A Mechanism for Complementation of the sodA sodBDefect in Escherichia coli by Overproduction of therbo Gene Product (Desulfoferrodoxin) fromDesulfoarculus baarsii *

Overexpression of rbo inEscherichia coli prevents the inactivation of the [4Fe-4S]-containing fumarases that otherwise occurs in the sodA sodB strain. It similarly protects against the increased sensitivity toward H2O2, which is imposed by the lack of SOD A and SOD B. These results would be explained on the basis of scavenging of O·̄2 within the cells by RBO. This interpretation was supported by measurements of intracellular scavenging of O·̄2 by the lucigenin luminescence method. Since SOD activity could not be detected in dilute extracts, of the RBO-overexpressing sodA sodB strain, we propose that RBO catalyzes the reduction of O·̄2 at the expense of cellular reductants such as NAD(P)H. A similar mechanism may apply to other instances of complementation of SOD defects by non-SOD genes.

. and thereby protect aerobic cells against the toxicity of this radical (1). The potential deleterious actions of O 2 . are manifold and include oxidation of the [4Fe-4S] clusters of dehydratases such as aconitase. This interferes with metabolism by inactivating these dehydratases, and it has the additional consequence of causing release of iron from the oxidized clusters (1)(2)(3). The released iron is then available to participate in Fenton chemistry with hydroperoxides, resulting in the production of dangerously reactive hydroxyl or alkoxyl radicals or of strongly oxidizing iron-oxygen complexes (2,3). When this Fenton chemistry is catalyzed by iron bound to DNA or to membranes, the result is apt to be a loss of viability (2)(3)(4)(5)(6). That this actually occurs in these cells has been both proposed (2,3) and verified in Escherichia coli (4,6). Pianzzola et al. (7) have reported that insertion and expression of the rbo gene from Desulfoarculus baarsii or Desulfovibrio vulgaris into sodA 1 sodB E. coli corrects the phenotypic defects of this mutant. They achieved this by coupling the rbo gene to the IPTG-inducible ptac promoter. There are two ways in which the RBO protein (desulfoferrodoxin) might have achieved this complementation. One is by catalyzing the elimination of O 2 . and another is by increasing the rate of reconstitution of the oxidized iron sulfur clusters thereby increasing the activities of the corresponding enzymes and decreasing the level of "free" iron. Since SOD activity could not be detected in extracts of the sodA sodB ϩ rbo strains, the second of these explanations was put forth (7). We now report that expression of RBO in E. coli does provide for the elimination of O 2 . within these cells as judged from the luminescence of lucigenin, and we propose how this might be accomplished.

EXPERIMENTAL PROCEDURES
Materials-Cytochrome c (III) and xanthine were from Sigma; glucose, H 2 O 2 , and salts were from Mallinckrodt Chemical Works; malate was from ICN; lucigenin from Aldrich; and yeast extract and bactotryptone from Difco. Xanthine oxidase from bovine cream was prepared by R. Wiley as described by Waud et al. (8). SOD activity was assayed by the xanthine oxidase-cytochrome c method (9). Fumarase was assayed according to Hill and Bradshaw (10), and protein by the method of Lowry et al. (11). E. coli were grown, extracted, and assayed for fumarases A ϩ B essentially as described previously (12).
The lethality of 2.5 mM H 2 O 2 was explored by using 30 min of exposure followed by dilution and plating essentially as described by Carlioz and Touati (13). Lucigenin luminescence was used to measure the scavenging of O 2 . within E. coli as described previously (14). Controls demonstrated that this luminescence was dependent upon the simultaneous presence of lucigenin, the E. coli cells, and an electron source such as glucose. Lucigenin was added, to 0.1 mM, to samples taken from growing cultures, and the luminescence was measured and expressed per A 600 nm . In some experiments 2.5% inocula were grown for 15 min and then for an additional 90 min Ϯ 2.0 mM IPTG. The cells were then collected, washed once with 50 mM potassium phosphate, 0.1 mM EDTA at pH 7.8 and then resuspended in this buffer to A 600 nm ϭ 5.0. These cells were then used for luminescence measurements at A 600 nm ϭ 0.50 in this buffer plus 0.25% glucose.
Growth of Cells-Cultures were stored on sealed LB agar plates containing the selective antibiotics at 4°C. Inocula taken from these plates were grown overnight at 37°C in LB medium containing the selective antibiotics. Aeration was maintained by shaking at 200 rpm. Cultures intended for experimental manipulation were subcultured in LB medium without antibiotics. The strains used were: QC2509 ϭ GC4468/pJF119EH; QC2510 ϭ GC4468 ⌬sodA sodB ⌬2/pJF119EH; QC3235 ϭ GC4468 ⌬sodA sodB ⌬2/pMJ25; QC3245 ϭ GC4468 sodA::cat sodB ⌬2/pJF119EH; QC3246 ϭ GC4468 sodA::cat sodB ⌬2/ pMJ25. GC4468 was the parental strain, and pJF119EH was the control plasmid, whereas pMJ25 was the rbo-containing plasmid. These strains and the antibiotics used in maintaining them in culture have been described (7,13). . both in vitro (15) and in vivo (16). Fig. 1 demonstrates that these fumarases are largely inactivated in the sodA sodB strain (compare bars 1 and 2). This was partially prevented when the rbo-bearing plasmid was present (bar 3) and was completely prevented when rbo expression was induced by IPTG (bar 4). A control demonstrated that IPTG had no effect on the fumarases A and B of the sodA sodB strain lacking rbo (data not shown). These results can be interpreted * This work was supported by grants from the Council for Tobacco Research-U.S.A., Inc., U. S. Army Medical Research, and the National Institutes of Health. 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.

Effect of RBO on Fumarase Activity-Fumarases
‡ To whom correspondence should be addressed. Tel.: 919-684-5122; Fax: 919-684-8885. 1 The abbreviations used are: sodA, the gene coding for the MnSOD; sodB, the gene coding for the FeSOD; SOD, superoxide dismutase; RBO, gene product of the rbo gene (desulfoferrodoxin); IPTG, isopropyl-1-thio-␤-D-galactopyranoside; LB, Luria-Bertani.  (18,19). Lucigenin luminescence has been used in this way to measure SOD activity in E. coli (14). Fig. 3 demonstrates that expression of RBO increased the rate of scavenging of O 2 . in E. coli. Thus the luminescence seen with the SOD-competent parent was much less than that seen with the sodA sodB mutant (compare bars 1 and 2). When rbo was present in the sodA sodB strain it decreased that luminescence (bar 3), and the increasing expression of rbo with IPTG decreased it further (bar 4). Comparison of bars 2 and 4 in Fig.  3 shows that IPTG had no effect in the absence of rbo. These experiments were done with cells in LB medium. They were then repeated with cells harvested from LB and then resuspended in glucose plus buffer to A 600 nm ϭ 0.5. These results shown in Fig. 4  and that activity can be observed in the absence of other reac-

FIG. 1. Effect of RBO expression of fumarases A and B.
Overnight cultures were diluted 40-fold in LB. After 15 min IPTG was added where indicated, and growth was continued for 105 min. Cells were collected, and extracts of the various strains were prepared and assayed for the unstable fumarases A and B as described under "Experimental Procedures." Bar 1, wild type (QC2509); bar 2, sodA sodB plus control plasmid (QC2510); bar 3, sodA sodB plus rbo-containing plasmid (QC3235); bar 4, as in bar 3 but cells were exposed to 2.0 mM IPTG for 105 min prior to extraction.   (20) This manganic porphyrin was actively taken into E. coli where it was kept reduced by cellular reductants. It complemented the sodA sodB strain and diminished the luminescence of lucigenin even though its activity could not be seen in extracts (14,20,21). It is possible that this sort of scavenging of O 2 . , by other than a dismutation reaction, may be used in organisms reportedly aerotolerant yet lacking in SOD. Neisseria gonorrhoea is one such organism (22). A similar explanation might apply to pseudoreversions of SOD negative mutants of yeast (23)(24)(25) and of other organisms and to the complementation of the sodA sodB defect by ruberythrin (26). Clearly, lucigenin luminescence carefully applied may prove to be a useful tool for investigating the intracellular scavenging of O 2 . in such instances.
Acknowledgment-We are grateful to Dr. Danièle Touati for generously providing the strains used in these studies and for a critical reading of the manuscript. Distinguishing between the proposed "direct" and "indirect" mechanisms of O 2 . scavenging cannot be done on the basis of the results presented.