MHYT, a new integral membrane sensor domain

Transcription

1 FEMS Microbiology Letters 205 (2001) 17^23 a d MHYT, a new integral membrane sensor domain Michael Y. Galperin a; *, Tatiana A. Gaidenko b, Armen Y. Mulkidjanian c, Michiko Nakano d, Chester W. Price b National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health, Bethesda, MD 20894, USA b Department of Food Science and Technology, University of California, Davis, CA 95616, USA c Division of Biophysics, University of Osnabru«ck, D Osnabru«ck, Germany Department of Biochemistry and Molecular Biology, Oregon Graduate Institute of Science and Technology, Beaverton, OR 97006, USA Received 26 July 2001; received in revised form 3 September 2001; accepted 4 September 2001 First published online 25 October 2001 Abstract MHYT, a new conserved protein domain with a likely signaling function, is described. This domain consists of six transmembrane segments, three of which contain conserved methionine, histidine, and tyrosine residues that are projected to lie near the outer face of the cytoplasmic membrane. In Synechocystis sp. PCC6803, this domain forms the N-terminus of the sensor histidine kinase Slr2098. In Pseudomonas aeruginosa and several other organisms, the MHYT domain forms the N-terminal part of a three-domain protein together with previously described GGDEF and EAL domains, both of which have been associated with signal transduction due to their presence in likely signaling proteins. In Bacillus subtilis YkoW protein, an additional PAS domain is found between the MHYT and GGDEF domains. A ykow null mutant of B. subtilis did not exhibit any growth alterations, consistent with a non-essential, signaling role of this protein. A model of the membrane topology of the MHYT domain indicates that its conserved residues could coordinate one or two copper ions, suggesting a role in sensing oxygen, CO, or NO. ß 2001 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords: Genome analysis; Protein domain; Sequence conservation; Anaerobic growth; Signal transduction; Metal binding 1. Introduction Sensor kinases of the two-component signal transduction systems are modular proteins that comprise several distinct domains. A typical sensor kinase contains an N- terminal periplasmic or extracytoplasmic ligand-binding sensor domain, anchored to the membrane by one or two transmembrane segments and followed by a cytoplasmic histidine kinase domain, which can be further divided into an ATPase and phosphoacceptor subdomains (reviewed in [1]). A detailed sequence analysis of sensor kinases and response regulators has shown that many of them have even more complex domain architectures and include additional ligand-binding and response output domains. Among the histidine kinases, several new domains were described, including the heme- and avin-binding PAS domain [2], the phytochrome- and cgmp-binding * Corresponding author. Tel.: +1 (301) ; Fax: +1 (301) address: galperin@ncbi.nlm.nih.gov (M.Y. Galperin). GAF domain [3], and the HAMP linker domain [4]. Socalled `hybrid' kinases were found to additionally contain the phosphotransfer HPt domain and the receiver CheYlike domain [1]. The downstream signal transduction module revealed an even greater diversity. In addition to the well-known response regulators formed by the association of the receiver CheY-like domain with a DNA-binding helix-turn-helix domain, and, in some cases, a c 54 -binding ATPase domain, additional output domains were described, such as GGDEF [5,6], EAL [6], and HD-GYP [7,8]. Although the involvement of these domains in signal transduction was originally proposed solely on the basis of their association with the CheY domain [5,7], it has now received experimental veri cation [8,9], demonstrating that these domains, indeed, can function as additional (or alternative) output domains (see [10] for a recent review). The growing number of complete genomes has helped to sort out even such a diverse group as the periplasmic (extracytoplasmic) sensor domains. The most common variant of such domains turned out to be homologous to the periplasmic solute-binding components of the ABC-type transporters [11]. Recently, another common type of extra / 01 / $20.00 ß 2001 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S (01)

2 18 M.Y. Galperin et al. / FEMS Microbiology Letters 205 (2001) 17^23 cellular ligand-binding domain has been described, referred to as Cache [12] or Esens [13] domain. In contrast to the extracytoplasmic sensor domains, membrane-bound sensor domains have not been systematically analyzed so far, probably due to their diversity and relatively low sequence conservation. Here we describe an extremely wellconserved integral membrane protein domain, which appears to function as a metal-containing sensor. While the nature of the environmental signal sensed by this domain remains unknown, it could be involved in sensing of oxygen, CO, or NO. 2. Materials and methods 2.1. Sequence analysis of YkoW protein Sequence database searches against the database of nished and un nished microbial genome sequences maintained at the NCBI ( nishedgenome.html) were performed using the BLAST program [14] with default parameters. The database hits identi ed in these searches were analyzed for domain composition using reverse position-speci c (RPS) BLAST searches ( cdd/wrpsb.cgi) against the Conserved Domain Database at the NCBI and hidden Markov model-based searches against SMART database, maintained at EMBL ( smart.embl-heidelberg.de [15]). The multiple alignment of the MHYT domains was constructed manually on the basis of the BLAST outputs. The sequence of Bacillus subtilis YkoW was extended by 51 N-terminal amino acid (aa) residues, based on the translation of the upstream region of the ykow gene using ORF nder ( The modi- ed sequence of the YkoW protein, as well as other MHYT domains, is available in the Clusters of Orthologous Groups of proteins (COG) database ( [16]) as COG3300. The MHYT domain listing in the ProDom database ( prodes.toulouse.inra.fr/prodom, [17]) as domain PD shows the original translation of the B. subtilis ykow gene Isolation and growth of the ykow null mutant The in-frame deletion in the B. subtilis ykow gene was obtained using a PCR-based method [18]. In the rst step, two ykow fragments were ampli ed using, for the rst fragment, primer 1: 5P-TTCACGGTCAGTTGAACT- AA-3P, which starts at position 3134 of the corrected coding region of the ykow gene, and primer 2: 5P-AT- TGTCCATTGTCATGCTGGCTGCAATCGCAATGA- TA-3P, ending at position +438, and, for the second fragment, primer 3: 5P-AGCATGACAATGGACAATC- TC-3P, starting at position +2020, and primer 4: 5P-TTA- TGGAGGCTGCGCAGTT-3P, ending at position These two fragments were linked in the second PCR reaction with primers 1 and 4, creating a single DNA fragment, corresponding to the ykow gene lacking the 1563 bp region between positions 456 and This latter fragment was cloned into the pcp115 vector [19] and integrated into the B. subtilis chromosome. The ykow duplication was resolved by spontaneous homologous recombination, screened by the loss of plasmid Cm R marker. The presence of the in-frame deletion was veri ed by PCR with primers 1 and 4; the wild-type strain gave a 2731 bp product, whereas the vykow strain gave a 1168 bp product. Aerobic growth of the vykow strain of B. subtilis was compared with the wild-type strain in rich medium (Luria broth) at 37³C and 50³C, and in a glucose^ammonia^minimal salts medium [20] at 37³C. Anaerobic growth was tested in 2UYT medium supplemented either with 1% glucose and 0.2% nitrate or with 1% glucose and 0.5% pyruvate [21]. The induction of c B activity in response to salt, ethanol, and stationary phase stresses was measured indirectly using a c B -dependent ctc-lacz transcriptional fusion [22]. 3. Results 3.1. Sequence and the phylogenetic distribution of the MHYT domain Sequence analysis of the signaling proteins encoded in complete bacterial genomes revealed complex domain architectures for many of them and allowed identi cation of a number of novel conserved domains (reviewed in [10]). Although the biochemical functions of some of these domains remain obscure, their association with various components of the signal transduction machinery strongly implies that they also participate in signal transduction. Analysis of the domain organization of various sensor proteins of the two-component signal transduction systems encoded in complete genomes showed that the histidine kinase Slr2098 from Synechocystis sp. PCC6803 contains a new conserved domain, which is also present in the uncharacterized protein YkoW from B. subtilis and several proteins from other recently sequenced bacterial genomes (Fig. 1). The multiple alignment of this domain, which we dubbed MHYT based on its conserved amino acid pattern, showed a remarkable degree of sequence conservation. In B. subtilis YkoW, however, V50 N-terminal amino acid residues appeared to be missing. Translation of the upstream region of the ykow gene using the ORF nder program allowed us to extend the ykow open reading frame by 51 residues that were very similar to those in other MHYT domains (Fig. 1). Sequence analysis of the MHYT domain using the PHDhtm [23] and TopPred [24] programs showed that it

3 M.Y. Galperin et al. / FEMS Microbiology Letters 205 (2001) 17^23 19 Fig. 1. Multiple alignment of MHYT domains. The proteins are listed under the names of their source organisms (left column); the numbers indicate positions of the rst and the last residues in each protein fragment and the distances between the aligned segments. Yellow shading indicates uncharged amino acid residues of the six transmembrane (TM) helices. The most conserved region of each TM helix is indicated by a bracket at the top; amino acid residues in this region are shown in bold, those likely involved in metal binding are in red. Conserved Pro, Gly, Ala, and Ser residues are shown in green; conserved positively charged residues (Arg and Lys) in cytoplasmic loops are shown in blue. The positions of the TM segments are as predicted by the PHDhtm program [23]. Species name abbreviations and protein gene identi cation (gi) numbers in the NCBI protein database are as follows: Bsubt, B. subtilis (YkoW, ); Syn, Synechocystis sp. (Slr2098, ); Ccre, C. crescentus (CC0091, ); Mlo, M. loti (Mlr3504, ; Mll4364, ); Ocarb, O. carboxidovorans (CoxC, ; CoxH, ); Paer, P. aeruginosa (PA1727, ; PA3311, ); Sc, S. coelicolor (SCG8A.15c, ). The three remaining proteins from B. pertussis, B. mallei, and S. typhi are not yet in GenBank. The white arrow indicates the N-terminal amino acid residues of the B. subtilis YkoW protein that were missed in the original translation of the genomic sequence. The black arrow indicates the portion of MHYT domain that was deleted in the vykow mutant. consists of six predicted transmembrane (TM) segments, connected by short arginine-rich cytoplasmic loops and periplasmic loops that are also rich in charged amino acid residues (Fig. 1). Three of its TM segments have a very similar amino acid sequence motif (Fig. 1) with highly conserved methionine and histidine residues located near the outer face of the cytoplasmic membrane (Fig. 2). The MHYT domain is encoded in a single copy in the genomes of B. subtilis, Caulobacter crescentus, and Synechocystis sp. and in two copies in the genomes of Mesorhizobium loti and Pseudomonas aeruginosa (Table 1). It is missing, however, in all other organisms whose complete genome sequences are currently available. Two copies of the MHYT domain are also found in Oligotropha carbox-

4 20 M.Y. Galperin et al. / FEMS Microbiology Letters 205 (2001) 17^23 Fig. 2. A model of the membrane orientation of the MHYT domain. The extended sequence of the B. subtilis YkoW protein was plotted according to the alignment from Fig. 1 and predictions of the PHDhtm [23] and TopPred [24] programs. Conserved residues, predicted to be involved in metal binding, are shaded yellow. Conserved positively charged residues (Arg and Lys) in cytoplasmic loops are shown in blue. The last transmembrane segment, leading to the cytoplasmic domains of the YkoW protein, is not shown. idovorans, where they ank the structural genes of CO dehydrogenase [25]. Search of the un nished genome sequences at the NCBI web site revealed the presence of this domain in L-proteobacteria Bordetella pertussis, Bordetella bronchiseptica, Burkholderia mallei, and Burkholderia pseudomallei and in Q-proteobacteria Pseudomonas putida and Pseudomonas syringae (Table 1). Although the MHYT domain is not encoded in the Escherichia coli K12 genome, it is present in Salmonella typhi, Salmonella paratyphi, and Salmonella enterica serovar typhimurium. Similar patchy distribution of this domain is found in bacilli, where, in contrast to B. subtilis, Bacillus cereus and Bacillus anthracis do not seem to encode the MHYT domain. In phylogenetic terms, the MHYT domain has been found in several representatives of the K, L, and Q subdivisions of proteobacteria, two Gram-positive bacteria, and in a cyanobacterium Domain organization of MHYT-containing proteins Until now, none of the proteins that contain the MHYT domain have been studied experimentally. However, the presence of this domain in the sensor histidine kinase Slr2098 of Synechocystis sp. and predicted histidine kinases of B. mallei and B. pseudomallei (Table 1) suggests that it can serve as a sensor for the two-component signal transduction system. It should be noted that the association of the MHYT domain with histidine kinases appears to be unusual; in most other organisms, MHYT forms a three-domain protein with GGDEF and EAL domains (Table 1), which are now believed to comprise alternative output domains of the two-component system [10]. In B. subtilis, C. crescentus, and one of the two M. loti MHYTcontaining proteins, the domain organization is even more complex and includes an additional PAS domain sandwiched between the MHYT and GGDEF domains (Table 1). Remarkably, homologs of the B. subtilis YkoW protein in B. cereus and B. anthracis have a PAS-GGDEF-EAL domain organization and do not contain the MHYT domain (not shown) E ects of the MHYT-less vykow allele of B. subtilis To characterize the cellular role of the MHYT domain, the e ect of MHYT de ciency in B. subtilis, which encodes only a single copy of MHYT in a MHYT-PAS- GGDEF-EAL combination (Table 1), was studied. Since an E. coli protein, lacking the MHYT domain but with an otherwise similar PAS-GGDEF-EAL domain organization, has been recently found to bind O 2, CO and NO with comparable a nities and suggested to serve as a direct oxygen sensor in that bacterium [26], the possibility that YkoW protein might serve as an oxygen sensor in B. subtilis was investigated.

5 M.Y. Galperin et al. / FEMS Microbiology Letters 205 (2001) 17^23 21 Table 1 Domain organization of the MHYT-containing proteins Organism a Protein name b Domain organization c Bacillus subtilis YkoW MHYT-PAS-GGDEF-EAL Caulobacter crescentus CC0091 MHYT-PAS-GGDEF-EAL Mesorhizobium loti Mlr3504 MHYT-PAS-GGDEF-EAL Mll4364 MHYT-GGDEF-EAL Oligotropha carboxidovorans CoxC MHYT-HTH CoxH MHYT-HTH Pseudomonas aeruginosa PA1727 MHYT-GGDEF-EAL PA3311 MHYT-GGDEF-EAL Synechocystis sp. Slr2098 MHYT-PAS-PAS-HisKin- CheY-CheY-HPt Stretomyces coelicolor SCG8A.15c Stand-alone MHYT domain SC3C9.02c Stand-alone MHYT domain Bordetella pertussis d ^ Stand-alone MHYT domain Bordetella parapertussis d ^ MHYT-GGDEF-EAL Bordetella bronchiseptica d ^ MHYT-GGDEF-EAL Burkholderia mallei d ^ MHYT-HisKin Burkholderia pseudomallei d ^ MHYT-HisKin Pseudomonas putida d ^ MHYT-GGDEF-EAL Pseudomonas syringae d ^ MHYT-GGDEF-EAL Salmonella typhi d ^ MHYT-GGDEF-EAL Salmonella paratyphi d ^ MHYT-GGDEF-EAL Salmonella enterica d ^ MHYT-GGDEF-EAL a MHYT domain is not encoded in the genomes of E. coli, Haemophilus in uenzae, Helicobacter pylori, Chlamydia trachomatis, Chlamydia pneumoniae, Mycoplasma genitalium, Mycoplasma pneumoniae, Ureaplasma urealyticum, Mycobacterium tuberculosis, Deinococcus radiodurans, Thermotoga maritima, Aquifex aeolicus, Vibrio cholerae, or in any archaeal or eukaryotic genomes sequenced so far. b Names of the proteins are taken from the original genome annotations. They can be used to retrieve corresponding sequences from the NCBI protein database ( c PAS, CheY, HPt, HisKin (histidine kinase), and HTH (helix-turn-helix) domains have been extensively described [1,2,13]. For a description of GGDEF and EAL domains, see [6,10]. d Based on un nished genome sequences obtained from the Sanger Centre ( TIGR ( and the Washington University ( gsc/projects/bacteria.shtml). Because ykow is the rst gene in an apparent three-gene operon in B. subtilis, a ykow alteration was designed to have little e ect on expression of the downstream ykov and ykou genes. The four-primer method [18] was used to make a large in-frame deletion in the ykow coding sequence, removing a 469 aa region extending from residue 152 to residue 622 of the corrected YkoW sequence. This mutant allele was then substituted for the wild-type chromosomal copy by a two-step allele replacement procedure [27]. The resulting strain was viable, indicating that ykow is not an essential gene under standard laboratory conditions. Loss of ykow function had no discernible e ect on aerobic growth of B. subtilis cells in rich or minimal medium at 37³C or at 50³C, or on sporulation timing or e ciency (data not shown). Moreover, loss of ykow function had no obvious e ect on two processes thought to be regulated by the redox state of the cell: energy-stress activation of the general stress transcription factor c B or anaerobic growth by means of either nitrate respiration or glucose fermentation (data not shown). In addition, the vykow allele appeared to have no e ect on aerotaxis (M. Alam, personal communication), which in B. subtilis is dependent upon a heme-based sensor [28]. These results indicate that YkoW function is not required for at least some physiological functions that involve oxygen sensing. However, it remains possible that YkoW function is redundant with one or more regulators whose action masks the loss of YkoW. 4. Discussion This work describes the rst conserved integral membrane domain found in a sensor histidine kinase. The MHYT domain has been found in several phylogenetically distant bacteria, either as a separate, single domain protein (Streptomyces coelicolor, B. pertussis), or fused to a VirRtype DNA-binding helix-turn-helix domain (O. carboxidovorans), or fused to signaling domains, such as His kinase, PAS, and GGDEF. These observations suggest that the MHYT domain serves as a sensor domain in some bacterial two-component signal transduction systems as well as in a variety of other bacterial proteins (Table 1). The intramembrane topology and the pattern of sequence conservation in the MHYT domain provide possible clues to the environmental signal to which this domain might respond. The conserved His residues of MHYT are located at approximately the same distance from the surface of the membrane (Fig. 2), indicating that they might cooperate, e.g. in binding a metal. The strict conservation of His and Met residues allows one to predict that if a metal is bound in this site, it may be copper [29]. There is a distinct possibility that three conserved His residues of MHYT, perhaps assisted by one of the conserved methionines, could form a three- or four-coordinated Cu-binding site or a binuclear copper center. A model of the orientation of the MHYT domain in the membrane (Fig. 2) shows that conserved Met, His, and Tyr residues can provide up to eight coordinating bonds, su cient for the coordination of one or two copper atoms in the membrane [29]. If the described motif indeed binds copper, its closest native counterpart is the intramembrane Cu B -binding center in the cytochrome c oxidase. Not surprisingly, the sequence of the rst Met-His-containing TM segment has some similarity to the His-containing, Cu B -binding helix of the cytochrome oxidase, although this similarity is not su ciently signi cant to suggest homology. But tellingly, the MHYT domain has an even closer functional counterpart, albeit an arti cial one. Schnepf et al. [30], investigating the metal-binding properties of cellulose-attached antiparallel four-helix bundles, found recently a set of Hiscontaining constructs with a distinct a nity to copper.

6 22 M.Y. Galperin et al. / FEMS Microbiology Letters 205 (2001) 17^23 Only three of the four helices were involved in the coordination of copper, providing two His and one Cys as ligands. The UV light absorption, resonance Raman and electron spin resonance spectra indicated a tetragonal, socalled type 1.5 coordination geometry of the copper atom with water as the fourth ligand [30]. Other recent studies indicate that those four-helix bundles are even able to bind binuclear copper centers. These observations suggest that the MHYT domain forms a metal, presumably copper-containing redox cluster involved in signaling of the redox state of the cytoplasm either directly, by undergoing redox changes, or, being binuclear, indirectly by binding O 2, CO, or NO. As studies with other copper-binding centers show, redox changes can be coupled to a rearrangement of protein ligands [31]. On binding of CO, or O 2, the ligand rearrangement is even larger [29,32]. In the case of MHYT domain, such a rearrangement might serve to transmit a signal, via relocation of TM helices, to the inner surface of the membrane. This type of signaling would be important for the cyanobacterium Synechocystis sp. which has to adjust its metabolism to autotrophic or heterotrophic growth conditions; this would explain the presence of MHYT domain in one of its sensor kinases. Consistent with this notion, the MHYT domain-containing proteins CoxC and CoxH in O. carboxidovorans are reportedly involved in some sort of regulation of chemolitoautotrophic utilization of CO by this organism [25]. It is perhaps not a coincidence that all the bacteria that encode a MHYT domain (Table 1) are either obligate aerobes or facultative anaerobes. In B. subtilis, transcriptional pro ling experiments designed to study the general stress response [33] showed that the MHYT-encoding ykow gene is expressed during logarithmic growth in rich medium. However, our analysis of a ykow null mutant found that this gene is not essential for growth under standard laboratory conditions, or for any of the more obvious anaerobic responses that would directly or indirectly require oxygen sensing. Along these same lines, one might note that the MHYT domain is apparently not encoded within the B. cereus or B. anthracis genomes, further arguing against an essential role for YkoW in a fundamental process that requires oxygen sensing. We therefore propose that MHYT might contribute to NO or CO sensing, as suggested by its potential to bind binuclear copper centers [30]. Further research of the possible function of this domain should include a direct testing of its metal content and its binding properties with respect to oxygen, NO, and particularly CO. Acknowledgements This research was supported in part by Public Health Service grant GM42077 (C.W.P.) and by a Deutsche Forschungsgemeinschaft travel grant (A.Y.M.). Analysis of un nished genome sequences has been made possible by generous submission to the public databases of preliminary sequence data by the Sanger Centre (B. pertussis, Bordetella parapertussis, B. bronchiseptica, B. pseudomallei, S. typhi), the Institute for Genomic Research (B. mallei, P. putida, P. syringae), and the Washington University Genome Sequencing Center (S. enterica, S. paratyphi). References [1] Stock, A.M., Robinson, V.L. and Goudreau, P.N. (2000) Two-component signal transduction. Annu. Rev. Biochem. 69, 183^215. [2] Taylor, B.L. and Zhulin, I.B. (1999) PAS domains: internal sensors of oxygen, redox potential, and light. Microbiol. Mol. Biol. Rev. 63, 479^506. [3] Aravind, L. and Ponting, C.P. (1997) The GAF domain: an evolutionary link between diverse phototransducing proteins. Trends Biochem. Sci. 22, 458^459. [4] Aravind, L. and Ponting, C.P. (1999) The cytoplasmic helical linker domain of receptor histidine kinase and methyl-accepting proteins is common to many prokaryotic signalling proteins. FEMS Microbiol. Lett. 176, 111^116. [5] Hecht, G.B. and Newton, A. (1995) Identi cation of a novel response regulator required for the swarmer-to-stalked-cell transition in Caulobacter crescentus. J. Bacteriol. 177, 6223^6229. [6] Tal, R., Wong, H.C., Calhoon, R., Gelfand, D., Fear, A.L., Volman, G., Mayer, R., Ross, P., Amikam, D., Weinhouse, H., Cohen, A., Sapir, S., Ohana, P. and Benziman, M. (1998) Three cdg operons control cellular turnover of cyclic di-gmp in Acetobacter xylinum: genetic organization and occurrence of conserved domains in isoenzymes. J. Bacteriol. 180, 4416^4425. [7] Galperin, M.Y., Natale, D.A., Aravind, L. and Koonin, E.V. (1999) A specialized version of the HD hydrolase domain implicated in signal transduction. J. Mol. Microbiol. Biotechnol. 1, 303^305. [8] Slater, H., Alvarez-Morales, A., Barber, C.E., Daniels, M.J. and Dow, J.M. (2000) A two-component system involving an HD-GYP domain protein links cell-cell signalling to pathogenicity gene expression in Xanthomonas campestris. Mol. Microbiol. 38, 986^1003. [9] Aldridge, P. and Jenal, U. (1999) Cell cycle-dependent degradation of a agellar motor component requires a novel-type response regulator. Mol. Microbiol. 32, 379^391. [10] Galperin, M.Y., Nikolskaya, A.N. and Koonin, E.V. (2001) Novel domains of the prokaryotic two-component signal transduction system. FEMS Microbiol. Lett. 203, 11^21. [11] Tam, R. and Saier Jr., M.H. (1993) Structural, functional, and evolutionary relationships among extracellular solute-binding receptors of bacteria. Microbiol. Rev. 57, 320^346. [12] Anantharaman, V. and Aravind, L. (2000) Cache ^ a signaling domain common to animal Ca 2 -channel subunits and a class of prokaryotic chemotaxis receptors. Trends Biochem. Sci. 25, 535^537. [13] Zhulin, I.B. (2001) The superfamily of chemotaxis transducers: from physiology to genomics and back. Adv. Microb. Physiol. 45, 157^198. [14] Altschul, S.F., Madden, T.L., Scha er, A.A., Zhang, J., Zheng, Z., Miller, W. and Lipman, D.J. (1997) Gapped BLAST and PSI-BLAST ^ A new generation of protein database search programs. Nucleic Acids Res. 25, 3389^3402. [15] Ponting, C.P., Schultz, J., Milpetz, F. and Bork, P. (1999) SMART: identi cation and annotation of domains from signalling and extracellular protein sequences. Nucleic Acids Res. 27, 229^232. [16] Tatusov, R.L., Galperin, M.Y., Natale, D.A. and Koonin, E.V. (2000) The COG database: a tool for genome-scale analysis of protein functions and evolution. Nucleic Acids Res. 28, 33^36. [17] Corpet, F., Gouzy, J. and Kahn, D. (1999) Recent improvements of the ProDom database of protein domain families. Nucleic Acids Res. 27, 263^267.

7 M.Y. Galperin et al. / FEMS Microbiology Letters 205 (2001) 17^23 23 [18] Ho, S.N., Hunt, H.D., Horton, R.M., Pollen, J.K. and Pease, L.R. (1989) Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77, 51^59. [19] Price, C.W. and Doi, R.H. (1985) Genetic mapping of rpod implicates the major sigma factor of Bacillus subtilis RNA polymerase in sporulation initiation. Mol. Gen. Genet. 201, 88^95. [20] Anagnostopoulos, C. and Spizizen, J. (1961) Requirements for transformation in Bacillus subtilis. J. Bacteriol. 81, 741^746. [21] Nakano, M.M., Dailly, Y.P., Zuber, P. and Clark, D.P. (1997) Characterization of anaerobic fermentative growth of Bacillus subtilis: identi cation of fermentation end products and genes required for growth. J. Bacteriol. 179, 6749^6755. [22] Boylan, S.A., Red eld, A.R., Brody, M.S. and Price, C.W. (1993) Stress-induced activation of the c B transcription factor of Bacillus subtilis. J. Bacteriol. 175, 7931^7937. [23] Rost, B., Fariselli, P. and Casadio, R. (1996) Topology prediction for helical transmembrane proteins at 86% accuracy. Protein Sci. 5, 1704^1718. [24] von Heijne, G. (1992) Membrane protein structure prediction. Hydrophobicity analysis and the positive-inside rule. J. Mol. Biol. 225, 487^494. [25] Santiago, B., Schubel, U., Egelseer, C. and Meyer, O. (1999) Sequence analysis, characterization and CO-speci c transcription of the cox gene cluster on the megaplasmid phcg3 of Oligotropha carboxidovorans. Gene 236, 115^124. [26] Delgado-Nixon, V.M., Gonzalez, G. and Gilles-Gonzalez, M.A. (2000) Dos, a heme-binding PAS protein from Escherichia coli, is a direct oxygen sensor. Biochemistry 39, 2685^2691. [27] Stahl, M.L. and Ferrari, E. (1984) Replacement of the Bacillus subtilis subtilisin structural gene with an in vitro-derived mutation. J. Bacteriol. 158, 411^418. [28] Hou, S., Larsen, R.W., Boudko, D., Riley, C.W., Karatan, E., Zimmer, M., Ordal, G.W. and Alam, M. (2000) Myoglobin-like aerotaxis transducers in Archaea and Bacteria. Nature 403, 540^544. [29] Fraüsto da Silva, J.J.R. and Williams, R.J.P. (1991) The Biological Chemistry of the Elements: the Inorganic Chemistry of Life. Oxford University Press, Oxford. [30] Schnepf, R., Ho«rth, P., Bill, E., Wieghardt, K., Hildebrandt, P. and Haehnel, W. (2001) De novo design and characterization of copper centers in synthetic four-helix-bundle proteins. J. Am. Chem. Soc. 123, 2186^2195. [31] Ralle, M., Verkhovskaya, M.L., Morgan, J.E., Verkhovsky, M.I., Wikstrom, M. and Blackburn, N.J. (1999) Coordination of Cu B in reduced and CO-liganded states of cytochrome bo 3 from Escherichia coli. Is chloride ion a cofactor? Biochemistry 38, 7185^7194. [32] Osborne, J.P., Cosper, N.J., Stalhandske, C.M., Scott, R.A., Alben, J.O. and Gennis, R.B. (1999) Cu XAS shows a change in the ligation of Cu B upon reduction of cytochrome bo 3 from Escherichia coli. Biochemistry 38, 4526^4532. [33] Price, C.W., Fawcett, P., Cërëmonie, H., Su, N., Murphy, C.K. and Youngman, P. (2001) Genome-wide analysis of the general stress response in Bacillus subtilis. Mol. Microbiol. 41, 757^774.