- Email: [email protected]
F0F1-ATPase Genes from an Archaebacterium,Methanosarcina barkeri
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ARTICLE NO.
241, 427–433 (1997)
RC977809
F0F1-ATPase Genes from an Archaebacterium, Methanosarcina barkeri Masato Sumi,* Masafumi Yohda,† Yosuke Koga,‡ and Masasuke Yoshida§,1 *Faculty of Pharmaceutical Sciences of Hokkaido University, Nishi-6, Kita-12, Kita-ku, Sapporo, Hokkaido 060; †Biochemical Systems Laboratory of the Institute of Physical and Chemical Research (RIKEN), Hirosawa 2-1, Wako, Saitama 351-01; ‡Department of Chemistry, University of Occupational and Environmental Health, 1-1, Ibugaoka, Yahatanishi-ku, Kitakyushu 807; and §Research Laboratory for Resources Utilization, Tokyo Institute of Technology, Nagatsuta 4259, Midori-ku, Yokohama 226, Japan
Received November 5, 1997
It has been known that an archaebacterium Methanosarcina barkeri strain MS (DSM 800) has a V-type ATPase (Inatomi, K., et al. (1989) J. Biol. Chem. 264, 10954–10959). Here, we report cloning of a cluster of F0F1-ATPase genes from the same organism, the first ever found in archaebacteria. The cluster and encoded subunits exhibit several unusual features such that a gene for d subunit is lacking, F0-b subunit is unusually large, and g subunit is split into two peptide fragments. Attempts to detect F0F1-ATPase proteins and mRNA have been unsuccessful and therefore it is not certain if this gene cluster is really expressed in the cell. q 1997 Academic Press
F0F1-ATPase and vacuolar(V)-type ATPase are two evolutionally related subclasses of a superfamily of proton translocating ATPases which do not form phosphorylated enzyme intermediates (1, 2). F0F1-ATPase are large (Ç500kDa) protein complexes in which a watersoluble catalytic sector, F1 , is associated to a relatively small integral membrane sector, F0 , which acts as a proton channel (3, 4). Subunit compositions of prokaryotic F1 and F0 are a3b3gde and ab2c12? , respectively. Eukaryotic F0F1-ATPases have additional subunits. The a and b subunits are homologous each other, and are catalytic and regulatory subunits, respectively. Recent achievement of crystal analysis (5) and single molecule observation (6) revealed that F0F1-ATPase is the 1 To whom correspondence should be addressed. E-mail: [email protected]. The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM Data Bank with accession number AF028006. Abbreviations: aa, amino acid residue; % identity, % of identical amino acid residues between aligned sequences of two proteins; PCR, polymerase chain reaction.
enzyme with a surprisingly novel mechanism; the centrally located g subunit rotates in the a3b3 hexagon during catalysis (3). Structure and mechanism of Vtype ATPase has been less understood but are thought to be essentially similar to those of F0F1-ATPase. F0F1ATPases are found in membranes of mitochondria, chloroplasts and eubacteria, and synthesize ATP using energy of proton translocation. V-type ATPases are responsible for acidification of eukaryotic vacuolar vesicles including lysosomes and chromaphin granules. They are also found in membranes of archaebacteria, such as Sulfolobus acidocaldarius (7-10), Methanosarcina barkeri (11), and Halobacterium salinarium (12). Agreement of taxonomic border to distribution of F0F1- and V-type ATPases in eubacteria and archaebacteria was taken as an indication of archaebacterial origin of eukaryotic vacuolar systems. However, it was recently found that distribution of V-type ATPase is not restricted within archaebacteria; a thermophilic eubacterium Thermus thermophilus has a V-type ATPase (13-15) and an eubacterium Enterococcus hirae has both types of ATPases (16-18). These facts raise a possibility that restricted distribution of F0F1-ATPase in eubacteria and V-type ATPase in archaebacteria is not as strict as it used to be thought and that some archaebacteria may have F0F1-ATPase. We report here the entire nucleotide sequence of a cluster of the F0F1ATPase genes in an archaebacterium, Methanosarcina barkeri. This paper contains the correction of our previous paper (19). MATERIALS AND METHODS Strain and medium. Methanosarcina barkeri strain MS (DSM 800) was obtained from the Deutsche Sammlung fu¨r Mikroorganismen und Zellkulturen (DSM) and grown under strictly anaerobic conditions (80 % N2 / 20 % CO2) in the medium (pH 6.85) containing
0006-291X/97 $25.00
427
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
BBRC 7809
/
6942$$$701
12-01-97 23:15:56
bbrcgs
AP: BBRC
Vol. 241, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
0.5 % methanol, 40 mM Na2 and other components as described in DSM 120. PCR reaction. Genomic DNA from M. barkeri was prepared as described in (20). Oligonucleotide primers for polymerase chain reaction (PCR) were as follows: Fs1, GGNGGNGCNGGNGTNGGNAARAC; Fs2, GARMGNACNMGNGARGGNAAYGA; Fa, TCNGTBAGRTCRTCNGCNGGNACRTA (NÅA, C, G or T; RÅA or G; YÅT or C; MÅA or C; BÅC, G, or T). The reaction mixture (100 mL) for PCR contained 200 ng of the template M. barkeri genomic DNA, 1 mM of each primer (Fs1 and Fa), 1 mM MgCl2 , 2.5 units Taq DNA polymerase (Promega) in the buffer provided by a Promega kit (50 mM KCl, 10 mM Tris HCl, pH 9.0 at 257C, 0.1% Triton X-100). The PCR reactions were carried out for 1 min at 947C, 1 min at 50 7C, 1 min at 74 7C, and were repeated for 30 cycles (the final step at 74 7C was extended to 10 min). The products were analyzed with 6 % polyacrylamide gel and visualized with ethidium bromide. The DNA fragment of interest (Ç490 bp DNA) was eluted from the gel and used as a template DNA for the second amplification which was performed to decrease non-specific products. The reaction mixture (100 mL) for the second amplification contained 2 ng of the template 490 bp DNA, 1 mM of each primer (Fs2 and Fa), 2.5 units Taq DNA polymerase in the buffer provided by a Promega kit. The product was electrophoresed in 2 % agarose gel, visualized with ethidium bromide, and the section containing a 380 bp fragment was cut out. For isolation of DNA fragments from the agarose gel, Geneclean II kit (Bio 101, Inc.) was used. The purified DNA fragment was ligated into the EcoRV site of pBluescript II SK0. The obtained recombinant plasmids were then introduced into Escherichia coli JM109. DNA sequence analysis. Southern hybridization of the genomic DNA and colony hybridization were carried out using the ECL random primer labeling and detection systems (Amersham). The PCR fragment was isolated from the plasmids by EcoRI and HindIII endonuclease treatment and was used as a template DNA to make labeled probes. A genomic BamHI library in the pBluescript II SK0 vector from M. barkeri was constructed. Colony hybridization was performed according to (21). Positive clones were selected and plasmids were isolated from the clones. The BamHI fragment was purified from the plasmids and broken into segments by sonication for random sequencing. DNA fragments of the desired size (500-800 bp) were isolated by gel electrophoresis and their termini were repaired with a blunting kit (TaKaRa). The DNAs were then ligated to SmaI site of pUC18 and then introduced into E. coli. Double-stranded plasmids prepared from colonies were used as templates for sequencing. All DNA sequences were confirmed by sequencing both strands. Protein sequence comparisons were carried out by using the Pileup program of University of Wisconsin Genetics Computer Group (GCG) packages.
RESULTS AND DISCUSSION Isolation of atp Gene Cluster To ensure the specificity of amplification, the segment of amino acid sequences, well conserved among F0F1-ATPases but not among V-type ATPases, were chosen to design PCR primers. Several F0F1-ATPase b (catalytic) subunit sequences from various organisms including E. coli (22), Rhodospirillum rubrum (23), Bacillus PS3 (24), bovine heart mitochondria (25), and spinach chloroplasts (26) were aligned. Three well-conserved regions in the sequence alignment were chosen, and used for design of two sense primers (Fs1 and Fs2) and an antisense primer (Fa), respectively.
A 490 bp DNA fragment was amplified by PCR using Fs1 and Fa primer from M. barkeri genomic DNA. The second PCR was carried out using Fs2, instead of Fs1, in order to confirm that the amplified DNA fragment of the first PCR really contained an internal sequence common to F0F1-ATPase b subunits. A 380 bp DNA fragment was amplified by the second PCR as expected. The 380 bp fragment was subcloned and sequenced. The deduced amino acid sequence from the DNA fragment was novel, and seemed to be a piece of F0F1-ATPase b subunit. We reported previously nucleotide sequence of another 380 bp PCR fragment that was obtained by the same methods from the same organism (19). Unexpectedly, the sequence of the fragment obtained this time was different from the previously reported one. Unfortunately the genomic DNA used in the previous study was used up completely and we could not examine why the difference occurred. In view of this trouble, we carefully tested the accuracy of PCR product and quality of genomic DNA. Using the same genomic DNA, existence of the gene encoding the catalytic subunit of Vtype ATPase reported by Inatomi et al (11) was confirmed by the same PCR method with appropriate primers. Southern analysis of M. barkeri genomic DNA using the 380 bp DNA fragment obtained this time as a probe was carried out independently in other laboratory and produced the same consistent result that we had (Mu¨ller, V., et al.; personal communication). The 380 bp fragment (Fig. 1A, white small box on B) hybridized with an 8 kbp BamHI fragment of M. barkeri genomic DNA. A plasmid library of the genomic DNA containing Ç8 kbp BamHI fragments (about 100 clones) was screened with the probe, and one positive clone (Fig. 1A, thick line B) was obtained. Sequence analysis revealed that the insert contained most atp genes but lacked 3* terminus of a subunit gene (and all g subunit gene). Then, using a Ç800 bp SacI fragment derived from 3* terminus of the BamHI fragment as a probe (Fig. 1A, white small box on E), another plasmid library made up from Ç4 kbp EcoRI fragments of genomic DNA (about 400 clones) was screened and four positive clones containing a 4 kbp EcoRI fragment were obtained (Fig. 1A, thick line E). Sequence analysis showed that the fragment contained the rest of atp genes and transcriptional termination region of the cluster. The nucleotide sequences has been submitted to the GenBankTM Data Bank with accession number AF028006. Organization of atp Genes The G/C content of the insert is 43 % which is close to the overall G/C content of this bacterium (36-43%) (27). In this clone, ten genes are arranged in the order atpD, -C, -I, urf2, atpB, -E, -F, -A, -G, and -J (Fig. 1A)
428
AID
BBRC 7809
/
6942$$$702
12-01-97 23:15:56
bbrcgs
AP: BBRC
Vol. 241, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. (A) Arrangement of genes and a restriction map of the atp gene cluster of M. barkeri chromosomal DNA. Open boxes indicate the open reading frames for atp genes. The letter A, B, C, D, E, F, G, I, and J in the open boxes indicated the genes of F0F1-ATPase subunits. The urf1, urf2, and urf3 indicate unknown putative open reading frames. The black thick lines, B and E, indicate the cloned 8,193 bp BamHI fragment and the 4,614 bp EcoRI fragment, respectively. The white boxes on these thick lines indicate the portion of PCR-amplified fragment (on B) and SacI fragment (on E) that was used for Southern and colony hybridization as probes (See Materials and Methods). Restriction sites: B, BamHI, E, EcoRI; H, HindIII; S, SacI; X, XhoI. (B) Comparison of arrangement of genes encoding F0F1-ATPase subunits in the atp gene cluster (operon) of M. barkeri, E. coli and Synechococcus 6301. The letters a, c, b, and b* represent genes for F0 subunits. Letters d, a, g, b, and e represent genes for F1 subunits. The letter I indicates the atpI gene. The letters gN and gC indicate N- and Cterminal halves of g subunit, respectively. X is a putative open reading frame whose deduced amino acid sequence has no similarity to any other proteins.
with the same transcriptional direction. There is no possible open reading frame in the immediate neighborhood at upstream of the 5* end of atpD and downstream of the 3 * end of atpJ. The sequence similar to the archaebacterial promoter element, which is homologous to the TATA box of eukaryotes, called box A, box B and box II (28), is found at upstream of atpD. Several tandem repeat sequences, which are likely to be a transcriptional terminator, exist at further upstream of the TATA box-like sequence and at downstream of atpJ. Therefore, this gene cluster is most likely an operon. All of open reading frames have an ATG as an initiation codon which is preceded by possible ribosome binding sequences at distances within 7-14 nucleotides. At upstream of atpC, four short repeat sequences exist and, interestingly, each of them includes ribosome binding sequence and an initiation codon. The putative termination codons are TAG for urf2 and atpG, TAA for atpA, and TGA for other genes. The 3* termini of some open reading frames are overlapped with 5* end of the next downstream open reading frame with 103 bp between atpG/atpJ being the longest.
Other open reading frames are closely arranged with short intergenic noncoding sequences except for the two relatively long ones, atpA/atpG (55 bp) and atpB/ atpE (150 bp). Sequence Similarity to F0F1-ATPase Deduced amino acid sequence of each gene in the cluster, except for the short possible open reading frame urf2 (312 bpÅ104 aa), shares a sequence similarity with each of F0F1-ATPase subunits. The AtpD (469
TABLE I
Unusual Features of M. barkeri F0F1-ATPase Genes Compared with Typical F0F1-ATPase Genes from Other Organisms 1. 2. 3. 4. 5. 6.
Sequences of a and b subunits are less conserved. a subunit is Ç 80 aa longer. d subunit gene is not found. F0-b subunit has additional Ç220 aa at C-terminus. F0-a subunit is 50 Ç 100 aa shorter. g subunit is split into two halves.
429
AID
BBRC 7809
/
6942$$$702
12-01-97 23:15:56
bbrcgs
AP: BBRC
Vol. 241, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Alignment of the amino acid sequences of two major subunits of F0F1-ATPase from various sources. (A) a subunits and (B) b subunit. From top; maize chloroplast (maize chl), maize mitochondrion (maize mit), bovine heart mitochondrion (bovine mit), E. coli, and M. barkeri. The amino acid residues with maximum identical number in each of positions are boxed. Dots denote gaps introduced to improve alignment.
aa), 0C (144 aa), 0B (190 aa), 0E (91 aa), and 0A (588 aa) exhibit significant overall sequence similarity to the b, e, F0-a, F0-c, and a subunits, respectively. As to AtpF (386 aa), sequence similarity to the F0-b subunit of typical F0F1-ATPase is found in its N-terminal domain (Ç120 aa). Almost all bacterial atp operons contain the gene encoding a 14 k protein with unknown function and atp gene cluster of M. barkeri also has its gene, AtpI (112 aa). The sequence similarity of AtpI to the 14 k protein from other organisms, for example,
Rhodospirillum rubrum AtpI, is 24 % identity (95 aa overlapped). When compared to atp operons of other organisms, it is clear that the arrangement of atp gene cluster of M. barkeri is mostly similar to the typical atp operons of other eubacteria including E. coli except that genes for b and e subunits precede the F0 subunit genes (Fig. 1B). Transposition of genes for b and e subunits has been also reported for Synechococcus 6301 where these two genes are in the second atp operon (29).
430
AID
BBRC 7809
/
6942$$$702
12-01-97 23:15:56
bbrcgs
AP: BBRC
Vol. 241, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2—Continued
Unusual Features of M. barkeri atp Gene Cluster There are several unusual features in atp gene cluster of M. barkeri and amino acid sequences deduced from the genes (Table I). In general, amino acid sequences of the a subunit and b subunit have been highly conservative during evolution, that is, more than 60 % of residues of the a (or b) subunit are identical among F0F1-ATPases from a variety of organisms. However, M. barkeri a and b subunits are distinct from those of other organisms in that degree of the sequence similarity of its a (or b) subunit to the a (or b) subunits from other organisms is Ç50 % identity, meaningfully lower than generally values (Fig. 2A, 2B). Nonetheless, the sequences important for function such as Walker’s motif A (GXGXXGKT) and motif B (ZZZZD; Z, hydrophobic residue) (30-32) are con-
served in the M. barkeri a and b subunits. A catalytic residue acting as a general base in ATP hydrolysis (bGlu188, bovine mitochondrial ATPase) is also conserved in the M. barkeri b subunit (Fig. 2B). Therefore, if these genes are expressed, these two subunits may be functional. Another unusual feature is that the a subunit of M. barkeri has additional polypeptide sequences at both N-terminus (Ç20 aa) and C-terminus (Ç60 aa) (Fig. 2A). A Gene for d Subunit Is Missing The atp gene cluster does not contain a gene for d subunit. The d subunit gene is usually located between genes of F0-b and a subunit in the atp operon of other prokaryotes, but, in the atp gene cluster of M. barkeri, that position is occupied by the sequence of the C-termi-
431
AID
BBRC 7809
/
6942$$$702
12-01-97 23:15:56
bbrcgs
AP: BBRC
Vol. 241, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
nal half of the AtpF. As mentioned above, the size of AtpF (386 aa) is Ç220 aa more larger than F0-b subunit from other organisms and sequence similarity to the F0-b subunit of typical F0F1-ATPase is found only in N-terminal domain of AtpF (Ç120 aa) including the membrane-spanning region. Because recent reports indicate the physical contact of F0-b subunit to d subunit in the structure of F0F1-ATPase (33), one can think of a possibility that, in M. barkeri atp gene cluster, d subunit is fused to the C-terminus of F0-b subunit to produce a single polypeptide. However, we could not find the sequence similar to the d subunit in the C-terminal region of the AtpF. Other Unusual Features of the Encoded Subunits The F0-a subunit of M. barkeri (190 aa) is shorter than typical F0-a subunit (240Ç290 aa) from other sources. The sequence of the AtpG (149 aa) protein is similar to that of N-terminal half of the g subunit of F0F1-ATPases from other sources, for example, 23 % identity to the Arabidopsis thaliana chloroplast F0F1ATPase g subunit (111 aa overlapped). However, its polypeptide length is short, approximately a half of the typical g subunit (Ç300 aa). Apparently complimenting the AtpG, the sequence of the next AtpJ (186 aa) resembles that of the C-terminal half of the g subunits from other sources, for example, 31 % identity to that of Bacillus subtilis (144aa overlapped). atpG and atpJ genes are overlapped with 110 bp (34 aa). Because a single error in the base sequence could have caused this kind of frame-shift like arrangement, we repeated sequencing carefully but the result was the same. We sequenced the re-isolated clone and obtained the same result. The presence of the split genes of the g subunit raises an interesting possibility that, if these genes are expressed in M. barkeri, N-terminal and C-terminal halves of the g subunit associate into a complex which can plays an equivalent function to the full-length g subunit. Is This Gene Cluster Expressed as a Functional ATPase? Since the amino acid sequences of F0F1-ATPase subunits deduced from the atp genes of M. barkeri are so deviated from those of the typical F0F1-ATPase that, if they are expressed, it may have considerable impact in the study of mechanism of F0F1-ATPase. We have attempted to detect mRNA and protein products of these genes under several growth conditions including low and high concentrations of sodium (11 mM and 400 mM, respectively) and high pH (pH 7.5). However, the positive result has not been obtained yet and this critical question is open to future study.
F0F1-ATPase in Archaebacteria In this report, we proved the presence of F0F1-ATPase genes in an archaebacteria, M. barkeri, even though it is not known if the genes are expressed or not. Thus, the distribution of F0F1-ATPase is not restricted within eubacteria. However, it is not known how widely F0F1-ATPase distribute in archaebacteria. Methanococcus jannashii (34), an archaebacterium whose whole genome was sequenced, has a V-type ATPase operon but not F0F1-ATPase genes. M. barkeri has a V-type ATPase operon which is constitutively expressed to produce a functional membrane ATPase (35). Probably V-type ATPase acts as ATP synthase in M. barkeri while F0F1-ATPase, if expressed, has an optional function required under special growth condition. Similar example has been reported for an eubacterium, Enterococcus hirae, which has also two types of ATPases (36); F0F1-ATPase transports protons and V-type ATPase does sodium. REFERENCES 1. Gogarten, J. P., et al. (1989) Proc. Natl. Acad. Sci. USA 86, 6661– 6665. 2. Konishi, J., et al. (1990) J. Biochem. (Tokyo) 108, 554–559. 3. Boyer, P. D. (1997) Annu. Rev. Biochem. 66, 717–749. 4. Weber, J., and Senior, A. E. (1997) Biochim. Biophys. Acta 1319, 19–58. 5. Abrahams, J. P., Leslie, A. G. W., Lutter, R., and Walker, J. E. (1994) Nature 370, 621–628. 6. Noji, H., Yasuda, R., Yoshida, M., and Kinosita, K. (1997) Nature 386, 299–302. 7. Denda, K., Konishi, J., Oshima, T., Date, T., and Yoshida, M. (1988) J. Biol. Chem. 263, 6012–6015. 8. Denda, K., Konishi, J., Hajiro, K., Oshima, T., Date, T., and Yoshida, M. (1990) J. Biol. Chem. 265, 21509–21513. 9. Denda, K., Konishi, J., Oshima, T., Date, T., and Yoshida, M. (1989) J. Biol. Chem. 264, 7119–7121. 10. Denda, K., Konishi, J., Oshima, T., Date, T.,and Yoshida, M. (1988) J. Biol. Chem. 263, 17251–17254. 11. Inatomi, K., Eya, S., Maeda, M., and Futai, M. (1989) J. Biol. Chem. 264, 10954–10959. 12. Ihara, K., and Mukohata, Y. (1991) Arch. Biochem. Biophys. 286, 111–116. 13. Tsutsumi, S., Denda, K., Yokoyama, K., Oshima, T., Date, T., and Yoshida, M. (1991) Biochim. Biophys. Acta 1098, 13–20. 14. Yokoyama, K., Oshima, T., and Yoshida, M. (1990) J. Biol. Chem. 265, 21946–21950. 15. Yokoyama, K., Akabane, Y., Ishii, N., and Yoshida, M. (1994) J. Biol. Chem. 269, 12248–12253. 16. Takase, K., Yamato, I., and Kakinuma, Y. (1993) J. Biol. Chem. 268, 11610–11616. 17. Kakinuma, Y., and Igarashi, K. (1994) J. Biochem. (Tokyo) 116, 1302–1308. 18. Takase, K., Kakinuma, S., Yamato, I., Konishi, K., Igarashi, K., and Kakinuma, Y. (1994) J. Biol. Chem. 269, 11037–11044. 19. Sumi, M., Sato, M. H., Denda, K., Date, T., and Yoshida, M. (1992) FEBS Lett. 314, 207–210.
432
AID
BBRC 7809
/
6942$$$702
12-01-97 23:15:56
bbrcgs
AP: BBRC
Vol. 241, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
20. Jarrell, K. F., Faguy, D., Hebert, A. M., and Kalmokoff, M. L. (1992) Can. J. Micobiol. 38, 65–68. 21. Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, New York. 22. Kanazawa, H., Kayano, T., Kiyasu, T., and Futai, M. (1982) Biochem. Biophys. Res. Commun. 105, 1257–1264. 23. Falk, G., Hampe, A., and Walker, J. E. (1985) Biochem. J. 228, 391–407. 24. Kagawa, Y., Ishizuka, M., Saishu, T., and Nakao, S. (1986) J. Biochem. (Tokyo) 100, 923–934. 25. Breen, G. A., Holmans, P. L., and Garnett, K. E. (1988) Biochemistry 27, 3955–3961. 26. Chen, Z., Spies, A., Hein, R., Zhou, X., Thomas, B. C., Richter, M. L., and Gegenheimer, P. (1995) J. Biol. Chem. 270, 17124– 17132.
27. Ferry, J. G. (1993) Methanogenesis, Chapman & Hall, Inc., New York. 28. Allmansberger, R., Knaub, S., and Klein, A. (1988) Nucleic Acids Res. 16, 7419–7436. 29. Cozens, A. L., and Walker, J. E. (1987) J. Mol. Biol. 194, 359– 383. 30. Higgins, C. F., et al. (1986) Nature 323, 448–450. 31. Walker, J. E., Saraste, M., Runswick, M. J., and Gay, N. J. (1982) EMBO J. 1, 945–951. 32. Fry, D. C., Kuby, S. A., and Mildvan, A. S. (1986) Proc. Natl. Acad. Sci. USA 83, 907–911. 33. Sawada, K., Kuroda, N., Watanabe, H., Moritani-Otsuka, C., and Kanazawa, H. (1997) J. Biol. Chem. 272, in press. 34. Bult, C. J., et al. (1996) Science 273, 1058–1073. 35. Inatomi, K. (1986) J. Bacteriol. 167, 837–841. 36. Kakinuma, Y., and Igarashi, K. (1989) J. Bioenerg. Biomembr. 21, 679–692.
433
AID
BBRC 7809
/
6942$$$702
12-01-97 23:15:56
bbrcgs
AP: BBRC
241, 427–433 (1997)
RC977809
F0F1-ATPase Genes from an Archaebacterium, Methanosarcina barkeri Masato Sumi,* Masafumi Yohda,† Yosuke Koga,‡ and Masasuke Yoshida§,1 *Faculty of Pharmaceutical Sciences of Hokkaido University, Nishi-6, Kita-12, Kita-ku, Sapporo, Hokkaido 060; †Biochemical Systems Laboratory of the Institute of Physical and Chemical Research (RIKEN), Hirosawa 2-1, Wako, Saitama 351-01; ‡Department of Chemistry, University of Occupational and Environmental Health, 1-1, Ibugaoka, Yahatanishi-ku, Kitakyushu 807; and §Research Laboratory for Resources Utilization, Tokyo Institute of Technology, Nagatsuta 4259, Midori-ku, Yokohama 226, Japan
Received November 5, 1997
It has been known that an archaebacterium Methanosarcina barkeri strain MS (DSM 800) has a V-type ATPase (Inatomi, K., et al. (1989) J. Biol. Chem. 264, 10954–10959). Here, we report cloning of a cluster of F0F1-ATPase genes from the same organism, the first ever found in archaebacteria. The cluster and encoded subunits exhibit several unusual features such that a gene for d subunit is lacking, F0-b subunit is unusually large, and g subunit is split into two peptide fragments. Attempts to detect F0F1-ATPase proteins and mRNA have been unsuccessful and therefore it is not certain if this gene cluster is really expressed in the cell. q 1997 Academic Press
F0F1-ATPase and vacuolar(V)-type ATPase are two evolutionally related subclasses of a superfamily of proton translocating ATPases which do not form phosphorylated enzyme intermediates (1, 2). F0F1-ATPase are large (Ç500kDa) protein complexes in which a watersoluble catalytic sector, F1 , is associated to a relatively small integral membrane sector, F0 , which acts as a proton channel (3, 4). Subunit compositions of prokaryotic F1 and F0 are a3b3gde and ab2c12? , respectively. Eukaryotic F0F1-ATPases have additional subunits. The a and b subunits are homologous each other, and are catalytic and regulatory subunits, respectively. Recent achievement of crystal analysis (5) and single molecule observation (6) revealed that F0F1-ATPase is the 1 To whom correspondence should be addressed. E-mail: [email protected]. The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM Data Bank with accession number AF028006. Abbreviations: aa, amino acid residue; % identity, % of identical amino acid residues between aligned sequences of two proteins; PCR, polymerase chain reaction.
enzyme with a surprisingly novel mechanism; the centrally located g subunit rotates in the a3b3 hexagon during catalysis (3). Structure and mechanism of Vtype ATPase has been less understood but are thought to be essentially similar to those of F0F1-ATPase. F0F1ATPases are found in membranes of mitochondria, chloroplasts and eubacteria, and synthesize ATP using energy of proton translocation. V-type ATPases are responsible for acidification of eukaryotic vacuolar vesicles including lysosomes and chromaphin granules. They are also found in membranes of archaebacteria, such as Sulfolobus acidocaldarius (7-10), Methanosarcina barkeri (11), and Halobacterium salinarium (12). Agreement of taxonomic border to distribution of F0F1- and V-type ATPases in eubacteria and archaebacteria was taken as an indication of archaebacterial origin of eukaryotic vacuolar systems. However, it was recently found that distribution of V-type ATPase is not restricted within archaebacteria; a thermophilic eubacterium Thermus thermophilus has a V-type ATPase (13-15) and an eubacterium Enterococcus hirae has both types of ATPases (16-18). These facts raise a possibility that restricted distribution of F0F1-ATPase in eubacteria and V-type ATPase in archaebacteria is not as strict as it used to be thought and that some archaebacteria may have F0F1-ATPase. We report here the entire nucleotide sequence of a cluster of the F0F1ATPase genes in an archaebacterium, Methanosarcina barkeri. This paper contains the correction of our previous paper (19). MATERIALS AND METHODS Strain and medium. Methanosarcina barkeri strain MS (DSM 800) was obtained from the Deutsche Sammlung fu¨r Mikroorganismen und Zellkulturen (DSM) and grown under strictly anaerobic conditions (80 % N2 / 20 % CO2) in the medium (pH 6.85) containing
0006-291X/97 $25.00
427
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
BBRC 7809
/
6942$$$701
12-01-97 23:15:56
bbrcgs
AP: BBRC
Vol. 241, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
0.5 % methanol, 40 mM Na2 and other components as described in DSM 120. PCR reaction. Genomic DNA from M. barkeri was prepared as described in (20). Oligonucleotide primers for polymerase chain reaction (PCR) were as follows: Fs1, GGNGGNGCNGGNGTNGGNAARAC; Fs2, GARMGNACNMGNGARGGNAAYGA; Fa, TCNGTBAGRTCRTCNGCNGGNACRTA (NÅA, C, G or T; RÅA or G; YÅT or C; MÅA or C; BÅC, G, or T). The reaction mixture (100 mL) for PCR contained 200 ng of the template M. barkeri genomic DNA, 1 mM of each primer (Fs1 and Fa), 1 mM MgCl2 , 2.5 units Taq DNA polymerase (Promega) in the buffer provided by a Promega kit (50 mM KCl, 10 mM Tris HCl, pH 9.0 at 257C, 0.1% Triton X-100). The PCR reactions were carried out for 1 min at 947C, 1 min at 50 7C, 1 min at 74 7C, and were repeated for 30 cycles (the final step at 74 7C was extended to 10 min). The products were analyzed with 6 % polyacrylamide gel and visualized with ethidium bromide. The DNA fragment of interest (Ç490 bp DNA) was eluted from the gel and used as a template DNA for the second amplification which was performed to decrease non-specific products. The reaction mixture (100 mL) for the second amplification contained 2 ng of the template 490 bp DNA, 1 mM of each primer (Fs2 and Fa), 2.5 units Taq DNA polymerase in the buffer provided by a Promega kit. The product was electrophoresed in 2 % agarose gel, visualized with ethidium bromide, and the section containing a 380 bp fragment was cut out. For isolation of DNA fragments from the agarose gel, Geneclean II kit (Bio 101, Inc.) was used. The purified DNA fragment was ligated into the EcoRV site of pBluescript II SK0. The obtained recombinant plasmids were then introduced into Escherichia coli JM109. DNA sequence analysis. Southern hybridization of the genomic DNA and colony hybridization were carried out using the ECL random primer labeling and detection systems (Amersham). The PCR fragment was isolated from the plasmids by EcoRI and HindIII endonuclease treatment and was used as a template DNA to make labeled probes. A genomic BamHI library in the pBluescript II SK0 vector from M. barkeri was constructed. Colony hybridization was performed according to (21). Positive clones were selected and plasmids were isolated from the clones. The BamHI fragment was purified from the plasmids and broken into segments by sonication for random sequencing. DNA fragments of the desired size (500-800 bp) were isolated by gel electrophoresis and their termini were repaired with a blunting kit (TaKaRa). The DNAs were then ligated to SmaI site of pUC18 and then introduced into E. coli. Double-stranded plasmids prepared from colonies were used as templates for sequencing. All DNA sequences were confirmed by sequencing both strands. Protein sequence comparisons were carried out by using the Pileup program of University of Wisconsin Genetics Computer Group (GCG) packages.
RESULTS AND DISCUSSION Isolation of atp Gene Cluster To ensure the specificity of amplification, the segment of amino acid sequences, well conserved among F0F1-ATPases but not among V-type ATPases, were chosen to design PCR primers. Several F0F1-ATPase b (catalytic) subunit sequences from various organisms including E. coli (22), Rhodospirillum rubrum (23), Bacillus PS3 (24), bovine heart mitochondria (25), and spinach chloroplasts (26) were aligned. Three well-conserved regions in the sequence alignment were chosen, and used for design of two sense primers (Fs1 and Fs2) and an antisense primer (Fa), respectively.
A 490 bp DNA fragment was amplified by PCR using Fs1 and Fa primer from M. barkeri genomic DNA. The second PCR was carried out using Fs2, instead of Fs1, in order to confirm that the amplified DNA fragment of the first PCR really contained an internal sequence common to F0F1-ATPase b subunits. A 380 bp DNA fragment was amplified by the second PCR as expected. The 380 bp fragment was subcloned and sequenced. The deduced amino acid sequence from the DNA fragment was novel, and seemed to be a piece of F0F1-ATPase b subunit. We reported previously nucleotide sequence of another 380 bp PCR fragment that was obtained by the same methods from the same organism (19). Unexpectedly, the sequence of the fragment obtained this time was different from the previously reported one. Unfortunately the genomic DNA used in the previous study was used up completely and we could not examine why the difference occurred. In view of this trouble, we carefully tested the accuracy of PCR product and quality of genomic DNA. Using the same genomic DNA, existence of the gene encoding the catalytic subunit of Vtype ATPase reported by Inatomi et al (11) was confirmed by the same PCR method with appropriate primers. Southern analysis of M. barkeri genomic DNA using the 380 bp DNA fragment obtained this time as a probe was carried out independently in other laboratory and produced the same consistent result that we had (Mu¨ller, V., et al.; personal communication). The 380 bp fragment (Fig. 1A, white small box on B) hybridized with an 8 kbp BamHI fragment of M. barkeri genomic DNA. A plasmid library of the genomic DNA containing Ç8 kbp BamHI fragments (about 100 clones) was screened with the probe, and one positive clone (Fig. 1A, thick line B) was obtained. Sequence analysis revealed that the insert contained most atp genes but lacked 3* terminus of a subunit gene (and all g subunit gene). Then, using a Ç800 bp SacI fragment derived from 3* terminus of the BamHI fragment as a probe (Fig. 1A, white small box on E), another plasmid library made up from Ç4 kbp EcoRI fragments of genomic DNA (about 400 clones) was screened and four positive clones containing a 4 kbp EcoRI fragment were obtained (Fig. 1A, thick line E). Sequence analysis showed that the fragment contained the rest of atp genes and transcriptional termination region of the cluster. The nucleotide sequences has been submitted to the GenBankTM Data Bank with accession number AF028006. Organization of atp Genes The G/C content of the insert is 43 % which is close to the overall G/C content of this bacterium (36-43%) (27). In this clone, ten genes are arranged in the order atpD, -C, -I, urf2, atpB, -E, -F, -A, -G, and -J (Fig. 1A)
428
AID
BBRC 7809
/
6942$$$702
12-01-97 23:15:56
bbrcgs
AP: BBRC
Vol. 241, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. (A) Arrangement of genes and a restriction map of the atp gene cluster of M. barkeri chromosomal DNA. Open boxes indicate the open reading frames for atp genes. The letter A, B, C, D, E, F, G, I, and J in the open boxes indicated the genes of F0F1-ATPase subunits. The urf1, urf2, and urf3 indicate unknown putative open reading frames. The black thick lines, B and E, indicate the cloned 8,193 bp BamHI fragment and the 4,614 bp EcoRI fragment, respectively. The white boxes on these thick lines indicate the portion of PCR-amplified fragment (on B) and SacI fragment (on E) that was used for Southern and colony hybridization as probes (See Materials and Methods). Restriction sites: B, BamHI, E, EcoRI; H, HindIII; S, SacI; X, XhoI. (B) Comparison of arrangement of genes encoding F0F1-ATPase subunits in the atp gene cluster (operon) of M. barkeri, E. coli and Synechococcus 6301. The letters a, c, b, and b* represent genes for F0 subunits. Letters d, a, g, b, and e represent genes for F1 subunits. The letter I indicates the atpI gene. The letters gN and gC indicate N- and Cterminal halves of g subunit, respectively. X is a putative open reading frame whose deduced amino acid sequence has no similarity to any other proteins.
with the same transcriptional direction. There is no possible open reading frame in the immediate neighborhood at upstream of the 5* end of atpD and downstream of the 3 * end of atpJ. The sequence similar to the archaebacterial promoter element, which is homologous to the TATA box of eukaryotes, called box A, box B and box II (28), is found at upstream of atpD. Several tandem repeat sequences, which are likely to be a transcriptional terminator, exist at further upstream of the TATA box-like sequence and at downstream of atpJ. Therefore, this gene cluster is most likely an operon. All of open reading frames have an ATG as an initiation codon which is preceded by possible ribosome binding sequences at distances within 7-14 nucleotides. At upstream of atpC, four short repeat sequences exist and, interestingly, each of them includes ribosome binding sequence and an initiation codon. The putative termination codons are TAG for urf2 and atpG, TAA for atpA, and TGA for other genes. The 3* termini of some open reading frames are overlapped with 5* end of the next downstream open reading frame with 103 bp between atpG/atpJ being the longest.
Other open reading frames are closely arranged with short intergenic noncoding sequences except for the two relatively long ones, atpA/atpG (55 bp) and atpB/ atpE (150 bp). Sequence Similarity to F0F1-ATPase Deduced amino acid sequence of each gene in the cluster, except for the short possible open reading frame urf2 (312 bpÅ104 aa), shares a sequence similarity with each of F0F1-ATPase subunits. The AtpD (469
TABLE I
Unusual Features of M. barkeri F0F1-ATPase Genes Compared with Typical F0F1-ATPase Genes from Other Organisms 1. 2. 3. 4. 5. 6.
Sequences of a and b subunits are less conserved. a subunit is Ç 80 aa longer. d subunit gene is not found. F0-b subunit has additional Ç220 aa at C-terminus. F0-a subunit is 50 Ç 100 aa shorter. g subunit is split into two halves.
429
AID
BBRC 7809
/
6942$$$702
12-01-97 23:15:56
bbrcgs
AP: BBRC
Vol. 241, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Alignment of the amino acid sequences of two major subunits of F0F1-ATPase from various sources. (A) a subunits and (B) b subunit. From top; maize chloroplast (maize chl), maize mitochondrion (maize mit), bovine heart mitochondrion (bovine mit), E. coli, and M. barkeri. The amino acid residues with maximum identical number in each of positions are boxed. Dots denote gaps introduced to improve alignment.
aa), 0C (144 aa), 0B (190 aa), 0E (91 aa), and 0A (588 aa) exhibit significant overall sequence similarity to the b, e, F0-a, F0-c, and a subunits, respectively. As to AtpF (386 aa), sequence similarity to the F0-b subunit of typical F0F1-ATPase is found in its N-terminal domain (Ç120 aa). Almost all bacterial atp operons contain the gene encoding a 14 k protein with unknown function and atp gene cluster of M. barkeri also has its gene, AtpI (112 aa). The sequence similarity of AtpI to the 14 k protein from other organisms, for example,
Rhodospirillum rubrum AtpI, is 24 % identity (95 aa overlapped). When compared to atp operons of other organisms, it is clear that the arrangement of atp gene cluster of M. barkeri is mostly similar to the typical atp operons of other eubacteria including E. coli except that genes for b and e subunits precede the F0 subunit genes (Fig. 1B). Transposition of genes for b and e subunits has been also reported for Synechococcus 6301 where these two genes are in the second atp operon (29).
430
AID
BBRC 7809
/
6942$$$702
12-01-97 23:15:56
bbrcgs
AP: BBRC
Vol. 241, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2—Continued
Unusual Features of M. barkeri atp Gene Cluster There are several unusual features in atp gene cluster of M. barkeri and amino acid sequences deduced from the genes (Table I). In general, amino acid sequences of the a subunit and b subunit have been highly conservative during evolution, that is, more than 60 % of residues of the a (or b) subunit are identical among F0F1-ATPases from a variety of organisms. However, M. barkeri a and b subunits are distinct from those of other organisms in that degree of the sequence similarity of its a (or b) subunit to the a (or b) subunits from other organisms is Ç50 % identity, meaningfully lower than generally values (Fig. 2A, 2B). Nonetheless, the sequences important for function such as Walker’s motif A (GXGXXGKT) and motif B (ZZZZD; Z, hydrophobic residue) (30-32) are con-
served in the M. barkeri a and b subunits. A catalytic residue acting as a general base in ATP hydrolysis (bGlu188, bovine mitochondrial ATPase) is also conserved in the M. barkeri b subunit (Fig. 2B). Therefore, if these genes are expressed, these two subunits may be functional. Another unusual feature is that the a subunit of M. barkeri has additional polypeptide sequences at both N-terminus (Ç20 aa) and C-terminus (Ç60 aa) (Fig. 2A). A Gene for d Subunit Is Missing The atp gene cluster does not contain a gene for d subunit. The d subunit gene is usually located between genes of F0-b and a subunit in the atp operon of other prokaryotes, but, in the atp gene cluster of M. barkeri, that position is occupied by the sequence of the C-termi-
431
AID
BBRC 7809
/
6942$$$702
12-01-97 23:15:56
bbrcgs
AP: BBRC
Vol. 241, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
nal half of the AtpF. As mentioned above, the size of AtpF (386 aa) is Ç220 aa more larger than F0-b subunit from other organisms and sequence similarity to the F0-b subunit of typical F0F1-ATPase is found only in N-terminal domain of AtpF (Ç120 aa) including the membrane-spanning region. Because recent reports indicate the physical contact of F0-b subunit to d subunit in the structure of F0F1-ATPase (33), one can think of a possibility that, in M. barkeri atp gene cluster, d subunit is fused to the C-terminus of F0-b subunit to produce a single polypeptide. However, we could not find the sequence similar to the d subunit in the C-terminal region of the AtpF. Other Unusual Features of the Encoded Subunits The F0-a subunit of M. barkeri (190 aa) is shorter than typical F0-a subunit (240Ç290 aa) from other sources. The sequence of the AtpG (149 aa) protein is similar to that of N-terminal half of the g subunit of F0F1-ATPases from other sources, for example, 23 % identity to the Arabidopsis thaliana chloroplast F0F1ATPase g subunit (111 aa overlapped). However, its polypeptide length is short, approximately a half of the typical g subunit (Ç300 aa). Apparently complimenting the AtpG, the sequence of the next AtpJ (186 aa) resembles that of the C-terminal half of the g subunits from other sources, for example, 31 % identity to that of Bacillus subtilis (144aa overlapped). atpG and atpJ genes are overlapped with 110 bp (34 aa). Because a single error in the base sequence could have caused this kind of frame-shift like arrangement, we repeated sequencing carefully but the result was the same. We sequenced the re-isolated clone and obtained the same result. The presence of the split genes of the g subunit raises an interesting possibility that, if these genes are expressed in M. barkeri, N-terminal and C-terminal halves of the g subunit associate into a complex which can plays an equivalent function to the full-length g subunit. Is This Gene Cluster Expressed as a Functional ATPase? Since the amino acid sequences of F0F1-ATPase subunits deduced from the atp genes of M. barkeri are so deviated from those of the typical F0F1-ATPase that, if they are expressed, it may have considerable impact in the study of mechanism of F0F1-ATPase. We have attempted to detect mRNA and protein products of these genes under several growth conditions including low and high concentrations of sodium (11 mM and 400 mM, respectively) and high pH (pH 7.5). However, the positive result has not been obtained yet and this critical question is open to future study.
F0F1-ATPase in Archaebacteria In this report, we proved the presence of F0F1-ATPase genes in an archaebacteria, M. barkeri, even though it is not known if the genes are expressed or not. Thus, the distribution of F0F1-ATPase is not restricted within eubacteria. However, it is not known how widely F0F1-ATPase distribute in archaebacteria. Methanococcus jannashii (34), an archaebacterium whose whole genome was sequenced, has a V-type ATPase operon but not F0F1-ATPase genes. M. barkeri has a V-type ATPase operon which is constitutively expressed to produce a functional membrane ATPase (35). Probably V-type ATPase acts as ATP synthase in M. barkeri while F0F1-ATPase, if expressed, has an optional function required under special growth condition. Similar example has been reported for an eubacterium, Enterococcus hirae, which has also two types of ATPases (36); F0F1-ATPase transports protons and V-type ATPase does sodium. REFERENCES 1. Gogarten, J. P., et al. (1989) Proc. Natl. Acad. Sci. USA 86, 6661– 6665. 2. Konishi, J., et al. (1990) J. Biochem. (Tokyo) 108, 554–559. 3. Boyer, P. D. (1997) Annu. Rev. Biochem. 66, 717–749. 4. Weber, J., and Senior, A. E. (1997) Biochim. Biophys. Acta 1319, 19–58. 5. Abrahams, J. P., Leslie, A. G. W., Lutter, R., and Walker, J. E. (1994) Nature 370, 621–628. 6. Noji, H., Yasuda, R., Yoshida, M., and Kinosita, K. (1997) Nature 386, 299–302. 7. Denda, K., Konishi, J., Oshima, T., Date, T., and Yoshida, M. (1988) J. Biol. Chem. 263, 6012–6015. 8. Denda, K., Konishi, J., Hajiro, K., Oshima, T., Date, T., and Yoshida, M. (1990) J. Biol. Chem. 265, 21509–21513. 9. Denda, K., Konishi, J., Oshima, T., Date, T., and Yoshida, M. (1989) J. Biol. Chem. 264, 7119–7121. 10. Denda, K., Konishi, J., Oshima, T., Date, T.,and Yoshida, M. (1988) J. Biol. Chem. 263, 17251–17254. 11. Inatomi, K., Eya, S., Maeda, M., and Futai, M. (1989) J. Biol. Chem. 264, 10954–10959. 12. Ihara, K., and Mukohata, Y. (1991) Arch. Biochem. Biophys. 286, 111–116. 13. Tsutsumi, S., Denda, K., Yokoyama, K., Oshima, T., Date, T., and Yoshida, M. (1991) Biochim. Biophys. Acta 1098, 13–20. 14. Yokoyama, K., Oshima, T., and Yoshida, M. (1990) J. Biol. Chem. 265, 21946–21950. 15. Yokoyama, K., Akabane, Y., Ishii, N., and Yoshida, M. (1994) J. Biol. Chem. 269, 12248–12253. 16. Takase, K., Yamato, I., and Kakinuma, Y. (1993) J. Biol. Chem. 268, 11610–11616. 17. Kakinuma, Y., and Igarashi, K. (1994) J. Biochem. (Tokyo) 116, 1302–1308. 18. Takase, K., Kakinuma, S., Yamato, I., Konishi, K., Igarashi, K., and Kakinuma, Y. (1994) J. Biol. Chem. 269, 11037–11044. 19. Sumi, M., Sato, M. H., Denda, K., Date, T., and Yoshida, M. (1992) FEBS Lett. 314, 207–210.
432
AID
BBRC 7809
/
6942$$$702
12-01-97 23:15:56
bbrcgs
AP: BBRC
Vol. 241, No. 2, 1997
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
20. Jarrell, K. F., Faguy, D., Hebert, A. M., and Kalmokoff, M. L. (1992) Can. J. Micobiol. 38, 65–68. 21. Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, New York. 22. Kanazawa, H., Kayano, T., Kiyasu, T., and Futai, M. (1982) Biochem. Biophys. Res. Commun. 105, 1257–1264. 23. Falk, G., Hampe, A., and Walker, J. E. (1985) Biochem. J. 228, 391–407. 24. Kagawa, Y., Ishizuka, M., Saishu, T., and Nakao, S. (1986) J. Biochem. (Tokyo) 100, 923–934. 25. Breen, G. A., Holmans, P. L., and Garnett, K. E. (1988) Biochemistry 27, 3955–3961. 26. Chen, Z., Spies, A., Hein, R., Zhou, X., Thomas, B. C., Richter, M. L., and Gegenheimer, P. (1995) J. Biol. Chem. 270, 17124– 17132.
27. Ferry, J. G. (1993) Methanogenesis, Chapman & Hall, Inc., New York. 28. Allmansberger, R., Knaub, S., and Klein, A. (1988) Nucleic Acids Res. 16, 7419–7436. 29. Cozens, A. L., and Walker, J. E. (1987) J. Mol. Biol. 194, 359– 383. 30. Higgins, C. F., et al. (1986) Nature 323, 448–450. 31. Walker, J. E., Saraste, M., Runswick, M. J., and Gay, N. J. (1982) EMBO J. 1, 945–951. 32. Fry, D. C., Kuby, S. A., and Mildvan, A. S. (1986) Proc. Natl. Acad. Sci. USA 83, 907–911. 33. Sawada, K., Kuroda, N., Watanabe, H., Moritani-Otsuka, C., and Kanazawa, H. (1997) J. Biol. Chem. 272, in press. 34. Bult, C. J., et al. (1996) Science 273, 1058–1073. 35. Inatomi, K. (1986) J. Bacteriol. 167, 837–841. 36. Kakinuma, Y., and Igarashi, K. (1989) J. Bioenerg. Biomembr. 21, 679–692.
433
AID
BBRC 7809
/
6942$$$702
12-01-97 23:15:56
bbrcgs
AP: BBRC