INFECTION AND IMMUNITY, Sept. 1992, p. 3814-3823 0019-9567/92/093814-10$02.00/0 Copyright X 1992, American Society for Microbiology

Vol. 60, No. 9

Coxiella burnetii Superoxide Dismutase Gene: Cloning, Sequencing, and Expression in Escherichia coli ROBERT A. HEINZEN,1 M. E. FRAZIER,2 AND LOUIS P. MALLAVIAl* Department of Microbiology, Washington State University, Pullman, Washington 99164-4233,1 and Office of Health and Environmental Research, Office of Energy Research, U.S. Department of Energ (GTN), Washington, D.C. 205452 Received 4 February 1992/Accepted 15 June 1992

A superoxide dismutase (SOD) gene from the obligate intracellular bacterium Coxiella burnetii has been cloned, and its DNA sequence has been determined and expressed in Escherichia coli. The gene was identified on pSJR50, a pHC79-derived genomic clone, by using the polymerase chain reaction with degenerate oligonucleotide primers corresponding to conserved regions of known SODs. Sequences resembling conventional E. coli ribosomal and RNA polymerase-binding sites preceded the C. burnetii 579-bp SOD open reading frame. An E. coli SOD-deficient double mutant (sodA sodB) that carried pSJR50 had growth and survival responses similar to those of the wild type when the transformant was challenged with 0.05 mM paraquat and 5 mM hydrogen peroxide, respectively. These observations indicated that the C. bumnetii gene was functionally expressed in E. coli. Staining of native polyacrylamide gels for SOD activity demonstrated that pSJR50 insert DNA codes for an SOD that comigrates with an SOD found in C. burnetii cell lysates. The enzyme was inactivated by 5 mM hydrogen peroxide, which is indicative of an iron-containing SOD. Additionally, the predicted amino acid sequence was significantly more homologous to known iron-containing SODs than to manganese-containing SODs. Isolation of the C. burnetii SOD gene may provide an opportunity to examine its role in the intracellular survival of this rickettsia.

The causative agent of human Q fever, Coxiella burnetii, is an obligate intracellular rickettsial parasite. Unlike other members of the family Rickettsiaceae, which replicate in the host cytoplasm or phagosomal vacuoles, C. burnetii proliferates in the harsh, acidic environment of the phagolysosome (14). The agent has been cultivated in vitro in professional and nonprofessional phagocytes, where it resists oxygen-dependent and -independent microbicidal activities induced upon phagocytosis (4). The oxygen-dependent arm corresponds with a phagocyte respiratory burst and produces toxic oxygen metabolites including superoxide anion (024, hydrogen peroxide (H202), and hydroxyl radicals (OH-) that play important microbicidal roles (8). The metalloenzyme superoxide dismutase (SOD; EC 1.15.1.1) detoxifies superoxide anion by converting it to hydrogen peroxide and oxygen (40). The detoxifying ability of SOD has established it as a virulence factor in some bacterial pathogens. For example, the ability of certain strains of Nocardia asteroides to escape killing by neutrophils is related to the production of SOD that is secreted; less-virulent strains do not secrete SOD (7, 9). In addition, a recent report describes a Shigella flexneri mutant that is unable to produce iron-containing SOD and is more sensitive than the wild type to killing by both macrophages and polymorphonuclear leukocytes (PMNs) (21). Superoxide dismutase from Escherichia coli has been thoroughly studied (for a review, see reference 22). The bacterium produces two distinct enzymes under different conditions. A manganese-containing enzyme (MnSOD) (33), encoded by the sodA gene (50, 52), is produced only under aerobic conditions (26). Conversely, an iron-containing enzyme (FeSOD) (53), encoded by the sodB gene (15, 30), is constitutively expressed during both anaerobiosis and aero*

biosis (26). In C. burnetii, SOD activity has been detected in cell lysates (1). The rickettsial holoenzyme harbors either Fe or Mn as its metal ligand and not Cu/Zn, which is associated, with rare exceptions, with SOD found in the cytosol of eucaryotic cells (22). Investigation of SOD from C. burnetii was initiated because of the agent's presumed ability to withstand the damaging effects of phagocyte-produced superoxide anion. In addition, the agent would produce, and consequently have to deal with, intracellular 02 generated by its own oxidative metabolism. The C. burnetii gene encoding SOD has been cloned by using the polymerase chain reaction (PCR) to directly amplify the gene from the rickettsial genome. Nucleotide sequencing and expression studies in E. coli were carried out to further characterize the gene. Our data indicate that C. bumetii produces one type of SOD with characteristics typical of iron-containing enzymes of other bacteria. (A preliminary account of these findings has been presented [29a].)

MATERIALS AND METHODS Bacterial strains and plasmids. All E. coli strains are derivatives of E. coli K-12. VCS257 (Stratagene, La Jolla, Calif.) is a subclone of DP50 and was used as the recipient of a C. burnetii cosmid gene library. Plasmid DNA manipulations used E. coli DH5ao (Bethesda Research Laboratories, Gaithersburg, Md.) as the host. SOD-deficient E. coli QC779 (sodA sodB) (16) and the parental strain GC4468 were used in complementation studies and were kindly provided by B. Bachmann, E. coli Genetic Stock Center, Yale University, New Haven, Conn. C. burnetii Nine Mile RSA493 isolate was obtained from Rocky Mountain Laboratories, National Institute of Allergy and Infectious Disease, Hamilton, Mont. Rochalimaea quintana ATCC VR-358 lot 2, strain Fuller,

Corresponding author. 3814

VOL. 60, 1992

was obtained from the American Type Culture Collection, Rockville, Md. The cosmid vector pHC79 (31) was used in C. bumnetii genomic library construction, and the plasmid vector pGEM7 (Promega, Madison, Wis.) was used in cloning of PCR products. Reagents, media, and growth conditions. Unless otherwise noted, all laboratory reagents and antibiotics were purchased from Sigma Chemical Co., St. Louis, Mo. LuriaBertani (LB) broth (37) (Difco Laboratories, Detroit, Mich.) was used as the complex growth medium for E. coli. Antibiotic selection was performed with ampicillin at a final concentration of 100 p,g/ml in solid media and 50 ,ug/ml in liquid media. Propagation and purification of C. bumnetii and R quintana have been described elsewhere (44, 45). Recombinant DNA and sequencing. Restriction endonuclease digestions and analyses, ligations, transformations, transductions, and plasmid DNA isolations were conducted by standard methods (37). T4 DNA ligase and restriction enzymes were purchased from Bethesda Research Laboratories. A C. bumetii genomic library was constructed by ligating Sau3A partially digested genomic DNA in the 20- to 40-kb range to BamHI-digested pHC79 previously treated with calf intestinal alkaline phosphatase (Boehringer Mannheim, Indianapolis, Ind.) pHC79 recombinants were packaged into lambda heads by using Gigapack Plus (Stratagene) packaging extract. The nucleotide sequence of both strands of the SOD coding region was determined by the dideoxychain termination method with the U.S. Biochemical Sequenase kit as suggested by the manufacturer. [a-35S]dATP was purchased from New England Nuclear, Boston, Mass. Sequencing reactions were primed by using synthetic oligonucleotides synthesized with a model 380A DNA synthesizer (Applied Biosystems Inc., Foster City, Calif.). Sequence data were compiled and analyzed by using the University of Wisconsin Genetics Computer Group Sequence Analysis package (17) managed by the VADMS Laboratory at Washington State University. Synthetic oligonucleotides for PCR. The following degenerate primers were synthesized as above for use in PCR:

SODLA, ATAGAATTCGA(TC)GC(GATC)CT(GATC)GA(G A)CC(GATC)CA; SOD1B, ATAGAATICGA(TC)GC(GA TC)TT(GA)GA(GA)CC(GATC)CA; SOD2, AA(GA)CA(TC )CA(TC)CA(GA)AC(GATC)TA(TC)GT; SOD3A, CG(GAT

C)TT(TC)GG(GATC)TC(GATC)GG(GATC)TGG; SOD3B, CG(GATC)TT(TC)GG(GATC)AG(TC)GG(GATC)TGG; SO D4, (GA)TA(GA)TA(GATC)GC(GA)TG(TC)TCCCA; and SOD5, ATAGGATCC(GA)TCCCA(GA)TT(GATC)AC(GA TC)AC(AG)TTCCA. To facilitate directional cloning of PCR products, the 5' ends of SODlA and SODlB primers specify an EcoRI site and the 5' end of SOD5 specifies a BamHI site. PCRs. A Cetus/Perkin-Elmer thermocycler (Cetus Corp., Emeryville, Calif.) and GeneAmp kit (Cetus Corp.) were used for PCR. Reaction mixtures (100 ,ul) contained 250 pmol of the 5' and 3' primers, all four deoxynucleoside triphosphates at 200 p,M each, 50 mM KCl, 10 mM Tris-HCl (pH 8.3), 1.5 mM MgCl2, 0.01% gelatin, 1 p.g of target DNA, and 2.5 U of Thermus aquaticus DNA polymerase (Cetus Corp.). The mixture was overlaid with 1 drop of mineral oil and subjected to 30 cycles consisting of a 1.5-min denaturation period at 94°C, a 2-min annealing period at 40°C, and a 1-min extension period at 72°C. PCR products were analyzed on 1.8% agarose gels and purified from the gel by using GeneClean (Bio 101, La Jolla, Calif.) when appropriate. Probe preparation and blot hybridizations. Southern analysis (46) was performed by using GeneScreen Plus (New

COXIELLA BURNETII SUPEROXIDE DISMUTASE

3815

England Nuclear). Degenerate oligonucleotide primers were prepared as probes by end labeling with [.y-32PJATP (New England Nuclear) and T4 DNA kinase. Other probes were prepared by using [a-32P]CTP (New England Nuclear) and a random primer extension kit (United States Biochemical). A pHC79-C. burnetii genomic library was screened for the full-length SOD gene by colony hybridization (37). Four high-stringency washes were performed in a solution of 0.3 x SSC (lx SSC is 0.15 M NaCl plus 0.015 M sodium citrate)-1% sodium dodecyl sulfate (SDS) at 680C. Preparation of cell lysates and SOD assays in polyacrylamide gels. Overnight cultures (-14 h) of E. coli (100 ml) grown in LB broth with ampicillin were pelleted, suspended in 1 ml of 50 mM potassium phosphate buffer (pH 7.8), and transferred to 1.5-ml microcentrifuge tubes. Cell lysates were obtained by using a model 250 Sonifier (Branson Co., Danbury, Conn.) and were cleared by centrifugation. Lysates of C. bumetii were prepared as above with 10 mg (wet weight) of yolk sac-purified organisms as the starting material. Protein concentrations were determined by the Bio-Rad assay (Bio-Rad Laboratories, Richmond, Calif.). SOD enzyme activity was visualized on 10% native polyacrylamide slab gels (10) by soaking gels in 0.2% nitroblue tetrazolium for 15 min at 37°C and then immersing them for 15 min at 37°C in a solution containing 0.028 M N,N,N',N'tetramethylethylenediamine (TEMED; Bio-Rad), 2.8 x 10-5 M riboflavin, and 0.036 M potassium phosphate (pH 7.8). The gels were then illuminated for 45 min. The presence of SOD corresponded to achromatic zones in a uniformly blue background. Purified E. coli MnSOD and FeSOD were obtained from Sigma. Hydrogen peroxide and Paraquat treatments. Sensitivities to hydrogen peroxide (H202) and Paraquat (1,1'-dimethyl4,4'-bipyridinium dichloride) were determined essentially as described by Carlioz and Touati (16). Sensitivity to H202 was assayed by measuring cell survival after exposure to 5 mM H202. Overnight E. coli cultures in LB broth plus ampicillin were diluted 1:50 by adding 0.5 ml to 25 ml of fresh LB broth plus ampicillin. Diluted cultures were grown (at 37°C with stirring at 250 rpm) to an optical density at 600 nm (OD6.) of approximately 0.5, at which point 142 p.l of 3% H202 was added for a final concentration of 5 mM. Immediately before addition of H202, a portion was drawn, diluted in ice-cold 10 mM MgSO4, and plated onto LB agar plates containing ampicillin. Additional portions were taken at 20, 40, and 60 min after H202 challenge, diluted in ice-cold 10 mM MgSO4, and plated. Colonies were counted after 24 h, and the percent cell survival was expressed as the percentage of original CFU remaining viable. To determine paraquat sensitivity, we monitored cell growth (OD6w) in cultures exposed to paraquat. A 1-ml portion of overnight cultures grown in LB broth plus ampicillin was diluted into 50 ml of fresh LB broth plus ampicillin and incubated for 1 h (at 37°C with stirring at 250 rpm). Paraquat was then added to a final concentration of 0.05 mM (0.5 ml of a 5 mM stock solution), and incubation was continued. The optical density was recorded hourly for 8 h. Nucleotide sequence accession number. The nucleotide sequence data reported in this paper will appear in the EMBL, GenBank, and DDBJ Nucleotide Sequence Data Bases under the accession number M74242. RESULTS PCR amplification. Amino acid sequences of several FeSODs and MnSODs have either been derived by direct

3816

INFECT. IMMUN.

HEINZEN ET AL.

coll (Fe) coil (Mn) H. halobium (Mn) P. l-iognathi (Fe) B. stearotherm. (Mn) S. cerevlaiae (Mn) Human liver(Mn) S. S.

Consensus Primers

Region 1 12 17 DALAPH DALEPH

Region 2 29 35 KHHQTYV KHHQTYV

Region 3 117 122 NFGSGW RFGSGW

Region 4 163 158 WEHAYY WEHAYY

DALEPH

THHQSYV KHHNTYV

DALEPH

KHHNTYV

GALEPY GALEPH

KHHQTYV KHHAAYV

GAASGW NFGSSW RFGSGW VQGSGW

WEHSYY

NALEPH

dALePh (SOD1A)

kHHqtYV (SOD2)

VQGSGW rfgSgW (SOD3A) (SOD3B)

(SOD1B)

WEHAYY WEHAYY WEHAYY WEHAYY

WEHaYY

(SOD4)

Region 5 178 184 WALVNWE WNVVNWD FEVIDWD WALVNWD WNVVNWD WNVVNWK WNVINWE wnvvnWd (SOD5)

FIG. 1. Conserved regions of SOD from various sources. E. coli iron (Fe) (15) and manganese (Mn) (47) SODS are compared with the Mn enzymes produced by H. halobium (38), B. stearothermophilus (13), Saccharomyces cerevisiae (18), and human liver (6) and the Fe enzyme of Photobacterium leiognathi (5). Degenerate PCR primers corresponding to the consensus sequence of each conserved region were synthesized. Regions 1, 2, and 3 were used in the design of 5' primers SOD1A/1B, SOD2, and SOD3A/3B, respectively. Regions 4 and 5 were used in the design of the 3' primers SOD4 and SOD5, respectively. To account for the six codons specifying leucine (SOD1) and serine (SOD3), A and B versions were synthesized and used as separate degenerate pools. The numbering reflects the amino acid position in E. coli SODB with the initiator methionine excluded. Capital letters in the derived consensus sequence represent residues conserved in all sequences, and lowercase letters represent residues present in the majority of sequences.

chemical means or deduced from the DNA sequence (5, 6, 13, 15, 18, 38, 47). A multiple-sequence alignment of these sequences demonstrated many clusters of conserved residues. With this high degree of sequence conservation, it was reasoned that a PCR approach to isolation of C. bumnetii SOD was feasible. The conserved SOD regions chosen for this study are illustrated in Fig. 1. Some of the depicted domains are intimately involved in metal-ligand binding. Of all regions, region 4 demonstrated the most obvious functional relevance since a highly conserved histidine residue involved in metal binding is present (43). Region 2 begins only 2 amino acids downstream from another conserved metal-binding histidine residue. Because regions 2 and 4 were the most highly conserved among SODs from the seven sources analyzed, they were used initially in the production of 5' (SOD2) and 3' (SOD4) degenerate primers, respectively (see Materials and Methods) (Fig. 1). When this primer pair and C. burnetii genomic DNA were used as template, PCR resulted in the amplification of a product with a predicted molecular size of -430 bp (Fig. 2A). Repeated attempts to clone this product into a plasmid vector by using blunt-end ligations proved unsuccessful. Therefore, new 5' and 3' degenerate primers, incorporating restriction sites into their 5' ends, were synthesized. Two different regions flanking a larger portion (-90%) of the enzyme were subsequently back-translated to degenerate DNA primers. SODlA and SODlB primers were synthesized with an EcoRI site as 5' primers and correspond to region 1. SOD5 was synthesized with a BamHI site as a 3' primer and corresponds to region 5. Additional degenerate primers corresponding to region 3 were synthesized for future use as end-labeled probes in Southern analysis (46) of PCR products (Fig. 1). A and B versions of degenerate primers SOD1 and SOD3 were synthesized as separate pools to accommodate the sixfold degeneracy in the amino acids leucine (SOD1) and serine (SOD3). When C. burnetii genomic DNA was amplified by using either SODlA or SODlB as the 5' primer and SOD5 as the 3' primer, the primary product was approximately 550 bp in size, as predicted (Fig. 2A). SODlA paired with the 3' primer SOD4 directed the amplification of a predicted 500-bp product. When E. coli DNA was used as the template in PCRs involving SOD2 and SOD4 primers, two bands of about equal intensity were visualized on agarose gels (Fig. 2A). By

estimating the coding capacity of these bands, it was concluded that the top band represents an amplified portion of the structural gene of MnSOD (sodA) (50, 52) while the bottom band represents an amplified portion of the structural gene for FeSOD (sodB) (15, 30). Region 2 and region 4 are identical in these proteins (Fig. 1), and amplification of the intervening sequences of both genes would lead to a lower-Mr product for sodB (15). When genomic DNA from R. quintana was used as a source of template, multiple low-Mr products were seen (Fig. 2A). To further confirm that PCR products represented amplified SOD coding sequences, we conducted Southern analysis (46) by using equal amounts of 32P-end-labeled SOD3A and SOD3B primers to probe PCR-generated products (Fig. 2B). These degenerate oligonucleotide pools correspond to an internal conserved region (Fig. 1; region 3) and should hybridize to the amplified SOD sequences. All PCR products synthesized by using C. burnetii target DNA and predicted to encode SOD produced a strong hybridization signal, except, surprisingly, the product directed by SOD2 and

A 1 2 3 4 0-6 7 891 0 B 1 2 3 4 5 6 7 8 9 1Q0 603. A1

2 71 234 194 11 18

__

72_ FIG. 2. Analysis of PCR products resulting from different 5' and 3' primer pairs. (A) Agarose gel (1.8% agarose) showing amplification products. Lanes: 1 and 8 contain 4X174 HaeIII-digested molecular size markers labeled in base pairs; 2 through 7 contain PCR products produced from C. bumetii DNA and the following primer pairs: lanes 2 and 3, SODlA/SOD5; lanes 4 and 5, SOD1B/ SOD5; lane 6, SOD1A/S0D4; lane 7, SOD2/SOD4. Lanes 9 and 10 show PCR products from R. quintana and E. coli DNA, respectively, and the SOD2/SOD4 primer pair. (B) Autoradiograph of DNA in panel A after transfer to GeneScreen Plus by the method of Southern (46) and hybridization with 32P-end-labeled SOD3A/3B primers. Washes were carried out in 2x SSC at 370C.

VOL. 60, 1992

SOD4 primers. When this degenerate primer pair was used with a PCR cycle involving a 33°C annealing temperature, substantially more product of the predicted size was formed, which did hybridize with the internal SOD3A/3B probe. On further review of the SOD2/SOD4-directed products produced at 33°C, it was noticed that a band resembling in size and intensity the product formed at 40°C migrated slightly below the primary, hybridizable product (data not shown). Hence, it appears that the product synthesized at 40°C was produced by non-SOD background amplification resulting in a product that was the approximate size predicted for SOD amplification. Both E. coli bands hybridized to SOD3A/3B primers, whereas none of the faint products produced from R. quintana DNA gave a signal. Weak signals were also seen with the 603- and 234-kb 4X174 HaeIII fragments. The PCR product whose synthesis was directed by SODlB and SOD5 primers was directionally cloned into the plasmid vector pGEM7. One recombinant, pSJR40, was selected and the entire 519 bp insert sequenced. An open reading frame (ORF) in the correct orientation had the potential to code for a peptide with 40 to 65% amino acid identity with known MnSODs and FeSODs. To further confirm the C. burnetii origin of pSJR40 insert DNA, we performed Southern analysis (46) by using 32P-radiolabeled insert DNA to probe rickettsial genomic and plasmid DNA. A single band was observed in lanes containing genomic DNA after a high-stringency wash, confirming that the cloned sequences are rickettsial in origin. We saw no additional weak bands of hybridization that might have indicated a coding region for an SOD isoenzyme or sod-like sequences (slg) such as those described for the archaebacterium Halobacterium cutirubrum (39). No signal was observed from lanes containing QpRS or QpHl plasmid DNA (45); consequently, the cryptic nature of these large rickettsial plasmids is maintained (data not shown). Isolation and nucleotide sequence of the full-length C. burnetii SOD gene. The coding potential of pSJR40 C. bumetii insert DNA strongly suggested that the cloned PCR product represents the majority (-90%) of the rickettsial SOD reading frame. To isolate the entire gene and flanking regions, insert DNA of pSJR40 was purified, radiolabeled with [a-32P]CTP by primer extension, and used as a hybridization probe to screen E. coli colonies harboring a C. burnetii pHC79 genomic library. A clone termed pSJR50 was isolated for further study. A total of 1,516 bases were sequenced on both insert strands of the 4.4-kb insert of pSJR50. We identified an ORF of 579 bases that could direct the synthesis of a 193-residue protein with a predicted Mr of 22,784 that is very similar to known manganese- and iron-containing SODs. (A preliminary report of the coding region has been published elsewhere [29].) As shown by the nucleotide sequence (Fig. 3), this ORF has an ATG initiation codon at position 241 and a TAA termination codon at position 820. The nucleotide sequence of the cloned PCR product of pSJR40 was identical to the sequence obtained for the same region of genomic clone pSJR50. Beginning 14 bases upstream from the ATG initiation codon is a polypurine-rich sequence of AGG AGGA that is a potential Shine-Dalgarno ribosome-binding site (23). Beginning 64 bases upstream of the predicted ribosome-binding site is the sequence CTGAAA-N18-TAT AAT, representing putative -35 and -10 RNA polymerasebinding sites (28). In E. coli, the gene product of fur (ferric uptake regulation) has been shown to act as a trans-acting repressor of sodA4 by binding to an iron box sequence in the promoter region (41). No significant homology to the iron

COXIELLA BURNETII SUPEROXIDE DISMUTASE

3817

box consensus sequence was observed upstream of the C. burnetii SOD coding region. No obvious rho-independent termination sequences were observed downstream of the SOD gene termination codon. However, starting 2 bases after the SOD gene TAA stop codon at position 824 was an ORF of 509 bases that was capable of coding for a polypeptide of 170 amino acids with a calculated molecular mass of 18.2 kDa (Fig. 3). The deduced amino acid sequence of this ORF displayed 34% identity when aligned with the first 194 residues of E. coli dipicolinate reductase (DapB) (12) (data not shown). This enzyme is a product of the dapB gene and is the second enzyme of the diaminopimelate-lysine biosynthetic pathway, where it reduces dihydrodipicolinate to tetrahydrodipicolinate. Within the alignment was a cluster of 20 amino acids with 15 perfect identities containing no gaps. No regions similar to the E. coli consensus for promoter and ribosomebinding sequences are observed immediately upstream of this ORF, and no predicted rho-independent terminator sequence is present downstream. In addition, the primary product formed from pSJR50 rickettsial DNA by in vitro transcription-translation had an apparent molecular mass of about 23 kDa, which is very near that calculated for the predicted amino acid sequence of C. burnetii SOD. A protein of 18.2 kDa whose synthesis is predicted to be directed by the ORF downstream of the SOD coding region was not detected by this procedure (data not shown). Features of the predicted SOD amino acid sequence. The amino acid sequence homology of C. bumetii SOD to other MnSODs and FeSODs from phylogenetically diverse sources is presented in Table 1. Percent identities were tabulated from pairwise alignments of the rickettsial sequence with iron-containing SODs from Photobacterium leiognathi (5), Pseudomonas ovalis (32), and E. coli (15), and manganese-containing SODs from Bacillus stearothermophilus (13), E. coli (47), Saccharomyces cerevisiae (18), and human liver (6). This analysis was extended to sequences from Mycobactenium leprae (51) and Halobacterium halobium (38), and, although the metal component of their respective enzymes has not been determined experimentally, they are regarded as MnSODs on the basis of amino acid sequence homology (43). The rickettsial SOD displayed significantly more amino acid identity with bacterial FeSODs than with bacterial and eucaryotic MnSODs (Table 1). Approximately 62 to 65% identity was seen with ironcontaining enzymes from Photobacterium leiognathi, P. ovalis, and E. coli. Excluding the MnSOD from B. stearothermophilus (51.8% identity), no MnSOD from either bacterial or eucaryotic sources displayed greater than 45% identity. These sequence comparisons also revealed that the rickettsial enzyme has retained homologs to three histidine residues and one aspartate residue known as ligands to Fe and Mn cofactors (43) (Fig. 3). Analysis of alignments of the C. bumetii enzyme to other SODs demonstrated that the enzyme contains four of the five residues considered primary indicators of an Fe-containing enzyme (43). These residues were Ala-69, Gln-70, Tyr-77, and Ala-142 (Fig. 3). An arginine residue was substituted for glycine at position 143, the final amino acid considered invariant for FeSODs. Native PAGE. To test whether the C. bumetii SOD gene specified by pSJR50 is expressed in vivo by E. coli and whether the gene product comigrates with a similar protein present in C bumnetii whole-cell lysates, we performed native polyacrylamide gel electrophoresis (PAGE) followed by staining for SOD activity. As shown in Fig. 4A, lysates from SOD-deficient E. coli QC779 (sodA sodB) (16) cells

3818

INFECT. IMMUN.

HEINZEN ET AL. 1

TTCGCGACCTACTAAAACAGCCATAATAACCTCCGCATTTCAAAATAGCACTAATTTAGT

60

61

CCTATATTAAAGGATTGTCAAATAAGCCTTCCAGAAAAATACCTTGAATGGCATCAGGCA

120

121

GAAGACACCCGCGGACACTTAAATAACCACCCAAACATCTCTCTGAAAATCTATAGAAAT ,RSS -10 TGAACCTATAATCGGGTATCGTGCGAAAGTCAAGATAAATGAAAAAAGGAGGAACTCCTA sod-owATGGCTTTTGAATTACCGGATTTGCCCTACAAACTCAACGCACTGGAACCGCATATCTCT

180

181 241 301

E

F

M A

P

L

D

L

P

K

Y

N

L

L

A

E

P

H

240 300

8

I

CAAGAAACGCTCGAATATCACCACGGAAAACACCATAGAGCTTATGTCAATAAACTCAAC E T L E Y H H G K H H R A Y V N K L N

360

Q

*

361

AAACTTATCGAAGGCACCCCTTTTGAAAAGGAACCTCTGGAAGAAATTATTCGAAAATCC K L I E G T P F E K E P L E E I I R K S

420

421

GACGGCGGAATCTTCAACAATGCAGCACAACATTGGAACCATACATTTTATTGGCACTGC D G G I F N N A A Q H W N H T F Y W H C

480

481

ATGAGCCCTGATGGCGGTGGAGATCCTTCTGGCGAATTGGCTTCAGCTATTGATAAAACT M S P D G G G D P S G E L A S A I D K T

540

541

TTTGGATCTTTAGAGAAATTTAAAGCGCTTTTTACCGACTCCGCAAATAATCATTTCGGC

600

+

G

F

601

L

E

K

F

K

L

A

T

F

S

D

A

N

N

G

W

W

A

K

V

L

N

D

K

G

N

L

E

L

V

S

A

R

P

N

M

G

E

T

K

K

P

C

T

M

L

V

D

W

720

E

*

+

CACGCCTATTACATCGATACCCGGAACGACCGCCCCAAATACGTTAATAACTTTTGGCAA H

660

V

T

AACGCTCGTAATCCCATGACTGAAGGCAAAAAACCCCTAATGACTTGTGATGTTTGGGAA N

721

G

F

H

TCGGGATGGGCTTGGCTCGTTAAAGATAACAATGGCAAATTAGAAGTCTTAAGCACTGTC S

661

S

+

*

+

A

Y

D

I

Y

T

R

D

N

R

P

V

Y

K

N

N

F

780

Q

W

GTGGTCAATTGGGATTTTGTGATGAAAAACTTCAAATCCTAACATGGCCATTAACGTTAT V N W D F V M K N F K S M A I N V I

840

841

TATTAATGGAATTAATGGCAAAATGGGCCGGGTGGTGAAAGAAAACATCACGGCCCAATC I N G I N G K M G R VV K E N I T A Q S

900

901

CGATCTTGAACTAGTGTCTGGAACCGGTCGCCAAGACGATTTGGCGAAAACTATTCAAAC D L E L V S G T G R Q D D L A K T I Q T

960

961

TACTCATGCGGATGTTGTCATCGATTTCTCGACGCCTCAATCGGTATTTCACAACGCCGA T H A D V V I D F S T P Q S V F H N A E

1020

1021

AATTATTATTCAATCCGGAGCAAGGCCTGTTATTGGTACGACGGGACTGACGTTAGAACA I II Q S G A R P V I G T T G L T L E Q

1080

1081

AATCGCTCTTTTAGATAAACAATGCCGTAATAAAAAGCTCGGCGCAATTGTGGCCCCTAA I A L L D K Q C R N K K L G A I V A P N

1140

1141

TTTTTCTGTCGGTGCGGTGTTAATGATGAAATACGCAAAAGAAGCAGCGCACTATTTTCC

1200

781

V

F

1201

S

V

G

A

V

L

M M

K

Y

A

K

E

A

A

H

Y

F

P

CGACGTAGAAATTATTGAAATGCATCATTCCCAAAAAATTGATGCCCCCTCAGGAACAGC D

V

E

I

I

E

M

H

H

S Q

K

I

D

A

P

S

G

T

1260

A

1261

GATCAAAACGGCACAAATGATAGGCGAAATGCGATCCCACCCCCATCAGGATGCGGCCGA I K T A Q M I G E M R S H P H Q D A A E

1320

1321

AAAAAGCAGCCGGTAAACTCAGAAAAGTAGCAAAAACGAAAGCGCCAACGGTGCATAAAA K S S R

1380

1381

GCGCGGACAAGGTAAGTACCCAGCGGGGATTATACCGGTCCATGATAACACCGGAAGGGA

1440

1441

TTTGCAAAGGCGTATAGGCATAATAATAAAAAGCCGAAATAAACCCAATCGTGGGCGCTC

1500

1501

CAACGTTGAATTGATC

1516

FIG. 3. Nucleotide sequence of the noncoding strand of C. bumetii sod and flanking regions. All nucleotide sequence data were confirmed by sequencing of both strands of DNA. The structural gene for sod begins with an ATG start codon at position 241 and ends with a TAA stop codon at position 820. Putative -35 and -10 promoter regions and a potential ribosome-binding site (RBS) are overlined. The translated SOD amino acid sequence is depicted under the nucleotide sequence. Asterisks show residues predicted to be ligands of iron. Plus signs show residues indicative of iron-containing SODs. An ORF with translation beginning with an ATG codon 2 bases downstream from the sod stop codon is also shown. This ORF ends with a TAA stop codon at position 1334.

harboring pSJR50 contained one SOD that migrated between E. coli MnSOD and FeSOD. This band of activity comigrated with a band obtained from C. burnetii cell lysates. No additional bands of activity were seen with the C. bumnetii lysate, suggesting that, unlike E. coli, the rickettsia produces a single type of SOD. Only the Fe enzyme was visible from lysates obtained from the wild-type E. coli GC4468 trans-

formed with the cloning vector pHC79. It is unclear why the MnSOD was not produced in sufficient quantity under the growth conditions imposed to be detected in this assay. However, GC4468/pHC79 cells did synthesize a detectable MnSOD when cells were grown overnight in LB broth with 0.05 mM paraquat. This redox cycling agent generates superoxide anion, which induces MnSOD synthesis in E.

VOL. 60, 1992

COXIELLA BURNETII SUPEROXIDE DISMUTASE

3819

TABLE 1. Percent amino acid identities between SODs % Identical amino acids with SOD fromab: SOD source

C. burnetii P. leiognathi (Fe) P. ovalis (Fe) E. coli (Fe) B. stearothermophilus (Mn) E. coli (Mn) M. leprae (Mn) H. halobium (Mn) S. cerevisiae (Mn) Human liver (Mn)

C burnetii

P. Ignathi (Fe)C

(100)

64.6 (100)

P.

ovalis E. coli

B stearo-

(Fe)

(Fe)

the(nophl'us

64.4 67.5 (100)

61.5 75.5 69.6 (100)

51.8 53.4 54.8 52.9 (100)

E. coli

M.

leprae

(Mn)

(Mn)

45.0 42.1 45.3 46.0 61.3 (100)

40.8 37.0 41.2 37.2 45.6 41.2 (100)

H. halobium (Mn)

36.0 29.7 31.1 31.0 37.5 33.3 30.9 (100)

S.

cerevisiae (Mn)

Human liver (Mn)

39.2 37.4 38.4 36.3 40.3 41.4 48.5 31.7

42.3 38.7 43.3 41.6 51.8 45.6 51.8 33.2 50.0 (100)

(100)

a Listed are percentages of identical residues as determined by the GAP program in the University of Wisconsin Genetics Computer Group sequence analysis package (9). b N-terminal methionines were excluded from the analysis. c Fe denotes iron-containing SOD. d Mn denotes manganese-containing SOD.

coli (27). Exposure to 0.05 mM paraquat had no demonstrable effect in enhancing the expression of C. bumnetii SOD in QC779/pSJR50 cells (data not shown). E. coli QC779 carrying the vector pHC79 produced no detectable SOD under any growth condition. Hydrogen peroxide (H202) inactivates iron-containing SODs by oxidizing tryptophan residues, resulting in probable modification of the active-site environment (11). To determine whether pSJR50-encoded SOD is sensitive to H202, a duplicate gel as described above was preincubated for 1 h in 5 mM H202 before being stained for SOD activity (Fig. 4B). Elimination of bands of activity with C. burnetii and QC779/pSJR50 lysates signified inactivation of a probable FeSOD. Similar results were observed with purified E. coli FeSOD and that obtained from GC4468/pHC79. Purified E. coli MnSOD was relatively weakly affected by this

deficient E. coli QC779 (sodA sodB) to paraquat and H202. Carlioz and Touati (16) have previously demonstrated that this strain is hypersensitive to these agents. E. coli QC779 harboring pSJR50 was considerably less sensitive to paraquat (Fig. 5). The presence of 0.05 mM paraquat had a negligible effect on the growth rate of QC779/pSJR50, and cultures reached about the same final OD as did cells grown in the absence of paraquat. Wild-type GC4468/pHC79 cells were similarly not affected by this treatment, whereas QC779/pHC79 transformants were inhibited. Increased resistance to paraquat by QC779/pSJR50 coincided with H202 insensitivity. The cloned C. burnetii SOD protected SOD-deficient E. coli from a toxic concentration of this oxidant (5 mM) (Fig. 6). The survival curve showed a resistance similar to that observed for the wild-type E. coli.

treatment. Functional expression of C. burnetii sod in E. coli. We tested the cloned rickettsial SOD for functional expression by determining whether it alleviates in vivo sensitivity of SOD-

DISCUSSION

A MnSOD

2 3 4

1 "q

i:

5

B 1

2

3

4

5

f IA 1:

FeSOD

FIG. 4. Native PAGE gel stained for SO]D activity. Crude cell extracts were prepared as described in Maiterials and Methods. Extracts were electrophoresed on a nondenaLturing 10% polyacrylamide gel and stained for SOD activity. (A) La nes: 1, 2.5 and 10.0 Ig of purified E. coli MnSOD and E. coli FeSOI 4 pLg of GC4469/pHC79 lysate; 3, 100 pLg of QC 82.5 ,ug of C. bumetii lysate; 5, 100 ,ug of QC'779/pHC79 lysate. (B) The gel is identical to that in panel A, but it Mvas immersed in 5 mM H202 before being stained for SOD activity. L,ocations of E. coli Mn and FeSOD are indicated.

D779/pSJRSOIysate;

SOD has been implicated as a virulence factor for a number of pathogenic bacteria that resist killing by oxygendependent mechanisms of professional phagocytes (8). The rickettsia C. burnetii is impressive in its resistance to both oxygen-dependent and oxygen-independent microbicidal activities. Indeed, it has a growth requirement for at least one physical attribute of the intraphagolysosomal milieu (low pH) normally considered deleterious to bacteria (24). The isolation and characterization of SOD from this pathogen provides a means of investigating its role in intracellular survival. The degenerate oligonucleotide primers (128-fold) used in the cloning procedure were highly specific for the C. bumetii SOD coding region with little background amplification. Examination of the predicted amino acid sequence of C. burnetii SOD offers an explanation of why PCR amplification, with SOD2 and SOD4 primer pools, yielded a product identified by hybridization analysis as encoding SOD at 33 but not 40°C. A comparison between the predicted amino acid sequence in region 2 (Fig. 3, residues 30 to 36) and the consensus sequence for the same region used in backtranslating to the SOD2 degenerate primer pool (Fig. 1) revealed that the amino acids arginine and alanine have been substituted for glutamine and tyrosine residues, respectively, in the rickettsial enzyme. Annealing of the primer

3820

INFECT. IMMUN.

HEINZEN ET AL. 1000

100I

GC4468/pHC79

-i 0 0 ID

-

Cl)

ci

-A

QC779/pHC79

-

-J u

-

QC779/pHC79+PQ2

QC779/pSJR5O

10

QC779/pSJR50

*

QC779/pSJR5O+PQ2

20

8

0C779/pHC79

C.

0 A

GC4468/pHC79

a

GC4468/pHC79+PQ2

40

60

10

HOURS AT 370 C

FIG. 5. Effect of paraquat (PQ2) on the growth of E. coli transformants. Sensitivity to paraquat was determined by growth inhibition of liquid cultures. Paraquat (0.05 mM) was added to log phase cultures (OD60o 0.2), and the OD was recorded hourly for 8 h. The E. coli transformants tested were GC4468/pHC79, QC779/ pHC79, and QC779/pSJR50. Lines connecting open datum points represent growth curves of cultures grown in the absence of paraquat. Lines connecting solid datum points represent growth curves of cultures grown in the presence of paraquat.

species within the SOD2 degenerate pool showing the greatest homology to C. burnetii sod would still result in a 3-bp mismatch. This event was presumably tolerated at 33°C but not at 40°C. The significance of the ORF beginning immediately downstream of C. burnetii sod is open to speculation. The

predicted 170-residue protein product was much smaller than the homologous 273-residue E. coli DapB (12). Therefore, it is unlikely that the identity between the two proteins coincides with a common function in lysine biosynthesis for the C. bumetii protein. It is possible that the predicted protein is a truncated version of the wild type that resulted from a pSJR50 deletion event. The rickettsial DNA insert (4.4 kb) is considerably smaller than the 20- to 40-kb genomic fragments used to construct the pHC79 genomic library. Nonetheless, two sequences of 100 amino acids or more exhibiting greater than 25% identity are considered to have common ancestry (19). It is conceivable that the predicted rickettsial protein has domains that function enzymatically in a reductase-type catalysis. The observation that no regions typifying RNA polymerase-binding sites were appropriately spaced upstream of the start of this ORF, coupled with the absence of a sequence predicted to function in rho-independent termination downstream of sod, may suggest that the two coding regions constitute a single transcriptional unit. However, it should be reemphasized that a

MINUTES AFTER H202 CHALLENGE

FIG. 6. Sensitivity of E. coli transformants to challenge by hydrogen peroxide. Log phase cultures (OD6J0 - 0.5) of E. coli transformants GC4468/pHC79, QC779/pHC79, and QC779/pSJR50 were subjected to challenge by 5 mM H202. Survival is expressed as the percentage of original CFU remaining viable and was calculated at 20, 40, and 60 min after H202 addition. Initial cell densities were 1 x 108 to 3 x 108 cells per ml.

translation product of the size predicted from the downstream ORF was not detected by in vitro transcriptiontranslation. It has been demonstrated that an SOD gene is partly cotranscribed with the photolyase gene in the archaebacterium H. halobium (49). Additional studies are required to resolve whether these C. bumetii reading frames are arranged in an operon. Three lines of evidence suggest that C. bumnetii synthesizes only one type of SOD. First, PCR amplification of C. bumetii target DNA by using primers corresponding to amino acids highly conserved in both FeSODs and MnSODs produced one amplified product. The same primer pair in conjunction with E. coli target DNA efficiently produced two bands as predicted. By molecular weight and Southern analysis (46), these bands are thought to represent coding regions for E. coli FeSOD and MnSOD isozymes. One would predict that if an additional sod homolog were present in C. bumetii, it would have been similarly detected. Second, Southern analysis (46) with the cloned, SOD-encoding PCR product to probe the rickettsial genome revealed one band of hybridization. Third, staining of native PAGE gels for SOD activity visualized a single band of enzyme of activity from C. bumetii cell lysate that comigrated with a single band of activity from the QC779/pSJR50 lysate. If C. burnetii produced additional SOD isozymes, they would have presumably been detected as electrophoretically distinct bands of activity in this assay. The sensitivity of the C. bumetii enzyme to inactivation by H202 and amino acid

VOL. 60, 1992

sequence homology showing greatest homology to FeSODs indicate that the locus designation for the structural gene should be sodB and not sodA. It should be noted that the presence of a copper- and zinc-containing SOD (Cu/Zn SOD), rarely found in procaryotes, would not have been detected by PCR or Southern approaches owing to a lack of any meaningful sequence similarity to FeSOD and MnSOD

(22). For a bacterium to utilize SOD as a means of detoxifying extracellular superoxide anion, the enzyme would probably have to be surface exposed or secreted into the medium. Superoxide anion is a negatively charged ion and, as such, requires a transport function to permeate the phospholipid bilayer of biological membranes (48). To our knowledge, the only channel of this type has been found in erythrocytes, in which superoxide anion transverses the plasma membrane through a system normally reserved for Cl- and HCO3 exchange (36). The human pathogens N. asteroides and M. tuberculosis both secrete SOD that apparently becomes surface associated (7, 34). Strains of N. asteroides unable to secrete SOD are less virulent as measured by increased killing by PMNs (7). In a related study with a murine model, organisms pretreated with monoclonal antibody to nocardial SOD were more efficiently cleared from the animal than were untreated organisms (9). The recent sequencing of the M. tuberculosis MnSOD gene (54) shows that secretion of the enzyme occurs despite the lack of an identifiable hydrophobic leader peptide preceding the structural gene. In normally nonpathogenic mycobacteria, such as M. smegmatis, SOD remains cytoplasmically localized (34). In at least one instance, secretion of SOD does not appear to be necessary for the enzyme to aid in bacterial virulence. A study with a sodB mutant of the enteric pathogen S. fle-xneri demonstrated that the organism was substantially more sensitive to killing by professional phagocytes than the wild type was (21). This outcome occurs despite the presence of functional sodA. Both genes presumably encode SODs that remain in the cytoplasm. The cellular location of C bumetii SOD is presently unknown. A hydrophobic leader peptide that would indicate possible membrane or extracellular partitioning was not predicted by the DNA sequence. Although the relationship of SOD production and/or secretion to increased virulence is established in the above cases, the exact mechanism by which the enzyme promotes survival and hence pathogenesis remains to be elucidated. The issue becomes further complicated when one examines instances of bacteria that are devoid of SOD but are still pathogenic. Neisseria gonorrhoeae, an obligate aerobe, lacks SOD, but viable cells are routinely recovered from purulent exudates filled with activated PMNs (42). A study by Archibald et al. (3) suggests that the high tolerance for extracellular superoxide anion and H202 may be explained in part by high levels of intracellular catalase and peroxidase. This study also demonstrated that Neisseria gonorrhoeae produces large amounts of the superoxide scavenger glutathione, which may compensate for the lack of SOD. The human pathogen Mycoplasma pneumoniae undergoes oxidative metabolism but produces neither SOD nor catalase (35). This phenotype appears to be of little consequence in terms of the virulence of this organism. One can speculate that in bacteria that withstand the oxidizing power of the phagolysosome, intracellular catalase may play at least as critical a role in detoxification as intracellular SOD does. Hydrogen peroxide, as an uncharged molecule, is freely permeable through biological

COXIELLA BURNETII SUPEROXIDE DISMUTASE

3821

membranes (25). Breakdown of H202 to 02 and H20 by catalase would lessen the probability of production of highly reactive hydroxyl radicals. These can be created in biological systems through iron-dependent decomposition of H202 and clearly cause cellular damage, which can manifest as lipid peroxidation and/or DNA strand breaks (25). The formation of OH- can occur solely in the presence of iron and H202 (Fenton reaction) or in a superoxide-driven process (Harber-Wiess reaction) (25). The latter reaction presents a scenario in which SOD and catalase would be of mutual benefit. The respiratory burst observed on phagocytosis involves a dramatic stimulation of the hexose monophosphate shunt. NADPH produced by the hexose monophosphate shunt acts as substrate with 02 for phagosomal membrane-bound NADPH oxidase. This reaction results in a flux of superoxide anion plus other toxic oxygen metabolites (e.g., H202). When lysosomal fusion occurs, myleoperoxidase is released, this enzyme can use H202 with halide ions (e.g., Cl-) to produce potent halogenating compounds (8). There are conflicting reports regarding the effect of phagocytosis of C. burnetii on the respiratory burst of PMNs. Ferencik and coworkers (20) reported that killed, avirulent phase II C. bumetii (truncated lipopolysaccharide 0 side chains) but not virulent phase I (full-length lipopolysaccharide 0 side chains) have a mild stimulatory effect on the hexose monophosphate shunt and the production of superoxide anion. If cells of either phase are pretreated with rabbit immune serum containing specific opsonizing antibodies, both hexose monophosphate shunt stimulation and superoxide anion production are enhanced to near zymosan control levels. In a parallel study, Akporiaye et al. (2) reported that neither killed nor viable phase I cells pretreated with specific opsonizing antibodies stimulate human superoxide production on initial phagocytosis by PMNs. Additional studies are clearly needed to explain these contradictory data. An accurate determination of the contribution of sod to C. burnetii virulence will be acquired only when sod strains are created and tested in tissue culture or a suitable animal model. Gene inactivation studies of rickettsiae require a system for introduction of DNA into the cell. Such a system was not available until recently, when the rickettsia R. quintana was transformed by electroporation (44). It is hoped that the same technology will lead to the development of a transformation system in C. burnetii. ACKNOWLEDGMENTS We express appreciation to Shirley Schmitt for her excellent technical assistance. The software cited is part of the VADMS Center, a campus-wide computer resource at Washington State University. This work was supported by grant AI20190 from the National Institutes of Health (NIAID); by the Northwest College and University Association for Science (University of Washington) under contract DE-AM06-76-RL02225 with the U.S. Department of Energy; and by Battelle, Pacific Northwest Laboratories.

REFERENCES 1.

Akporiaye, E. T., and 0. G. Baca. 1983. Superoxide anion

production and superoxide dismutase and catalase activities in Coxiella bumetii. J. Bacteriol. 154:520-523. 2. Akporiaye, E. T., D. Stefanovich, V. Tsosic, and 0. G. Baca. 1990. Coxiella burnetii fails to stimulate human neutrophil superoxide anion production. Acta Virol. 34:64-70. 3. Archibald, F. S., and M.-N. Duong. 1986. Superoxide dismutase and oxygen toxicity defenses in the genus Neisseria. Infect. Immun. 51:631-641. 4. Baca, O. G., and D. Paretsky. 1983. Q-fever and Coxiella

3822

INFEcr. IMMUN.

HEINZEN ET AL.

burnetii: a model for host-parasite interactions. Microbiol. Rev. 47:127-149. 5. Bara, D., M. E. Schinina, W. H. Banister, J. V. Bannister, and F. Bossa. 1987. The primary structure of iron-superoxide dismutase from Photobacterium leiognathi. J. Biol. Chem. 262:1001100gc. 6. Bara, D., M. E. Schinina, M. Simmaco, J. V. Bannister, W. H. Banister, G. Rotilia, and F. Bossa. 1984. The primary structure of human liver manganese superoxide dismutase. J. Biol. Chem. 259:12595-12601. 7. Beaman, B. L., L. M. Black, F. Doughty, and L. Beaman. 1985. Role of superoxide dismutase and catalase as determinants of pathogenicity of Nocardia asteroides: importance in resistance to microbicidal activities of human polymorphonuclear neutrophils. Infect. Immun. 47:135-141. 8. Beaman, L., and B. L. Beaman. 1984. The role of oxygen and its derivatives in microbial pathogenesis and host defense. Annu. Rev. Microbiol. 38:27-48. 9. Beaman, L., and B. L. Beaman. 1990. Monoclonal antibodies demonstrate that superoxide dismutase contributes to protection of Nocardia asteroides within the intact host. Infect. Immun. 58:3122-3128. 10. Beauchamp, C., and I. Fridovich. 1971. Superoxide dismutase: improved assays and an assay applicable to acrylamide gels. Anal. Biochem. 44:276-287. 11. Beyer, W. F., Jr., and I. Fridovich. 1987. Effect of hydrogen peroxide on the iron-containing superoxide dismutase of Escherichia coli. Biochemistry 26:1251-1257. 12. Bouvier, J., C. Richaud, F. Richaud, J.-C. Patte, and P. Straigier. 1984. Nucleotide sequence and expression of the Escherichia coli dapB gene. J. Biol. Chem. 259:14829-14834. 13. Brook, C. J., and J. E. Walker. 1980. Superoxide dismutase from Bacillus stearothermophilus. Complete amino acid sequence of a manganese enzyme. Biochemistry 19:2873-2882. 14. Burton, P. R., J. Stuckemann, R. M. Welsh, and D. Paretsky. 1978. Some ultrastructural effects on persistent infections by the rickettsia Coxiella bumetii in mouse L-cells and green monkey kidney (VERO) cells. Infect. Immun. 21:556-566. 15. Carlioz, A., M. L. Ludwig, W. C. Stallings, J. A. Fee, H. M. Steinman, and D. Touati. 1988. Iron superoxide dismutase. Nucleotide sequence of the gene from Escherichia coli K12 and correlations with crystal structures. J. Biol. Chem. 263:15551562. 16. Carlioz, A., and D. Touati. 1986. Isolation of superoxide dismutase mutants in Escherichia coli: is superoxide dismutase necessary for aerobic life? EMBO J. 5:623-630. 17. Devereux, J., P. Haeberli, and 0. Smithies. 1984. A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res. 12:387-395. 18. Ditlow, C., J. T. Johansen, B. M. Martin, and I. B. Svendsen. 1982. The complete amino acid sequence of manganese-superoxide dismutase from Saccharomyces cerevisiae. Carlsberg Res. Commun. 47:81-91. 19. Doolittle, R. F. 1990. Searching sequence databases. Methods Enzymol. 183:99-110. 20. Ferencik, M., S. Schramek, J. Kazar, and J. Stefanovich. 1984. Effect of Coxiella bumetii on the stimulation of hexose monophosphate shunt and on superoxide anion production in human polymorphonuclear leukocytes. Acta Virol. 28:246-250. 21. Frazon, V. L., J. Arondel, and P. J. Sansonetti. 1990. Contribution of superoxide dismutase and catalase activities to Shigella flexneri pathogenesis. Infect. Immun. 58:529-535. 22. Fridovich, I. 1986. Superoxide dismutases. Adv. Enzymol. 58:61-97. 23. Gold, L., D. Pribnow, T. Schneider, S. Shinedling, B. Singer, and G. Stromo. 1981. Translation initiation in procaryotes. Annu. Rev. Microbiol. 35:365-407. 24. Hackstadt, T., and J. C. Williams. 1981. Biochemical stratagem for obligate parasitism of eucaryotic cells by Coxiella burnetii. Proc. Natl. Acad. Sci. USA 78:3240-3244. 25. Halliwell, B., and J. M. Gutteridge. 1990. Role of free radicals and catalytic metal ions in human disease: an overview. Methods Enzymol. 186:1-85.

26. Hassan, H. M., and I. Fridovich. 1977. Enzymatic defense against the toxicity of oxygen and of streptonigrin in Escherichia coli. J. Bacteriol. 129:1574-1583. 27. Hassan, H. M., and I. Fridovich. 1977. Regulation of the synthesis of superoxide dismutase in Escherichia coli. Induction by methyl viologen. J. Biol. Chem. 252:7667-7672. 28. Hawley, D. K., and W. R. McClure. 1983. Compilation and analysis of Escherichia coli promoter DNA sequences. Nucleic Acids Res. 11:2237-2255. 29. Heinzen, R. A., M. E. Frazier, and L. P. Mallavia. 1990. Nucleotide sequence of Coxiella bumetii superoxide dismutase. Nucleic Acids Res. 18:6437. 29a.Heinzen, R. A., M. E. Frazier, and L. P. Mallavia. 1991. Abstr. 91st Gen. Meet. Am. Soc. Microbiol. 1991, abstr. D-163, p. 105. 30. Hiroshi, S., and D. Touati. 1984. Cloning of the iron superoxide dismutase gene (sodB) in Escherichia coli K-12. J. Bacteriol.

159:418-420.

31. Hohn, B., and J. Collins. 1980. A small cosmid for efficient cloning of large DNA fragments. Gene 11:291-298. 32. Isobe, R., Y. I. Fang, D. Muno, T. Okuyama, D. Ohomori, and F. Yamakura. 1987. Amino acid sequence of iron-superoxide dismutase from Pseudomonas ovalis. FEBS Lett. 223:92-96. 33. Keele, B. B., Jr., J. M. McCord, and I. Fridovich. 1970. Superoxide dismutase from Escherichia coli B. A new manganese containing enzyme. J. Biol. Chem. 245:6179-6181. 34. Kusunose, E., E. Ichihara, Y. Noda, and M. Kusunose. 1976. Superoxide dismutase from Mycobacterium tuberculosis. J. Biochem. 80:1343-1352. 35. Lynch, R. E., and B. C. Cole. 1980. Mycoplasma pneumoniae: a procaryote which consumes oxygen but which lacks superoxide dismutase. Biochem. Biophys. Res. Commun. 96:98-105. 36. Lynch, R. E., and I. Fridovich. 1978. Permeation of the erythrocyte stroma by superoxide radical. J. Biol. Chem. 253:46964699. 37. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 38. Marvin, S. L., M. V. Duke, D. Oerterhelt, and D. Ma. 1988. Cloning and determination of the nucleotide sequence of the Mn-containing superoxide dismutase gene of Halobacterium halobium. Gene 70:153-159. 39. May, B. P., and P. P. Dennis. 1990. Unusual evolution of a superoxide dismutase-like gene from the extremely halophilic archaebacterium Halobacterium cutirubrum. J. Bacteriol. 172:

3725-3729.

40. McCord, J. M., and I. Fridovich. 1969. Superoxide dismutase. An enzymic function for erythrocuprein (Hemocuprein). J. Biol. Chem. 244:6049-6055. 41. Niederhoffer, E. C., C. M. Naranjo, K. L. Bradly, and J. A. Fee. 1990. Control of Escherichia coli superoxide dismutase (sodA and sodB) genes by the ferric uptake regulation (fur) locus. J. Bacteriol. 172:1930-1938. 42. Norrod, P., and S. A. Morse. 1979. Absence of superoxide dismutase in some strains of Neisseria gonorrhoeae. Biochem. Biophys. Res. Commun. 90:1287-1294. 43. Parker, W. M., and C. C. F. Blake. 1988. Iron- and manganesecontaining superoxide dismutases can be distinguished by analysis of their primary structures. FEBS Lett. 229:337-382. 44. Reschke, D. K., M. E. Frazier, and L. P. Mallavia. 1990. Transformation of Rochalimaea quintana, a member of the family Rickettsiaceae. J. Bacteriol. 172:5130-5134. 45. Samuel, J. E., M. E. Frazier, and L. P. Mallavia. 1985. Correlation of plasmid type and disease caused by Caxiella burnetii. Infect. Immun. 49:775-779. 46. Southern, E. M. 1975. Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98:503-517. 47. Steinman, H. M. 1978. The amino acid sequence of manganosuperoxide dismutase from E. coli B. J. Biol. Chem. 253:87088720. 48. Takahashi, M.-A., and K. Asada. 1983. Superoxide anion permeability of phospholipid membranes and chloroplast thylakoids. Arch. Biochem. Biophys. 226:558-566.

VOL. 60, 1992 49. Takao, M., T. Kobayashi, A. Oikawa, and A. Yasui. 1989. Tandem arrangement of photolyase and superoxide dismutase genes in Halobacterium halobium. J. Bacteriol. 171:6323-6329. 50. Takeda, Y., and H. Avila. 1986. Structure and gene expression of the E. coli Mn superoxide dismutase gene. Nucleic Acids Res. 14:4577-4589. 51. Thangaraj, H. S., F. I. Lamb, E. 0. Davis, P. J. Jenner, L. H. Jeyakumar, and M. J. Colston. 1990. Identification, sequencing, and expression of Mycobactenium leprae superoxide dismutase, a major antigen. Infect. Immun. 58:1937-1942.

COXIELLA BURNETII SUPEROXIDE DISMUTASE

3823

52. Touati, D. 1983. Cloning and mapping of the manganese superoxide dismutase gene (sodA) of Escherichia coli K-12. J. Bacteriol. 155:1078-1085. 53. Yost, F. J., Jr., and I. Fridovich. 1973. An iron containing superoxide dismutase from E. coli. J. Biol. Chem. 248:49054908. 54. Zhang, Y., R. Lathigra, T. Garbe, D. Catty, and D. Young. 1991. Genetic analysis of superoxide dismutase, the 23 kilodalton antigen of Mycobacterium tuberculosis. Mol. Microbiol. 5:381391.