The genetic basis of tetrathionate respiration in ... ?· The genetic basis of tetrathionate respiration…

  • Published on
    27-Jul-2018

  • View
    212

  • Download
    0

Transcript

  • The genetic basis of tetrathionate respiration inSalmonella typhimurium

    Michael Hensel,1 Andrew P. Hinsley,2 ThomasNikolaus,1 Gary Sawers3 and Ben C. Berks2*1Lehrstuhl fur Bakteriologie, Max von Pettenkofer-Institutfur Hygiene und Medizinische Mikrobiologie,Pettenkoferstr. 9a, D-80336 Munich, Germany.2Centre for Metalloprotein Spectroscopy and Biology,School of Biological Sciences, University of East Anglia,Norwich NR4 7TJ, UK.3Nitrogen Fixation Laboratory, The John Innes Centre,Norwich NR4 7UH, UK.

    Summary

    A range of bacteria are able to use tetrathionate as aterminal respiratory electron acceptor. Here we reportthe identification and characterization of the ttrRSBCAlocus required for tetrathionate respiration in Salmo-nella typhimurium LT2a. The ttr genes are locatedwithin Salmonella pathogenicity island 2 at centisome30.5. ttrA, ttrB and ttrC are the tetrathionate reductasestructural genes. Sequence analysis suggests thatTtrA contains a molybdopterin guanine dinucleotidecofactor and a [4Fe4S] cluster, that TtrB binds four[4Fe4S] clusters, and that TtrC is an integral mem-brane protein containing a quinol oxidation site. TtrAand TtrB are predicted to be anchored by TtrC to theperiplasmic face of the cytoplasmic membrane imply-ing a periplasmic site for tetrathionate reduction. It isinferred that the tetrathionate reductase, together withthiosulphate and polysulphide reductases, make up apreviously unrecognized class of molybdopterin-dependent enzymes that carry out the reductive clea-vage of sulphursulphur bonds. Cys-256 in TtrA isproposed to be the amino acid ligand to the molyb-dopterin cofactor. TtrS and TtrR are the sensor andresponse regulator components of a two-componentregulatory system that is absolutely required fortranscription of the ttrBCA operon. Expression of anactive tetrathionate reduction system also requiresthe anoxia-responsive global transcriptional regulatorFnr. The ttrRSBCA gene cluster confers on Escheri-chia coli the ability to respire with tetrathionate aselectron acceptor.

    Introduction

    The ability to respire tetrathionate is characteristic of certaingenera of Enterobacteriaceae including Salmonella, Citro-bacter and Proteus (Richard, 1977; Barrett and Clark,1987). The differential ability of Enterobacteriaceae torespire tetrathionate is exploited in the clinical diagnosticlaboratory where tetrathionate broth is used as a stan-dard enrichment medium for Salmonella species. Similarly,Richard (1977) has pioneered the use of the tetrathionaterespiration phenotype in the classification of Enterobac-teriaceae. The occurrence of tetrathionate respiration hasnot been so extensively studied in other bacterial families,but it is clear that the metabolism is phylogenetically wide-spread (Barrett and Clark, 1987), allowing bacteria totake advantage of the availability of tetrathionate in manyenvironments (Barrett and Clark, 1987; Barbosa-Jeffersonet al., 1998).

    In Salmonella enterica serovar Typhimurium (hereafterS. typhimurium) the tetrathionate reductase (Ttr) is a mem-brane-bound enzyme containing molybdopterin guaninedinucleotide cofactor (MGD) as a prosthetic group (Casseet al., 1972; Hinojosa-Leon et al., 1986). Ttr catalyses thetwo-electron reduction of tetrathionate to two moleculesof thiosulphate (Table 1, reaction 1). When tetrathionatein the growth medium has been fully reduced the bacter-ium can utilize the thiosulphate produced as an additionalelectron acceptor. Thiosulphate is reductively cleaved tosulphite plus sulphide by the enzyme thiosulphate reduc-tase (Phs; production of hydrogen sulphide), a membrane-bound, MGD-binding enzyme that utilizes menaquinol aselectron donor (Table 1, reaction 2). Sulphite is furtherreduced by the cytoplasmic enzyme sulphite reductase(Asr; anaerobic sulphite reductase) (Table 1, reaction 3).

    In Proteus mirabilis a single enzyme is reported to cata-lyse both tetrathionate and thiosulphate reduction (Oltmannet al., 1974). Like the Ttr and Phs enzymes of S. typhimur-ium, the bifunctional reductase from P. mirabilis is a mem-brane-bound molybdenum protein (Oltmann et al., 1979).

    The presence of a molybdopterin cofactor in tetrathio-nate and thiosulphate reductases deserves comment. Themolybdopterin cofactor is normally regarded as carryingout reactions in which either (i) an oxygen atom is trans-ferred between substrate and water via the molybdenumco-ordination sphere, or (ii) the molybdenum atomabstracts a hydrogen atom from the substrate or (iii)these reactions are combined to effect a hydroxylation of

    Molecular Microbiology (1999) 32(2), 275287

    Q 1999 Blackwell Science Ltd

    Received 29 October, 1998; revised 7 January, 1999; accepted 11January, 1999. These authors contributed equally to this work.*For correspondence. E-mail b.berks@uea.ac.uk; Tel. (44) 1603592186; Fax (44) 1603 592250.

  • the substrate (Hille, 1996; Khangulov et al., 1998). Clearlythe reactions catalysed by Ttr, Phs or the MGD-dependentpolysulphide reductase (Psr) of Wolinella succinogenes(Table 1, reaction 4) do not fall into one of these categories.The Phs and Psr reactions can be described as sulphuratom transfers. Alternatively, the reactions catalysed bythese enzymes as well as that performed by Ttr can beviewed as the reductive cleavage of a sulphursulphurbond. In either case, the reaction would represent a pre-viously unrecognized function of the molybdopterin cofactorin biology.

    Although tetrathionate reduction is an important bacter-ial anaerobic respiratory process with intriguing catalyticchemistry our knowledge of this metabolism at the molecularlevel is rudimentary. Taking advantage of the sophisticatedsystems of genetic analysis available for S. typhimurium,we have initiated a molecular study of tetrathionate reduc-tion in this organism. We report here the identification andcharacterization of the genetic locus required for tetrathio-nate respiration.

    Results and discussion

    Isolation of mutants with defects in tetrathionaterespiration

    A mutagenesis approach was used to identify the geneticloci in S. typhimurium specifically required for tetrathionaterespiration. A pool of random miniTn10 transposon inser-tions was screened for the ability to reduce tetrathionate(Ttr phenotype) using the EMB-tetrathionate indicatoragar of Le Minor et al. (1970). When cultured on this med-ium cells respiring tetrathionate develop a metallic sheen.In addition, the large quantities of acid produced duringtetrathionate respiration turns the agar around the coloniesfrom red to deep purple. Cells that are unable to reducetetrathionate fail to develop the metallic sheen and thesurrounding agar remains red. From a screen of 15 000bacteria harbouring mini-Tn10 insertions we isolated nineTtr strains.

    The Ttr strains were subject to secondary screens toidentify those mutants in which the lesion specificallyaffects tetrathionate respiration. Mutants that had a Ttr

    phenotype as a consequence of a pleiotropic molybdop-terin cofactor deficiency were identified by their resistanceto chlorate, which, in cofactor-producing cells, would bemetabolized to toxic chlorite by the molybdopterin-contain-ing enzyme nitrate reductase (Stewart and MacGregor,1982). Mutant strains were also screened for a functionalthiosulphate reduction pathway (Phs) using peptone ironagar plates. This test was used to assess whether theTtr phenotypes represented generalized, possibly regu-latory, defects in which subsequent steps in the thionatereduction pathway are also affected. Seven of the Ttr

    strains were chlorate resistant and Phs (thiosulphatereductase is a molybdoenzyme) and were assumed tohave defects in molybdopterin cofactor biosynthesis. Thetwo remaining strains, BCB4 and BCB6, were chloratesensitive and Phs and were analysed in more detail.

    Mutant strains BCB4 and BCB6 are incapable of reduc-ing tetrathionate when cultured anaerobically on a complex,fermentable carbon source (Table 2). In contrast, the par-ental strain LT2a shows stoichiometric conversion of tetra-thionate to thiosulphate while tetrathionate is in excess(Table 2). The parental strain LT2a, but neither BCB4 norBCB6, can grow anaerobically on minimal medium con-taining glycerol as the non-fermentable carbon sourceand tetrathionate as the sole respiratory electron acceptor.These results confirm that BCB4 and BCB6 are com-pletely defective in physiological tetrathionate reduction.

    To ascertain whether strains BCB4 and BCB6 lack theenzyme tetrathionate reductase, we directly assayed theactivity of the reductase using dithionite-reduced methylviologen radical as a non-physiological electron donor (seeExperimental procedures). The rationale for this assay isthat viologen radicals are capable of donating electronsdirectly to the catalytic site of molybdopterin-dependentenzymes (e.g. McEwan et al., 1991; Guigliarelli et al.,1996). We found that strains BCB4 and BCB6 retain onlyabout 10% of the viologen-linked tetrathionate reductaseactivity of the parental strain (Table 2). Subsequent geneticanalysis of the Ttr system (below) confirms that this residualactivity is not associated with the tetrathionate respiratorypathway. It has been reported that the thiosulphate reduc-tase of P. mirabilis exhibits viologen-linked tetrathionatereductase activity (Oltmann et al., 1974). To test whetherthe residual viologen-linked tetrathionate reductase activ-ity in mutant strains BCB4 and BCB6 was due to thio-sulphate reductase, a phs mutation was moved into BCB4and BCB6. No change in viologen-linked tetrathionatereductase activity was observed in the resultant doublemutants (Table 2). Thus, under the growth conditionsused, thiosulphate reductase does not contribute signifi-cantly to the total methyl viologen-linked tetrathionatereductase activity.

    The large decrease in viologen-linked tetrathionatereductase activity and complete absence of physiological

    Q 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 275287

    Table 1. Reactions of some enzymes that require inorganic sulphurcompounds.

    1. Tetathionate reductase (Ttr)O3S-S-S-SO3

    2e 2 S-SO3

    2. Thiosulphate reductase (Phs)S-SO3

    MQH2 HSO3 HS MQ3. Sulphite reductase (Asr)

    HSO3 3NAD(P)H 3H HS 3NAD(P) 3H2O

    4. Polysulphide reductase (Psr)S-(S)n-S

    2e H S-(S)n 1-S HS

    276 M. Hensel et al.

  • tetrathionate reduction found in the BCB4 and BCB6 strainsindicate that these strains cannot form a functional tetra-thionate reductase enzyme.

    Mapping of transposon insertions in mutantsdefective in tetrathionate respiration

    The approximate chromosomal locations of the transposoninsertions in mutants BCB4 and BCB6 were establishedusing the Benson and Goldman Mud-P22 prophage map-ping set (Benson and Goldman, 1992). The insertions inboth mutants were complemented by phages from strainTT15244 in which DNA is packaged clockwise from aroD.aroD is located at centisome 30.5 on the current S. typhi-murium genetic map (Sanderson et al., 1996). The transpo-sons in BCB4 and BCB6 co-transduced with the TT15244aroD marker at frequencies of 21% (BCB6) and 0.5%(BCB4). From these frequencies it can be inferred (Sander-son and Roth, 1988) that the transposon insertions are

  • Two open reading frames upstream, and divergently tran-scribed from, ttrBCA code for the sensor and responseregulator of a two-component transcriptional regulatorysystem (Hoch and Silhavy, 1995). This genetic organiza-tion is typical of regulators controlling a structural geneoperon, and so a function of these open reading framesin regulation of tetrathionate reductase expression wasanticipated. The genes for the sensor and response regu-lator, designated ttrS and ttrR, respectively, overlap by26 bp, suggesting that the two genes form an operon.

    We identified the sites of the transposon insertions in theTtr mutant strains BCB4 and BCB6. The transposon in theBCB4 strain is inserted at nucleotide 777 of ttrS and that inBCB6 at nucleotide 3002 of ttrA (Fig. 1). These observa-tions provide experimental evidence that the ttr genesidentified by sequence inspection are involved in tetrathio-nate respiration. They also show that a Ttr phenotype canarise from insertions in either the ttrBCA or ttrSR units.We constructed a second ttrS mutant (strain MvP256)in which transcriptional readthrough from the antibioticresistance gene of an inserted marker cassette wouldallow ttrR expression. This mutant has the same Ttr phe-notype as strain BCB4 (ttrS ::Tn10 ) (Table 2), demonstrat-ing that the Ttr phenotype of the transposon mutant is notsolely a consequence of a failure to transcribe ttrR. Similarnon-polar insertional mutagenesis of ttrB (strain MvP257)also produced a Ttr phenotype (Table 2), indicating that

    TtrB, a protein with an apparent electron transfer ratherthan catalytic function, is required for a functional tetrathio-nate reductase.

    A plasmid (pAH26) bearing only ttrRSBCA confers onE. coli the ability to synthesize tetrathionate reductaseand to respire tetrathionate (Table 2). This plasmid alsoconfers on E. coli the ability to grow anaerobically on tetra-thionate-containing minimal media with glycerol as the solecarbon source. These observations indicate that additionalgenes at the centisome 30.5 locus are not required fortetrathionate respiration. In agreement with this observa-tion mutants with insertions in orf242 (MvP256) or thessa, ssc, sse or ssr clusters of SPI2 virulence genes (Hen-sel et al., 1997b; Hensel et al., 1999) remain Ttr (notshown). A plasmid (p1-7) bearing only ttrBCA is ableto complement S. typhimurium ttrA and ttrB mutants butdoes not produce a Ttr phenotype in E. coli (Table 2).This confirms that expression of functional tetrathionatereductase requires the TtrSR two-component regulatorysystem. We conclude that ttrRSBCA constitute the ttrlocus of S. typhimurium.

    The accuracy of the ttr ORF assignments was assessedby independently expressing ttrBCAorf408 and ttrSR froma T7 promoter element (Fig. 2 and Table 3). Polypeptidesof masses appropriate for TtrA, TtrB, TtrR and Orf408could be identified. A protein corresponding to the highlyhydrophobic TtrC protein could not be clearly identified

    Q 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 275287

    Fig. 1. Genetic organization of the ttr locus of S. typhimurium. The positions of mini-transposon Tn10 (mTn10 ) insertions and of the aphTcassette in various mutant strains are indicated, and ttr genes are indicated by hatched arrows (A). Partial restriction map and positions ofplasmids used for cloning, complementation and expression of ttr genes (B). B, BamHI; Bb, BbsI; E, EcoRI; V, EcoRV; P, Pst I; S, SmaI;H, HindIII; C, ClaI; (C), dam methylation-sensitive ClaI site; II, HincII (only the sites used to construct MvP256 and MvP265 are indicatedfor this enzyme).

    278 M. Hensel et al.

  • but might be the diffuse, anomalously migrating band of23 kDa. A product corresponding to TtrS was not detected.

    Sequence analysis of the MGD-binding protein TtrA

    The N-terminal region of the TtrA precursor protein containsa sequence (T-R-R-x-x-L-K) with high similarity to the con-sensus motif (S/T-R-R-x-F-L-K) of twin-arginine transferpeptides (Fig. 3A; Berks, 1996). Such transfer peptidesfunction to direct Sec-independent export of periplasmicallylocated proteins binding cofactors including the molybdop-terin cofactor (Berks, 1996; Santini et al., 1998; Weineret al., 1998; Sargent et al., 1998). A twin-arginine transfersequence suggests that TtrA, and hence the site of tetra-thionate reduction, is located at the periplasmic side of themembrane.

    Immediately after the putative transfer sequence of TtrA,there is a cluster of four cysteine residues. These cysteineresidues are conserved in many other MGD-binding pro-teins and in two cases have been shown to bind a [4Fe4S] cluster (Breton et al., 1994; Boyington et al., 1997).It is therefore probable that TtrA binds a [4Fe4S] cluster.The most likely function of this cluster is to mediate elec-tron transfer from the ironsulphur clusters in TtrB to theMGD cofactor at the active site.

    Analysis of MGD-binding polypeptides has identified aregion that exhibits sequence similarity only between

    those enzymes with similar substrates (Segment III ofBerks et al., 1995). This sequence region can thereforebe used as a fingerprint for the catalytic specificity of theenzyme. Examination of the crystal structures of MGD-binding proteins (Schindelin et al., 1996; Boyington et al.,1997) suggests a structural basis for this sequence specifi-city correlation. The polypeptide region contributes a rela-tively large number of residues to the enzyme active sitepocket and substrate access channel including an aminoacid ligand (serine, selenocysteine or cysteine) to themolybdenum atom of the cofactor. Currently, three func-tional groupings are recognized for this specificity region.These are the S- and N-oxide reductases, the formatedehydrogenases and soluble nitrate reductases, and themembrane-bound nitrate reductases (Berks et al., 1995).On the basis of the alignments in Fig. 4, it is proposed thata fourth specificity grouping can be defined that containsS. typhimurium thiosulphate reductase and the polysul-phide reductase of W. succinogenes. Close inspection indi-cates that TtrA also belongs to this grouping. This suggeststhat the catalytic mechanism of tetrathionate reductase maybe related to that of thiosulphate and polysulphide reduc-tases. On the basis of this inferred mechanistic similarity,we propose that the function of the molybdopterin cofactorin the three enzymes is to cleave reductively a sulphursulphur bond. As noted in the introduction, such a catalyticreaction would represent a previously unrecognized activ-ity of the molybdopterin cofactor in biology.

    Sequence similarity in the specificity-defining regioncan be further used to propose possible candidates forthe amino acid ligand to the molybdenum atom of theMGD cofactor. We suggest that for the sulphursulphurreductases a conserved cysteine residue, Cys-256 inTtrA, is the molybdenum ligand (Fig. 4)

    TtrA shows significantly higher sequence identity(32.8%) to the product of open reading frame AF0159 ofthe archaeal hyperthermophilic sulphate-reducer Archaeo-globus fulgidus (Klenk et al., 1997) than it does to otherMGD-binding proteins. Like ttrA, AF0159 is transcribed

    Q 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 275287

    Fig. 2. Expression of ttr genes under the control of the T7promoter. SDSPAGE analysis of the 35S-labelled polypeptidesproduced by E. coli strain BL21 (DE3) harbouring plasmidspET21() (A), p1-7-T7 (B) and pttrSR-T7 (C) after induction,radiolabelling and processing as described in Experimentalprocedures. The molecular masses in kDa of marker proteins areshown to the left of the gel and the assignment of the inducedbands indicated to the right of the gel.

    Table 3. Properties of the Ttr polypeptides. The experimental deter-mination of molecular mass is shown in Fig. 2.

    Amino Calculated Experimental PredictedProtein acids Mw (Da) Mw (kDa) pI

    TtrA 1020 110 992 104 8.35(107 344)* (7.63)*

    TtrB 250 27 710 26 7.25(24 208)* (7.25)*

    TtrC 340 38 394 23 10.48TtrS 559 61 724 ND 8.56TtrR 206 22 692 19 6.90Orf408 408 44 265 43 8.73

    * Values for the mature forms of TtrA and TtrB produced by pro-cessing at the sites shown in Fig. 3A and B.ND, not detected.

    Salmonella tetrathionate reductase 279

  • after the genes coding for the ferredoxin and integral mem-brane protein subunits of the enzyme. Further, the AF0159gene product resembles TtrA in possessing an N-terminaltwin-arginine transfer sequence (Fig. 3A) and the two pro-teins have sequence similarity in the specificity-determin-ing region (Fig. 4). It is therefore possible that the AF0159gene product is the catalytic subunit of an archaeal tetra-thionate reductase.

    TtrB analysis

    TtrB is a member of a large family of proteins binding four[4Fe4S] clusters (16Fe ferredoxins; Berks et al., 1995).All the cysteines that act as the ironsulphur cluster ligandsare conserved in TtrB. The TtrB ironsulphur clustersprobably function to transfer electrons from TtrC to TtrA.

    The N-terminus of TtrB has sequence similarity to theknown (Keon and Voordouw, 1996) or predicted twin-argi-nine transfer peptides of other 16Fe-ferredoxins (Fig. 3B).The putative TtrB transfer peptide departs from the twin-arginine consensus motif by substituting a lysine residuefor the first of the two totally conserved arginines of themotif. The limited experimental evidence available sug-gests that such a substitution would drastically impair the

    ability of the transfer peptide to direct export by the Sec-independent pathway (Niviere et al., 1992; Chaddock etal., 1995). To exclude the possibility that this substitutionis a cloning artefact, the ttrB region was amplified fromthe S. typhimurium chromosome and the sequence con-firmed. Several possibilities exist to explain the lysine forarginine substitution in what appears to be a twin-argininetransfer peptide. It is conceivable, but unlikely, that thesequence similarity between the N-terminus of TtrB andthe twin-arginine transfer peptides of 16Fe-ferredoxins(Fig. 3B) is coincidental. It is possible that the requirementfor the first arginine in the consensus sequence is relaxedin the context of the TtrB signal sequence. Finally, as it isproposed that export of the 16Fe-ferredoxin subunitsof MGD-dependent enzymes is normally mediated bythe transfer peptide on the MGD-binding subunit (Berks,1996), the N-terminus of TtrB may represent a degenerat-ing evolutionary relic, the function of which has been super-seded by the TtrA transfer peptide.

    TtrC analysis

    TtrC is predicted to be an integral membrane protein con-taining nine transmembrane helices (amino acids 2240,

    Q 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 275287

    Fig. 3. Comparative alignment of the predicted twin-arginine transfer peptide of (A) TtrA with those found on a selection of other MGD-bindingproteins (B) TtrB with other 16Fe ferredoxins, which have, or are proposed to have, twin-arginine transfer peptides. The twin-arginineconsensus motif is (S/T)-R-R-x-F-L-K, where the arginine residues are invariant (Berks, 1996). Amino acids corresponding to the consensusmotif are indicated in bold. Sites of leader peptidase processing are experimentally determined (EX) or predicted (SP) by the program SIGNALP(Nielsen et al., 1997). SWISSPROT database accession numbers for the proteins shown are SPTREMBL O30078 (A. fulgidus AF0159),P18775 (E. coli DmsA; N-terminus reassigned according to Berks, 1996), SPTREMBL P77374 (E. coli O808a), SPTREMBL P77783 (E. coliO808b), P37600 (S. typhimurium PhsA), P31075 (W. succinogenes PsrA), SPTREMBL O30286 (A. fulgidus AF2384), P33225 (E. coli TorA),P46923 (E. coli BisZ), P24183 (E. coli FdnG), P32176 (E. coli FdoG), P33937 (E. coli NapA), SPTREMBL O29751 (A. fulgidus AF0499),P32708 (E. coli NrfC), P45015 (H. influenzae NrfC), SPTREMBL P77375 (E. coli F239), P33389 (D. vulgaris HmcB) and P37179 (E. coliHybA).

    280 M. Hensel et al.

  • 5877, 94112, 132156, 161182, 200215, 230253,267287 and 309327) with a topological organizationthat places the N-terminus of the protein in the periplasm.TtrC exhibits statistically significant sequence similarity tothe AF2386 gene product of A. fulgidus (Klenk et al., 1997),which in turn shows statistically significant sequence simi-larity to the PsrC protein of W. succinogenes polysulphidereductase (Krafft et al., 1992) and the NrfD proteins of theperiplasmic nitrite reductase systems of E. coli and Hae-mophilus influenzae (Hussain et al., 1994; Fleischmannet al., 1995). A small C-terminal extension in TtrC relativeto the other proteins of the PsrC/NrfD/AF2386 family ispredicted to contain an extra transmembrane helix. PsrCand NrfD most likely function to oxidize menaquinol withthe electrons passing to periplasmically located 16Fe-ferre-doxins (Krafft et al., 1992; Hussain et al., 1994; Berks et al.,1995; Tyson et al., 1997; Hedderich et al., 1999). By ana-logy we propose that TtrC transfers electrons from mem-brane quinol to TtrB bound at its periplasmic face. In thiscontext it is notable that the regions of highest sequencesimilarity between Ttr and the PsrC/NrfD family proteins

    are all found at the periplasmic ends of the predicted trans-membrane helices. An arginine residue (Arg-280 in TtrC)towards the periplasmic end of proposed helix eight isthe only invariant charged residue predicted to be buriedin the membrane bilayer and could be involved in the quinoloxidation reaction. There are no conserved histidinesamong the TtrC/PsrC/NrfD family proteins that could pro-vide haem iron axial ligands. We therefore predict that TtrCdoes not bind haem, a proposal that is consistent with areport that haem-deficient S. typhimurium mutants arestill able to respire tetrathionate (Barrett and Clark, 1987).

    Regulation of Ttr synthesis

    TtrS and TtrR have the standard domain organization oftwo-component transcriptional regulators. The responseregulator TtrR and the C-terminal signal kinase domainof the TtrS sensor protein have highest sequence simi-larity to the FixJL subfamily of two-component regulators.However, the N-terminal stimulus-receiving domain of TtrSshows no sequence similarity to other proteins. Although

    Q 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 275287

    Fig. 4. The polypeptide region of MGD-dependent enzymes providing a ligand to the molybdenum atom (Segment III of Berks et al., 1995).These fingerprint regions of representative MGD enzymes are displayed in substrate specificity groups. The secondary structure assignmentsfor the S- and N-oxide reductases and for the formate dehydrogenase/soluble nitrate reductase groupings are taken from the crystalstructures of R. sphaeroides DmsA (Schindelin et al., 1996) and E. coli FdhF (Boyington et al., 1997) respectively. For the reductasescleaving sulphursulphur bonds the secondary structure predictions are those of the program PHD (Rost and Sander, 1994) at the nominal72% confidence level. The molybdenum ligand, or proposed molybdenum ligand, is shown in bold. Residues in R. sphaeroides DmsA involvedin forming the active site or substrate channel are indicated by . PhsA is the catalytic subunit of S. typhimurium thiosulphate reductase. PsrAis the catalytic subunit of Wolinella succinogenes polysulphide reductase. SWISSPROT database accession numbers for the proteins shownare as in Fig. 3 with the addition of P20099 (E. coli BisC), Q57366 (R. sphaeroides DmsA) and P07658 (E. coli FdhF). The membrane-boundnitrate reductase specificity grouping is not shown as it is not crystallographically defined and does not exhibit sequence similarity to thesulphursulphur bond reductases in this region.

    Salmonella tetrathionate reductase 281

  • this sensing domain probably ends in a transmembranehelix, an unambiguous topological model for the remain-der of the domain could not be derived from the sequence.

    The complementation experiments described aboveand in Table 2 show that TtrSR are required for formationof functional tetrathionate reductase and sequence analy-sis of TtrSR indicates that these proteins are most likely tobe involved in control of ttrBCA transcription. To test thisidea ttrB transcript levels were measured in cells grownanaerobically in the presence of tetrathionate (Fig. 5A).A single transcript starting 81 basepairs before the ttrBstart codon was detected in the parental S. typhimuriumstrain (Fig. 5A, lane 1) and in E. coli carrying a plasmidbearing the full ttrRSBCA locus (Fig. 5A, lane 5). However,no transcript was detectable in an S. typhimurium ttrSmutant (Fig. 5A, lane 2) or in E. coli containing a plasmidbearing the ttrBCA and not the ttrRS genes (Fig. 5A,lane 4). These experiments demonstrate that the TtrSRsystem is absolutely required for ttrB transcription. Two-component transcriptional regulators interact with s70-containing RNA polymerase. Reasonable 10 and 35sequences for a s70-dependent promoter are presentupstream of the ttrB transcriptional start site (Fig. 5B), sug-gesting that direct regulation of the ttrB promoter by TtrSRis plausible.

    Tetrathionate reductase expression is induced by its sub-strate (Pollock and Knox, 1943) and the TtrSR system isan obvious candidate for the tetrathionate-sensing path-way. Additional environmental factors may also contributeto the control of Ttr expression, most notably a require-ment for anaerobiosis (Barrett and Clark, 1987). Oxygensensing in enteric bacteria is usually mediated by theFnr global transcriptional regulator. An S. typhimuriumstrain carrying an fnr (formerly oxrA) mutation was unableto respire tetrathionate or produce viologen-linked tetra-thionate reductase activity (Table 2), indicating that Fnris required for Ttr expression. Tetrathionate respiratoryactivity was not restored by growing the strain in the pre-sence of formate (Table 2), demonstrating that Fnr is notregulating Ttr expression indirectly by affecting formate pro-duction. A sequence with high (9 out of 10 bases) similarityto the consensus Fnr binding site TTGATxxxxATCAA iscentred at 71.5 relative to the ttrB transcriptional startsite (Fig. 5B). Fnr binding sites are normally centred at41.5 relative to the transcriptional start site formingclass II promoters at which both subunits of the Fnr homo-dimer interact with RNA polymerase. However, experimentswith synthetic promoters have established that Fnr is alsoable to function at sites further upstream of the transcrip-tional start, including at 71 (Wing et al., 1995). In thesecases class I promoters are formed in which only thedownstream subunit of Fnr binds to the RNA polymerase.Naturally occurring class I Fnr-activated promoters haverecently been identified for the E. coli napF (Darwin et al.,

    1998) and hlyE operons (Ralph et al., 1998), and it is there-fore possible that the ttrB promoter is another rare exampleof this mechanism of Fnr action. Alternatively, the Fnr bind-ing site in the ttrS/ttrB intergenic region could be involvedin regulating ttrS expression. Note, however, that if the Fnrbinding site is positioned at 41.5 relative to the ttrS tran-script start the ttrS promoter lacks recognizable RNA poly-merase 10 and 35 binding sites.

    Q 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 275287

    Fig. 5. A. Primer extension analysis of ttrB transcription in totalRNA from various strains grown anaerobically in the presence oftetrathionate. Lane 1, S. typhimurium LT2a; lane 2, S. typhimuriumBCB4 (ttrS ::mTn10 ); lane 3, E. coli DH5a; lane 4, E. coli DH5ap1-7 (ttrBCA); lane 5, E. coli DH5a pAH26 (ttrRSBCA). Lanes14 contain samples prepared from 2.3 mg of RNA, whereas lane 5used a 0.6 mg RNA sample. The DNA sequence was generatedusing the same labelled oligonucleotide used to perform the primerextension experiments. The arrow indicates the location in the DNAsequence of the longest transcript.B. Sequence of the ttrS/ttrB intergenic region. The bent arrowindicates the transcriptional start site identified in (A). Putative ttrBtranscript 10 and 35 RNA polymerase binding sites, ttrBribosome binding site (RBS) and a possible Fnr binding site areindicated. The numbering corresponds to that of the databaseentry.

    282 M. Hensel et al.

  • Occurrence of ttr genes

    We tested representative strains of all S. enterica/S. bon-gori subspecies in the SARC reference collection (Boyd etal., 1996) and found them to be both Ttr (data not shown)and to exhibit hybridization between their chromosomalDNA and ttrA, ttrC and ttrS probes (Hensel et al., 1999).The ability to respire tetrathionate is therefore likely tobe significant within the life cycle of Salmonella spp. Initialstudies indicate that an inability to respire tetrathionatedoes not significantly affect the virulence of S. typhimur-ium in mice (Hensel et al., 1999). Tetrathionate respirationmay therefore be most important when the bacterium is inthe free living state in tetrathionate-containing environmentssuch as soil or decomposing carcasses.

    In conclusion, we have reported the first genetic identi-fication of a tetrathionate-reducing system together witha two-component regulatory system required for its synth-esis. The genetic characterization described will provide abasis for future biochemical, bioenergetic and regulatorystudies of this important respiratory pathway.

    Experimental procedures

    Bacterial strains, plasmids, phages and cultureconditions

    Bacterial strains, phages and plasmids used in this study are

    described in Table 4. Unless otherwise stated, bacteria werecultured on Luria broth (LB) or LB agar at 378C (Sambrook etal., 1989). Antibiotics were used at the following concentra-tions: tetracycline, 10 mg ml1; ampicillin, 50 mg ml1; kanamy-cin 50 mg ml1. Where required, potassium tetrathionate wasadded to media directly before inoculation from a freshly pre-pared 1 M filter-sterilized stock solution. For anaerobic growthon liquid media, the bacteria were cultured in completely filledMcCartney bottles. For anaerobic growth on solid media,plates were incubated in gas jars under a nitrogen/carbondioxide atmosphere generated using the Anaerogen system(Oxoid). Tests for growth on the non-fermentable carbonsource glycerol used the minimal medium described by Popeand Cole (1982) but lacking the nutrient broth componentand supplemented with 30 mM glycerol as the carbon sourceand 30 mM potassium tetrathionate as the respiratory elec-tron acceptor.

    Phenotypic tests on solid media for Ttr activity used EMB-tetrathionate plates (Le Minor et al., 1970). Chlorate resis-tance was assessed on solid media as described by Stewartand MacGregor (1982). Peptone-iron agar plates were usedto assess Phs phenotype and contained 15 g of agar, 15 gof bactopeptone, 5 g of proteose peptone, 0.5 g of ferricammonium citrate, 1 g of sodium glycerophosphate and 10 gof sodium thiosulphate per litre.

    Genetic techniques

    P22-mediated transductions were performed as detailed inMaloy et al. (1996) using P22 HT int. Marked Ttr-defective

    Q 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 275287

    Table 4. Strains and plasmids used in this study.

    Strains Relevant genotype/characteristic Source or reference

    S. typhimuriumLT2a Wild type B. N. AmesEB222 phs-101::Mu d1 (Aprlac) Clark and Barrett (1987)TT10423 LT2/pNK972 (Tn10 transposase overexpresser) Way et al. (1984)TT10427 proAB471 F8 128 (pro lac zzf-1831::Tn10D16D17 ) Way et al. (1984)TT15244 aroD561::MudP Benson and Goldman (1992)TN2336 LeuBCD485 pepT7::MudJ fnr[oxrA2 ::Tn10 allele] C. Miller via SGSCBCB4 ttrS ::mTn10 This studyBCB6 ttrA::mTn10 This studyAPH6 ttrS ::mTn10 phs-101::Mu d1 (Apr lac) Recipient BCB4, donor P22 lysate EB222APH13 ttrA::mTn10 phs-101::Mu d1 (Apr lac) Recipient BCB6, donor P22 lysate EB222APH17 fnr[oxrA2 ::Tn10 allele] Recipient LT2a, donor TN2336MvP256 DttrS ::aphT deletioninsertion between HindII sites in ttrS This studyMvP257 ttrB::aphT insertion at EcoRV site in ttrB This studyMvP265 orf242 ::aphT insertion at HindII site in orf242 Hensel et al. (1999)

    E. coliDH5a F8, endA1 hsdR17(rK

    mK ) supE44 thi-1 recA1 gyrA(Nalr )

    relA1 D(lacIZYAargF )U169 deoR (f80 dLacD(lacZ )M15 )Gibco BRL

    S17-1 lpir lpir lysogen thi pro hsdR recA RP4-2[Tcr ::Mu][Kmr::Tn7 (TprSmr)] de Lorenzo and Timmis (1994)

    PlasmidspSU41 Kmr Bartolome et al. (1991)pKS() Ampr StratagenepET21() Ampr T7 expression vector NovagenpAH26 ttrRSBCA on 7.7kb fragment bounded by BbsI and EcoRI sites in pSU41 This studyp1-7 ttrS 8BCA orf408 orf2458 on 8.2 kb ClaI fragment in pBluescript KS This studyp1-7-T7 ttrS 88 BCA orf408 orf2458 on 7.6 kb Sal INot I fragment from p1-7 in pET21() This studypttrSR-T7 ttSR in pET21() This study

    Salmonella tetrathionate reductase 283

  • mutants were selected using the transposase-defective tetra-cycline-resistant mini-transposon Tn10D16D17 (Way et al.,1984). A pool of random mutations was constructed in strainTT10427 (which contains a plasmid overexpressing Tn10transposase) by transducing to tetracycline resistance witha P22 lysate grown on the mini-Tn10-containing strainTT10423. A fresh P22 lysate prepared on the pool of cellswith mini-Tn10 insertions was then used to transduce strainLT2a, and transductants were screened directly for the Ttrphenotype on tetracycline-containing EMB-tetrathionate plates.Linkage of insertion and the Ttr phenotype was confirmed bytransduction into a fresh LT2a background. Rapid mapping ofthe ttr ::Tn10 insertions was carried out using the Benson andGoldman excision-defective Mud-P22 prophage set (Bensonand Goldman, 1992) using EMB-tetrathionate agar to ascer-tain the Ttr phenotype.

    DNA techniques

    Standard recombinant DNA techniques were as described bySambrook et al. (1989). The ttr locus was sequenced usingsubclones from phage l1 (Shea et al., 1996) predominantlyby a primer-walking strategy. The sequencing reactions wererun on an ABI 377 sequencer and assembled using the pro-grams ASSEMBLYLIGN and MACVECTOR (Oxford). All regions weresequenced at least twice on each strand.

    Sequence analysis used the University of WisconsinGenetics Computing Group package (Version 8.1) with theGCGEMBL extension (Version 8.1.0) (Devereux et al., 1984)and the BLAST server at the NCBI (Altschul et al., 1990).Sequence similarities were regarded as indicating definitestructural similarity, and designated statistically significant inthe main text, if pairwise alignment of the sequences providesa binary comparison score at least nine standard deviationsgreater than that of randomized sequences of the sameamino acid composition (Marger and Saier, 1993). The pre-sence of leader peptidase sites was assessed using the pro-gram SIGNALP version 1.1 (Nielsen et al., 1997). The presence,location and topological organization of potential transmem-brane helical regions was assessed for single sequences bythe program DAS (Cserzo et al., 1997) in combination with thepositive-inside rule (von Heijne, 1992), and for multiple align-ments of homologous proteins by the programs TMAP (Perssonand Argos, 1994) and PHD (Rost et al., 1996). When predic-tions from multiple methods were available, the predictionswere in complete agreement.

    The location of the mini-transposon insertions was deter-mined by directly sequencing the specific product of PCRreactions using a Tn10-specific primer (58-AAGGATCCTT-TGGTCACCAACGCTTTTCCC-38 or 58-GGAATTGATTGCT-GTTGACAAAGGGAATC-38) and one of a panel of ttr-specificprimers.

    Directed ttr mutations were constructed by inserting a kana-mycin resistance gene lacking a transcriptional terminator(the aphT cassette of Galan et al., 1992) in the same orienta-tion as the Salmonella gene. These constructs allow expres-sion of genes downstream of the insertion site. For the ttrBmutation the aphT cassette was inserted in the EcoRV siteof ttrB and for the orf242 mutation in the HindII site oforf242. For the ttrS mutation, a HindII fragment of 117 bpwas deleted within ttrS and replaced by the aphT cassette.

    The disrupted genes were cloned into the lpir-dependent sui-cide vector pKAS32 (Skorupski and Taylor, 1996) and resultingconstructs were electroporated into E. coli S17-1 lpir (de Lor-enzo and Timmis, 1994). Plasmids were then conjugationallytransferred from E. coli S17-1 lpir to S. typhimuriumMvP100 (Strepr) by filter matings. Recipients that had inte-grated the suicide vector were selected by resistance to kana-mycin and chloramphenicol. Subsequently, clones that hadundergone replacement of wild-type alleles by alleles dis-rupted by the aphT cassette were enriched by culturing inLB containing 250 mg ml1 streptomycin as previouslydescribed (Skorupski and Taylor, 1996). The positions ofthe aphT cassettes were confirmed by Southern hybridizationanalysis of total DNA of exconjugants, and the insertions weretransferred to LT2a-derived backgrounds by P22-mediatedtransduction.

    Plasmid pAH26 was constructed by blunt-ending the BbsIfragment bearing ttrSRBCA8 and cloning this into HincII-cutpSU41 then replacing the fragment of the resultant plasmidextending from the Pst I site in ttrS through ttrBCA8 to theEcoRI site of the multiple cloning site by the ttrS8BCA Pst I/EcoRI fragment from p1-7. Plasmid p1-7-T7 for T7-dependentexpression was constructed by subcloning the 7.5 kb Sal I/Not Ifragment of p1-7 in pET21(). For the construction of plasmidpttrSR-T7, PCR was performed using primers ttrS-For (58-TCCATATTGACTCCCGTCC-38) and ttrR-Rev (58-ACTTTT-TACTCGCCAG-38). The resulting product of 2.5 kb wascloned in pT7-Blue using the PerfectlyBlunt cloning system(Novagen), and the insert was subsequently transferred topET21(). Plasmids in which ttrSR were transcribed fromthe T7 promoter were identified by restriction analysis andtransferred to BL21(DE3).

    Total RNA was isolated from cultures grown to mid-expo-nential phase, and primer extension analysis was carriedout as described by Sawers and Bock (1989). Analysis of thettrB transcription start site was performed with oligonucleotideTtrb-1 (58-CCAAGCTGCTGGAGAAATTGCC-38) and the DNAsequence ladder was generated using pAH26 as template(Sanger et al., 1977).

    Analytical methods

    Tetrathionate reduction was assessed in cells cultured accord-ing to a standard protocol. LB was inoculated with 1% concen-tration of an overnight culture of aerobically grown cells. Thecells were then cultured anoxically until mid-logarithmic phase(A600 0.3), 30 mM potassium tetrathionate was added tothe culture, and growth continued for another 2.5 h. After2.5 h cells were pelleted by centrifugation. The tetrathionateand thiosulphate concentrations in the supernatant weredetermined, while the pelleted cells were washed threetimes in ice-cold 10 mM sodium phosphate buffer pH 7.4 andthen assayed for methyl viologen-linked tetrathionate reduc-tase activity. Culture on a fermentable carbon source allowedanalysis of strains with defects in tetrathionate respiration.However, on such media induction of tetrathionate reductasevaried somewhat with the growth phase, hence the adoptionof a standardized induction protocol.

    Tetrathionate and thiosulphate concentrations were deter-mined by cyanolysis according to the protocols of Kelly andWood (1994). Methyl viologen-linked tetrathionate reduction

    Q 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 275287

    284 M. Hensel et al.

  • was measured by a modification of the nitrate reductaseassay of Bell et al. (1990). Assays were carried out in glasscuvettes, sealed with a butyl rubber stopper and renderedanoxic by bubbling with oxygen-free nitrogen. The assay solu-tion contained 10 mM sodium phosphate buffer pH 7.4, 2 mMNa2EDTA and 1 mM methyl viologen. The methyl viologenwas reduced by titration with a freshly prepared solution ofsodium dithionite (

  • Devereux, J., Haeberli, P., and Smithies, O. (1984) A com-prehensive set of sequence analysis programs for theVAX. Nucleic Acids Res 12: 387395.

    Fleischmann, R.D., Adams, M.D., White, O., Clayton, R.A.,Kirkness, E.F., Kerlavage, A.R., et al. (1995) Whole-gen-ome random sequencing and assembly of Haemophilusinfluenzae Rd. Science 269: 496511.

    Galan, J.E., Ginocchio, C., and Costeas, P. (1992) Molecularand functional characterization of the Salmonella invasiongene invA: homology of InvA to members of a new proteinfamily. J Bacteriol 174: 43384349.

    Guigliarelli, B., Magalon, A., Asso, M., Bertrand, P., Frixon,C., Giordano, G., and Blasco, F. (1996) Complete coordi-nation of the four Fe-S centers of the b subunit fromEscherichia coli nitrate reductase. Physiological, biochemi-cal, and EPR characterization of site-directed mutantslacking the highest and lowest potential [4Fe-4S] clusters.Biochem 35: 48284836.

    Hedderich, R., Klimmek, O., Kroger, A., Dirmeier, R., Keller,M., and Stetter, K.O. (1998) Anaerobic respiration withsulfur, polysulfide and disulfide. FEMS Microbiol Rev 22:353381.

    von Heijne, G. (1992) Membrane protein structure prediction hydrophobicity analysis and the positive-inside rule. J MolBiol 225: 487494.

    Hensel, M., Shea, J.E., Baumler, A.J., Gleeson, C., Blattner,F., and Holden, D.W. (1997a) Analysis of the boundaries ofSalmonella Pathogenicity Island 2 and the correspondingchromosomal region of Escherichia coli K-12. J Bacteriol179: 11051111.

    Hensel, M., Shea, J.E., Raupach, B., Monack, D., Falkow, S.,Gleeson, C., and Holden, D.W. (1997b) Functional analy-sis of ssaJ and the ssaK/U operon, 13 genes encodingcomponents of the type III secretion apparatus of Salmo-nella pathogenicity island 2. Mol Microbiol 24: 155167.

    Hensel, M., Nikolaus, T., and Egelseer, C. (1999) Molecularand functional analysis indicates a mosaic structure ofSalmonella pathogenicity Island 2. Mol Microbiol 31:489498.

    Hille, R. (1996) The mononuclear molybdenum enzymes.Chem Rev 96: 27572816.

    Hinojosa-Leon, M., Dubourdieu, M., Sanchez-Crispin, J.A.,and Chippaux, M. (1986) Tetrathionate reductase of Sal-monella typhimurium: a molybdenum containing enzyme.Bioch Biophys Res Comm 136: 577581.

    Hoch, J.A., and Silhavy, T.J. (eds). (1995) Two-ComponentSignal Transduction. Washington DC: American Societyfor Microbiology Press.

    Hussain, H., Grove, J., Griffiths, L., Busby, S., and Cole, J.(1994) A seven-gene operon essential for formate-depen-dent nitrite reduction to ammonia by enteric bacteria. MolMicrobiol 12: 153163.

    Kelly, D.P., and Wood, A. (1994) Synthesis and determina-tion of thiosulphate and polythionates. Methods Enzymol243: 475501.

    Keon, R.G., and Voordouw, G. (1996) Identification of theHmcF and topology of the HmcB subunit of the Hmc com-plex of Desulfovibrio vulgaris. Anaerobe 2: 231238.

    Khangulov, S.V., Gladyshev, V.N., Dismukes, G.C., and Stadt-man, T.C. (1998) Selenium-containing formate dehydro-genase H from Escherichia coli : a molybdopterin enzyme

    that catalyzes formate oxidation without oxygen transfer.Biochem 37: 35183528.

    Klenk, H.P., Clayton, R.A., Tomb, J.-F., White, O., Nelson,K.E., Ketchum, K.A., et al. (1997) The complete genomesequence of the hyperthermopholic, sulphate-reducingarchaeon Archaeoglobus fulgidus. Nature 390: 364370.

    Krafft, T., Bokranz, M., Klimmer, O., Schroder, I., Fahren-holz, F., Kojro, E., and Kroger, A. (1992) Cloning andnucleotide sequence of the psrA gene of Wolinella succi-nogenes polysulphide reductase. Eur J Biochem 206:503510.

    Le Minor, L., Chippaux, M., Pichinoty, F., Coynault, C.,and Piechaud, M. (1970) Methodes simples permettantde rechercher la tetrathionate reductase en culturesliquides ou sur colonies isolees. Ann Inst Pasteur 119:733737.

    de Lorenzo, V., and Timmis, K.N. (1994) Analysis and con-struction of stable phenotypes in Gram-negative bacteriawith Tn5-and Tn10-derived minitransposons. MethodsEnzymol 235: 386405.

    McEwan, A.G., Ferguson, S.J., and Jackson, J.B. (1991)Purification and properties of the dimethyl sulfoxide reduc-tase from Rhodobacter capsulatus. Biochem J 274: 305307.

    Maloy, S.R., Stewart, V.J., and Taylor, R.K. (1996) GeneticAnalysis of Pathogenic Bacteria. Cold Spring Harbor,NY: Cold Spring Harbor Laboratory Press.

    Marger, M.D., and Saier, M.H., Jr. (1993) A major super-family of transmembrane facilitators that catalyse uni-port, symport and antiport. Trends Biochem Sci 18:1320.

    Nielsen, H., Engelbrecht, J., Brunak, S., and von Heijne, G.(1997) Identification of prokaryotic and eukaryotic signalpeptides and prediction of their cleavage sites. Prot Eng10: 16.

    Niviere, V., Wong, S.-L., and Voodouw, G. (1992) Site-direc-ted mutagenesis of the hydrogenase signal peptide consen-sus box prevents export of a b-lactamase fusion protein.J Gen Microbiol 138: 21732183.

    Oltmann, L.F., Schoenmaker, G.S., and Stouthamer, A.H.(1974) Solubilization and purification of a cytoplasmic mem-brane bound enzyme catalyzing tetrathionate and thiosul-fate reduction in Proteus mirabilis. Arch Microbiol 98:1930.

    Oltmann, L.F., Claasen, V.P., Kastelein, P., Reijinders,W.N.M., and Stouthamer, A.H. (1979) The influence oftungstate on the formation and activities of four reductaseof Proteus mirabilis: identification of two new molybdo-enzymes, chlorate reductase and tetrathionate reductase.FEBS Lett 106: 4346.

    Persson, B., and Argos, P. (1994) Prediction of transmem-brane segments in proteins utilising multiple sequencealignments. J Mol Biol 237: 182192.

    Pollock, M.R., and Knox, R. (1943) Bacterial reduction of tetra-thionate. Biochem J 37: 476481.

    Pope, N.R., and Cole, J.A. (1982) Generation of a membranepotential by one of two independent pathways for nitritereduction by Escherichia coli. J Gen Microbiol 128: 219222.

    Ralph, E.T., Guest, J.R., and Green, J. (1998) Altering theanaerobic transcription factor FNR confers a hemolytic

    Q 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 275287

    286 M. Hensel et al.

  • phenotype on Escherichia coli K12. Proc Natl Acad SciUSA 95: 1044910452.

    Richard, C. (1977) La tetrathionate-reductase (TTR) chez lesbacilles a gram negatif: interet diagnostique et epidemilo-qique. Bull Inst Pasteur 75: 369382.

    Rost, B., and Sander, C. (1994) Combining evolutionaryinformation and neural networks to predict protein second-ary structure. Proteins 19: 5572.

    Rost, B., Fariselli, P., and Casadio, R. (1996) Topology pre-diction for helical transmembrane proteins at 86% accuracy.Protein Sci 7: 17041718.

    Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989) MolecularCloning. A Laboratory Manual. Cold Spring Harbor, NY:Cold Spring Harbor Laboratory,

    Sanderson, K.E., and Roth, J.R. (1988) Linkage map of Sal-monella typhimurium, Edition VIII. Microbiol Rev 52: 485532.

    Sanderson, K.E., Hessel, A., Liu, S.-L., and Rudd, K. (1996)The genetic map of Salmonella typhimurium Edition VIII. InEscherichia Coli and Salmonella Cellular and MolecularBiology, 2nd edn. Neidhardt F.C. (ed.). Washington DC,USA: American Society for Microbiology Press.

    Sanger, F., Nicklen, S., and Coulsen, A.R. (1977) DNA sequ-encing with chain terminating inhibitors. Proc Natl Acad SciUSA 74: 54635467.

    Santini, C.-L., Ize, B., Chanal, A., Muller, M., Giordano, G.,and Wu, L.-F. (1998) A novel Sec-independent periplasmicprotein translocation pathway in Escherichia coli. EMBO J17: 101112.

    Sargent, F., Bogsch, E.G., Stanley, N.R., Wexler, M., Robin-son, C., Berks, B.C., and Palmer, T. (1998) Overlappingfunctions of components of a bacterial Sec-independentprotein export pathway. EMBO J 17: 36403650.

    Sawers, G., and Bock, A. (1989) Novel transcriptional controlof the pyruvate formate-lyase gene: upstream regulatorysequences and multiple promoters regulate anaerobicexpression. J Bacteriol 174: 24852498.

    Schagger, H., and von Jagow, G. (1987) Tricine-sodiumdodecyl sulfate-polyacrylamide gel electrophoresis for

    the separation of proteins in the range from 1 to 100 kDa.Anal Biochem 166: 368379.

    Schindelin, H., Kisker, C., Hilton, J., Rajagopalan, K.V., andRees, D.C. (1996) Crystal structure of DMSO reductase:redox-linked changes in molybdopterin coordination.Science 272: 16151621.

    Shea, J.E., Hensel, M., Gleeson, C., and Holden, D.W. (1996)Identification of a virulence locus encoding a second typeIII secretion system in Salmonella typhimurium. Proc NatlAcad Sci USA 93: 25932597.

    Skorupski, K., and Taylor, R.K. (1996) Positive selectionvectors for allelic exchange. Gene 169: 4752.

    Smith, P.K., Krohn, R.I., Hermanson, G.T., Maliia, A.K., Gart-ner, F.H., Provenzano, M.D., et al. (1985) Measurementsof protein using bicinchoninic acid. Anal Biochem 150:7685.

    Stewart, V., and MacGregor, C.H. (1982) Nitrate reductase inEscherichia coli K-12: involvement of chlC, chlE and chlGloci. J Bacteriol 151: 788799.

    Tabor, S., and Richardson, C.C. (1985) A bacteriophageT7 polymerase/promoter system for controlled exclusiveexpression of specific genes. Proc Natl Acad Sci USA82: 10741078.

    Tyson, K., Metheringham, R., Griffiths, L., and Cole, J.(1997) Characterisation of Escherichia coli K-12 mutantsdefective in formate-dependent nitrite reduction: essentialroles for hemN and the menFDBCE operon. Arch Microbiol168: 403411.

    Way, J.C., Davis, M.A., Morisato, D., Roberts, D.E., andKleckner, N. (1984). New Tn10 derivatives for transposonmutagenesis and for the construction of lacZ operonfusions by transposition. Gene 32: 369379.

    Weiner, J.H., Bilous, P.T., Shaw, G.M., Lubitz, S.P., Frost, L.,Thomas, G.H., et al. (1998) A novel and ubiquitous systemfor membrane targeting and secretion of cofactor-contain-ing proteins. Cell 93: 93101.

    Wing, H.J., Williams, S.M., and Busby, S.J. (1995) Spacingrequirements for transcriptional activation by Escherichiacoli FNR protein. J Bacteriol 177: 67046710.

    Q 1999 Blackwell Science Ltd, Molecular Microbiology, 32, 275287

    Salmonella tetrathionate reductase 287