The unorthodox histidine kinases BvgS and EvgS are responsive to the oxidation status of a quinone electron carrier Andreas Bock and Roy Gross Lehrstuhl fu ¨ r Mikrobiologie, Biozentrum, Universita ¨ tWu ¨ rzburg, Am Hubland, Wu ¨ rzburg, Germany The purified soluble forms of the histidine kinases BvgS and EvgS of Bordetella pertussis and Escherichia coli,respect- ively, are shown to be responsive to oxidized ubiquinone-0 (Q-0) in vitro. The oxidized ubiquinone is a strong inhibitor of kinase activity of both enzymes with half maximal inhi- bition occurring at 11 l M (BvgS) and 4 l M (EvgS). Reduced Q-0 has no effect on the histidine kinases. Kinase activity can reversibly be switched off and on by changing the oxidation status of the quinone. This inhibitory effect is due to a decrease of the kinase activity of BvgS rather than an increase of intrinsic phosphatase activities. Other electron carriers such as menadione (MK-3), NAD or FAD did not have a significant effect on the kinase activities of BvgS and EvgS. Nicotinic acid and sulfate ions, known to inhibit the histidine kinases in vivo, did not affect the purified truncated sensor proteins lacking their periplasmic domains in vitro. Mutations introduced by site-directed mutagenesis into the putative PAS domain of BvgS caused a weak decrease of qui- none-dependent inhibition of autophosphorylation. These data suggest that BvgS and EvgS are connected with the oxi- dation status of the cell via the link to the ubiquinone pool. Keywords: BvgS; EvgS; histidine kinase; signal transduction; PAS domain; phosphorelay; ubiquinone. Two-component signal transduction systems are wide- spread among prokaryotes, but they are also found in eukaryotic organisms such as fungi and plants [1,2]. The BvgAS two-component system regulates the virulence properties of Bordetella pertussis, the etiological agent of whooping cough, and of close relatives such as B. parap- ertussis and B. bronchiseptica that cause similar respiratory diseases in man and animals, respectively [3,4]. It controls expression of a huge virulence regulon including the genes encoding the pertussis and adenylate cyclase toxins and several adhesins. Much less is known about the EvgAS system of Escherichia coli, but it appears to be involved in the transcriptional regulation of drug efflux pumps [5]. The two systems are highly related in their primary structure and domain composition. The histidine kinases BvgS and EvgS belong to the small group of the so-called unorthodox sensor proteins because they harbour an intermediate receiver and at the C-terminus a histidine-containing phosphotransfer (HPt) domain in addition to the classical transmitter domain [6–9]. Under permissive conditions, the BvgS and EvgS histidine kinases autophosphorylate at a conserved histidine residue in their transmitter domains. Autophosphorylation is followed by an obligate intramo- lecular phosphorelay to an aspartic acid in their receiver and a histidine in their HPt domains. Finally, the phosphate from the HPt domain is transferred with high specificity to an aspartic acid in the cognate response regulators BvgA and EvgA, respectively [7,10]. Both sensor proteins have large periplasmic domains which were believed to be the major signal input sites. Very little is known so far about the relevant signals perceived by these proteins, although temperature and compounds such as nicotinic acid and sulfate ions were recognized to modulate kinase activity in vivo [9,11]. However, attempts to identify specific binding of modulating compounds to B. pertussis cells remained inconclusive [12]. Recently, the presence of a PAS domain in BvgS was predicted; it is located in the cytoplasmic portion between the transmembrane and transmitter domains [13]. PAS is an acronym formed from the names of the proteins in which imperfect repeat sequences (PAS domains) were first identified, i.e. the Drosophila period clock protein (PER), vertebrate aryl hydrocarbon receptor nuclear translocator (ARNT), and Drosophila single-minded protein (SIM) [13]. PAS domains are signalling domains that function as input modules able to perceive oxygen, redox potential, light, and some other stimuli [13]. The part of the BvgS protein containing the putative PAS domain was previously called linker region and several mutations have been identified within this domain which led either to constitutive kinase activity or to an inactivation of the protein [14–17]. This suggests that the BvgS protein may perceive intracellular signals characteristic for PAS domains such as the energy status of the cell or the oxygen concentration. In this respect it is interesting to note that the BvgAS system does not only control expression of virulence related functions but also several housekeeping genes including cytochrome c and cytochrome d oxidases of the respiratory chain and several putative dehydrogenases [4,18,19]. Remarkably, in contrast to the virulence genes, many of these metabolic functions are negatively controlled by the BvgAS system. Several properties of the BvgAS regulon are similar to those of another unorthodox two-component system, the ArcAB system of E.coli[20,21]. The unorthodox histidine Correspondence to R. Gross, Lehrstuhl fu ¨ r Mikrobiologie, Biozen- trum, Universita ¨ tWu ¨ rzburg, Am Hubland, D-97074 Wu ¨ rzburg, Germany. Fax: + 49 931 888 4402, Tel.: + 49 931 888 4403, E-mail: roy@biozentrum.uni-wuerzburg.de Abbreviations: HPt, histidine-containing phosphotransfer; Q-0, ubiquinone-0; Q-8, ubiquinone-8; MK-8, menaquinone-8. (Received 15 February 2002, revised 7 May 2002, accepted 30 May 2002) Eur. J. Biochem. 269, 3479–3484 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03029.x kinase ArcB together with the response regulator ArcA regulates the expression of many genes which are involved in the adaptation of the bacteria during the switch between aerobic and anaerobic growth conditions. The ArcAB regulon comprises factors involved in electron transport, the citrate cycle, the glyoxylate shunt and lipid metabolism [21]. Recently, it was shown that the oxidized form of quinones such as ubiquinone and menadione (MK-3) specifically interfere with the kinase activity and switch the enzyme off, whereas the reduced quinones remained without effect [22]. Interestingly, the unorthodox ArcB protein shows a domain structure that is very similar to those of the BvgS and EvgS proteins, but it lacks the extensive periplasmic domain. Because of to their striking structural similarities to ArcB and to the genes regulated by these two-component systems, in the present study we investigated whether quinones may also be a relevant signal for the BvgS and EvgS histidine kinases. MATERIALS AND METHODS Phosphorylation assays Over-expression and purification of the histidine kinases BvgS and EvgS was described recently [7]. Briefly, both proteins were expressed with an N-terminal His-tag which is fused in front of the putative PAS domains of the two proteins thus substituting for the transmembrane and periplasmic domains [7]. Prior to their analysis the protein solutions were dialysed to remove dithiothreitol and phenylmethanesulfonyl fluoride from the reaction mix- tures. The phosphorylation assays were carried out as described previously using [c- 32 P]ATP (specific activity > 5000 CiÆmmol )1 ) [23,24]. The purified proteins were incubated at concentrations of 1 l M (BvgS and EvgS) or between 4.7 and 7.1 l M (BvgS-H643A, BvgS-Del, BvgS-C4) at room temperature in reaction buffer containing 50 m M Tris pH 7.5, 50 m M KCl, 10 m M MgCl 2 ,50 n M [c- 32 P]ATP, 0.2 m M ATP. Reactions were stopped by the addition of sample buffer (60 m M Tris pH 6.8, 10% glycerol, 2% SDS, 5% 2-mercaptoethanol, 0.05% Bromophenol blue). After separation of the phosphorylated proteins by SDS/PAGE [25] the gels were dried at 80 °C for 90 min and exposed for several hours at room temperature. Quantitative densitom- etry was performed using a PhosphorImager (Molecular Dynamics) according to standard protocols. Relative phos- phorylation values were calculated with respect to the strongest signal on a gel set as 100%. To investigate the effect of inhibitory compounds, the kinases were incubated for 2.5 min in the absence or presence of increasing concentrations of ubiquinone-0 (Q-0) (1 l M to 5 m M ), MK-3 (1 l M to 1 m M ), oxidized and reduced FAD and NAD (1 l M to 5 m M ), nicotinic acid (1 m M to 10 m M )orMgSO 4 (1 l M to 50 m M ). In the case of menadione, the reaction mixtures contained 15% (v/v) ethanol, because MK-3 was dissolved in ethanol as a stock solution. All compounds were obtained from Sigma. For the determination of the effect of Q-0 on the autophosphorylation kinetics of BvgS and EvgS the kinases were incubated for 15 s, 30 s, 1 min, 2 min or 4 min before the addition of sample buffer. Reaction mixtures contained 20 l M Q-0 and/or 500 l M dithionite (sodium hydrosulfite). For experiments investigating the reversibility of the quinone effect on BvgS, the kinase was first incubated in the presence of Q-0 (5 m M ). Then, after an additional 4.5 min, Q-0 was reduced by the addition of 25 m M dithionite. For characterization of the autodephosphorylation of BvgS, the kinase was first incubated for 5 min under standard conditions in the presence of [c- 32 P]ATP as described above. Then 5 m M Q-0 and/or 16 m M nonradio- active ATP were added. After the addition of Q-0 and/or ATP samples were taken immediately and after 0.5, 1, 2, 4, 8, 16 and 32 min; the reaction was stopped by the addition of sample buffer. The half-life of phosphorylated BvgS was determined as described previously [24]. All experiments were performed at least three times independently. Statistical analysis was performed using Microsoft EXCEL 97 and SIGMAPLOT for Windows, version 5.00. Site-directed mutagenesis The BvgS mutants were constructed using the Quik- Change TM site-directed mutagenesis kit (Stratagene). In the case of BvgS-C4 a previously described constitutive BvgS mutant was reconstructed [16]. In the case of BvgS-H and BvgS-Del the putative quinone binding site in the BvgS PAS domain was the target for mutagenesis [26]. The previously described plasmid pQE-BvgS-TRO [7] was used as a template and the following oligonucleotides were synthes- ized for mutagenesis.: For the construction of the mutant bvgS alleles the oligonucleotides bvgS-C4-a (5¢-GCTGGG CGAACTCAAGAGCATCATCGGCGG-3¢) and bvgS- C4-b (5¢-CCAGCCGCCGATGATGCTCTTGAGTTCG CC-3¢), bvgS-Qmut1 (5¢-GCCCGCGAAATGGCCGAG TTCCTGCTC-3¢) and bvgS-Qmut2 (5¢-GAGCAGGAAC TCGGCCATTTCGCGGGC-3¢) and bvgS-Del-a (5¢-GA CCCGGCGCTGACGCGCGTGGCC-3¢) and bvgS-Del-b (5¢-GGCCACGCGCGTCAGCGCCGGGTC-3¢)were used. All proteins used in this study are shown schematically in Fig. 1. Hpt N BvgS C RT PAS TM BvgS* C BvgS-H643A C H643A F583 BvgS-C4 C G688S BvgS-Del C ∆ 639-647 > Fig. 1. Structural organization of BvgS derivatives used in this study. On the top the domain structure of the wild-type BvgS protein is shown. BvgS* shows the organization of the soluble truncated BvgS derivative referred to as Ôcontrol proteinÕ throughout this study. The positions of the mutations introduced in the three BvgS derivatives BvgS-H643A, BvgS-Del and BvgS-C4 are shown. TM, transmem- brane domain; PAS, PAS domain; T, transmitter; R, receiver; Hpt, histidine-containing phosphotransfer domain. 3480 A. Bock and R. Gross (Eur. J. Biochem. 269) Ó FEBS 2002 RESULTS Effect of quinones on the activity of BvgS and EvgS In the genus Bordetella, ubiquinone-8 (Q-8) is the predom- inant quinone [27], whereas the major quinones synthesized by E.coliare Q-8, menaquinone-8 (MK-8) and demethyl- menaquinone-8 [28]. The incubation of the BvgS and EvgS histidine kinases with the soluble Q-8 analogue Q-0 resulted in a strong inhibition of autophosphorylation of both enzymes (Fig. 2). In fact, half maximal inhibition of BvgS and EvgS occurred at about 11 l M and 4 l M , respectively (Fig. 3). Reduction of the quinone by dithionite entirely abolished its negative effect on BvgS and EvgS (Fig. 2). Whereas the autophosphorylation activity of EvgS could be inhibited completely by the addition of  100 l M Q-0, BvgS activity could not be entirely abrogated by Q-0 reaching a minimal basal activity of  10% (Fig. 3). The soluble MK-8 analogue MK-3 did not affect the kinase activities of BvgS and EvgS, which contrasts with the results reported previously for the ArcB histidine kinase of E.coli (data not shown). The effect of Q-0 on the enzyme activities of the histidine kinases is probably reversible because autophos- phorylation of BvgS could be reactivated when oxidized Q-0 had been reduced by addition of dithionite (Fig. 4), although we can not entirely exclude a Q-0 independent action of dithionite on the autokinase activity of BvgS. Ubiquinone acts on the histidine kinase but not on intrinsic phosphatase activities The negative effect on autophosphorylation observed after addition of the oxidized ubiquinone to the histidine kinases may either be the consequence of a decrease in the histidine kinase activity, or, alternatively, of an increase in an intrinsic autophosphatase activity present in the BvgS and EvgS proteins [7,10]. To investigate these two alternatives we characterized the dephosphorylation kinetics of BvgS-P. The half-life of phosphorylated BvgS under the experimen- tal conditions was estimated to be  12.3 min. As shown in Fig. 5, the addition of oxidized Q-0 did not cause an increase in the dephosphorylation of BvgS-P, in contrast, a slight increase in the stability of the phosphoprotein could be noted. Moreover, no significant change in the stability of BvgA-P could be observed in the presence of BvgS and oxidized Q-0 (data not shown). These data suggest that in the case of BvgS probably its kinase activity but not its intrinsic phosphatases are affected by the presence of the quinone. Effect of FAD, NAD or modulating agents on the BvgS histidine kinase activity To investigate whether other electron carriers abundant in the cell have an impact on the activity of the BvgS and EvgS proteins, we added up to 5 m M FAD or 5 m M NAD to the reaction mixtures. FAD was shown previously to bind to the PAS domain of the membrane signal transducer protein Aer of E.coli[29]. However, using physiologically relevant concentrations neither coenzyme had any significant effect on the autophosphorylation activity of the two histidine kinases independent of their oxidation status (data not shown). Moreover, nicotinic acid and sulfate ions (which have long been known to inhibit BvgS and EvgS) at in vivo relevant concentrations did not influence the autophospho- rylation properties of the purified truncated BvgS and EvgS kinases in vitro (data not shown). Effects of mutations in the BvgS PAS domain on quinone sensing Similar to the ArcB histidine kinase of E.coli, the presence of a PAS domain was recently predicted in front of the transmitter domain of the BvgS protein [13]. Similarly, the EvgS protein contains a sequence with characteristics of a PAS domain between its transmembrane and transmitter domains (data not shown). Previously, this part of the BvgS protein was termed linker and its relevance for BvgS function was recognized because mutations in this region were isolated which caused either an inactivation of the protein time (min) 01234 relative level BvgS-P 0 20 40 60 80 100 120 14 0 BvgS BvgS +D t BvgS +Q-0 BvgS +Q-0 +D t +Q-0 +Dt +Q-0 +Dt none A 0.25 0.50 1 2 4 (min) +Q-0 +Dt +Q-0 +Dt none 0.25 0.50 1 2 4 (min) time (min) 01234 relative level EvgS-P 0 20 40 60 80 100 120 140 EvgS EvgS +D t EvgS +Q-0 EvgS +Q-0 +D t B Fig. 2. Effect of (Q-0) on the rate of auto- phosphorylation of the BvgS (A) and EvgS (B) histidine kinases. The purified proteins were incubated with [c- 32 P]ATP in the presence or absence of Q-0 and/or dithionite. On the bottom panel autoradiograms of representa- tive SDS/PAGE are shown. In the top panel, the increase with time of BvgS-P and EvgS-P in the presence or absence of Q-0 and/or dithionite is shown. For this presentation the relative amount of the phosphorylated histi- dine kinases was determined by Phosphor- Image analysis of the respective gels shown in the bottom panel. The maximal phosphory- lation levels of BvgS and EvgS in the absence of Q-0 and/or dithionite were taken as 100% which correspond to P/protein values of 28.5% and 15.5%, respectively. Ó FEBS 2002 Quinone sensing by histidine kinases (Eur. J. Biochem. 269) 3481 (mutations in the PAS core domain) or a lack of response to modulating compounds such as temperature, nicotinic acid or sulfate ions (mutations in the b-scaffold domain or the helical connector) [14–17]. Interestingly, in the so-called helical connector of the BvgS PAS domain a sequence with similarities to a recently described quinone binding motif is present consisting of a weak structural element of a triad of residues [aliphatic-(X) 3 -H-(X) 2,3 -(L/T/S)] with the central histidine residue on one side of an a-helical stretch [26]. To investigate the functional relevance of the PAS domain for quinone sensing we generated three BvgS derivatives by site- directed mutagenesis carrying mutations either in the putative quinone binding motif (BvgS-H643A, BvgS-Del) or containing a previously described mutation in the b-scaffold part of the PAS domain (BvgS-C4) (Fig. 1). The latter mutation was found after a screen for B. pertussis mutants exhibiting a ÔconstitutiveÕ phenotype, i.e. not responding anymore to modulating compounds in vivo [16]. However, all three mutations caused only mild effects on quinone sensing (Fig. 3). The most prominent effect was observed with the mutant BvgS-Del protein lacking the entire putative quinone binding site. Even at Q-0 concen- trations > 100 l M the autophosphorylation activity of BvgS-Del could only be inhibited to  20% of its maximal activity, whereas, for comparison, the activity of the control protein BvgS decreased to  10% under the same conditions (Fig. 3). DISCUSSION The data presented here provide strong evidence for a direct perception of redox signals by the BvgAS and EvgAS two- component signal transduction systems, as both sensory histidine kinases strongly respond to the presence of oxidized but not reduced ubiquinone (Q-0). The fact that menadione had no effect on BvgS is not surprising, because in contrast with the situation in E.coliMK-8 does not play aroleinB. pertussis [27]. However, in E.coli ArcB time (min) 0 5 10 15 20 25 30 35 relative level BvgS-P 20 30 40 50 60 70 80 90 100 BvgS +Q0 BvgS -Q0 0 0.5 1 2 4 816 32 (min) 0 0.5 1 2 4 816 32 (min) +Q-0 -Q-0 Fig. 5. Autodephosphorylation of BvgS in the presence or absence of Q-0. BvgS was phosphorylated by the addition of [c- 32 P]ATP. After 5 min Q-0 (final concentration 5 m M ) and/or a 75-fold excess of nonradioactive ATP was added (time point 0 in the figure). At the indicated time points samples were taken and analysed. On the x-axis time is shown in minutes; on the y-axis the log of the relative phosphorylation level at the various time points is shown. On the right the autoradiographies of representative gels are shown. 0 1 2.5 5 10 20 50 100 200 (µM Q-0) BvgS-H643A BvgS-Del BvgS-C4 BvgS EvgS Q-0 [µM] 0 50 100 150 200 relative phosphorylation (%) 0 20 40 60 80 100 120 BvgS BvgS-H643A BvgS-Del BvgS-C4 EvgS Fig. 3. Inhibition of autophosphorylation of BvgS, EvgS and of mutated BvgS derivatives (BvgS-H643A, BvgS-Del, BvgS-C4) in the presence of different Q-0 concentrations. After SDS/PAGE analysis the relative amount of the phosphorylated histidine kinases was determined by Phos- phorImage analysis. The figure shows representative autoradiographs of the samples after SDS/PAGE (right panel) and the results of their quantification by the PhosphorImager (left panel). time (min) 0 1020304 00 5 relative level BvgS-P 0 20 40 60 80 100 addition of Dt 2 4 6.5 8.5 10.5 16 31 51 (min) BvgS Fig. 4. Dependence of the kinase activity of BvgS on the redox state of ubiquinone-0 (Q-0). BvgS was incubated with [c- 32 P]ATP in the pres- ence of 5 m M Q-0 for 4.5 min Then dithionite was added to a final concentration of 25 m M . The insert shows the autoradiography of a representative gel. 3482 A. Bock and R. Gross (Eur. J. Biochem. 269) Ó FEBS 2002 autophosphorylation was inhibited by both oxidized qui- nones [22], whereas EvgS was found to be responsive only to Q-0 but not to MK-3. This differential behaviour of the two histidine kinases of E.colimay indicate interesting differ- ences in the fine tuning of their activity under different growth conditions. For example, it is known that high aeration and logarithmic growth of the bacteria leads to a predominance of Q-8, whereas under anaerobic conditions and in stationary phase MK-8 is the predominant quinone [30]. In contrast with the ArcB protein, the BvgS and EvgS proteins contain extensive periplasmic domains. This indi- cates that via these domains both proteins may also be engaged in the perception of additional stimuli possibly derived from the environment such as different modulating agents (e.g. sulfate ions and nicotinic acid). In fact, the data presented here demonstrate that such compounds do not interfere directly with the cytoplasmic portion of the BvgS histidine kinase. As already pointed out by Georgellis and coworkers [22], the quinones are membrane localized electron transporters which are interesting candidates as redox indicators for sensor proteins such as ArcB, BvgS and EvgS. As the quinones are the only components of the respiratory chain which apparently are free in their movement within the membrane, they may easily come into close contact with membrane anchored sensor proteins. It is not yet known which parts of the senor proteins interact with the quinones, although this interaction must involve the cytoplasmic domains present in the truncated proteins used for these studies. The PAS domains, which have previously been proposed to be localized in the ArcB and BvgS linker regions close to the transmembrane domains, appear to be suitable devices for such quinone interaction sites [13]. PAS domains are known to monitor changes in redox potential, oxygen, the overall level of energy of a cell, and it was recently shown that they can bind small ligands such as FAD and ATP [13,29,31]. In fact, mutations in this domain have been reported which cause either the inactivation of the BvgS protein or its constitutive activity under modulating conditions induced by temperature, nicotinic acid or sulfate ions [14–17]. Interestingly, the BvgS PAS domain contains a sequence motif that comprises essential features of a recently described quinone binding motif [26]. However, mutations in this motif or in the b-scaffold domain caused only relatively mild effects on quinone sensing by BvgS. This indicates that the sequence motifs described here may contribute to quinone sensing but there may be additional quinone interaction sites. In fact, in the case of BvgS a second putative quinone binding motif is predicted in the transmitter domain centred around the autophosphoryla- tion site at His729 (data not shown). Interestingly, the ArcB orthologue of Haemophilus influenzae lacks a PAS domain, but nevertheless can complement an E.coliarcBmutant to the wild-type phenotype [32]. Therefore, the function of the PAS domains of these proteins in quinone sensing remains obscure. It is interesting to note that the PAS domain of the phosphorelay histidine kinase A of Bacillus subtilis was recently shown to be a catalytic ATP-binding domain [31]. As the PAS domain of BvgS contains a putative ATP binding site and mutations in this motif were previously shown to inactivate the protein [14] it is likely that there are alternative intracellular signals which are perceived by BvgS via its PAS domain. It is not known yet which stimuli are relevant for BvgS mediated regulation of virulence genes during infection [3,4]. Due to the lack of relevant changes of the virulence properties of bvgS mutants not responsive anymore to modulating compounds such as nicotinic acid and sulfate ions in infection experiments it was recently suggested that BvgS-mediated modulation of the virulence regulon may not be relevant at least in the respective animal models [33]. Because such a ÔconstitutiveÕ BvgS derivative (BvgS-C4) insensitive to modulating compounds is shown here to exhibit biochemical properties very similar to those of the control protein BvgS with regard to its responsiveness to oxidized quinones, it may well be that in the previously reported infection experiments modulation of BvgS activity by relevant intracellular stimuli still occurred. Therefore, in the future for ÔconstitutiveÕ bvgS mutants that are not responsive to quinones it will be necessary to investigate whether modulation of BvgS activity is relevant for the infectious process. Quinone sensing by BvgS brings our attention to the BvgAS regulated housekeeping genes, the characterization of which has been largely neglected as compared with that of the virulence genes [19]. In fact, sensing of redox signals by the BvgS protein provides an interesting link to the previously reported observations that cytochrome d and c oxidases are regulated in a reciprocal manner by the BvgAS system [4,18] and may have important implications for the scenario of virulence evolution in the genus Bordetella. The strong link of the master regulator of virulence to basic metabolism of the bacteria suggests that this system may have already been used by an ancestor of the pathogenic bacteria. Virulence genes acquired later in evolution may have been connected with the BvgAS regulon because their expression may be required under the same conditions which also lead to expression of cytochrome d oxidase and repression of cytochrome c oxidase. In this respect it is interesting to note that a recently isolated environmental organism probably closely related to a common ancestor of the pathogenic Borde- tellae, Bordetella petrii, is the first member of the genus with a facultatively anaerobic metabolism [27]. In con- trast, the pathogenic Bordetellae are characterized by an obligate aerobic metabolism with their growth opti- mum under microaerophilic conditions which may be a consequence of their permanent association with host organisms. ACKNOWLEDGEMENTS We would like to thank V. Weiss for discussions and D. Beier and J. Gross for carefully reading of the manuscript. This work was supported by a grant from the Deutsche Forschungsgemeinschaft as part of the priority program ÔRegulatory networks in bacteriaÕ and by the Fonds der Chemischen Industrie. REFERENCES 1. Loomis, W.F., Shaulsky, G. & Wang, N. (1997) Histidine kinases in signal transduction pathways of eukaryotes. J. Cell Sci. 110, 1141–1145. 2. Perraud, A L., Weiss, V. & Gross, R. (1999) Signalling pathways in two-component phosphorelay systems. Trends Microbiol. 7, 115–120. Ó FEBS 2002 Quinone sensing by histidine kinases (Eur. J. Biochem. 269) 3483 3. Bock, A. & Gross, R. (2001) The BvgAS two-component system of Bordetella spp. a versatile modulator of virulence gene expres- sion. Int. J. Med. Microbiol. 291, 119–130. 4. Cotter, P.A. & DiRita, V.J. (2000) Bacterial virulence gene regulation: an evolutionary perspective. Ann. Rev. Microbiol. 54, 519–565. 5. Nishino, K. & Yamaguchi, A. (2001) Overexpression of the response regulator EvgA of the two-component signal transduc- tion system modulates multidrug resistance conferred by multi- drug resistance transporters. J. Bacteriol. 183, 1455–1458. 6. Beier, D., Schwarz, B., Fuchs, T.M. & Gross, R. (1995) In vivo characterization of the unorthodox BvgS two-component sensor protein of Bordetella pertussis. J. Mol. Biol. 248, 596–610. 7. Perraud, A L., Kimmel, B., Weiss, V. & Gross, R. (1998) Speci- ficity of the BvgAS and EvgAS phosphorelay systems is mediated by the C-terminal HPt domains of the sensor proteins. Mol. Microbiol. 27, 875–887. 8. Uhl, M.A. & Miller, J.F. (1994) Autophosphorylation and phos- photransfer in the Bordetella pertussis BvgAS signal transduction cascade. Proc. Natl Acad. Sci. USA 91, 1163–1167. 9. Utsumi, R., Katayama, S., Taniguchi, M., Horie, T., Ikeda, M., Igaki, S., Nakagawa, H., Miwa, A. & Noda, M. (1994) Newly identified genes involved in the signal transduction of Escherichia coli K-12. Gene 140, 73–77. 10. Uhl, M.A. & Miller, J.F. (1996) Central role of the BvgS receiver as a phosphorylated intermediate in a complex two-component phosphorelay. J. Biol. Chem. 271, 33176–33180. 11. Weiss, A.A. & Falkow, S. (1984) Genetic analysis of phase change in Bordetella pertussis. Infect. Immun. 43, 263–269. 12. Melton, A.R. & Weiss, A.A. (1993) Characterization of environ- mental regulators of Bordetella pertussis. Infect. Immun. 61, 807–815. 13. Taylor, B.L. & Zhulin, I.B. (1999) PAS domains: internal sensors of oxygen, redox potential, and light. Microbiol. Mol. Biol. Rev. 63, 479–506. 14. Beier, D., Deppisch, H. & Gross, R. (1996) Conserved sequence motifs in the unorthodox BvgS two-component sensor protein of Bordetella pertussis. Mol. Gen. Genet. 252, 169–176. 15. Manetti, R., Arico, B., Rappuoli, R. & Scarlato, V. (1994) Mutations in the linker region of BvgS abolish response to environmental signals for the regulation of the virulence factors in Bordetella pertussis. Gene 150, 123–127. 16. Miller,J.F.,Johnson,S.A.,Black,W.J.,Beattie,D.T.,Mekalanos, J.J. & Falkow, S. (1992) Constitutive sensory transduction mutations in the Bordetella pertussis bvgS gene. J. Bacteriol. 174, 970–979. 17. Goyard, S., Mireau, H. & Ullmann, A. (1995) Mutations which result in constitutive expression of the Bordetella pertussis filamentous haemagglutinin gene. Res. Microbiol. 146,363– 370. 18. Ezzell, J.W., Dobrogosz, W.J., Kloos, W.E. & Manclark, C.R. (1981) Phase-shift markers in the genus Bordetella: loss of cyto- chrome d-629 in phase IV variants. Microbios 31, 171–181. 19. Schneider, B., Stu ¨ bs, D. & Gross, R. (2002) Identification and genomic organization of gene loci negatively controlled by the virulence regulatory BvgAS two-component system in Bordetella bronchiseptica. Mol. Gen. Genom. 267, 526–535. 20. Sawers, G. (1999) The aerobic/anaerobic interface. Curr. Opin. Microbiol. 2, 181–187. 21. Unden, G. & Bongaerts, J. (1997) Alternative respiratory path- ways of Escherichia coli: energetics and transcriptional regulation in response to electron acceptors. Biochim. Biophys. Acta 1320, 217–234. 22. Georgellis, D., Kwon, O. & Lin, E.C.C. (2001) Quinones as the redox signal for the arc two-component system of bacteria. Science 292, 2314–2316. 23. Bock,A.,Bantscheff,M.,Perraud,A L.,Rippe,K.,Weiss,V., Glocker, M.O. & Gross, R. (2001) Rational design and molecular characterization of a chimaeric response regulator protein. J. Mol. Biol. 310, 283–290. 24. McCleary, W.R. (1996) The activation of PhoB by acetylphos- phate. Mol. Microbiol. 20, 1155–1163. 25. Laemmli, U.K. (1970) Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature 227, 9441– 9456. 26. Fisher, N. & Rich, P.R. (2000) A motif for quinone binding sites in respiratory and photosynthetic systems. J. Mol. Biol. 296, 1153–1162. 27. von Wintzingerode, F., Schattke, A., Siddiqui, R.A., Ro ¨ sick, U., Go ¨ bel, U.B. & Gross, R. (2001) Bordetella petrii sp. nov., isolated from an anaerobic bioreactor, and emended description of the genus Bordetella. Int. J. Syst. Evol. Microbiol. 51, 1257–1265. 28. Meganathan, R. (1996) Biosynthesis of the isoprenoid quinones menaquinone (vitamin K 2 ) and ubiquinone (coenzyme Q). In Escherichia coli and Salmonella typhimurium (Neidhardt, C.F., ed), pp. 642–656. ASM Press, Washington DC. 29. Bibikov, S.I., Barnes, L.A., Gitin, Y. & Parkinson, J.S. (2000) Domain organization and flavin adenine dinucleotide-binding determinants in the aerotaxis signal transducer Aer of Escherichia coli. Proc. Natl Acad. Sci. USA 97, 5830–5835. 30. Gennis, R.B. & Stewart, V. (1996) Respiration. In Escherichia Coli and Salmonella Typhimurium (Neidhardt, C.F., eds), pp. 217–261. ASM Press, Washington DC. 31. Stephenson, K. & Hoch, J.A. (2001) PAS-A domain of phos- phorelay sensor kinase A: a catalytic ATP-binding domain in- volved in the initiation of development in Bacillus subtilis. Proc. Natl Acad. Sci. USA 98, 15251–15256. 32. Georgellis, D., Kwon, O., Lin, E.C., Wong, S.M. & Akerley, B.J. (2001) Redox signal transduction by the ArcB sensor kinase of Haemophilus influenzae lacking the PAS domain. J. Bacteriol. 183, 7206–7212. 33. Martinez de Tejada, G., Cotter, P.A., Heininger, U., Camilli, A., Akerley, B.J., Mekalanos, J.J. & Miller, J.F. (1998) Neither the Bvg- phase nor the vrg6 locus of Bordetella pertussis is required for respiratory infection in mice. Infect. Immun. 66, 2762–2768. 3484 A. Bock and R. Gross (Eur. J. Biochem. 269) Ó FEBS 2002 . The unorthodox histidine kinases BvgS and EvgS are responsive to the oxidation status of a quinone electron carrier Andreas Bock and Roy Gross Lehrstuhl. signal for the BvgS and EvgS histidine kinases. MATERIALS AND METHODS Phosphorylation assays Over-expression and purification of the histidine kinases BvgS