Regulation of manganese-containing superoxide dismutase in ...

2 downloads 0 Views 4MB Size Report
Dec 15, 2015 - richia coli (Moody, C. S., and Hassan, H. M. (1984) J. Biol. Chem. 259 ... a multicopy plasmid carrying the manganese-super- oxide dismutase .... Ultrasonics W-370 sonicator, equipped with a cup horn, operated at a 60-watt ...
THEJOURNAL OF BIOLOGICAL CHEMISTRY

Vol. 262, No. 35, Issue of December 15, pp. 17173-17177,1987 Printed in U.S.A.

Q 1987 by The American Society for Biochemistry and Molecular Biology, Inc.

Regulation of Manganese-containing Superoxide Dismutasein Escherichia coli ANAEROBICINDUCTION

BY NITRATE* (Received for publication, May 28, 1987)

Hosni M. HassanS and CarmellaS. Moody8 From the Departments of Food Science and Microbiology, North Carolina State University, Raleigh,North Carolinn 27695-7624

We have previously demonstrated that iron plays an involving the variationof the growth rate (4) and metabolism important regulatory role in the biosynthesis of man- (5) indicated that 0; is the inducing substrate and notdioxganese-containingsuperoxidedismutasein Esche- ygen. Furthermore, redox active compounds such as paraquat, richia coli (Moody, C. S . , and Hassan, H. M. (1984) J. which enter the cell, become reduced, and react rapidly with Biol. Chem. 259, 12821-12825). In this study, we dioxygen t o generate 0 2 , also increase the rateof biosynthesis demonstrated that the effect of ironis at the transcrip- of manganese-superoxide dismutase (6, 7). An exception to tional/translational level, whereas the effect of man- the rule of induction by 0; was seen when nalidixic acid, a ganese is at the post-translational level. The anaerobic compound incapableof generating 0 5 , caused a2-fold increase additions of nitrate or nitrate plus paraquat caused a in the level of the manganese-superoxide dismutase in E. coli positive change in the redox potential of the growth (8).Later studies in our laboratory (9) indicate that induction medium and concomitant induction of the manganese- by nalidixic acid was not related to its ability to inhibit DNA superoxide dismutase in the cells. By using “Fe, we biosynthesis but rather to its ability to chelate ferrous iron. were able to identify two unique proteins that were These findingsled us to study the role of iron in the regulation constitutively made,butcontainedirononlyunder of manganese-superoxide dismutase (10).Removal of iron conditions where the synthesis of manganese-superoxide dismutase was fully repressed. The presence of from thegrowth medium by Chelex 100 causedan increase in a multicopy plasmid carrying the manganese-super- the level of manganese-superoxide dismutase, and repletion oxide dismutase gene resulted in the anaerobic expres- of the medium with iron abolished the inductive effect. We further demonstrated that chelators specific for Fez+(i.e. 2,2’sion of the gene presumably by neutralizing the limited dipyridyl and 1,lO-phenanthroline) cause a greater induction number of repressor molecules found in the cells. In toto, the data support our previously proposed model of manganese-superoxide dismutase. A most striking and a novel finding was theinduction of manganese-superoxide for a negatively controlled operon where the repressor molecule is envisioned as an allosteric redox sensing dismutase in the absenceof molecular oxygen in response t o protein. Conditions known tooxidize or to deplete the Fez+chelation.Thesefindings led us to propose that the synthesis of manganese-superoxide dismutase in E. coli is ironare also found to causeinductionofthemanganese-superoxidedismutasealbeittheabsence of regulated by a negative control operon, where the repressor protein (RP)’ contains iron (10). This iron-containing represdioxygen. sor may exist either in the ferric (RP-Fe3+) or the ferrous (RP-Fez+) state dependingon the redox state of the cell. We proposed that the activerepressor contains Fez+,whereas the Superoxide dismutases area ubiquitous classof metalloen- inactive repressor contains either Fe3+or noiron. Accordingly, zymes found in most oxygen-exposed organisms (1,2). These in the absence of dioxygen, the active form of the repressor enzymes protectcells against the toxicityof the first product will be predominant andno manganese-superoxide dismutase of univalent reductionof oxygen, the superoxide radical(0;). will be made. On the other hand, conditions known to oxidize There are three isoenzymatic forms of superoxide dismutase Fez+ to Fe3+ (i.e. oxygen or oxyradicals) or to remove Fez+ in Escherichia coli (3). Under normal growth conditions, an from the cells (ie.chelators) will generate the inactive forms iron-containing superoxide dismutase is synthesized both aerof the repressor and will result in the synthesis of the enzyme, obically and anaerobically, whereas a manganese-containing even in the absence of dioxygen. A similar study confirmed isoenzyme and a hybrid-isoenzyme are synthesized only under our findings that iron chelatorsinduce the manganese-superaerobic conditions. oxide dismutase in E. coli, but the data were explained by The regulation of the manganese-superoxide dismutase in another hypothetical model based on autogenous regulation E. coli has been extensively explored. Physiological studies (11). In this report,we present data that substantiate our previ* This work was supported in part by Grant DMB-8609239 from ously proposed model and suggest that theredox state of the the National Science Foundation. This is Paper 11129 of the Journal cell plays an important role in the regulation of manganeseSeries of the North Carolina Agricultural Research Service, Raleigh, superoxide dismutase in E . coli.

NC 27695. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ T o whom correspondence should be addressed Dept. of Food Science and Microbiology, North Carolina State University, Box 7624, Raleigh, NC 27695-7624. 8 Present address: Pharmacology Dept., Duke University Medical Center, Durham, NC 27710.

MATERIALS AND METHODS Chemicals-Ferrous ammonium sulfate, manganese sulfate, and potassium nitrate were purchased from Fisher. Methyl viologen (pal The abbreviations used are: RP, repressor protein; TSY, trypticase-soy yeast extract; GMM, glucose minimal medium; PQ*+,paraquat.

17173

Anaerobic Induction of Manganese-Superoxide Dismutase by Nitrate

17174

raquat, l,l'-dimethyl-4,4'-bipyridinium dichloride) was purchased from Sigma. Ampholytes used for isoelectric focusing were purchased from Bio-Rad. Radiolabeled ferrous sulfate (59Fe,9.91 mCi/mg) and manganese dichloride ("n, 21.31 mCi/mg) were purchased from Du Pont-New EnglandNuclear. Bacterial Strains-E. coliK12AB2463 (pHC79) and AB2463 (pDT1-5) (12) (a generous gift of J. S. Siwecki, North Carolina State University, Veterinary School) and E. coli BB12(ATCC 29682) were used throughout this study. Media and Growth Conditions-A glucose minimal medium (GMM) containing, per liter, 0.2 g of MgSOI.7H,0, 2.0 g of citric acid/HzO, 10 gof K2HP04,3.5 g of NaNH4HP04,0.02 mg of vitamin B12,and 5.0 g of glucose was used to grow the overnight cultures (4). A trypticase-soy yeast extract (TSY) medium containing, per liter, 30 g of trypticase broth (Difco) and 5.0 g of yeast extract (Baltimore Biological Laboratories) was used throughout except where otherwise indicated. Overnight cultures grown in glucose minimal medium for 15-17 h were used to inoculate TSY medium to an initial optical density a t 600 nm equal to 0.05. The cultures were allowed to grow for one doubling, before any treatments, toensure they were in the logarithmic phase of growth. To maintain sufficient aeration, liquid cultures were shaken at 200 rpm where the flask-to-culture volume ratio was 51. Anaerobic growth experiments and protocols were carried out in a Coy anaerobic chamber (Coy Laboratory Products,Ann Arbor, MI) as previously described (10). Chloramphenicol (150 pg/ml) was added to the cultures at the end of the specified anaerobic growth period, and the cultures were allowed a further 15-min incubation before they were removed from the anaerobic chamber and poured onto crushed ice. The addition of chloramphenicol and the quick cooling step were essential inorder to prevent protein synthesis and induction of superoxide dismutase by oxygen. 59Fe (2.5 pCi/ml) and =Mn (3.67 pCi/ml) were added as indicated. In all cases, cells were allowed to grow for four to five generations before they were harvested and assayed. Assays-Cells were harvested by centrifugation a t 4 "C and 10,000 X g for 20 min. The cells pellets were resuspended in 0.05 M potassium phosphate buffer, pH 7.8, containing 0.1 mM EDTA (KPJEDTA buffer) and disrupted by sonication for 6 min using a Heat Systems Ultrasonics W-370 sonicator, equipped with a cup horn, operated a t a 60-watt output. Temperature was maintained a t 4 "C by using a circulating refrigerated water bath, and thesonifier power was applied intermittently a t 45-s pulses. Unbroken cells and cell debris were removed by centrifugation a t 4 'C and 27,000 X g for 30 min. Cellfree extracts were dialyzed for 48 h against three changes of KPi/

a ' 1

2

h -

B

A

3

4 " 5

6

7

RESULTS

Induction of Manganese-Superoxide Dismutase: 59Fe/ S4Mn-Growth of E. coli in the presence of 0.2 mM paraquat, an intracellular generatorof 02 (6,7), resulted in asignificant increase in thespecific activity of total superoxide dismutase with most of the increase being due to the manganese- and hybrid-isoenzymes (Fig. la, lanes 1 and 5 versus 2 and 6). The observed increase in manganese-superoxide dismutase activity was due to increased protein biosynthesis as shown in Fig. l b (lanes 1 and 5 versus 2 and 6). These data indicate de novo induction of the manganese-enzyme in response to the addition of paraquat as previously concluded from studies using inhibitors of protein biosynthesis (7, 10). The addition of excess Mn2+( 1 mM) resulted in a 10% increase in the activity of manganese-superoxide dismutase (Fig. la, lane 1 versus 3; and data notshown), whereas the addition of excess Mn2+in the presence of paraquat caused a 25% increase in the activities of the manganese- and hybrid-isoenzymes (Fig. la, lane 2 versus 4; and data not shown). On the other hand, the addition of excess iron ( 1 mM) resulted in more than a 20% decrease in theactivity of the manganese-isoenzyme (Fig. la, lanes 5 and 6 versus 7 and 8). In the presence of 0.2 mM paraquat, the addition of excess Mn2+resulted in acompacted protein band (Fig. lb, lanes 2 versus 4 ) , whereas the addition of excess iron caused a significant decrease in the intensity of B

A

' I

8 '

EDTA buffer. Dialysis of cell-free extracts from samples labeled with 59Feor %Mnwas achieved by using an Amicon Centricon-10 (Amicon Corp.). Protein was assayed by the method of Lowry et al. (13), using bovine serum albumin as astandard. Superoxide dismutase was assayed by the cytochrome c method (14). Superoxide dismutase isoenzymes and proteins were separated by electrophoresis on 10% polyacrylamide gels (15). Superoxide dismutases were visualized in the gels by using an activity stain (16) and were quantitated by linear scanning densitometry (3). Gels were stained for protein using Coomassie Brilliant Blue R-250. Isoelectric focusing at pH 5-7 in polyacrylamide gels was performed according to Righetti and Drysdale (17). Radiolabeled gels were dried under vacuum and subjected to autoradiography a t -70 'C with Kodak XAR-5 film and a Cronex Lightning Plus intensifying screen (Du Pont-New England Nuclear).

2

3

4"5

6

7

8'

" M n -

C

Mn

-Hy

- HY - Fe

I

A B I 2 3 4"5 6 7 8 '

- Fe -

-

*-Mn

FIG. 1. Effectsof excess MnCl, and Fe(NH,),SO, on the biosynthesis of superoxide dismutase isoenzymes in E. coli (ATCC 29682). Overnight (15 h) cultures grown in the appropriate GMM were used to inoculate flasks containing TSY medium (20 ml) supplemented as follows. Radiolabeled iron (59Fe)was added to group A (lanes 1 4 ) , whereas radiolabeled manganese ("Mn) was added to group B (lanes 5-8). Excess MnC1, (1 mM) was added to lanes 3 and 4, whereas excess Fe(NH4),S04(1 mM) was added to lanes 7 and 8. Paraquat (0.2 mM) was added to lanes 2, 4, 6, and 8. Lanes I , 2, 5, and 6 were inoculated with overnight culture grown in GMM, lanes 3 and 4 were inoculated with culture grown in GMM plus MnC12 (1 mM); and lanes 7 and 8 were inoculated with culture grown in GMM plus Fe(NH,),SO, (1mM). All cultures were allowed to grow logarithmically for four generations before dialyzed cell-free extracts were prepared and assayed for superoxide dismutase and protein contents. Samples containing 50 pg of protein were electrophoresed using 10% polyacrylamide slab gels (1.5 mm thick). a, gel stained for superoxide dismutase activity; b, gel stained for protein using.Coomassie Brilliant Blue R250; c, unstained gel dried and exposed to x-ray film for 13 ha t -70 OC. Hy, hybrid.

.- H - Fe

Anaerobic Induction of Manganese-Superoxide Dismutase by Nitrate TABLE I Effectsof nitrate and paraquat on the anaerobic biosynthesis of mnnganese-superoxide dismutase Anaerobic overnight cultures of E. coli (ATCC 29682) grown in TSY medium with and without KN03 (0.1 M) were used to inoculate TSY medium containing the specified chemicals. Paraquat (1 mM), KN03 (0.1 M), or the combination of PQ" and NO: was added when the cultures reached ODsoo= 0.1. The cultures were allowed to grow for four generations before final measurements were made, and dialyzed cell-free extracts were prepared and assayed as described under "Materials and Methods." E,, was measured using an Ingold Electrodes redox Drobe connected to Accumet Model 620 DHmeter. Final Final Final OD, pH Eo

Conditions

TSY TSY + PQ" TSY NO; TSY + PQz++ NO;

+

1.0

6.0 0.95 6.0 1.47 6.17 0.82 6.28

mV -325 -330 -234 -226

Superoxide dismutase Man- Hybrid Iron ganese unitsfmg

7.3 7.0 16.6 19.1

0 0 0 1.0

0 0 1.5 1.9

7.3 7.0 15.1 16.2

the protein band (Fig. lb, lane 6 versus 8 ) . These results may indicate that the presence of excess iron affects the de novo synthesis of the manganese-isoenzyme. Data in Fig. ICshow that iron (59Fe)is present in all three isoenzymatic forms of superoxide dismutase, whereas manganese ("n) is found mainly in the manganese- and hybridisoenzymes with no or very little 54Mn found in the ironsuperoxide dismutase. Induction of manganese-superoxide dismutase by paraquat caused a large increase in the amounts of 59Feand 54Mnincorporated in the protein (Fig. IC,lanes 2 and 6, respectively). In thepresence of paraquat, the addition

a ' 1

B

A 2 '

' 3

4

b 5'

A 11

-Mn

"HY - Fe

of excess unlabeled Mn2+ ( 1 mM) resulted in a significant decrease in the amount of 59Feincorporated in the manganeseisoenzyme (notice the disappearance of one 59Fe band;Fig. IC, lane 2 versus 4 ) . At the same time, the addition of excess iron ( 1 mM) resulted in asignificant decrease in 54Mnincorporated in the manganese-superoxide dismutase (Fig. IC,lane 6 versus 8). It is clear, therefore, that under aerobic conditions, in the absence or in the presence of PQ2+,both manganese and iron compete for the active site of manganese-superoxide dismutase, whereas the manganese does not compete with the iron for the active site of iron-superoxide dismutase. Effect of Nitrute-The redox status of the cell is influenced by the concentration of dissolved oxygen, the presence of redox cycling compounds (e.g. paraquat and quinones), the NAD(P)H/NAD(P)+ ratio, and a number of cellular reductants and enzymes which catalyze redox changes during cellular metabolism. The model proposedby Moody and Hassan (10) suggests that an iron-containing protein regulates the synthesis of manganese-superoxide dismutase in E. coli. The state of the iron in thisregulatory protein (i.e. Fe2+:Fe3+ ratio) will reflect the redox status of the cell. Therefore, the presence or absence of oxygen (as well as other oxidants) couldbe reflected by the relative proportions of ferric and ferrous species of a protein whose conformation changes with the oxidation state of the iron in a fashion similar to that found in hemoglobin. In this part of the study, we investigated the effect of nitrate, which can be usedanaerobically by E. coli as an electron acceptor, on the biosynthesis of manganese-superoxide dismutase. The reduction of nitrate to nitrite has a midpoint potential of +420 mV at pH 7.0 (18). We also included paraquatto serve as an electron sink andasa

B 21'3

17175

4

5

C

'

-

B

A

'1

2 "3

4

5

'

" "

lMn

F

e e

4-

4-

- HY 4- Fe FIG.2. Identification of uniqueiron-proteins. An overnight culture of E. coli (ATCC 29682)grown aerobically in GMM was used to inoculate aerobic media (groupA ) : lane 1,TSY, lane 2, TSY plus 0.2 mM paraquat. An anaerobic overnight culture was used to inoculate anaerobic media (group B ) : lane 3, TSY; lane 4, TSY plus 0.1 M KNO3; lane 5,TSY plus 0.1 M KNO, and 1.0 mM paraquat. Cells were allowed to grow for four generations in the presence of radiolabeled 'Te (2.5 pCi/ml) before dialyzed cell-free extracts were prepared and assayed as and iron-superoxide dismutases were identified described for Fig. 1. The locations of manganese-, hybrid (Hy)-, from gels run under identical conditions and stained for enzyme activity. a, autoradiogram showing samples (50 pg of protein/lane) electrophoresed on 10% polyacrylamide gels which weredried and exposed to x-ray film for 21 h a t -70 "C; b, the same samples shown in a subjected to isoelectric focusing at pH5-7 in polyacrylamide gels and stained for protein using Coomassie Brilliant Blue; c, an autoradiogram of the gel shown in b.

17176

Anaerobic Induction of Manganese-Superoxide Dismutase by Nitrate

caused a %fold increase in the amount of iron-superoxide dismutase. This phenomenon is under further investigation. In Search ofa Unique Iron-containing Protein--In this part of the study, we used 50Feto label and examine the patterns of the different cellular iron-proteins produced by E. coli when grown under different growth conditions known to affect the synthesis of manganese-superoxide dismutase. We reasoned that RP-Fe” is stable and persists under anaerobic growth conditions, whereas RP-Fe3+ is unstable and dissociates to RP plus Fe3+ under aerobic or inducing growth conditions. Fig. 2a shows that cell-free extracts prepared from cells grown under different inducing/noninducing conditions contained several iron-proteins. Under anaerobic conditions, however, I two iron-proteins ( R , = 0.52 and 0.59) wereseen (Fig. 2u, lune 3 ) that were not seen in extracts prepared from aerobically grown cells (Fig. 2u, lanes 1and 2 ) . Furthermore, the intensity of 59Fe in these two iron-protein bands decreased when the manganese/hybrid-superoxide dismutases were induced anaerobically by nitrate or nitrate plus paraquat (Fig. 2u, lanes 4 and 5). Thesame samples were also subjected to isoelectric focusing/polyacrylamide gel electrophoresis at pH 5-7, and the gels were stained for proteins using Coomassie Brilliant Blue (Fig. 2b). The dried gels were exposed to x-ray films (Fig. 2c) to identify the protein bands that possessed labeled 59Fe. Upon examination of the autoradiograms (Fig. 2c), two iron-proteins were noticed (indicated by arrows) that lost their 59Fe upon exposing the cells to inducing conditions. When these gels were compared to the same gel stained for protein, it became clear that these two proteins were present under all growth conditions (Fig. 2b, bands indicated by arrows). The results indicate that these two unique proteins FIG.3. Densitometric scans showing the anaerobic induc- were made constitutively but contained iron only under antion of manganesebybrid (Hy)-superoxide dismutases in a strain containing a multicopy plasmid for the sodA gene. An aerobic (repressed) conditions. These two iron-proteins are anaerobic overnight culture of E. coli AB2463 (pDT1-5)was used to good candidates for the proposed regulatory protein(s), and inoculate TSY medium k 0.2 M KNO, ( A ) or TSY medium supplefurther studies areunderway to isolate and characterize them. mented with 1mM MnClz & 0.2 M KNO, ( B ) .Ampicillin (250 sg/ml) Neutralization of the Repressor Protein-Recently, the was added t o all growth media. The cultures were allowed to grow manganese-superoxide dismutase gene (sodA)has been cloned anaerobically for four generations before dialyzed cell-free extracts were prepared, electrophoresed on 10%polyacrylamide gels, (constant (12, 20) and sequenced (20). The DNAsequence indicates protein, 100 pg/gel), stained for superoxide dismutase activity, and that the gene constitutes a single normal operon having two possible promoters, but only one of them appearsactive under scanned using a Gilford 2600 spectrophotometer equipped with gel scanner. Upper trace, cells grown in TSY medium; lower trace, cells normal growth conditions (20). In this experiment, we reagrown in TSY medium plus NO,. soned that if the sodA gene is negatively regulated by a transacting element capable of binding the operator site, then the mediator for donating electrons to nitrate reductase (19). presence of a multicopy plasmid carrying the sodA gene may Data in Table I clearly demonstrate that under anaerobic titrate and neutralize the amount of the repressor protein conditions nitrate alone was able to induce the hybrid-super- present in the cells and therefore result in the induction of oxide dismutase. This indicated that the manganese-super- manganese-superoxide dismutase anaerobically. In this exoxide dismutase gene (sodA) was turned on and some of the periment, we used the strain AB2463 (pDT1-5) to determine manganese-superoxide dismutase subunits were made and if the presence of a multicopy plasmid containing the sodA interacted with subunits of the iron-isoenzyme to form the gene can neutralize the repressor protein. Strain AB2463 hybrid-isoenzyme. The addition of 1 mM paraquat plus 0.1 M (pHC79) was used as a control. The data in Fig. 3 clearly nitrate further promoted the formation of manganese-super- show that strain AB2463 (pDT1-5) synthesized the hybridoxide dismutase anaerobically (Table I). It is interesting to superoxide dismutase anaerobically and that this strain was note that the addition of nitrate stimulated the anaerobic very responsive to the presence of nitrate. Furthermore, the growth yield by -47% over the fermentative metabolism and addition of excess manganese (1mM) augmented the activities also raised the final pH by 0.17 unit. This was probably due of manganese- and hybrid-superoxide dismutases. The same to the utilization of organic acids via the nitrate-dependent effects were not seen in strainAB2463 (pHC79), which carries the same plasmid but not sodA (12). electron transportchain (18). We also noted that NO, changed the final redox potential of the medium by +91 mV. DISCUSSION The addition of paraquat plus NO, changed the redox by approximately +lo0 mV and raised the pH by 0.28 unit. It is The datapresented support our previously proposed model also interesting to note that thepresence of paraquat signifi- (10) for the role of iron in the regulation of manganesecantly reduced the final cell yield, which ismore probably due superoxide dismutase biosynthesis in E. coli. The increase in to thediversion of electrons from the electron transport chain the activity of manganese-superoxide dismutase caused by the by PQ’+ and the direct donation of the electrons to nitrate addition of PQ’+ was associated with a parallel increase in its reductase (19). Table Ialso shows that thepresence of nitrate protein. The addition of excess manganese resulted in more

Anaerobic Induction of Manganese-Superoxide Dismutase

by Nitrate

17177

The most convincing argument in support of a negatively compacted protein/activity bands of manganese-superoxide dismutase, in contrast to excess iron which caused a signifi- controlled operon (10) comes from recent studies by Touati cant decrease in the intensity of protein/activity bands of this and Carlioz (24) in which they found that P-galactosidase is enzyme. These results may suggest that the manganeseeffect induced by paraquat ina straincarrying sodA::lac fusion. is at the post-translationallevel, whereas the ironeffect is a t These data clearly demonstrate that the induction of manthe transcriptional/translational level. The dataclearly dem- ganese-superoxide dismutase by paraquat is at the transcriponstrate that 59Fe bindstoallthedismutase isoenzymes, tional/translational level. The exact identity of the transwhereas 54Mn is more specific for manganese/hybrid-isoen- acting regulatory element that responds tochanges in flux of zymes. The autoregulatory model (11)states that 0; oxidizes OF, Fe2+concentration, and the redox state of the cell is not Mn2+ to Mn3+ butreduces Fe3+ to Fe2+ and that Mn3+ hasa yet known.However, we were able to identify two unique candidate iron-proteins that may qualify for the repressor greater affinity than Fez+ for the apoproteins of both ironfunction. We found that thesetwo proteins were loaded with and manganese-superoxide dismutases. Accordingly, in the 59 Fe only under repressive conditions (i.e. anaerobically), but presence of 0; generator ( i e . paraquat), one expects to see lost their iron under inducing conditions ( i e . aerobically or more 54Mn in both the iron/manganese-isoenzymes. On the contrary, our data showed that in the presence of paraquat in the presence of NO;/PQZ') (Fig. 2). These two candidate repressor proteins were made constitutively under all growth more 59Fe was incorporated into manganese-superoxide disconditions; however, they contained 59Fe only anaerobically. mutase,whereasno 54Mnwas incorporated into the ironThere was also a progressive decrease in 59Fe of these two isoenzyme (Fig.1).The data also indicate that under inductive stress ( i x . inthepresence ofPQ") the cells make more proteins as thecells were subjected to progressively inducing manganese-apoprotein and that iron can bind to it to yield conditions ( i e . NO, alone uersus NO; plus PQ"). Work is in regulatory an inactive enzyme. However, the presence of excess man- progress to isolatecells withmutationsinthe gene(s) which will enable us to characterize further the reguganese can reduce the amountof the inactiveenzyme by more lation of manganese-superoxide dismutase. effectively competing for the active site than iron. FurtherIn conclusion, the repressor protein is envisioned as an more, the inactive form of the manganese-enzyme ( i e . the allosteric redox sensing protein that can existin at least two one containing iron) appeared not to havea regulatory funcconformations: an active form (RP-Fe2') which can bind to tion, as evidentby the fact that its abundant presence did not the operator and an inactive form (RP-Fe3' or RP) which stop the inductionby paraquat (Fig. 1). cannot bind the operator. RP-Fe3' is unstable anddissociates The anaerobic induction of manganese/hybrid-superoxide to RP plus Fe3+. The presence of oxygen or other oxidants dismutases by compounds like nitrate and nitrate plus para- (21) canchangethe valence of theironandthus cause quat (Table I) or by other oxidants (21)2 may indicate that conformational changes in therepressor, which decreases its the regulation of the sodA gene is dependent on the redox affinity for theoperator.Ontheotherhand, ferrous iron state of the cell. Therefore, we propose that the iron present chelators (9, 10) can remove the iron from the repressor and in the repressor protein acts asa redox sensing device which cause conformational changes as well. The exact mechanism enables the cell to switch on the synthesis of manganese- for the regulation of manganese-superoxide dismutase is cursuperoxidedismutase even inthepresence of avery low rently under investigation. concentration of oxygen (3). REFERENCES The amount of any given repressor protein made by the 1. I. (1975) Annu. Reu. Biochem. 4 4 , 147-159 Fridovich, For example, the wild-type E. coli is usuallyverysmall. 2. Hassan, H. M.,and Fridovich, I. (1980) in Enzymatic Basis ojDetoxication concentration of the Lac repressor is estimated to be in the (Jakoby, W. B., ed) Vol. 1, pp. 311-332, Academic Press, New York Hassan, H. M., and Fridovich, I. (1977) J . Bacteriol. 1 2 9 , 1574-1583 order of lo-' M (22,23). Thus,a wild-type E. coli carrying one 3. 4. Hassan, H. M., and Fridovich, I. (1977) J. Bacteriol. 130,805-811 repressor-producing i gene would contain 10-20 Lac repressor 5. Hassan, H. M., and Fridovich, I. (1977) J. Bacteriol. 1 3 2 , 505-510 6. Hassan, H. M., and Fridovich, I. (1977) J. Biol. Chem. 2 5 2 , 7667-7672 molecules. By analogy, we assumed that the cell contains a 7. Hassan, H. M.,and Fridovich, I. (1979) Arch. Biochem. Biophys. 196,38584.5 -_small number of repressor molecules for the sodA gene and and Fridovich, I. (1980) Abst.Annu.Meet.Am. SOC. 8. Hassan,H.M., testedthe effect of thepresence of a multicopy plasmid Microbiol. 142, K96 9. Hassan, H. M., and Moody, C. S. (1984) FEMS Microbiol. Lett. 2 5 , 233carryingthe sodA gene ontheexpression of manganese236 superoxide dismutase in the absence of oxygen (Fig. 3). The 10. Moody, C. S., and Hassan, H. M. (1984) J. Biol. Chem. 2 5 9 , 12821-12825 Pugh, S. Y. R., and Fridovich, I. (1985) J. Bacteriol. 1 6 2 , 196-202 11. results showed thatstrain AB2463 (pDT1-5) was ableto 12. Touati, D. (1983) J. Bacteriol. 1 5 5 , 1078-1087 express hybrid-superoxide dismutase when grown anaerobi- 13. Lowrv. 0.H.. Rosebrouch. N. J.. Farr. A. L.. and Randall. R. J. (1951) J. Biil: Chem: 193,265-275 ' cally. These data indicate that the sodA gene was partially 14. McCord, J. M., and Fridovich, I. (1969) J. Biol. Chem. 2 4 4 , 6049-6055 induced anaerobically since the hybrid enzyme is madeof one 15. Davis. B. J. (1964) A n n . N . Y. Acad. Sci. 121,404-427 Beauchamp, C., and Fridovich, I. (1971) Anal.. Biochem.44,276-287 subunit of each of the manganese- and iron-isoenzymes. These 16. 17. Righetti, P. G., and Drysdale, J. W. (1976) in Laboratory Techniques in Biochemistry and Molecular Biology (Work, T. S., and Work, E., eds) data also suggest that the multicopy plasmid (12) contained Vol. 5, pp. 341-590, North-Holland Publishing Co., Amsterdam the operator site that was able to neutralize the effect of a 18. Ingledew, W. J., and Poole, R. K. (1984) Microbiol. Reu. 4 8 , 222-271 trans-acting element. Recent DNA sequence analysis of the 19. Jones, R. W., and Graland, P. B. (1977) Biochem. J. 1 6 4 , 199-211 20. Takeda, Y., and Avila, H. (1986) Nucleic Acids Res. 1 4 , 4577-4589 sodA gene (20) has indeed demonstrated the presence of an 21. Schiavone, J. R., and Hassan, H. M. (1987) Abst. Annu. Meet. Am. Soe. Microbiol. 8 7 , K163 almost perfect 19-base palindrome at the -35 region which 22. Muller-Hill, B. (1975) Prog. Biophys. Mol. Biol. 30, 227-252 represents a potential binding site. 23. Muller-Hill, B., Beyreuther, K., and Gilbert, W. (1971) Methods Enzymol. '

J. R. Schiavone and H. M. Hassan, manuscript submitted for publication.

'

21D,483-487 24. Touati, D., and Carlioz, A. (1986) in Superoxide and Superoxide Dismutase in Chemistry, Biology and Medicine (Rotilio, G., ed) pp. 287-292, Elsevier Scientific Publishing Co., Amsterdam