Vol. 172, No. 7

JOURNAL OF BACTERIOLOGY, JUlY 1990, p. 3725-3729

0021-9193/90/073725-05$02.00/0 Copyright © 1990, American Society for Microbiology

Unusual Evolution of a Superoxide Dismutase-Like Gene from the Extremely Halophilic Archaebacterium Halobacterium cutirubrum BRUCE P. MAY AND PATRICK P. DENNIS*

Department of Biochemistry, University of British Columbia, Vancouver, British Columbia, Canada V6T I W5 Received 27 November 1989/Accepted 10 April 1990

The archaebacterium Halobacterium cutirubrum contains a single detectable, Mn-containing superoxide dismutase, which is encoded by the sod gene (B. P. May and P. P. Dennis, J. Biol. Chem. 264:12253-12258, 1989). The genome of H. cutirubrum also contains a closely related sod-like gene (slg) of unknown function that has a pattern of expression different from that of sod. The four amino acid residues that bind the Mn atom are conserved, but the flanking regions of the two genes are unrelated. Although the genes have 87% nucleotide sequence identity, the proteins they encode have only 83% amino acid sequence identity. Mutations occur randomly at the first, second, and third codon positions, and transversions outnumber transitions. Most of the mutational differences between the two genes are confined to two limited regions; other regions totally lack differences. These two gene sequences are apparently in the initial stage of divergent evolution. Presumably, this divergence is being driven by strong selection at the molecular level for either acquisition of new functions or partition and refinement of ancestral functions in one or both of the respective gene products.

ancestral gene; the two sequences are now undergoing divergent evolution driven by selection at the molecular level for different functions.

Archaebacteria are a diverse group of procaryotes that differ fundamentally from eubacteria. Comparisons of rRNA sequences indicate that they diverged from the eubacterial and eucaryotic lineages over 3.5 x 109 years ago (26). The extremely halophilic archaebacteria are facultative anaerobes that inhabit salt brines containing 1.5 to 5 M NaCl (7). Halobacterium cutirubrum and its close relative Halobacterium halobium are distinguished from other halobacteria by production of proton- and chloride ion-pumping proteins (bacteriorhodopsin and halorhodopsin, respectively) and by a high frequency of genome rearrangements mediated by insertion elements (2, 17). During their evolution from the anaerobic methanogens, the halophilic archaebacteria have acquired oxygen tolerance (26). The enzyme superoxide dismutase (SOD) is an important component of oxygen tolerance because it catalyzes the conversion of reactive superoxide radicals (02) to peroxide and dioxygen (4). H. halobium and H. cutirubrum are independent isolates of the same species, Halobacterium salinarium (7). Both H. cutirubrum and H. halobium contain a single detectable, Mn-containing SOD (9, 15). The SOD of H. cutirubrum is 200 amino acid residues in length, functions optimally in 2 M KCI, is encoded by the sod gene on a 1.1-kilobase-pair (kbp) Sau3AI restriction fragment, and is homologous to the Mn- and Fe-containing SODs of eubacteria and unrelated to the Cu-Zn-containing SODs of eucaryotes (9, 10). During verification of the structure of the sod gene clone by Southern hybridization to genomic DNA, a second sequence representing a closely related gene, here designated slg, was discovered (10). Two separate reports of a sequence from H. halobium identical to slg have been published recently (14, 25). In both reports, the slg-equivalent sequence was imputed to encode SOD activity. This appears to be incorrect, because the amino acid sequence of the purified SOD differs substantially from the sequence predicted for the protein product of the slg gene (9, 15). No second SOD activity that could correspond to this protein product has been detected. We suggest that sod and slg were produced by the duplication of an *

MATERIALS AND METHODS Culture conditions. For isolation of DNA, H. cutirubrum was grown in rich medium containing salts, yeast extract, and Casamino Acids (9). Cultures treated with paraquat were grown in a defined medium containing salts, glycerol, and amino acids (1). H. cutirubrum, H. salinarium, and H. halobium are independent isolates of a single species, H. salinarium (7). Analysis of DNA. Genomic DNA was isolated as described previously (18). Oligonucleotides were synthesized on an Applied Biosystems 380B DNA synthesizer and were purified on polyacrylamide sequencing gels. The sequences of the oligonucleotides are: 5'-TGGAGGATGTGCCCCGAGCC-3' (sod specific) 5'-TCGGCGCTGTTGTGGCCGTC-3' (sig specific) Molecular cloning, Southern blotting, and labeling of DNA were performed by standard methods (8). Cloned DNA was sequenced by dideoxy chain termination using the Klenow enzyme (or T7 DNA polymerase) and M13 phage vectors (11, 16, 24). Analysis of RNA. Total cellular RNA was purified by lysis in buffer containing sodium dodecyl sulfate and by subsequent centrifugation through CsCl as described previously (10). Previously described methods were used for primer extension of DNA along RNA templates and for nuclease S1 analysis of RNA-DNA hybrids (3, 12). Restriction fragments used as probes in Southern or Northern (RNA) hybridization were labeled by the mixed-primer method, as described previously (10). RESULTS AND DISCUSSION Identification and sequence of the slg gene. When restriction enzyme-digested genomic DNA was probed with the 1.1-kbp Sau3AI fragment that contains the sod gene, a

Corresponding author. 3725

3726

MAY AND DENNIS

kbcL

11.5-

4.5-

2.8-0

0.5-

FIG. 1. Southern blot of genomic DNA of H. cutirubrum probed with the cloned sod gene. The lane contains 1 p.g of PstI-digested DNA. The blot was hybridized to the 1.1-kbp Sau3AI fragment that carries sod and was washed at high stringency (8).

strong hybridization signal and a weak

one were seen (Fig. 1). In a PstI-digested DNA, the major signal was from a 3.0-kbp fragment that contained the 1.1-kbp Sau3AI fragment and carries the authentic sod gene. The minor signal was from a 1.8-kbp fragnent. This fragment was cloned into pUC13, and both strands of overlapping subfragments were sequenced. The sequence was found to contain an open reading frame, designated slg (sod-like gene), that encodes a protein of 200 amino acid residues (Fig. 2). The slg gene has 87% nucleotide sequence identity with the sod gene, and the putative protein that it encodes has 83% amino acid sequence identity with SOD (Fig. 2). When compared to the N-terminal amino acid sequence of SOD purified from H. cutirubrum, the predicted slg-encoded protein differs at 10 of 56 residues, including the Ser and Glu residues at positions 2 and 52, respectively, that are also different in the protein predicted from the sod gene sequence. These latter two differences appear to be the result of protein sequencing errors (9, 10). Likewise, when compared with the purified SOD of H. halobiuin, the predicted slg-encoded protein differs at 8 of 26 residues (14, 15, 25). The putative protein encoded by sig retains the four residues, His-29, His-76, Asp-158, and His-162, that correspond to those in the eubacterial enzyme that bind the Mn atom (22). The 5' and 3' flanking regions of sod and slg appear to be unrelated. Differentil expression of sod and sig. Expression of the sod gene has been shown to respond to oxidative stress (10). When cultures of H. cutirubrum gowing aerobically in a defined medium were exposed to the redox cycling drug paraquat, sod mRNA levels and SOD activity were elevated four to fivefold. The amounts of sod and slg mRNA were quantitatefd by using total RNA from untreated and paraquat-treated cells in a primer extension assay (Fig. 3). Two separate primers, one specific for sod mRNA and one specific for slg mRNA, were used (see Fig. 2 for binding sites of the primers). The slg transcript has a unique 5' end at position 162 that yields a 5' untranslated leader of 13 nucleotides. The sod mRNA initiates at positions 172 to 173

J. BACTERIOL.

and has a leader of only two or three nucleotides. Most importantly, the level of sod mRNA responds to paraquat, whereas the level of slg mRNA is apparently unaffected. No cross-pnming by either oligonucleotide on the heterologous transcript was detected. Halobacterial promoters contain an AT-rich element and a sequence resembling TTCGA about 25 to 30 nucleotides and 35 to 40 nucleotides, respectively, upstream of the transcription initiation site (13, 20). Both sod and slg contain such elements in the appropriate positions. The slg promoter matches the consensus somewhat more closely than does the sod promoter (Fig. 2). The sequences and mechanisms responsible for the differential expression of the two genes have not yet been identified. The 3' ends of slg transcripts were mapped by nuclease S1 analysis to a T tract at position 795 to 801 (Fig. 2). Such T tracts are typical of halobacterial transcription termination sites (20). Again, no increase in slg mRNA was seen in response to paraquat (Fig. 3C). Recently, Takao et al. (25) have found that the slg homolog in H. halobium is located 3' to the photolyaseencoding gene. Their Northern hybridization results indicate that the low-abundance transcripts of the photolyase gene extend through the slg gene. Our unpublished results confirm the presence of the longer transcript. On the basis of the sensitivity of our primer extension, our S1 nuclease protection and Northern hybridization results, and the Northern hybridization results of Takao et al., we conclude that the monocistronic slg transcript is about 100-fold more abundant than the longer polycistronic transcript. The TGA translation termination codon of the photolyase gene is located at position 108 in the slg sequence illustrated in Fig. 2. The 65-nucleotide-long intergenic space (positions 111 to 174) containing the major sig promoter noticeably lacks an extended T tract sequence normally associated with halophilic transcription terminators; the significance of this low level of cotranscription remains to be explained. Evolutinary divergence. The function of the putative slg product is an enigma. Only a single SOD activity has been detected in and purified from cell extracts of H. cutirubrum and H. halobium (9, 15). The SOD activity of H. cutirubrum is encoded by the sod gene (10). The related slg gene produces an easily detectable transcript that is at least as abundant as the sod transcript but is not inducible by paraquat. The putative protein encoded by slg retains the residues used for Mn binding, but it also contains a single cysteine, a residue notably lacking in Fe- and Mn-containing SODs (23). The slg gene either continues to encode an SOD activity that has escaped detection or, more likely, encodes a different activity. Also, the unlikely possibility that the mRNA is not translated cannot yet be ruled out. Even a perfunctory comparison indicates that the sod and slg coding sequences do not exhibit the pattern of random drift expected for duplicated genes whose product proteins are under selection for conservation of function. Selection for conservation would produce occasional mutational differences at the silent third position of codons, less-frequent differences resulting in conservative amino acid substitution, and only rarely, differences that produce nonconservative amino acid substitutions. Rather, sod and slg appear to be under strong selection at the molecular level to produce proteins with different functions; mutational differences between slg and sod occur almost randomly among the first, second, and third codon positions, transversions outnumber transitions, and most nucleotide substitutions result in amino

SUPEROXIDE DISMUTASE-LIKE GENE

VOL. 172, 1990

3727

sod GRT CGCGCG TTGT TCG TCGCT GRGT TCGRRG TCCRTGTGG TRT CRT CCRRCRT RCR TCR TGRRRRR TRC T TGTRRRCGCT CGGCR TCRCT TCCGCGGT TGCCG TCCCGTCRCGRRCCCRR so * 00 e * *-@@@@@@@e *so@ -@0 *@@esooe g osees 0 age go-e@@@s*e *@@se* eo**--oo e@s -see *-@@@@s-ese TGT CCGRGCGCCGCCGRCRCGCCCCGGRG TR TCCGGRCCCCRT CG TGGRCCRCRGCCRGCGCCGCGRGGRCGCGRTCGCGRT G TTCGRGCG TGCGCGCGGCGRCGRGTGRGCCCRCGRTC L S E R R R H R P E Y P D P I U D H S Q R R E D R I R M F E R R R G D E TER 20 60 80 100 120 slg 10 10 20 -40 -30 n S E Y E L P P L P V DY D R L E P H I S E CT GRRRCT GCRTT CCGGRRRCCRCCRT RRGCRGCGCCGRCGTRCGRCRCRC TGTRT GTCCGRRTRCGRRC TCCCRCCGCTGCCGT RCGRC TRCGRCGCGCT CGRRCCRCRCRT CRGCGRG

CT CRCGRRCRT U RRCRT GRCGCCGCG TGRTCRC TGRTCCGG TGGR TTCCRCCGRT GRGCCRGCRCGRRCTCCCRTCGC TGCCGT RCGRC TRCGRCGCRCT CGRRCCRCRCRT CRG TGRG -10 -30 8 Q S - - - - - - - - - - - - - 110 160 220 240 I 60 200

%

30 40 50 60 H D T H H Q G V U N G U N D R E E T L R E N R E T G D H A S T R G A Q U L T U CRGGTG C TCRCGT GGCRTCRCGRCRCCCRCCRT CRGGGCT RCGTGRRCGGCT GGRRCGRCGCCGRGGRGRCRC TCGCGGRGRRCCGTGRGRCCGGCGRCCRCGCCTCGRCRGCCGGCGCG

[|

CRGGT GGTCRCGTGGCRCCRCGRCRCCCRCCRCCRGRGC TRCGT GGRCGGCC TCRRCRGEGCCGRGGRGRCGCTGGCGGRGRRCCG TGRGRCCGGCGRCCRCGC TTCGRCTGCCGGCGCG --0 - L - S - - - - - - - - - - - - - - - - - - - U -- _ 260 300 280 320 340 360

70 90 00 100 L G D U T H N G S G H I T 0 F U 0 S n s P R G G D E P S G A L R D R I A R D CT CGGGGRCGTCRCGCRCRRCGGCTCGGGGCRCRT CCTCCRCRCGCT GTTC TGGCRGTCCRTGRGCCCGGCGGGCGGCGRCGRGCCG TCCGGGGCGC TCGCCGRCCGCR TCGCGGCGGRC

L[3

**

0*

*

*

so

**

CT CGGG GRCGT CRCGCRCRRCGGCTG TGGGCRC TRCC TCCRCRCGR TGT TC TGGGRGCRCRTGRG TCCCGRCGGGGGCGGCGAGCCGT CCGGGGCGCT CGCCGRCCGCRT CGCGGCGGRC V E H ---D --G - - - - - - - - - - - - --- - -c -M 1 380 400 110 120 460 180

110 120 130 110 F G S V E N U R R E F E R R R S R R S G U R L L U VD S H S N T L R N U R U 0 T TCGGC TCC TRCGRGRRCT GGCGGGCCGRGT TCGRGGCCGCCGCCRGCGCGGECCRGCGGCTGGGCGCT GCTCGT CTRCGRC TCCCRC RGCRRCRCGCTCCGGRRCG TGGCCGTGGRCRRC

RCGRT;C;;CGGCCRRGCRGCTCCGGRRCGTGGCCGTGGRCRRC

T TCG GCT CCT RCGRGRRC TGGCGGGCT GRRT TCGRGG TGGCGGCCGGCGCGGCCRGCGGCT GGGCGC TGCT CG TCT - U -- G - - - - - - - - - - - P U R K Q - 500 520 510 560 580

-

-

-

-

-

-

600

150 160 170 180 H 0 E G R L U G S H P I L A L QiIU U E jES V v Y D V G P 0 R G S F U 0 R F F E U CRCGRCGRGGGCGCGC TCTGGGGCRGCCRCCCCRTCCTCGCGCTCGRCGTCTGGGRGCRCTCCTRCTRC TRCGRC TRCGGTCCCGRCCGCGGCRGCTTCGT CGRCGCCT TCT TCGRGG TC CRC GRCGRGGGCGCGCT CT GGGGCRGCCRCCCCRT CCTCGC;CCTCGRCGT CTGGGRGCRCT CCT RCTRCTRCGRCTRCGGCCCCGRCCGCGGCRGCT TCG TCGRCGCCT TCTTCGRGGTG

610

620

660

680

700

720

190 200 U 0 U 0 E P T E R F E Q R A E R F E 00000 GTC GRCT GGGRCG,RGC,CCRCCG.R.G.CGCT TCGRGCRGGCGGCCGRGCGCT TCGRG TRRCGCCCCGCCCGCGGGGRCRCCC TGRRRCGCCRCGCT T TTTCGCCGTGTRGCGT TCCGRTRGCT . .... ..........: MM:: ...........

"::::::..:::

C GCGTRTGCGT;;CRT GTG;;GCTGGCGT GRGGCG CGRCTGGGRCC;;CCRC;GCGG;CTRCCGRCGRCG;GG;GTCGCTGTTCGRGTGRC;;CGR CRCGC;CCGTGT T T T T T CT --

RT I - -- P

I

R R N Y 0D 740

U U S L 760

860

880

00000

780

800

820

810

FIG. 2. Nucleotide and amino acid sequences of sod and slg. The sod (top) and slg (bottom) nucleotide and amino acid sequences are aligned over their open reading frames. Nonidentical nucleotides are indicated (0); only nonidentical amino acids are indicated in the slg protein sequence. Nucleotide and amino acid positions are indicated below and above each line, respectively. The four boxed amino acids in SOD are residues used for Mn2" binding. Sites of initiation (solid boxes with arrows) and termination (0) are indicated for both the sod and slg transcripts. The putative -30 and -40 promoter elements preceding the transcription initiation sites are indicated. The binding sites for the slg- and sod-gene-specific 20-mer oligonucleotides used in the primer extension analysis are underlined. The EcoRI site in the sig gene is at position 508. The multiple and overlapping direct repeat sequences unique to the 3' end of the sod gene are indicated (stippled underline). The carboxy-terminal amino acid sequence of the photolyase protein is illustrated below the corresponding gene sequence; the translation termination codon is at position 108 (25).

acid substitutions (Table 1). Only 12 of the 35 amino acid changes would normally be considered conservative. Two other pairs of evolutionarily related coding regions from H. halobium have been analyzed (Table 1). The closely related gas vacuole genes (c-vac and p-vac) encode the major protein of the gas vacuoles that are responsible for buoyancy of cells (6). The two sequences have differences typical of genes with a conserved function; most changes occur at the third position of codons and are consequently silent, and transitions outnumber transversions (6). The more distantly related bacterio-opsin and halo-opsin genes (bop and hop) have diverged substantially since duplication and encode proteins with different functions (2). During phototrophic

growth, the former protein pumps protons outward to produce an electrochemical gradient for the generation of ATP and the latter protein pumps chloride ions inward for maintenance of intracellular isotonic conditions (19). Changes between the bop and hop genes are distinctive and characteristic of selection for different functions; differences are evenly distributed among codon positions, and transversions outnumber transitions (2). Whereas in the bop-hop example the genes have clearly reached a near-terminal stage in divergence and functional differentiation, the sod-slg example appears to provide an intriguing and rare glimpse of a much earlier stage in this process. Unfortunately, it is not possible to explicitly enun-

J. BACTERIOL.

MAY AND DENNIS

3728

C

A

AGCT 1 2

B

l 2IA;1

I,

AGC T 1 2 ...

_

: _.

NW

:, i

_

aft

kb

!!i.

.--

u.

_,; ...

FIG. 3. Mapping and quantitation of the transcripts of slg and sod. (A) One nanogram of a 5'-end-labeled 20-mer oligonucleotide complementary to codons 38 to 44 of slg was annealed to 10 ,ug of RNA from untreated (lane 1) or paraquat-treated (lane 2) cells and used to prime reverse transcription. The sequencing ladder was made by using the slg-specific 20-mer to prime chain termination reactions on a single-stranded M13 clone of slg. (B) One nanogram of a 5'-end-labeled 20-mer oligonucleotide complementary to codons 70 to 76 of sod was used to prime reverse transcription of 10 ,ug of RNA from untreated (lane 1) or paraquat-treated (lane 2) cells. The sequencing ladder was made by using the sod-specific 20-mer to prime chain termination reactions on a single-stranded M13 clone of sod. (C) A 1-kbp EcoRI-PstI fragment (145 ng) spanning the 3' end of slg was labeled at the 3' end of the EcoRI site (positions 508 to 513), hybridized to 10 ,ug of RNA from untreated (lane 1) or paraquat-treated (lane 2) cells, and digested with nuclease Si (3). Markers (lane M) were Sau3AI fragments of plasmid pUC13. nt, Nucleotides.

ciate the mechanism that resulted in the duplication of the primordial sod-slg sequence, the time at which this duplication occurred, or the forces that are driving the divergent evolution. Curiously, the positions of divergence between the two gene sequences are not randomly distributed; they are concentrated in two clusters and result in extended regions of nonidentical amino acids in the two proteins. The first is between positions 561 and 575; the second, longer cluster is between positions 733 and 768. Other portions of the coding sequence are bereft of even silent mutations (e.g., positions 577 to 639 of Fig. 2). There are two possible explanations for this mutational clustering. The regions containing the multiple differences may encode important functional domains

upon which selection is operating in one or both of the proteins. A mutational event in such a region that improves function would rapidly spread through the population, displacing unmutated or randomly mutated gene copies. If such displacements occur frequently, neutral mutations in functionally unimportant regions will have a diminished probability for fixation. Alternatively, regions representing the distinctive functional domains in the contemporary sod and slg proteins may be much more subtle and the function of both proteins may depend on a common catalytic mechanism, possibly involving the prosthetic Mn2" ion. Conservation of the regions encoding these common domains and their protection against mutational obliteration may be of paramount importance; this could be ensured by gene conversion-type mechanisms between the homologous regions of the sod and slg genes. Within the two genes, frequently converted regions would maintain sequence identity, whereas regions not subject to conversion would exhibit sequence divergence. A similar mechanism has been proposed to explain conservation of sequence within and between the protocatechuate and catechol oxygenase genes of Acinetobacter calcoaceticus (5). Another interesting feature is apparent within the unique 3' end of the sod gene. This region contains four identical direct repeats and one nearly identical direct repeat of the pentanucleotide CGAGC at positions 732, 741, 750, 762, and 771. Each of these is centered by a GAG glutamic acid codon. Overlapping and encompassing these pentanucleotide repeats are two longer 14-nucleotide direct repeats at positions 740 and 761. Each encodes the tetrapeptide Glu Arg Phe Glu that is repeated in the unique carboxy terminus of the SOD. Such repeated sequences might protect this region from inactivating mutations by a slippage repair process (again, see reference 5). In summary, the duplicated sod and slg genes of H. cutirubrum exhibit a highly unusual pattern of nucleotide sequence divergence. This divergence is apparently being driven by strong selective forces operating at the molecular level to produce two proteins with different, although possibly related, functions. Nucleotide differences between sod and slg are clustered and occur randomly at first, second, and third codon positions such that almost every nucleotide difference results in an amino acid substitution. Most of these amino acid substitutions are considered to be nonconservative. This pattern of sequence evolution clearly contrasts with the pattern observed when the molecular function of the gene product is being conserved (6, 21). At the transcription level, the sod and slg genes appear to be regulated independently and their respective 5' and 3' flanking regions lack detectable nucleotide sequence homology.

TABLE 1. Comparison of pairs of related genes from H. halobium and H. cutirubrum Protein gene

Nucleotide

Amino acid

paira

identity (%)b

identity (%)c

Transitions/ transversions

c-vac, p-vac bop, hop sod, slg

81 50 87

97 30 83

1.5 (21:14) 0.53 (133:249) 0.64 (30:47)

Distribution of differences'

Evolutionary

1

2

3

function

5 141

3 131

36 138

Conserved

26

21

31

Changed ?

a c-vac and p-vac, Gas vacuole genes from the chromosome and plasmid, respectively, of H. halobium (6); bop and hop, bacterio-opsin and halo-opsin genes from H. halobium (2); sod and slg, SOD and SOD-like genes from H. cutirubrum. b Nucleotide identity is the percentage of nucleotides that are identical within the coding regions in the respective aligned sequences. c Amino acid identity is the percentage of amino acids that are identical in the aligned coding regions, excluding deletions. d The distribution between first, second, or third codon position of mutational differences between the diverging sequences being compared.

VOL. 172, 1990

ACKNOWLEDGMENTS We thank Lawrence C. Shimmin, Craig H. Newton, and W. Ford Doolittle for helpful discussions and Deidre deJong-Wong for technical assistance. This work was supported by the Medical Research Council of Canada and the U.S. Office of Naval Research. Patrick P. Dennis is a fellow of the Canadian Institute for Advanced Research. Bruce P. May was supported by a Medical Research Council studentship. LITERATURE CITED 1. Bayley, S. T. 1971. Protein synthesis systems from halophilic bacteria. Methods Mol. Biol. 1:89-100. 2. Blanck, A., and D. Oesterhelt. 1987. The halo-opsin gene. II. Sequence, primary structure of halorhodopsin and comparison with bacteriorhodopsin. EMBO J. 6:265-273. 3. Dennis, P. P. 1985. Multiple promoters for the transcription of the ribosomal RNA gene cluster in Halobacterium cutirubrum. J. Mol. Biol. 186:457-461. 4. Fridovich, I. 1986. Superoxide dismutases. Adv. Enzymol. Relat. Areas Mol. Biol. 58:68-97. 5. Hartnett, C., E. L. Neidle, K.-L. Ngal, and L. N. Ornston. 1990. DNA sequences of genes encoding Acinetobacter calcoaceticus protocatechuate 3,4-dioxygenase: evidence indicating shuffling of genes and of DNA sequences within genes during their evolutionary divergence. J. Bacteriol. 172:956-966. 6. Horne, M., C. Englert, and F. Pfeifer. 1988. Two genes encoding gas vacuole proteins in Halobacterium halobium. Mol. Gen. Genet. 213:459-464. 7. Larsen, H. 1984. Family V. Halobacteriaceae, p. 261-267. In N. R. Krieg and J. G. Holt (ed.), Bergey's manual of systematic bacteriology, vol. 1. The Williams & Wilkins Co., Baltimore. 8. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 9. May, B. P., and P. P. Dennis. 1987. Superoxide dismutase from the extremely halophilic archaebacterium Halobacterium cutirubrum. J. Bacteriol. 169:1417-1422. 10. May, B. P., and P. P. Dennis. 1989. Evolution and regulation of the gene encoding superoxide dismutase from the archaebacterium Halobacterium cutirubrum. J. Biol. Chem. 264:1225312258. 11. Messing, J. L. 1983. New M13 vectors for cloning. Methods Enzymol. 101:20-79. 12. Newman, A. 1987. Specific accessory sequences in Saccharomyces cerevisiae introns control assembly of pre-mRNAs into

SUPEROXIDE DISMUTASE-LIKE GENE

3729

spliceosomes. EMBO J. 6:3833-3839. 13. Reiter, W.-D., P. Palm, and W. Zillig. 1988. Analysis of transcription in the archaebacterium Sulfolobus indicates that archaebacterial promoters are homologous to eucaryotic polll promoters. Nucleic Acids Res. 12:7949-7959. 14. Salin, M. L., M. V. Duke, D. Oesterhelt, and D.-P. Ma. 1988. Cloning and determination of the nucleotide sequence of the Mn-containing superoxide dismutase gene from Halobacterium halobium. Gene 70:153-159. 15. Salin, M. L., and D. Oesterhelt. 1988. Purification of a manganese-containing superoxide dismutase from Halobacterium halobium. Arch. Biochem. Biophys. 260:806-810. 16. Sanger, F., S. Nicklen, and A. R. Coulsen. 1977. DNA sequencing with chain terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463-5467. 17. Sapienza, C., and W. F. Doolittle. 1982. Unusual physical organization of the halobacterial genome. Nature (London) 295:384-389. 18. Schnabel, H., W. Zillig, M. Pfaffie, R. Schnabel, H. Michel, and H. Delius. 1982. Halobacterium halobium phage phiHl. EMBO J. 1:87-92. 19. Schobert, B., and J. K. Lanyi. 1982. Halorhodopsin is a lightdriven chloride pump. J. Biol. Chem. 257:10306-10313. 20. Shimmin, L. C., and P. P. Dennis. 1989. Characterization of the Lii, Li, L10, and L12 equivalent ribosomal protein gene cluster of the halophilic archaebacterium Halobacterium cutirubrum. EMBO J. 8:1225-1235. 21. Smith, T. F., A. Srinivasan, G. Schechetman, M. Marcus, and G. Myers. 1988. The phylogeneic history of immunodeficiency viruses. Nature (London) 333:573-575. 22. Stallings, W. C., K. A. Pattridge, R. K. Strong, and M. L. Ludwig. 1984. Manganese and iron superoxide dismutases are structural homologs. J. Biol. Chem. 259:10695-10699. 23. Steinman, H. M. 1982. Superoxide dismutases: protein chemistry and structure-function relationships, p. 11-68. In L. W. Oberley (ed.), Superoxide dismutase, vol. 1. CRC Press, Inc., Boca Raton, Fla. 24. Tabor, S., and C. C. Richardson. 1987. DNA sequence analysis with modified bacteriophage T7 DNA polymerase. Proc. Natl. Acad. Sci. USA 84:4767-4771. 25. 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. 26. Woese, C. R. 1987. Bacterial evolution. Microbiol. Rev. 51: 221-271.