa species from the 25-kDa species appear to result from conversion of a subunit containing free sulfhydryl our method. Human Mn-SOD is a relatively stable enzyme and retains groups to one having an intrachain disulfide. This result again suggests that no intrasubunit disulfide most of its activity and tetrameric structure in pH 8.0, 0.1 M Tris/HCl containing 8 M urea at 25°C. At acidic pH in the bonds are present in the native enzyme. The reactive cysteine presence of 8 M urea or 6M guanidinium chloride, the enzyme residue was found to be Cys196. One might speculate that rapidly loses activity concomitant with loss of manganese as nascent polypeptide chains of Mn-SOD are folded so as to described [26, 271. In a 1% SDS solution at 25"C, as shown prevent intrasubunit disulfide formation. In fact most of proin Fig. 5, human Mn-SOD rapidly loses activity [28] and teins are likely to start folding in order just after translation gradually dissociates to monomers. This indicates that the from mRNA, but some protein requires C-terminal peptides Mn-SOD was inactivated prior to dissociation in SDS and to complete its structure [33, 341. Mn-SOD synthesized in finally converts a tetramer to a monomer, some via the 25-kDa cytosol must be transported into mitochondria where it posform. Following repeated freezing and thawing, the Mn-SOD sibly becomes the active tetrameric form. This transport underwent inactivation and dissociation in SDS, even at 25 "C system and subsequent folding and oligomerization mechanfor a 20-min incubation (data not shown). Heating in SDS or ism in mitochondria is still unclear. To clarify this point, urea also converts the native enzyme to monomers, forming including the role of the C-terminal peptides of the enzyme, predominantly intrachain disulfide bonds. The likely forma- further study will be required. The result of the present study indicates that human Mntion of intrasubunit rather than intersubunit disulfide linkages may be explained by the distance between two cysteine resi- SOD is a tetrameric enzyme consisting of four identical subdues [30]; that is, actual distance between two intrachain units, in which one cysteine (Cys196) is readily reactive, and cysteines would be very close compared to those of intersub- the other (Cysl40) is apparently buried in the molecule. Under unit cysteines. Crystallographic studies of human Mn-SOD physiological conditions, Cys196 of Mn-SOD may be highly are necessary, but to date this data has not been completely reactive to intracellular sulfhydryl compounds such as reported 1181. A structure deduced from X-ray analysis of glutathione. It is well known that various proteins form a mixed disulfide with glutathione and undergo functional or bacterial Mn-SOD [31] seems to support our results. The HPGCiLALLS is a method for molecular mass deter- structural changes (for review see [35]). Whether the mixed mination based on hydrodynamics. Utilizing this we have disulfide formation of Mn-SOD with glutathione occurs in obtained an apparent molecular mass of 88.6 kDa, a subunit the tissues is an interesting problem. In fact we found that in molecular mass of 21.3 kDa and a molecular absorption coef- liver tissues two forms of Mn-SOD exist, and one of them has ficient of 1.926 x lo5 M-' cm-l. These values are close to a an acidic isoelectric point [36]. We are now studying whether subunit molecular mass of 22.2 kDa and an E of 18.1 x lo4 or not this heterogeneity is due to a mixed disulfide formation.

720 The authors are indebted to Mr Y. Sakamoto and Mr N. Koike of Central Laboratory for Research and Education, Osaka University Medical School for performing the amino acid and sequence analysis. This work was supported by a grants-in-aid for Cancer Research from the Ministry of Education, Science and Culture, and Welfare of Japan, and of the Yasuda Memorial Foundation.

REFERENCES 1. Keele, B. B. Jr, McCord, J. M. & Fridovich, I. (1971) J. Biol. Chem. 246,2875-2880. 2. McCord, J. M., Keele, B. B. Jr & Fridovich, I. (1971) Proc. Nut1 Acad. Sci. USA 68, 1024- 1027. 3. McCord, J. M. & Fridovich, I. (1969) J . Biol. Chem. 244, 60496055. 4. Carrico, R. J. & Deutsch, H. F. (1969) J . Biol. Chem. 244,60876093. 5. Carrico, R. J. & Deutsch, H. F. (1970) J . Biol. Chem. 245,723 727. 6. Weisinger, R. A. & Fridovich, 1. (1973) J. Biol. Chem. 248, 35823592. 7. Rotilio, G., Calabrese, L., Finazzi-Agro, A,, Argento-Ceru, M. P., Autuori, F. & Mondovi, B. (1973) Biochim. Biophys. Acta 321,98-102. 8. Marklund, S. L. (1982) Proc. Nut1 Acad. Sci. USA 79, 76347638. 9. Kawaguchi, T., Noji, S., Uda, T., Nakashima, Y., Takeyasu, A., Kawai, Y., Takagi, H., Tohyama, M. & Taniguchi, N. (1989) J. Biol. Chem. 264, 5762 - 5767. 10. Jabusch, J. R., Farb, D. L., Kerschensteiner, D. A. & Deutsch, H. F. (1980) Biochemistry 19,2310-2316. 11. Richardson, J. S., Thomas, K. A., Rubin, B. H. & Richardson, D. C. (1975) Proc. Natl Acad. Sci. USA 72, 1349-1353. 12. Barra, D., Schinina, M. E., Simmaco, M., Bannister, J. V., Bannister, W. H., Rotilio, G. & Bossa, F. (1984) J . Biol. Chem. 259, 12 595-12601. 13. Beck, Y., Oren, R., Amit, B., Levanon, A., Gorecki, M. & Hartman, J. R. (1987) Nucleic Acids Res. 15, 9076. 14. Ho, Y.-S. & Crapo, J. D. (1988) FEBS Lett. 229,256-260. 15. Salin, M. L., Day, E. D. Jr & Crapo, J. D. (1978) Arch. Biochem. Biophys. 187,223- 228.

16. McCord, J. M., Boyle, J. A., Day, E. D. Jr, Rizzolo, L. J. & Salin, M. L. (1977) Superoxide and superoxide dismutase (Michelson, A. M., McCord, J. M. & Fridovich, I., eds) pp. 129-138, Academic Press, London. 17. Beck, Y., Bartfeld, D., Yavin, Z., Levanon, A,, Gorecki, M. & Hartman, J. R. (1988) Bio-technology 6,930-935. 18. Wagner, U. G., Weber, M. M., Beck, Y., Hartman, J. R., Frolow, F. & Sussman, J. L. (1989) J . Mol. Biol. 206, 787-788. 19. Beauchamp, C. & Fridovich, I. (1971) Anal. Biochem. 44, 276287. 20. Bradford, M. M. (1976) Anal. Biochem. 72,248-254. 21. Laemmli, U. K. (1970) Nature 227,680-685. 22. Takagi, T. (1985) in Progress in HPLC (Parvez, H., Kato, Y. & Parvez, S., eds) vol. 1, pp. 27 -41, VNU Science Press, Utrecht. 23. Ellman, G. L. (1959) Arch. Biochem. Biophys. 82, 70-77. 24. Kawaguchi, T., Suzuki, K., Matsuda, Y., Nishiura, T., Uda, T., Ono, M., Sekiya, C., Ishikawa, M., Iino, S., Endo, Y. & Taniguchi, N. (1990) J. Immunol. Methods 127,249-254. 25. Sugiura, Y., Kawabe, H. & Tanaka, H. (1980) Am. J . Chem. SOC. 102,6581-6586. 26. Misra, H. P. & Fridovich, I. (1977) J. Biol. Chem. 252, 64216423. 27. Ose, D. E. & Fridovich, I. (1979) Arch. Biochem. Biophys. 194, 360 - 364. 28. Geller, B. L. & Winge, D. R. (1983) Anal. Biochem. 128, 86-92. 29. Goto, Y. & Hamaguchi, K. (1981) J. Mol. Biol. 198, 321 -340. 30. Stallings, W. C., Pattridge, K. A,, Strong, R. K. & Ludwig, M. L. (1985) J. Biol. Chem. 260, 16424-16432. 31. Wispe, J. R., Clark, J. C., Burhans, M. S., Kropp, K. E., Korfhagen, T. R. & Whitsett, J. A. (1989) Biochim. Biophys. Acts 994, 30 - 36. 32. Taniuchi, H. & Anfinsen, C. B. (1969) J . Biol. Chem. 244, 38643875. 33. Taniuchi, H. (1970) J. Biol. Chem. 245, 5459-5468. 34. Sies, H., Brigelius, R. & Ackerboom, T. P. M. (1983) in Functions of glutathione: biochemical, physiological, toxicological and clinical Aspects (Larsson, A,, Orrenius, S., Holmgren, A. & Mannervik, B., eds) pp. 231 -240, Raven Press, New York. 35. Matsuda, Y., Ishikawa, T., Higashiyama, S., Sugiyama, T., Kawata, S., Tarui, S. & Taniguchi, N. (1989) in Medical, biochemical and chemical aspects of free radicals (Hayaishi, O., Niki, E., Kondo, M. & Yoshikawa, T., eds) pp. 751 -754, Elsevier Science Publishers, B. V., Amsterdam.