Eur. J. Biochem. 73, 373-381 (1977)
Superoxide Dismutase from Thermus aquaticus Isolation and Characterisation of Manganese and Apo Enzymes Showbu SAT0 and J. Ieuan HARRIS Medical Research Council, Laboratory of Molecular Biology, Cambridge (Received September 21, 1976)
Superoxide dismutase has been isolated and characterised from the extreme thermophile Thermus aquaticus. The pure enzyme is a reddish-purple manganese-containing protein with a molecular weight of z 80000 f 5000. Combination of gel electrophoresis in dodecylsulphate and amino acid analysis shows that it is composed of four identical subunit polypeptide chains consisting of approximately 186 amino acids. The tetrameric protein contains two atoms of manganese. A stable manganese-free apoprotein has been prepared by treatment with EDTA in 8 M urea at acidic pH. The apoprotein regains the tetrameric structure in the absence of manganese but is inactive. Reconstitution of active Mn-enzyme was achieved by addition of Mn2+ to apoprotein in 8 M urea at acid pH. Reconstitution was monitored by absorption spectroscopy, manganese analysis and regain of activity and by these criteria the reconstituted enzyme with two atoms Mn per mole is indistinguishable from the native enzyme. The enhanced stability of the thermophile apoenzyme and Mn-enzyme is of advantage for studies of the structure and mechanism of action of superoxide dismutase. The N-terminal amino acid sequence to the 40th residue of the subunit was determined by automated Edman degradation. The sequence has a close resemblance to that of the dimeric Mnenzyme from another thermophile, Bacillus stearothermophilus. Superoxide dismutase which is widely distributed among oxygen-metabolising organisms catalyses the dismutation of superoxide radicals to hydrogen peroxide and molecular oxygen (for reviews see [1,2]). The enzyme is a metalloprotein and appears to occur in two evolutionarily distinct forms. Cu/Zn-enzymes found mainly in the cytoplasm of eukaryotic organisms are dimers comprising two identical polypeptide chains each containing one atom of copper and of zinc. Mn or Fe-containing enzymes found principally in prokaryotes are either dimers or tetramers that are also composed of identical subunits. The two classes of superoxide dismutase are believed to be of independent origin [3,4]. The amino acid sequence and three-dimensional structure of the Cu/Zn enzyme from bovine erythrocytes has been determined [5] and the protein ligands to the two metals in each subunit have been identified. In order to establish structural and functional relationships between the two classes of superoxide dismutase it is necessary to determine the threedimensional structure of a manganese or iron-con~.
Enzjwies. Superoxide dismutase (EC 1.15.1.1); xanthine oxidase
(EC 1.2.3.2).
taining enzyme. X-ray diffraction analysis of manganese-containing enzymes from Escherichia coli [6] and Bacillus stearothermophilus [7] are in progress and in order to complement extend and these studies we have sought to study manganese-containing superoxide dismutases from thermophilic microorganisms. Attempts to prepare a Mn-free (or Fe-free) apoenzyme for independent study have hitherto been unsuccessful owing to the instability of the putative apoenzyme [8,9]. It has already been shown that the greater stability of thermophile enzymes can be utilised to advantage in studies of enzyme structure and mechanism (see, for example, [10-12]). Enzymes from Thermus aquaticus are particularly stable to heat and to protein-denaturing solvents [12,13] and we now report the isolation and properties of a tetrameric manganese-containing superoxide dismutase from this extreme thermophile. MATERIALS Xanthine oxidase (milk) and cytochrome c type 111 from horse heart were obtained from Sigma. T . aquaticus cells (strain ATCC 25104) were obtained from the Microbiological Research Establishment (Porton
374
Superoxide Dismutase (Mn) from Thermus aquaticus
approximately equally distributed in the 0.1 M and 0.4 M NaCl fractions (Table 1). Since the 0.4 M NaCl fraction was required for the preparation of glyceraldehyde-3-phosphate dehydrogenase [12] and phosphofructokinase [13] the 0.1 M NaCl fraction was used initially to develop a method for the purification of pure superoxide dismutase. The pooled fraction (E 1000 ml) was brought to 40% saturation by addition of solid ammonium sulphate. After 1 h the precipitate was removed by centrifugation and the supernatant brought to 70% saturation with ammonium sulphate. The precipitate, containing most of the dismutase activity, was redissolved in 50 ml 10 mM Tris-HCI (pH 7.5), dialysed against several changes of the buffer and applied to a column of DEAE-cellulose (DE-52, 1.4 x 26 cm) in the same buffer. The column was developed with a linear gradient (1000 ml to 0.2 M) of NaCl in 10 mM Tris-HC1 (pH 7.5). Active fractions (1 - 70, 240 ml, Fig. 1) were pooled, reduced in volume to 30 ml by pressure dialysis through a Diaflo PM-10 membrane, dialysed against 10 mM potassium phosphate (pH 5.5) and applied to a column of CM-cellulose (CM-52, 1.2 x 13 cm). The column was developed with a linear gradient (400 ml to 0.2 M) of NaCl in the same buffer. Active fractions (61 - 75, Fig. 2) were pooled, concentrated to 2 ml by pressure dialysis through a Diaflo PM-10 membrane, and gel-filtered through Sephadex (G-100, fine grade, 2.4 x 50 cm) in 50 mM Tris-HC1 (pH 7.5) containing 0.1 M KC1 to remove last traces of contaminating proteins. Effluent fractions (2 ml) analysed for absorbance at 280 nm and for superoxide dismutase activity showed only one major protein peak. Alternate fractions were also examined by dodecylsulphate-gel electrophoresis and those showing only one protein band (cf. Fig. 3) were pooled and the reddish purple solution of superoxide dismutase was stored frozen at - 20 "C. The purification scheme is summarized in Table 1. Pure superoxide dismutase with a specific activity of
Down, Wiltshire, England) and stored frozen at -20 "C. METHODS AND RESULTS Protein concentrations were determined by the method of Lowry et al. [14] using crystalline bovine serum albumin as standard. Superoxide dismutase assays were carried out using xanthine oxidase and cytochrome c according to the method of McCord and Fridovich [15], and on occasion by the pulse radiolysis method [16]. 'Native' and dodecylsulphatepolyacrylamide gel electrophoresis were carried out as described previously [17,18]. Manganese was estimated by neutron activation analysis. Amino acid analyses were carried out with a Durrum-D5OO automatic analyser and N-terminal sequences were determined with a Beckman 890B automated sequencer (cf. [41). Enzyme Purification Frozen cells (100 g) were thawed in 100 ml buffer A (10 mM Tris-HCI, 5 mM 2-mercaptoethanol, 0.1 mM EDTA, pH 7.5) and disrupted by passage through a French pressure cell at 110- 140 kg/cm2. The resulting dark orange viscous extract was diluted to 200ml with buffer A and kept at room temperature for 45 min following addition of 2 ml M MgClz and 600 pg DNAase I. It was then treated batchwise with DEAEcellulose (400 ml Whatman DE23 equilibrated with buffer A) according to the method of Hocking and Harris [12]. After 1 h the resin was collected and washed with four 250-ml portions of bufferA to remove as much as possible of the pigmented 'slime' present in extracts of T. aquaticus [19]. The washed resin was then extracted with buffer A (four portions of 250 ml) containing successively 0.1 M, 0.4 M and 1.0 M NaCl. Assay of the pooled fractions for superoxide dismutase activity showed that the enzyme was
Table 1. Purification of superoxide dismutase from T. aquaticus The activity is expressed in units as defined by McCord and Fridovich [15]. The fraction eluted with 0.4 M NaCl was not included in the subsequent steps Purification step
Homogenate DEAE-cellulose (DE-23): 0.1 M NaCl 0.4 M NaCl Ammonium sulfate (40-70 % saturation) DEAE-cellulose (DE-52) chromatography CM-cellulose (CM-52) chromatography Sephadex (3-100 gel-filtration
Volume
Protein concn
Total protein
x Total activity
Specific activity at 23 'C
ml
mg/ml
mg
units
unitsjmg
'i:
40.6
8120
140
2
100
200 1000 1000 58 240 45 6
1.2 2.6 5.1 0.35 0.22 1.45
1200 2600 290 83 10 8.7
60 60 53 38 25 23
5 2.5 185 400 2530 2700
Yield
42 42 38 21 18 17
S. Sato and J. I. Harris
315 20
- -
1 M NaCl
Gradient
F
8 N
;i; 10 a,
m c
f
::
rl
a
0 20
40
60 Fraction number
80
100
120
Fig. 1. Chromatography on DEAE-cellulose (DE-52). The enzyme fraction salted out in 40- 70 % saturated ammonium sulfate solution was dialysed against 10 mM Tris-HC1 (pH 7.5) and applied to a column (1.4 x 26 cm) of DEAE-cellulose (DE-52). The column was eluted with Absorbance at 280 nm; (e-0) superoxide dismutase activity a linear gradient from 0. 0.2 M in NaCl in a total volume of 1 1. (--O) in 1 pl of each fraction (12 ml)
I
Gradient
0
Fig. 2. Chromatography on CM-cellulo.se. The pooled active fraction (51 -70, Fig. 1) was concentrated, dialysed against 10 mM potassium phosphate p H 5.5 and applied onto a column (1.2 x 13 cm) of CM-cellulose (CM-52). Elution was performed with a 400-mI linear gradient ) dismutase activity in 1 p1 of each fraction (2.5 ml) of NaCl (0-0.2 M). (0- -0)Absorbance at 280 nm; (O-- ~ - 0 superoxide
2700 units/mg at 23 "C was obtained in a yield of approximately 10 mg (based on an A @ nm value of 16.2 cm-') from 100 g frozen cells. A similar amount with the same specific activity was subsequently obtained from the fraction that eluted from DEAEcellulose with 0.4 M NaCl (cf. Table 1).
the protein band. The enzyme was also homogeneous as judged by electrophoresis on cellulose acetate strips at various pH. The results of isoelectric pH determination are presented in Fig. 4. The distance of migration was plotted against the pH value of the buffer and the isoelectric point was found to be 4.9.
Homogeneity
Molecular Weight
The homogeneity of the enzyme preparation was examined by polyacrylamide gel electrophoresis in the absence and presence of 1 % sodium dodecylsulphate. It gave a single protein band in both systems (Fig. 3). In the absence of dodecylsulphate, the band staining for enzymic activity was superimposed on
The estimation of molecular weight of the enzyme was carried out by gel-filtration on a column of Sephadex G-100 (fine grade) [20]. As shown in Fig. 5 , the molecular weight was determined to be about 80000. The enzyme was dissociated into subunits when it was heated in 1% sodium dodecylsulphate
376
Superoxide Dismutase (Mn) from Thermus aquuticus I
A 1
I
I
2
-5 .-c .-ts, L 0
E
e
L
a,
c m + .-
n
i
I
I
I
4
5
6
PH
Fig. 4. Determinution of isoelectric point. Electrophoresis was carried out on a strip of cellulose acetate (Cellogel) in sodium acetate buffer of constant ionic strength (0.05 M) at different pH values at 250 V for 30 min. The protein was stained with Ponceau S. The absolute mobility of the protein was plotted against pH
20
1c
e c
Fig. 3 . Polyacrylanzide gel electrophoresis of the purified T. aquaticus superoxide dismutase. (A) Electrophoresis in a 7.5'j: gel at pH 8.4 [17]. (1) Stain with Coomassie brilliant blue (5 pg of protein); (2) stain for enzyme activity by a photochemical method [25] (0.5 pg pf protein). (B) Dodecylsulphdte gel electrophoresis. The sample ( 5 pg) was heated in 1 % sodium dodecylsulphate in boiling water for 10 min and run on a 12.5% slab gel [I81
L .m
E
a,
3
-m L
-3
a,
4
r" b
\t'
L
with or without 2-mercaptoethanol. The molecular weight of the subunit was estimated to be 21000 by dodecylsulphate gel electrophoresis [18] (Fig. 6). These results show that the enzyme is composed of four subunit polypeptide chains of identical molecular weight, that are not covalently linked. Amino Acid Composition and Manganese Content Duplicate samples for amino acid analysis were hydrolysed in vucuo for 24, 48, and 72 h at 105 "C in 6 M HCl containing 0.1 o/, phenol. The analyses were carried out with a Durrum (D-500) amino acid analyser. The amounts of threonine, serine and methionine were extrapolated to zero time of hydrolysis and the 72-h values were adopted for valine and isoleucine. For other amino acids, the values at three
50
75 100 ELution vo(urne (rnL)
125
Fig. 5. Determination of the molecular. iveight of T. aquaticus superoxide dismutase. Gel-filtration was performed on a column (2.4 x 48 cm) of Sephadex G-100 (fine grade) [20]. Standard proteins: (1) bovine serum albumin dimer (130000); (2) bovine serum albumin monomer (65000); ( 3 ) egg albumin (46000); (4) myoglobin (17000). The arrow indicates the elution volume of the T. aquaticus enzyme
different hydrolysis times were averaged. Tryptophan content was determined by the spectrophotometric method of Edelhoch [21]. The results are given in Table 2 together with the amino acid compositions of other manganese superoxide dismutases. The T. aquaticus enzyme lacks half-cysteine residues but
311
S. Sato and J. I. Harris
Table 2. Amino acid composition of supero.uide di.rmuta,w fronl T. aquaticus Calculations for the T. aquaticus enzyme are based o n a subunit molecular weight of 21 000. Results for the B. stearothrrmophil~r.v enzyme are from unpublished data of M. J. Runswick and J. I. Harris. Results for the E. coli enzyme were obtained from Keele et al. [9] Amino acid
Amount in Mn-enzyme from -~
-~
~P
T aquaticus
1
c
0
0.25
0.50
0.75
Relative mobility
Fig. 6. Subunit nioleculur bveiglit OfT. aquaticus superoxide dismutase hjl dodecylsulphate gel electrophoresis. Electrophoresis was carried out in 12.5 y:, acrylamide gel. Standard proteins: (1) bovine serum albumin (65000); (2) egg albumin (46000); (3) carboxypeptidase A (34000); (4) trypsin (24000); (5) myoglobin (17000); (6) cytochrome c (12000). The arrow indicates the mobility of T. ayuaticus superoxide dismutase
contains six residues of tryptophan which accounts for the A ? F n mvalue of 16.7 cm-I. The amino acid composition gives a molecular weight of 21016 for the subunit chain in excellent agreement with the value obtained by dodecylsulphate gel electrophoresis. Determination of manganese was carried out by neutron activation analysis. Sample solutions (200400 pg/ml) were extensively dialysed against 5 mM Tris-HC1 buffer (pH 7.5) containing 0.5 mM of EDTA. The enzyme contained 0.139 (k 0.004 w/w) of manganese. This corresponds to 0.53 atom manganese per subunit, equivalent to two atoms manganese per molecule of native tetramer.
Aspartic acid Threonine Serine Glutamic acid Proline Glycine Alanine Half-cysteine Valine Methionine Isoleucine Leucine Tyrosine Phenylalanine Histidine Lysine Arginine Tryptophan Total residues
mol/mol subunit -~ ~ _ _ _ 15.7 (16) 7.0 (7) 4.1 (4) 21.7 (22) 5.9 (6) 16.6 (17) 22.5 (23) 0.0 ( 0 ) 11.6 (12) 3.0 (3) 8.2 (8) 19.1 (19) 9.8 (10) 7.0 (7) 8.2 (8) 10.7 (11) 7.0 (7) 6.3 (6) 186
B stearothermophrlur
E c olr
~-
~~
23 10 10 20 10 15 19 0 6 2 8 17 7 7 7 10 6 5
21
182
178
10
11 19 8 13 24
10 2 7 19 6 9 6 14 5
with 1600 units/mg for the E. coli enzyme [22]). By the pulse radiolysis assay [16] the specific activity (expressed as the second-order rate constants for the turnover of 0; under conditions of exponential decay) was 4.0 x lo8 M-' SP1. p H Stability
Absorption Spectrum
The enzyme gives a reddish-purple solution and as shown in Fig. 7, it has a broad absorption band with a maximum at 478 nm and a shoulder near 600 nm. This spectrum is very similar to that of other manganese superoxide dismutases [1,9] and the value of A f g n m was estimated to be 0.238 c m P t . The ultraviolet absorption spectrum has a peak at 280 - 282 nm and the value of was 16.2 cm-' by the Lowry method [14] in good agreement with the value of 16.7 cm-l based on amino acid analysis. SpeciJic Activity When assayed by the xanthine oxidase method [I53 T. aquaticus superoxide dismutase gave a specific activity of 2700 units/mg at 23 "C (this compares
When the enzyme (0.35 mg/ml) was incubated for up to 12 h at room temperature (23 "C) at pH values in the range 4.0-10.5 in 0.1 M acetate or Tris buffers, with or without 5 mM EDTA, no loss of activity was observed. Heat Stability
The enzyme in 50 mM Tris, 5 mM EDTA (0.35 mg/ ml) was heated in sealed tubes at temperatures up to 100 "C. After 10-min intervals the heated solutions were cooled in ice and assayed by the xanthine oxidase method. No detectable loss of activity was noted at temperatures up to 95 "C and in this respect superoxide dismutase resembles T. aquaticus glyceraldehyde-3-phosphate dehydrogenase [121 and phosphofructokinase [13].
378
Superoxide Dismutase (Mn) from Therrnus aquaticus Table 3. Reconstitution of active M n superoxide dismutase Mn content was determined by neutron activation analysis as mol Mn atoms per mol enzyme tetramer ( M , 84000). Specific activity was determined by pulse radiolysis assay [16] and expressed as the second-order rate constant k per enzyme dimer ( M , 40000) (assuming two active sites per tetramer) for the turnover of 0 2 under conditions of exponential decay ~~
"
400
500 Wavelength (nm)
600
Fig. 7. Absorption spectra q f T. aquaticus superoxide dismutase. Measurements were performed in 5 mM Tris-HC1 buffer (pH 7.5) containing 0.5 mM EDTA (10-mm light path). (-) Native enzyme (4.28 mg/ml); (. . . . . .) reconstituted enzyme (3.20 mg/ml)
I
I
I
I
I
~
Enzyme
Protein concn
Mn content
Specific activity
Native APO Reconstituted
0.34 0.30 0.30
2.1 < 0.2 1.9
4.0 x 10' 4.15 x 10' 3.15 x lo8
these conditions at neutral pH but dissociation with concomitant loss of manganese and of activity occurs at acid pH (see below). The enzyme is also inactivated in 1 % sodium dodecylsulphate at pH 7.5 (Fig. 8). Inactivation is accompanied by dissociation to the monomer as observed by dodecylsulphate-gel electrophoresis. In the presence of 5 mM EDTA dissociation and inactivation were accelerated as is shown in Fig. 8.
I Preparation of Apoenzyme
Fig. 8. Inactivation of T. aquaticus superoxide dismutase in dodecylsulphate. Enzyme solutions (0.35 mg protein/ml) in 25 mM TrisHCI, 1 % sodium dodecylsulphate (pH 7.5) were incubated at 40 "C with or without additions. At intervals, 1 p1 of the solution was pipetted out and immediately assayed. (0---O) No addition; (o--o) 10 mM NaCl; (O----O) 5 mM EDTA; (0-0) + 100 mM NaCl and 5 mM EDTA
Incubation of T. aquaticus superoxide dismutase ( 3 mg/ml) in 8 M urea containing 10 mM EDTA and 1 % acetic acid (pH 3.7) for up to 16 h at 23 "C led to the disappearance of the purple-red colour and to irreversible loss of activity (i.e. when a sample of the solution was diluted into neutral buffer containing 10 mM MnClz no activity was regained). The solution was gel-filtered on Sephadex G-25 in 8 M urea11 % acetic acid in the absence of EDTA, and subsequently dialysed successively against 25 mM Tris-HC1 (pH 7.5) with, and without 8 M urea. The resulting slightly turbid solution was centrifuged and the clear colourless supernatant was found to contain less than 0.01 mol of manganese per mol of protein, and to possess less than 10% of the activity of the native enzyme (Table 3). The apoprotein did not absorb at 478 nm. It was, however, indistinguishable from the native enzyme when it was analysed by acrylamide gel electrophoresis at pH 8.9 (Fig. 9) or by gel-filtration on Sephadex G-75 (Fig. 10) showing that the tetrameric structure had been reformed in the absence of manganese.
Stability in Denaturing Solvents
Reconstitution of Active Mn-Enzyme
Enzyme solutions (0.35 mgiml) in 50 mM TrisHCl, 1 mM EDTA (pH 7.5) containing 8 M urea or 6 M guanidinium chloride were found to be completely stable for periods of up to 16 h at 23 "C. The native tetrameric structure appears to be stable under
Reconstitution of active enzyme does not occur by addition of manganese chloride to the apoenzyme in buffer (or to buffer containing 8 M urea) at neutral pH. Total reconstitution of fully active enzyme was, however, found to occur following the exposure of
6t 0
1 20
40 60 Time (min)
+
80
100
+
319
S. Sato and J . I. Harris
t I I,
Fraction number Fig. 10. Gel-filtration oJ' T. aquaticus supero.uide disnzutase on a column of Sephadex G-100 (superfine grade, 2.4 x 50 cm) ; 3.3-ml fractions. (-0) Native enzyme (z300 Fg) blue dextran; (O----O)apoprotein ( z300 kg) blue dextran. Molecular weight, calculated according to Andrews [20], is z 80000 (cf: Fig. 5)
+
B
A
Fig. 9. Polyacrylarnide gel electrophoresis of T. aquaticus supero.xide dismutase. (A) Native Mn-enzyme; (B) apo-protein 7.5 % gel at pH 8.4 [17] stained with Coomassie brilliant blue
the apoprotein to manganese chloride in 8 M urea at acid pH. In a typical experiment the apoprotein (2 mg/ml) in 25 mM Tris-HC1 (pH 7.5) was dialysed successively for periods of at least 4 h at 4 "C against : (a) 8 M urea/l % acetic acid/lO mM MnC12 (pH 3.7), (b) 8 M urea/25 mM Tris-HCl/lO mM MnClz (pH 7.5), (c) 25 mM Tris-HCl/l mM MnClz (pH 7.5). Finally, the resulting purple-red solution was dialysed exhaustively against several changes of 25 mM Tris-HC1 containing 0.5 mM EDTA (pH 7.5). The extent of reconstitution was monitored by measurement of absorbance at 478 nm, manganese content, and regain of superoxide dismutase activity. As shown in Fig. 7 the absorption spectrum of the reconstituted enzyme is indistinguishable from that of the native enzyme; moreover as shown in Table 3 it contains 1.9 mol Mn per tetramer ( M , 84000) and is fully active when assayed by the pulse-radiolysis met hod. N- Terminal Sequence
Sequence analysis by automated Edman degradation was performed in a Beckman 890B sequencer using the standard Quadrol double-cleavage pro-
+
gramme. Native protein (8 mg, 400 nmol) was dried under vacuum in the reaction cup and degradations were commenced at the second acid-cleavage step so as to stabilise the protein film in the cup. Phenylthiohydantoins were identified by a combination of gas liquid chromatography, thin-layer chromatography and amino acid analysis following regeneration by hydrolysis with HI (cf. [4]). All residues were identified by at least two of these methods in each of two separate sequencer runs. Repetitive yield was calculated as 95% between cycle 6 (Leu-4) and cycle 16 (Leu-14). The sequence of 40 residues from the N-terminus was unambiguously determined and is given together with the corresponding N-terminal sequence of the E. coli Mn-enzyme [3] and B. stearothermophilus Mn-enzyme [4] in Table 4.
DISCUSSION Thermus aquaticus appears to contain only one active superoxide dismutase as evidenced by the appearance of a single band of enzymic activity (with an isoelectric point of 4.9) when a crude extract was examined by polyacrylamide disc-gel electrophoresis at pH 8.4. The appearance of two active fractions from DEAE-cellulose was caused by interference in the fractionation procedure by the viscous slime present in cell extracts of T aquaticus, rather than to the presence of two different active superoxide dismutases as has been found in other microorganisms [l]. Thermus aquaticus superoxide dismutase is a manganese-containing metalloprotein composed of
380
Superoxide Dismutase (Mn) from Thermus aquuricus
Table 4.N-terminal sequences of Mn-supero.de dismutases
Organism
Sequence
T aquaticuy
Pro-Tyr-Pro-Phe-Lys-Leu-Pro-Glu-Leu-Gly-Tyr -Pro -Tyr- Glu-AId -Leu-Glu-Pro- HIS-1le -Acp-Ah-Arg-Thr-
a
i
B strar othermophilus
Glu Ser -Tyr -Thr
E tolr ~~
~
T uquuticus
Pro
Ser
Pro
~
Lys Glu
ASP Ala
Asp
- _ _ ~
Phe _ _ _ _~
Lys Glu ~ ~ ___
25 10 35 40 Met-Glu-I le -His -His -Gln-Lys -H is -His -Gly-Ald-Tyr -Val -Thr-Asn-Leu-Asn-Ald
B stearothermophilus
Asn
Thr
E coli
Glu
1-
a
Ald
20
15
10
5
1
l
Asn-Thr
Ab
Insertion in T. uquuticus enzyme. Not determined incoli enzyme.
four identical polypeptide chains. In this respect it resembles the chicken liver mitochondria1 enzyme [22] and differs from the Mn-containing dimeric enzymes hitherto isolated from microorganisms such as E. coliB [9] and B. stearothermophilus [4]. In common with the latter two enzymes the T. aquaticus dismutase appears to contain one atom of manganese per two polypeptide chains. The tetrameric metalloenzyme is extremely stable over a wide range of pH (4-10) even in the presence of EDTA. It is also stable in dissociating solvents such as 8 M urea and 6 M guanidinium-HCI if the pH is maintained above 6.0. In order to dissociate the tetramer and to obtain a manganese-free apoprotein it is necessary to expose the native Mnenzyme to 8 M urea at acid ( < 4 ) pH. Removal of the dissociating reagent at neutral pH gave rise to a native tetrameric Mn-free apoprotein showing that the protein subunits are capable of forming a stable tetramer in the absence of manganese. Moreover, to reconstitute active Mn-enzyme it was found necessary to add the metal to apoprotein in acid 8 M urea suggesting that dissociation of the tetramer is an essential step, and that the metal may be shared between two subunits since the reconstituted enzyme, like the native enzyme, contained only two atoms of manganese per tetramer. This is the first time that it has proved possible to prepare a stable apoprotein from a Mn-containing superoxide dismutase, and to reconstitute fully active enzyme from it by addition of inorganic manganese. Similar results have since been obtained with the dimeric manganese-containing superoxide dismutase from another thermophile, B. stearothermophilus [22a]. It has not so far been possible to isolate a stable dimeric apoprotein from the E. coli Mn-enzyme although reversible metal exchange does occur in situ [23]. The colour and nuclear magnetic resonance properties of the E. coli enzyme suggest that the constituent manganese is trivalent in the resting state [24]. Since
reconstitution of the two thermophile Mn-enzymes was achieved following addition of divalent manganese it is assumed that Mn2+ is oxidised to Mn3+ during the process of reconstitution. The ability to prepare stable apoproteins and to reconstitute fully active Mn-containing enzymes from T. aquaticus and B. stearothermophilus dismutases opens new possibilities for the study of the specificity and mode of binding of the metal and of its role in catalysis. In common with other enzymes from T. aquaticus, the superoxide dismutase is extremely stable to heat. Thus it was not inactivated at temperatures of up to 95 "C, whereas enzyme from the more moderate thermophile, B. stearothermophilus, under the same conditions of heating was inactivated at 75 "C. Numerous attempts have been made to correlate differences in the amino acid compositions of proteins with their thermal stabilities. Comparison of the amino acid compositions of superoxide dismutase from T. aquaticus, B. stearothermophilus and E. coli does not, however, reveal any consistent trend. It is perhaps of interest that the thermophile enzymes do not contain cysteine residues that might be a source of instability in organisms that grow aerobically at temperatures in excess of 70 "C (cf. [12]). Comparison of the N-terminal sequences of the T. aquaticus, B. stearothermophilus and E. coli enzymes reveals a very high degree of sequence homology for enzymes isolated from such diverse genera. Over the first 26 residues 13 (50%) are common to all three proteins. The two thermophile enzymes are remarkably homologous, 30 of the first 40 residues (75 %) being common to the two proteins. There is no particular reason to suppose that B. stearothermophilus and T. aquaticus are closely related organisms. The former is a gram-positive spore-former whereas the latter is a gram-negative non-sporulating organism that could belong to any one of several orders of procaryotes. The significance of the N-terminal sequence homology between the
_
_
38 1
S. Sato and J. I . Harris
two Mn-dismutases must therefore await further study of their structures and of the relationship between three-dimensional structure and catalytic activity. We thank D r M. E. McAdam (Department of Physics, Institute of Cancer Research, Sutton) for superoxide dismutase assays by the pulse radiolysis method, Mr J. Herrington (Aldermaston Weapons Research Establishment, Aldermaston) for manganese estimations by neutron activation analysis, and Mr F. Northrop for .assistance in determining the N-terminal sequence.
REFERENCES 1. Fridovich, 1. (1975) Annu. Rev. Biochem. 44, 147- 157. 2. Fridovich, 1. (1974) Adv. Enzymol. 41, 35-97. 3. Steinman, H. M . & Hill, R. L. (1973) Proc. Nut/ Acad. Sci. U.S.A. 70,3725 - 3729. 4. Bridgen, J., Harris, J. I. & Northrop, F. (1975) FEBS Left. 49,392- 395. 5. Richardson, J. S., Thomas, K. A,, Rubin, B. H. & Richardson, D. C. (1975) Proc. Natl Acad. Sci. U . S . A . 72, 1349-1353. 6. Beem, K . M., Richardson, J. S. & Richardson, D. C. (1976) J . Mol. Biol. 105, 327 - 332. 7. Bridgen, J., Harris, J. I. & Kolb, E. (1976) J . Mol. B i d . 105, 333 - 335. 8. Puget, K . & Michelson, A. M . (1974) Biochem. Biophys. Res. Comnzun. 58, 830- 838.
9. Keele, B. B., Jr, McCord, J . M . & Fridovich, I. (1970) J . B i d . Cl7em. 245,6176-6181. 10. Suzuki, I. & Harris, J. I. (1971) FEBS L e f f . 13, 217-220. 11. Hill, H. A. O., Lobb, R . R., Sharp. S. L., Stokes, A. M., Harris, J . I. & Jack, R. S. (1976) Biochem. J . 153, 551 - 560. 12a. Hocking, J. D. & Harris, J. I. (1973) FEBS L e t / . 34, 280284. 12b. Hocking, J. D. & Harris, J . I . (1976) in Symposiumon €n:yrncs and Proteins from Thermophilic Microorganisnis (Zuber, H ., ed.) pp. 121- 133, Birkhauser-Verlag, Basel. 13. Hengartner, H . & Harris, J . I. (1975) FEBS L e t / . 55, 272-285. 14. Lowry, 0. H., Rosebrough, N . J., Farr, A. W. & Randall, R. J. (1951) J . Biol. Chem. 193, 265-275. 15. McCord, J. M. & Fridovich, I. (1969) J . Bid. Chmz. 244. 6049 - 6055. 16. Rotilio, G., Bray, R. C . & Fielden, E. M. (1972) Biocliim. Biophys. Acta, 268, 605 -609. 17. Davis, B. J. (1964) Ann. N . Y . Acad. Sci. 121, 404-427. 18. Laemmli, U..K. (1970) Nurure (Lond.) 227, 680-685. 19. Brock, T. D. & Freeze, H. (1969) J . Bucteriol. 98, 289-298. 20. Andrews, P. (1964) Biochem. J . 91, 222-233. 21. Edelhoch, H. (1967) Biochemistry, 6, 1948- 1954. 22. Weisiger, R. A. & Fridovich, I. (1973) J . B i d . Chem. 248, 3582- 3592. 22a. Brock, C. J., Harris, J. I. & Sato, S. (1976) J . Mol. Biol. 107. 175- 178. 23. Ose, D. E. & Fridovich, I. (1973) J . B i d . Chem. 251, 12171218. 24. Villafranca, J. J., Yost, F. J., J r & Fridovich, 1. (1974) J . Biol. Chem. 249,3532- 3536. 25. Beauchamp, P. C. & Fridovich, 1. (1971) Anal. Biocheni. 44. 276 - 287.
S. Sato and J. I. Harris, M.R.C. Laboratory of Molecular Biology, Postgraduate Medical School, University of Cambridge, Hills Road, Cambridge, Great Britain, CB2 2QH
Superoxide Dismutase from Thermus aquaticus Isolation and Characterisation of Manganese and Apo Enzymes Showbu SAT0 and J. Ieuan HARRIS Medical Research Council, Laboratory of Molecular Biology, Cambridge (Received September 21, 1976)
Superoxide dismutase has been isolated and characterised from the extreme thermophile Thermus aquaticus. The pure enzyme is a reddish-purple manganese-containing protein with a molecular weight of z 80000 f 5000. Combination of gel electrophoresis in dodecylsulphate and amino acid analysis shows that it is composed of four identical subunit polypeptide chains consisting of approximately 186 amino acids. The tetrameric protein contains two atoms of manganese. A stable manganese-free apoprotein has been prepared by treatment with EDTA in 8 M urea at acidic pH. The apoprotein regains the tetrameric structure in the absence of manganese but is inactive. Reconstitution of active Mn-enzyme was achieved by addition of Mn2+ to apoprotein in 8 M urea at acid pH. Reconstitution was monitored by absorption spectroscopy, manganese analysis and regain of activity and by these criteria the reconstituted enzyme with two atoms Mn per mole is indistinguishable from the native enzyme. The enhanced stability of the thermophile apoenzyme and Mn-enzyme is of advantage for studies of the structure and mechanism of action of superoxide dismutase. The N-terminal amino acid sequence to the 40th residue of the subunit was determined by automated Edman degradation. The sequence has a close resemblance to that of the dimeric Mnenzyme from another thermophile, Bacillus stearothermophilus. Superoxide dismutase which is widely distributed among oxygen-metabolising organisms catalyses the dismutation of superoxide radicals to hydrogen peroxide and molecular oxygen (for reviews see [1,2]). The enzyme is a metalloprotein and appears to occur in two evolutionarily distinct forms. Cu/Zn-enzymes found mainly in the cytoplasm of eukaryotic organisms are dimers comprising two identical polypeptide chains each containing one atom of copper and of zinc. Mn or Fe-containing enzymes found principally in prokaryotes are either dimers or tetramers that are also composed of identical subunits. The two classes of superoxide dismutase are believed to be of independent origin [3,4]. The amino acid sequence and three-dimensional structure of the Cu/Zn enzyme from bovine erythrocytes has been determined [5] and the protein ligands to the two metals in each subunit have been identified. In order to establish structural and functional relationships between the two classes of superoxide dismutase it is necessary to determine the threedimensional structure of a manganese or iron-con~.
Enzjwies. Superoxide dismutase (EC 1.15.1.1); xanthine oxidase
(EC 1.2.3.2).
taining enzyme. X-ray diffraction analysis of manganese-containing enzymes from Escherichia coli [6] and Bacillus stearothermophilus [7] are in progress and in order to complement extend and these studies we have sought to study manganese-containing superoxide dismutases from thermophilic microorganisms. Attempts to prepare a Mn-free (or Fe-free) apoenzyme for independent study have hitherto been unsuccessful owing to the instability of the putative apoenzyme [8,9]. It has already been shown that the greater stability of thermophile enzymes can be utilised to advantage in studies of enzyme structure and mechanism (see, for example, [10-12]). Enzymes from Thermus aquaticus are particularly stable to heat and to protein-denaturing solvents [12,13] and we now report the isolation and properties of a tetrameric manganese-containing superoxide dismutase from this extreme thermophile. MATERIALS Xanthine oxidase (milk) and cytochrome c type 111 from horse heart were obtained from Sigma. T . aquaticus cells (strain ATCC 25104) were obtained from the Microbiological Research Establishment (Porton
374
Superoxide Dismutase (Mn) from Thermus aquaticus
approximately equally distributed in the 0.1 M and 0.4 M NaCl fractions (Table 1). Since the 0.4 M NaCl fraction was required for the preparation of glyceraldehyde-3-phosphate dehydrogenase [12] and phosphofructokinase [13] the 0.1 M NaCl fraction was used initially to develop a method for the purification of pure superoxide dismutase. The pooled fraction (E 1000 ml) was brought to 40% saturation by addition of solid ammonium sulphate. After 1 h the precipitate was removed by centrifugation and the supernatant brought to 70% saturation with ammonium sulphate. The precipitate, containing most of the dismutase activity, was redissolved in 50 ml 10 mM Tris-HCI (pH 7.5), dialysed against several changes of the buffer and applied to a column of DEAE-cellulose (DE-52, 1.4 x 26 cm) in the same buffer. The column was developed with a linear gradient (1000 ml to 0.2 M) of NaCl in 10 mM Tris-HC1 (pH 7.5). Active fractions (1 - 70, 240 ml, Fig. 1) were pooled, reduced in volume to 30 ml by pressure dialysis through a Diaflo PM-10 membrane, dialysed against 10 mM potassium phosphate (pH 5.5) and applied to a column of CM-cellulose (CM-52, 1.2 x 13 cm). The column was developed with a linear gradient (400 ml to 0.2 M) of NaCl in the same buffer. Active fractions (61 - 75, Fig. 2) were pooled, concentrated to 2 ml by pressure dialysis through a Diaflo PM-10 membrane, and gel-filtered through Sephadex (G-100, fine grade, 2.4 x 50 cm) in 50 mM Tris-HC1 (pH 7.5) containing 0.1 M KC1 to remove last traces of contaminating proteins. Effluent fractions (2 ml) analysed for absorbance at 280 nm and for superoxide dismutase activity showed only one major protein peak. Alternate fractions were also examined by dodecylsulphate-gel electrophoresis and those showing only one protein band (cf. Fig. 3) were pooled and the reddish purple solution of superoxide dismutase was stored frozen at - 20 "C. The purification scheme is summarized in Table 1. Pure superoxide dismutase with a specific activity of
Down, Wiltshire, England) and stored frozen at -20 "C. METHODS AND RESULTS Protein concentrations were determined by the method of Lowry et al. [14] using crystalline bovine serum albumin as standard. Superoxide dismutase assays were carried out using xanthine oxidase and cytochrome c according to the method of McCord and Fridovich [15], and on occasion by the pulse radiolysis method [16]. 'Native' and dodecylsulphatepolyacrylamide gel electrophoresis were carried out as described previously [17,18]. Manganese was estimated by neutron activation analysis. Amino acid analyses were carried out with a Durrum-D5OO automatic analyser and N-terminal sequences were determined with a Beckman 890B automated sequencer (cf. [41). Enzyme Purification Frozen cells (100 g) were thawed in 100 ml buffer A (10 mM Tris-HCI, 5 mM 2-mercaptoethanol, 0.1 mM EDTA, pH 7.5) and disrupted by passage through a French pressure cell at 110- 140 kg/cm2. The resulting dark orange viscous extract was diluted to 200ml with buffer A and kept at room temperature for 45 min following addition of 2 ml M MgClz and 600 pg DNAase I. It was then treated batchwise with DEAEcellulose (400 ml Whatman DE23 equilibrated with buffer A) according to the method of Hocking and Harris [12]. After 1 h the resin was collected and washed with four 250-ml portions of bufferA to remove as much as possible of the pigmented 'slime' present in extracts of T. aquaticus [19]. The washed resin was then extracted with buffer A (four portions of 250 ml) containing successively 0.1 M, 0.4 M and 1.0 M NaCl. Assay of the pooled fractions for superoxide dismutase activity showed that the enzyme was
Table 1. Purification of superoxide dismutase from T. aquaticus The activity is expressed in units as defined by McCord and Fridovich [15]. The fraction eluted with 0.4 M NaCl was not included in the subsequent steps Purification step
Homogenate DEAE-cellulose (DE-23): 0.1 M NaCl 0.4 M NaCl Ammonium sulfate (40-70 % saturation) DEAE-cellulose (DE-52) chromatography CM-cellulose (CM-52) chromatography Sephadex (3-100 gel-filtration
Volume
Protein concn
Total protein
x Total activity
Specific activity at 23 'C
ml
mg/ml
mg
units
unitsjmg
'i:
40.6
8120
140
2
100
200 1000 1000 58 240 45 6
1.2 2.6 5.1 0.35 0.22 1.45
1200 2600 290 83 10 8.7
60 60 53 38 25 23
5 2.5 185 400 2530 2700
Yield
42 42 38 21 18 17
S. Sato and J. I. Harris
315 20
- -
1 M NaCl
Gradient
F
8 N
;i; 10 a,
m c
f
::
rl
a
0 20
40
60 Fraction number
80
100
120
Fig. 1. Chromatography on DEAE-cellulose (DE-52). The enzyme fraction salted out in 40- 70 % saturated ammonium sulfate solution was dialysed against 10 mM Tris-HC1 (pH 7.5) and applied to a column (1.4 x 26 cm) of DEAE-cellulose (DE-52). The column was eluted with Absorbance at 280 nm; (e-0) superoxide dismutase activity a linear gradient from 0. 0.2 M in NaCl in a total volume of 1 1. (--O) in 1 pl of each fraction (12 ml)
I
Gradient
0
Fig. 2. Chromatography on CM-cellulo.se. The pooled active fraction (51 -70, Fig. 1) was concentrated, dialysed against 10 mM potassium phosphate p H 5.5 and applied onto a column (1.2 x 13 cm) of CM-cellulose (CM-52). Elution was performed with a 400-mI linear gradient ) dismutase activity in 1 p1 of each fraction (2.5 ml) of NaCl (0-0.2 M). (0- -0)Absorbance at 280 nm; (O-- ~ - 0 superoxide
2700 units/mg at 23 "C was obtained in a yield of approximately 10 mg (based on an A @ nm value of 16.2 cm-') from 100 g frozen cells. A similar amount with the same specific activity was subsequently obtained from the fraction that eluted from DEAEcellulose with 0.4 M NaCl (cf. Table 1).
the protein band. The enzyme was also homogeneous as judged by electrophoresis on cellulose acetate strips at various pH. The results of isoelectric pH determination are presented in Fig. 4. The distance of migration was plotted against the pH value of the buffer and the isoelectric point was found to be 4.9.
Homogeneity
Molecular Weight
The homogeneity of the enzyme preparation was examined by polyacrylamide gel electrophoresis in the absence and presence of 1 % sodium dodecylsulphate. It gave a single protein band in both systems (Fig. 3). In the absence of dodecylsulphate, the band staining for enzymic activity was superimposed on
The estimation of molecular weight of the enzyme was carried out by gel-filtration on a column of Sephadex G-100 (fine grade) [20]. As shown in Fig. 5 , the molecular weight was determined to be about 80000. The enzyme was dissociated into subunits when it was heated in 1% sodium dodecylsulphate
376
Superoxide Dismutase (Mn) from Thermus aquuticus I
A 1
I
I
2
-5 .-c .-ts, L 0
E
e
L
a,
c m + .-
n
i
I
I
I
4
5
6
PH
Fig. 4. Determinution of isoelectric point. Electrophoresis was carried out on a strip of cellulose acetate (Cellogel) in sodium acetate buffer of constant ionic strength (0.05 M) at different pH values at 250 V for 30 min. The protein was stained with Ponceau S. The absolute mobility of the protein was plotted against pH
20
1c
e c
Fig. 3 . Polyacrylanzide gel electrophoresis of the purified T. aquaticus superoxide dismutase. (A) Electrophoresis in a 7.5'j: gel at pH 8.4 [17]. (1) Stain with Coomassie brilliant blue (5 pg of protein); (2) stain for enzyme activity by a photochemical method [25] (0.5 pg pf protein). (B) Dodecylsulphdte gel electrophoresis. The sample ( 5 pg) was heated in 1 % sodium dodecylsulphate in boiling water for 10 min and run on a 12.5% slab gel [I81
L .m
E
a,
3
-m L
-3
a,
4
r" b
\t'
L
with or without 2-mercaptoethanol. The molecular weight of the subunit was estimated to be 21000 by dodecylsulphate gel electrophoresis [18] (Fig. 6). These results show that the enzyme is composed of four subunit polypeptide chains of identical molecular weight, that are not covalently linked. Amino Acid Composition and Manganese Content Duplicate samples for amino acid analysis were hydrolysed in vucuo for 24, 48, and 72 h at 105 "C in 6 M HCl containing 0.1 o/, phenol. The analyses were carried out with a Durrum (D-500) amino acid analyser. The amounts of threonine, serine and methionine were extrapolated to zero time of hydrolysis and the 72-h values were adopted for valine and isoleucine. For other amino acids, the values at three
50
75 100 ELution vo(urne (rnL)
125
Fig. 5. Determination of the molecular. iveight of T. aquaticus superoxide dismutase. Gel-filtration was performed on a column (2.4 x 48 cm) of Sephadex G-100 (fine grade) [20]. Standard proteins: (1) bovine serum albumin dimer (130000); (2) bovine serum albumin monomer (65000); ( 3 ) egg albumin (46000); (4) myoglobin (17000). The arrow indicates the elution volume of the T. aquaticus enzyme
different hydrolysis times were averaged. Tryptophan content was determined by the spectrophotometric method of Edelhoch [21]. The results are given in Table 2 together with the amino acid compositions of other manganese superoxide dismutases. The T. aquaticus enzyme lacks half-cysteine residues but
311
S. Sato and J. I. Harris
Table 2. Amino acid composition of supero.uide di.rmuta,w fronl T. aquaticus Calculations for the T. aquaticus enzyme are based o n a subunit molecular weight of 21 000. Results for the B. stearothrrmophil~r.v enzyme are from unpublished data of M. J. Runswick and J. I. Harris. Results for the E. coli enzyme were obtained from Keele et al. [9] Amino acid
Amount in Mn-enzyme from -~
-~
~P
T aquaticus
1
c
0
0.25
0.50
0.75
Relative mobility
Fig. 6. Subunit nioleculur bveiglit OfT. aquaticus superoxide dismutase hjl dodecylsulphate gel electrophoresis. Electrophoresis was carried out in 12.5 y:, acrylamide gel. Standard proteins: (1) bovine serum albumin (65000); (2) egg albumin (46000); (3) carboxypeptidase A (34000); (4) trypsin (24000); (5) myoglobin (17000); (6) cytochrome c (12000). The arrow indicates the mobility of T. ayuaticus superoxide dismutase
contains six residues of tryptophan which accounts for the A ? F n mvalue of 16.7 cm-I. The amino acid composition gives a molecular weight of 21016 for the subunit chain in excellent agreement with the value obtained by dodecylsulphate gel electrophoresis. Determination of manganese was carried out by neutron activation analysis. Sample solutions (200400 pg/ml) were extensively dialysed against 5 mM Tris-HC1 buffer (pH 7.5) containing 0.5 mM of EDTA. The enzyme contained 0.139 (k 0.004 w/w) of manganese. This corresponds to 0.53 atom manganese per subunit, equivalent to two atoms manganese per molecule of native tetramer.
Aspartic acid Threonine Serine Glutamic acid Proline Glycine Alanine Half-cysteine Valine Methionine Isoleucine Leucine Tyrosine Phenylalanine Histidine Lysine Arginine Tryptophan Total residues
mol/mol subunit -~ ~ _ _ _ 15.7 (16) 7.0 (7) 4.1 (4) 21.7 (22) 5.9 (6) 16.6 (17) 22.5 (23) 0.0 ( 0 ) 11.6 (12) 3.0 (3) 8.2 (8) 19.1 (19) 9.8 (10) 7.0 (7) 8.2 (8) 10.7 (11) 7.0 (7) 6.3 (6) 186
B stearothermophrlur
E c olr
~-
~~
23 10 10 20 10 15 19 0 6 2 8 17 7 7 7 10 6 5
21
182
178
10
11 19 8 13 24
10 2 7 19 6 9 6 14 5
with 1600 units/mg for the E. coli enzyme [22]). By the pulse radiolysis assay [16] the specific activity (expressed as the second-order rate constants for the turnover of 0; under conditions of exponential decay) was 4.0 x lo8 M-' SP1. p H Stability
Absorption Spectrum
The enzyme gives a reddish-purple solution and as shown in Fig. 7, it has a broad absorption band with a maximum at 478 nm and a shoulder near 600 nm. This spectrum is very similar to that of other manganese superoxide dismutases [1,9] and the value of A f g n m was estimated to be 0.238 c m P t . The ultraviolet absorption spectrum has a peak at 280 - 282 nm and the value of was 16.2 cm-' by the Lowry method [14] in good agreement with the value of 16.7 cm-l based on amino acid analysis. SpeciJic Activity When assayed by the xanthine oxidase method [I53 T. aquaticus superoxide dismutase gave a specific activity of 2700 units/mg at 23 "C (this compares
When the enzyme (0.35 mg/ml) was incubated for up to 12 h at room temperature (23 "C) at pH values in the range 4.0-10.5 in 0.1 M acetate or Tris buffers, with or without 5 mM EDTA, no loss of activity was observed. Heat Stability
The enzyme in 50 mM Tris, 5 mM EDTA (0.35 mg/ ml) was heated in sealed tubes at temperatures up to 100 "C. After 10-min intervals the heated solutions were cooled in ice and assayed by the xanthine oxidase method. No detectable loss of activity was noted at temperatures up to 95 "C and in this respect superoxide dismutase resembles T. aquaticus glyceraldehyde-3-phosphate dehydrogenase [121 and phosphofructokinase [13].
378
Superoxide Dismutase (Mn) from Therrnus aquaticus Table 3. Reconstitution of active M n superoxide dismutase Mn content was determined by neutron activation analysis as mol Mn atoms per mol enzyme tetramer ( M , 84000). Specific activity was determined by pulse radiolysis assay [16] and expressed as the second-order rate constant k per enzyme dimer ( M , 40000) (assuming two active sites per tetramer) for the turnover of 0 2 under conditions of exponential decay ~~
"
400
500 Wavelength (nm)
600
Fig. 7. Absorption spectra q f T. aquaticus superoxide dismutase. Measurements were performed in 5 mM Tris-HC1 buffer (pH 7.5) containing 0.5 mM EDTA (10-mm light path). (-) Native enzyme (4.28 mg/ml); (. . . . . .) reconstituted enzyme (3.20 mg/ml)
I
I
I
I
I
~
Enzyme
Protein concn
Mn content
Specific activity
Native APO Reconstituted
0.34 0.30 0.30
2.1 < 0.2 1.9
4.0 x 10' 4.15 x 10' 3.15 x lo8
these conditions at neutral pH but dissociation with concomitant loss of manganese and of activity occurs at acid pH (see below). The enzyme is also inactivated in 1 % sodium dodecylsulphate at pH 7.5 (Fig. 8). Inactivation is accompanied by dissociation to the monomer as observed by dodecylsulphate-gel electrophoresis. In the presence of 5 mM EDTA dissociation and inactivation were accelerated as is shown in Fig. 8.
I Preparation of Apoenzyme
Fig. 8. Inactivation of T. aquaticus superoxide dismutase in dodecylsulphate. Enzyme solutions (0.35 mg protein/ml) in 25 mM TrisHCI, 1 % sodium dodecylsulphate (pH 7.5) were incubated at 40 "C with or without additions. At intervals, 1 p1 of the solution was pipetted out and immediately assayed. (0---O) No addition; (o--o) 10 mM NaCl; (O----O) 5 mM EDTA; (0-0) + 100 mM NaCl and 5 mM EDTA
Incubation of T. aquaticus superoxide dismutase ( 3 mg/ml) in 8 M urea containing 10 mM EDTA and 1 % acetic acid (pH 3.7) for up to 16 h at 23 "C led to the disappearance of the purple-red colour and to irreversible loss of activity (i.e. when a sample of the solution was diluted into neutral buffer containing 10 mM MnClz no activity was regained). The solution was gel-filtered on Sephadex G-25 in 8 M urea11 % acetic acid in the absence of EDTA, and subsequently dialysed successively against 25 mM Tris-HC1 (pH 7.5) with, and without 8 M urea. The resulting slightly turbid solution was centrifuged and the clear colourless supernatant was found to contain less than 0.01 mol of manganese per mol of protein, and to possess less than 10% of the activity of the native enzyme (Table 3). The apoprotein did not absorb at 478 nm. It was, however, indistinguishable from the native enzyme when it was analysed by acrylamide gel electrophoresis at pH 8.9 (Fig. 9) or by gel-filtration on Sephadex G-75 (Fig. 10) showing that the tetrameric structure had been reformed in the absence of manganese.
Stability in Denaturing Solvents
Reconstitution of Active Mn-Enzyme
Enzyme solutions (0.35 mgiml) in 50 mM TrisHCl, 1 mM EDTA (pH 7.5) containing 8 M urea or 6 M guanidinium chloride were found to be completely stable for periods of up to 16 h at 23 "C. The native tetrameric structure appears to be stable under
Reconstitution of active enzyme does not occur by addition of manganese chloride to the apoenzyme in buffer (or to buffer containing 8 M urea) at neutral pH. Total reconstitution of fully active enzyme was, however, found to occur following the exposure of
6t 0
1 20
40 60 Time (min)
+
80
100
+
319
S. Sato and J . I. Harris
t I I,
Fraction number Fig. 10. Gel-filtration oJ' T. aquaticus supero.uide disnzutase on a column of Sephadex G-100 (superfine grade, 2.4 x 50 cm) ; 3.3-ml fractions. (-0) Native enzyme (z300 Fg) blue dextran; (O----O)apoprotein ( z300 kg) blue dextran. Molecular weight, calculated according to Andrews [20], is z 80000 (cf: Fig. 5)
+
B
A
Fig. 9. Polyacrylarnide gel electrophoresis of T. aquaticus supero.xide dismutase. (A) Native Mn-enzyme; (B) apo-protein 7.5 % gel at pH 8.4 [17] stained with Coomassie brilliant blue
the apoprotein to manganese chloride in 8 M urea at acid pH. In a typical experiment the apoprotein (2 mg/ml) in 25 mM Tris-HC1 (pH 7.5) was dialysed successively for periods of at least 4 h at 4 "C against : (a) 8 M urea/l % acetic acid/lO mM MnC12 (pH 3.7), (b) 8 M urea/25 mM Tris-HCl/lO mM MnClz (pH 7.5), (c) 25 mM Tris-HCl/l mM MnClz (pH 7.5). Finally, the resulting purple-red solution was dialysed exhaustively against several changes of 25 mM Tris-HC1 containing 0.5 mM EDTA (pH 7.5). The extent of reconstitution was monitored by measurement of absorbance at 478 nm, manganese content, and regain of superoxide dismutase activity. As shown in Fig. 7 the absorption spectrum of the reconstituted enzyme is indistinguishable from that of the native enzyme; moreover as shown in Table 3 it contains 1.9 mol Mn per tetramer ( M , 84000) and is fully active when assayed by the pulse-radiolysis met hod. N- Terminal Sequence
Sequence analysis by automated Edman degradation was performed in a Beckman 890B sequencer using the standard Quadrol double-cleavage pro-
+
gramme. Native protein (8 mg, 400 nmol) was dried under vacuum in the reaction cup and degradations were commenced at the second acid-cleavage step so as to stabilise the protein film in the cup. Phenylthiohydantoins were identified by a combination of gas liquid chromatography, thin-layer chromatography and amino acid analysis following regeneration by hydrolysis with HI (cf. [4]). All residues were identified by at least two of these methods in each of two separate sequencer runs. Repetitive yield was calculated as 95% between cycle 6 (Leu-4) and cycle 16 (Leu-14). The sequence of 40 residues from the N-terminus was unambiguously determined and is given together with the corresponding N-terminal sequence of the E. coli Mn-enzyme [3] and B. stearothermophilus Mn-enzyme [4] in Table 4.
DISCUSSION Thermus aquaticus appears to contain only one active superoxide dismutase as evidenced by the appearance of a single band of enzymic activity (with an isoelectric point of 4.9) when a crude extract was examined by polyacrylamide disc-gel electrophoresis at pH 8.4. The appearance of two active fractions from DEAE-cellulose was caused by interference in the fractionation procedure by the viscous slime present in cell extracts of T aquaticus, rather than to the presence of two different active superoxide dismutases as has been found in other microorganisms [l]. Thermus aquaticus superoxide dismutase is a manganese-containing metalloprotein composed of
380
Superoxide Dismutase (Mn) from Thermus aquuricus
Table 4.N-terminal sequences of Mn-supero.de dismutases
Organism
Sequence
T aquaticuy
Pro-Tyr-Pro-Phe-Lys-Leu-Pro-Glu-Leu-Gly-Tyr -Pro -Tyr- Glu-AId -Leu-Glu-Pro- HIS-1le -Acp-Ah-Arg-Thr-
a
i
B strar othermophilus
Glu Ser -Tyr -Thr
E tolr ~~
~
T uquuticus
Pro
Ser
Pro
~
Lys Glu
ASP Ala
Asp
- _ _ ~
Phe _ _ _ _~
Lys Glu ~ ~ ___
25 10 35 40 Met-Glu-I le -His -His -Gln-Lys -H is -His -Gly-Ald-Tyr -Val -Thr-Asn-Leu-Asn-Ald
B stearothermophilus
Asn
Thr
E coli
Glu
1-
a
Ald
20
15
10
5
1
l
Asn-Thr
Ab
Insertion in T. uquuticus enzyme. Not determined incoli enzyme.
four identical polypeptide chains. In this respect it resembles the chicken liver mitochondria1 enzyme [22] and differs from the Mn-containing dimeric enzymes hitherto isolated from microorganisms such as E. coliB [9] and B. stearothermophilus [4]. In common with the latter two enzymes the T. aquaticus dismutase appears to contain one atom of manganese per two polypeptide chains. The tetrameric metalloenzyme is extremely stable over a wide range of pH (4-10) even in the presence of EDTA. It is also stable in dissociating solvents such as 8 M urea and 6 M guanidinium-HCI if the pH is maintained above 6.0. In order to dissociate the tetramer and to obtain a manganese-free apoprotein it is necessary to expose the native Mnenzyme to 8 M urea at acid ( < 4 ) pH. Removal of the dissociating reagent at neutral pH gave rise to a native tetrameric Mn-free apoprotein showing that the protein subunits are capable of forming a stable tetramer in the absence of manganese. Moreover, to reconstitute active Mn-enzyme it was found necessary to add the metal to apoprotein in acid 8 M urea suggesting that dissociation of the tetramer is an essential step, and that the metal may be shared between two subunits since the reconstituted enzyme, like the native enzyme, contained only two atoms of manganese per tetramer. This is the first time that it has proved possible to prepare a stable apoprotein from a Mn-containing superoxide dismutase, and to reconstitute fully active enzyme from it by addition of inorganic manganese. Similar results have since been obtained with the dimeric manganese-containing superoxide dismutase from another thermophile, B. stearothermophilus [22a]. It has not so far been possible to isolate a stable dimeric apoprotein from the E. coli Mn-enzyme although reversible metal exchange does occur in situ [23]. The colour and nuclear magnetic resonance properties of the E. coli enzyme suggest that the constituent manganese is trivalent in the resting state [24]. Since
reconstitution of the two thermophile Mn-enzymes was achieved following addition of divalent manganese it is assumed that Mn2+ is oxidised to Mn3+ during the process of reconstitution. The ability to prepare stable apoproteins and to reconstitute fully active Mn-containing enzymes from T. aquaticus and B. stearothermophilus dismutases opens new possibilities for the study of the specificity and mode of binding of the metal and of its role in catalysis. In common with other enzymes from T. aquaticus, the superoxide dismutase is extremely stable to heat. Thus it was not inactivated at temperatures of up to 95 "C, whereas enzyme from the more moderate thermophile, B. stearothermophilus, under the same conditions of heating was inactivated at 75 "C. Numerous attempts have been made to correlate differences in the amino acid compositions of proteins with their thermal stabilities. Comparison of the amino acid compositions of superoxide dismutase from T. aquaticus, B. stearothermophilus and E. coli does not, however, reveal any consistent trend. It is perhaps of interest that the thermophile enzymes do not contain cysteine residues that might be a source of instability in organisms that grow aerobically at temperatures in excess of 70 "C (cf. [12]). Comparison of the N-terminal sequences of the T. aquaticus, B. stearothermophilus and E. coli enzymes reveals a very high degree of sequence homology for enzymes isolated from such diverse genera. Over the first 26 residues 13 (50%) are common to all three proteins. The two thermophile enzymes are remarkably homologous, 30 of the first 40 residues (75 %) being common to the two proteins. There is no particular reason to suppose that B. stearothermophilus and T. aquaticus are closely related organisms. The former is a gram-positive spore-former whereas the latter is a gram-negative non-sporulating organism that could belong to any one of several orders of procaryotes. The significance of the N-terminal sequence homology between the
_
_
38 1
S. Sato and J. I . Harris
two Mn-dismutases must therefore await further study of their structures and of the relationship between three-dimensional structure and catalytic activity. We thank D r M. E. McAdam (Department of Physics, Institute of Cancer Research, Sutton) for superoxide dismutase assays by the pulse radiolysis method, Mr J. Herrington (Aldermaston Weapons Research Establishment, Aldermaston) for manganese estimations by neutron activation analysis, and Mr F. Northrop for .assistance in determining the N-terminal sequence.
REFERENCES 1. Fridovich, 1. (1975) Annu. Rev. Biochem. 44, 147- 157. 2. Fridovich, 1. (1974) Adv. Enzymol. 41, 35-97. 3. Steinman, H. M . & Hill, R. L. (1973) Proc. Nut/ Acad. Sci. U.S.A. 70,3725 - 3729. 4. Bridgen, J., Harris, J. I. & Northrop, F. (1975) FEBS Left. 49,392- 395. 5. Richardson, J. S., Thomas, K. A,, Rubin, B. H. & Richardson, D. C. (1975) Proc. Natl Acad. Sci. U . S . A . 72, 1349-1353. 6. Beem, K . M., Richardson, J. S. & Richardson, D. C. (1976) J . Mol. Biol. 105, 327 - 332. 7. Bridgen, J., Harris, J. I. & Kolb, E. (1976) J . Mol. B i d . 105, 333 - 335. 8. Puget, K . & Michelson, A. M . (1974) Biochem. Biophys. Res. Comnzun. 58, 830- 838.
9. Keele, B. B., Jr, McCord, J . M . & Fridovich, I. (1970) J . B i d . Cl7em. 245,6176-6181. 10. Suzuki, I. & Harris, J. I. (1971) FEBS L e f f . 13, 217-220. 11. Hill, H. A. O., Lobb, R . R., Sharp. S. L., Stokes, A. M., Harris, J . I. & Jack, R. S. (1976) Biochem. J . 153, 551 - 560. 12a. Hocking, J. D. & Harris, J. I. (1973) FEBS L e t / . 34, 280284. 12b. Hocking, J. D. & Harris, J . I . (1976) in Symposiumon €n:yrncs and Proteins from Thermophilic Microorganisnis (Zuber, H ., ed.) pp. 121- 133, Birkhauser-Verlag, Basel. 13. Hengartner, H . & Harris, J . I. (1975) FEBS L e t / . 55, 272-285. 14. Lowry, 0. H., Rosebrough, N . J., Farr, A. W. & Randall, R. J. (1951) J . Biol. Chem. 193, 265-275. 15. McCord, J. M. & Fridovich, I. (1969) J . Bid. Chmz. 244. 6049 - 6055. 16. Rotilio, G., Bray, R. C . & Fielden, E. M. (1972) Biocliim. Biophys. Acta, 268, 605 -609. 17. Davis, B. J. (1964) Ann. N . Y . Acad. Sci. 121, 404-427. 18. Laemmli, U..K. (1970) Nurure (Lond.) 227, 680-685. 19. Brock, T. D. & Freeze, H. (1969) J . Bucteriol. 98, 289-298. 20. Andrews, P. (1964) Biochem. J . 91, 222-233. 21. Edelhoch, H. (1967) Biochemistry, 6, 1948- 1954. 22. Weisiger, R. A. & Fridovich, I. (1973) J . B i d . Chem. 248, 3582- 3592. 22a. Brock, C. J., Harris, J. I. & Sato, S. (1976) J . Mol. Biol. 107. 175- 178. 23. Ose, D. E. & Fridovich, I. (1973) J . B i d . Chem. 251, 12171218. 24. Villafranca, J. J., Yost, F. J., J r & Fridovich, 1. (1974) J . Biol. Chem. 249,3532- 3536. 25. Beauchamp, P. C. & Fridovich, 1. (1971) Anal. Biocheni. 44. 276 - 287.
S. Sato and J. I. Harris, M.R.C. Laboratory of Molecular Biology, Postgraduate Medical School, University of Cambridge, Hills Road, Cambridge, Great Britain, CB2 2QH
