Differential Regulation of the Bordetella bipA Gene: Distinct Roles

JOURNAL OF BACTERIOLOGY, Dec. 2002, p. 6942–6951 Vol. 184, No. 240021-9193/02/$04.00�0 DOI: 10.1128/JB.184.24.6942–6951.2002Copyright © 2002, American Society for Microbiology. All Rights Reserved.

Differential Regulation of the Bordetella bipA Gene: Distinct Roles forDifferent BvgA Binding Sites

Rajendar Deora*

Department of Microbiology, Immunology and Molecular Genetics, David Getten University of California—Los Angeles, School ofMedicine, Los Angeles, California 90095-1747

Received 15 January 2002/Accepted 18 March 2002

The BvgAS signal transduction system of Bordetella controls an entire spectrum of gene expression states inresponse to differences in environmental conditions. In particular, the Bordetella Bvg-intermediate-phase genebipA displays a complex regulatory pattern in response to various concentrations of modulators. Expression ofbipA is low in the absence of modulating signals, maximal at intermediate concentrations of modulators, andnear background levels at high concentrations of modulators. bipA is regulated at the transcriptional level, andthe bipA promoter contains multiple BvgA binding sites present both upstream and downstream of thetranscriptional initiation site. In vivo transcriptional analyses, utilizing several mutant promoter fusions to thereporter enzyme �-galactosidase, suggest that the upstream binding site IR1 is essential for expression andthat the downstream binding sites IR2 and IR3 are involved in transcriptional repression. Mutations of IR2or IR3 convert the expression profile of bipA from that of a Bvg-intermediate-specific-phase gene to that of aBvg�-phase gene. To gain insight into the mechanism responsible for differential bipA regulation, DNase Iprotection studies were conducted with various mutant promoters. These analyses suggest that IR1 and IR2function as core binding sites and are the primary determinants for the phosphorylation-induced oligomer-ization of BvgA to the adjacent regions.

The survival and replication of bacteria in a particular nichewithin a host or in a constantly changing environment outsidethe host require continuous monitoring of surrounding condi-tions and the ability to generate an adaptive response. In gen-eral, bacteria respond to environmental cues by controllinggene expression, resulting in the upregulation and/or down-regulation of appropriate genes. For most bacteria, a majormechanism for eliciting adaptive responses to subtle changes intheir environment is a signal transduction cascade, mediatedby the so-called two-component systems. In their simplest formthese systems are comprised of two regulatory proteins, a sen-sor kinase and a transcriptional activator/repressor, also knownas the response regulator (22, 37). The sensor is an autophos-phorylating protein kinase, and the response regulator under-goes a phosphorylation-induced conformational change, therebyeliciting a specific response. Two-component systems are alsofound in some lower eukaryotes, reflecting the diversity of thismethod of regulating gene expression (29).

The Bordetella BvgAS locus encodes a two-component sys-tem that regulates the expression of nearly all of the knownvirulence factors synthesized by these gram-negative respira-tory pathogens and plays an important role in their survivalstrategy. The bvgAS loci of Bordetella pertussis, Bordetella para-pertussis, and Bordetella bronchiseptica are 96% identical at thenucleotide level (6) and are functionally interchangeable dur-ing infection (32). BvgA and BvgS are members of a class oftwo-component systems that communicate via a four-step His-Asp-His-Asp phosphorelay (22, 46). BvgA is a DNA-bindingresponse regulator, and BvgS is a transmembrane sensor pro-tein. Using the �-phosphoryl group of ATP, BvgS is able to

autophosphorylate and then transphosphorylates BvgA (10,46–47). Phosphorylated BvgA has an increased affinity for Bvg-activated promoters and activates transcription of a variety ofgenes (11–12, 23–25, 30, 38–40, 43, 49).

In response to environmental signals, Bordetella can switchbetween distinct phenotypic phases (27). In an extensive anal-ysis, Lacey showed that members of the Bordetella species canalternate between three distinct phenotypic modes, designatedX, I, and C, in response to changes in temperature or in theconcentrations of specific ions in the culture medium (27).Results from several studies have now clearly demonstratedthat the BvgAS signal transduction system controls the transi-tion between at least three distinct phenotypic phases, Bvg�,Bvgi, and Bvg� (for a review, see reference 34). The Bvg�

phase appears to correspond to Lacey’s X mode, and the Bvg�

phase apparently corresponds to the C mode. When BvgAS isactive, Bordetella organisms are in the Bvg� phase and expressa variety of putative adhesins and toxins (34), such as filamen-tous hemagglutinin, fimbriae (Fim2 and Fim3), pertactin(Prn), adenylate cyclase toxin (CyaA) (28), and dermonecrotictoxin (Dnt) (34). The other potential virulence determinantsinclude secreted factors of the type III secretion system (48),tracheal colonization factor (TcfA) (20), and a serum resis-tance factor, BrkAB (18). Inactivation of BvgAS by mutationor the presence of modulating signals transforms Bordetella tothe Bvg� or avirulent phase. The Bvg� phase is characterizedby expression of motility and other coregulated phenotypes inB. bronchiseptica (2, 3). In contrast, B. pertussis and B. para-pertussis strains are nonmotile although these strains containthe motility genes. The Bvg� phase in B. pertussis is character-ized by the expression of several outer membrane proteins ofunknown function encoded by the vrg loci (7, 26, 35, 44).Experiments with phase-locked and ectopic expression mu-

* Mailing address: 715 Gayley Ave. 309, Los Angeles, CA 90024.Phone: (310) 208-8205. E-mail: [email protected].

6942

on April 14, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

tants demonstrated that Bvg� phase is necessary and sufficientfor respiratory tract colonization by B. pertussis and B. bron-chiseptica (4, 13, 31).

Isolation of a mutant displaying phenotypes intermediatebetween those characteristic of the Bvg� and Bvg� phaseschanged the traditional view of the bvg regulon (14). Com-pared to its Bvg� phase-locked parent, this mutant (designatedBvgi for Bvg intermediate) displayed increased resistance tonutrient limitation and a decreased ability to colonize the re-spiratory tract. Molecular analyses indicated that the Bvgi mu-tant had lost the ability to express a subset of Bvg� phasefactors and expressed factors unique to the Bvgi phase. Inter-estingly, the phenotype of this particular mutant appears to besimilar to that of the I mode, described by Lacey (14, 27). It isnow becoming increasingly clear that as opposed to mediatinga biphasic transition, the Bvg regulatory system appears tocontrol an entire spectrum of distinct states of gene expression.

The discovery of Bvgi-phase-specific antigens predicted theexistence of a class of genes that are maximally expressed in theBvgi phase. That such a class exists was confirmed by theidentification of bipA, a Bvg-regulated gene that is maximallyexpressed in the intermediate phase, induced by either muta-tion (bvgS-I1) or growth under semimodulating conditions (16,45). BipA protein shares sequence similarity with intimin ofenteropathogenic and enterohemorrhagic Escherichia coli andinvasin of Yersinia spp. Comparison of the wild-type B. bron-chiseptica strain and the �bipA strain did not reveal a require-ment for bipA in colonization of the respiratory tract (45). It ispossible that BipA performs important functions in the Borde-tella infectious cycle other than colonization, such as in aerosoltransmission.

Analyses of DNA binding demonstrated that BvgA specifi-cally binds to a number of sites on the bipA promoter with a

hierarchy of binding affinities. Three of these sites (IR1, HS1,and HS2) are present upstream of the transcription initiationsite, whereas two of the sites (IR2 and IR3) are present down-stream (Fig. 1). The binding sites also vary with regard to therequirement for the phosphorylation state of BvgA. Whereasnonphosphorylated BvgA is able to bind to IR1 and HS2, onlyBvgA-P binds to IR2, IR3, and HS1 (Fig. 1). It was hypothe-sized that the unique expression pattern of the bipA gene is aconsequence of the concentration-dependent differential occu-pancy of BvgA-P at the various binding sites (16).

To gain insight into the complex regulation of bipA geneexpression, I have examined the nature and relative contribu-tions of the different BvgA-binding sites to the Bvgi-phase-specific expression pattern of the bipA gene. Mutational anal-yses of the different BvgA binding sites in the bipA promotershow that the upstream binding site IR1 is essential for acti-vation, and the downstream binding sites IR2 and IR3 areinvolved in efficient repression of transcription. Further, it isshown that the complex signal-dependent response profile ofthe bipA gene can be altered by changes in the promoterregion.

MATERIALS AND METHODS

Bacterial strains and plasmids. The bacterial strains and plasmids used in thisstudy are listed in Table 1.

PCR and oligonucleotide primers. The conditions for PCRs were as describedpreviously (16). Pfu DNA polymerase (Stratagene) was used for all PCRs. Se-quences of the oligonucleotide primers are listed in Table 2.

Media and growth conditions. All Bordetella strains were grown on BG agar(Becton Dickinson Microbiology Systems) supplemented with 7.5% sheep bloodand incubated at 37°C for 48 h. For RNA extraction and �-galactosidase assays,cells were grown in SS broth (42) and harvested at mid-exponential phase. E. colistrains were grown in Luria-Bertani medium at 37°C with shaking at 250 rpm. Asneeded, the culture media were supplemented with ampicillin (100 �g/ml),

FIG. 1. Summary of the occupancy of BvgA and BvgA-P at the different binding sites present on the bipA promoter. The arrangement andboundaries of the BvgA binding sites relative to the transcriptional start site (�1) are depicted. Sequences of the different binding sites are shown.N represents one of the four nucleotides, and the subscript digits denote the number of nucleotides that separate the half-sites of IR2 and IR3.Schematics of the DNase I footprinting patterns of BvgA and BvgA-P as determined previously (16) and in Fig. 4A are indicated by bars. Therelative thicknesses of the bars represent the relative affinities of binding to the different sites.

VOL. 184, 2002 REGULATION OF BORDETELLA bipA GENE 6943


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

gentamicin (40 �g/ml), kanamycin (40 �g/ml), chloramphenicol (40 �g/ml), andstreptomycin (40 �g/ml).

Construction of the promoter deletion strain. The promoter deletion strainRKD 100 was constructed as follows. An XbaI-HindIII fragment containingsequences 5� to the bipA promoter, spanning �296 to �85 relative to thetranscription start site, was amplified from the chromosome of RB50 usingprimers 351-5 and 353-3. A second HindIII-KpnI fragment containing sequences3� to the promoter, from �125 to �519, was also amplified using primers 352-5and 352-3. The XbaI-HindIII-digested fragment was ligated with the HindIII-KpnI-digested fragment and the XbaI-KpnI-digested allelic exchange vectorpRE118 (Kmr) (17) in a three-way ligation resulting in plasmid pRD588. Thisplasmid was transformed into DH5�pir (M. Liu, unpublished data). DH5�piris a derivative of the E. coli DH5� strain that allows replication of plasmidscontaining the conditional R6Kori. Plasmid pRD588 was mobilized fromDH5�pir into B. bronchiseptica using the plasmid pRK2013 by triparentalmating (19). After conjugation, cointegrants were selected on BG blood agarplates containing kanamycin and streptomycin. Colonies arising from second

recombination events were selected on Luria-Bertani agar containing 7.5% su-crose as described previously (17). The genotype of RKD100 was confirmed byPCR and primer extension assays.

Construction of �IR2-3 strain. RKD 101 containing deletion of the bindingsites IR2 and IR3 was constructed using the strategy described above. A 5�XbaI-HindIII fragment was amplified using primers 351-5 and 351-3, spanning aregion from �296 to �11 of the bipA promoter. The 3� PCR fragment was thesame HindIII-KpnI fragment as described above, spanning a region from �125to �519. The two fragments were cloned as described above in the KpnI-XbaI-digested pRE118, resulting in pRD578. pRD578 was used to delete the regionfrom �12 to �122, which includes IR2 and IR3, from the chromosome of RB50.The protocol for selection of the first and second recombination events was sameas described above. The genotype of RKD101 was confirmed by both PCR andprimer extension.

Mutagenesis of the bipA promoter. The Quick-change site-directed mutagen-esis kit (Stratagene) was utilized to delete IR2, following the manufacturer’sprotocol. The mutagenic oligonucleotide primers 341-5 and 341-3 were used,

TABLE 1. Bacterial strains and plasmids

Strain or Plasmid Relevant feature(s) Source or reference

DH5� E. coli strain, high-efficiency transformation 36RB50 Wild-type B. bronchiseptica isolate 14RB53i bvgS-C3, bvgS-I1, Bvgi phase locked derivative 14RB54 �bvgS, Bvg� phase locked derivative 14RKD100 Derivative of RB50, containing genomic deletion of bipA promoter from �84 to �124 This studyRKD101 Derivative of RB50, containing deletion of IR2 and IR3 This studypcDNA3 Cloning vector; Ampr InvitrogenpEGZ Transcriptional fusion vector 32pEGZ421 bipA promoter fragment cloned into EcoRI-BamHI site of pEGZ 16pRE118 Allelic exchange vector; Kmr 17pRK2013 Mobilization helper plasmid 19pGMT18 prn-lacZ fusion vector 32pEG112 frl-lacZ fusion vector 32pRD303 Contains DNA sequence spanning the promoter region and the partial open reading frame of the

bipA gene16

pRD403 �120 to �357 region of the bipA promoter cloned into EcoRI-BamHI site of pcDNA3.0 16pRD404 �60 to �357 region of the bipA promoter cloned into EcoRI-BamHI site of pcDNA3.0 This studypRD408 �60 to �271 region of the bipA promoter cloned into EcoRI-BamHI site of pcDNA3.0 This studypRD425 �121 to �66 region of the bipA promoter lacking IR3 cloned into EcoRI-BamHI site of pcDNA3.0 This studypRD555 �121 to �197 region of the bipA promoter cloned into EcoRI-BamHI site of pUC19spf This studypRD572 pcDNA3.0 derivative containing a �120 to �197 region of bipA promoter having deletion of the

IR2 binding siteThis study

pRD578 pRE118 derivative, IR2-IR3 deletion plasmid This studypRD588 pRE118 derivative, bipA promoter deletion plasmid This studypRD590 bipA promoter assay vector This studypRD593 pRD590 derived; �120 to �143 region of the bipA promoter, wt-bipA-lacZ This studypRD597 pRD590 derived; �120 to �143 region of the bipA promoter containing deletion of IR2, �IR2-lacZ This studypRD610 pRD590 derived; �60 �143 region of the bipA promoter, �IR1-lacZ This studypRD622 �121 to �66 region of the bipA promoter, �IR3-lacZ This study

TABLE 2. Oligonucleotide primers

Designation Sequence and/or reference

BipA101 16BipA104 16BipAExt Oligonucleotide for primer extension analyses of bipA; 16301-11 5�-CGCGGATCCAAGGAGTGGTCGAAAAAGGCG-3�305-5 5�-CCGGAATTCAGATACATGGTATTTGGCGAC-3�341-5 5�-CCTCCTGTGTTCGGAGATATGTCACTCCTTTATCCAAATCGA-3�341-3 5� TCGATTTGGATAAAGGAGTGACATATCTCCGAACACAGGAGG-3�351-5 5�-CTAGTTCTAGAGCTTGTTCGAATGTGTTGCCGGCATTG-3�351-3 5�-CCCAAGCTTCTCCGAACACAGGAGGGAACTTTATCGA-3�352-5 5�-CCCAAGCTTAGCGGGCAGTGCGCCGAACATTCAG-3�352-3 5�-CGGGGTACCGCGCTTCACGCTTGAGATAGTC-3�353-3 5�-CCCAAGCTTCCTCACCTACAACCAGGGCAAAGG-3�401-5 5�-CCCGCAATTGCTTGTTCGAGGAATGTGTTGCCGGCAT-3�401-3 5�-CCGGAATTCGGCTGGATCACATCAATCGGGCAC-3�402-3 5�-CGCGGATCCGTTCGGCGCACTGCCCGCTAA-3�

6944 DEORA J. BACTERIOL.


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

resulting in deletion of the region from �17 to �59 of the bipA promoter. Theplasmid pRD303 was used as the DNA template.

Construction of the bipA promoter assay vector. The suicide plasmid pEGZ(32) has unique EcoRI and BamHI sites upstream of the promoterless lacZ gene.An MfeI-EcoRI, 177-nucleotide-region, spanning �298 to �121, was amplifiedfrom the genome of RB50 using the primers 401-5 and 401-3 and cloned into theEcoRI site of pEGZ. Digestion with MfeI results in cohesive ends that arecompatible with EcoRI. The clones were checked by restriction digest, and aclone where the EcoRI-MfeI hybrid site is distal to the BamHI site was selected,resulting in the plasmid pRD590. Since the EcoRI-MfeI hybrid site cannot berecleaved by EcoRI, cloning of EcoRI-BamHI fragments in EcoRI-BamHI-digested pRD590 places the fragments upstream of the promoterless lacZ gene.These plasmid derivatives were then integrated into the Bordetella genome by asingle crossover at the bipA locus.

LacZ transcriptional fusions and �-galactosidase assays. Plasmids pGMT18,pEG112, and pEGZ421 were used for measuring the activities of the prn, frl, andbipA promoters, respectively (16, 32). Different bipA promoter derivatives wereamplified using the primers BipA101 and 402-3. These primers amplify a regionspanning �120 to �143 of the bipA promoter. For amplification of the wild-typebipA promoter, �IR2, and �IR1, the plasmids pRD403, pRD572, and pRD404were used, respectively. For generation of the �IR1 derivative of the promoter,the primer 305-5 was used instead of the BipA101 primer, resulting in a fragmentspanning �60 to �143 of the bipA promoter. For the amplification of the �IR3derivative, the 3� primer 301-11 was used. The different promoter fragments werecloned into the EcoRI-BamHI site of pRD590, resulting in the correspondinglacZ fusion plasmids, pRD593 (wild type), pRD597 (�IR2), pRD610 (�IR1),and pRD622 (�IR3). Delivery of the different suicide plasmids into the wild-typeB. bronchiseptica strain RB50 or the bipA promoter deletion strain RKD100followed by selection for gentamicin resistance resulted in integration of theplasmids in the chromosome. �-galactosidase assays were performed as de-scribed previously (32).

RNA extraction, primer extension, and reverse transcriptase PCR (RT-PCR)analyses. Extraction of RNA from different Bordetella strains, primer extension,and RT-PCR analyses were performed as described previously (15, 16). Theoligonucleotide BipAExt was used for primer extension and DNA sequencing(16). The primers 352-5 and 352-3 were used for amplification of the bipA gene.Mock RT reactions (RT absent) did not result in any PCR product (data notshown).

DNase I footprinting assays. The plasmids pRD555 (wild type), pRD408(�IR1), pRD572 (�IR2), and pRD425 (�IR3) were utilized to generate thevarious promoter fragments by first digesting them with EcoRI and then endlabeling them with T4 polynucleotide kinase followed by digestion with BamHIto release the promoter fragments. DNase I footprinting assays were carried outas described previously (16).

RESULTS

A combination of sequence comparisons and DNA-proteininteraction analyses of the bipA promoter region revealed theconservation of three pairs of heptanucleotide inverted repeats(IR1, IR2, and IR3) that interact with BvgA and exhibit sim-ilarity to the consensus BvgA binding sequence TTTC(C/T)TA(Fig. 1) (16). This sequence has been previously shown bygenetic and biochemical means to be important for promoteractivity of a number of Bvg-regulated promoters (8, 11–12, 23,25). To characterize the role of the multiple BvgA-bindingsequences in expression of the bipA promoter, a series ofplasmid-borne bipA-lac transcriptional fusions was generatedin pRD590, a variant of the suicide plasmid pEGZ (32), whichcarries a fragment extending from �298 to �121 upstream ofthe bipA promoter. Different promoter variants were clonedupstream of the lac operon in pRD590 (Fig. 2) and thenintegrated in the B. bronchiseptica strain RKD100 by singlecrossover at the bipA locus. RKD100 is a derivative of the B.bronchiseptica strain RB50, containing deletions of the regionbetween �84 and �122 of the bipA promoter from the chro-mosome.

The bipA promoter contains both activation and repression

sites. Introduction of the wild-type bipA promoter fragment inRKD100, as described above, did not significantly alter tran-scriptional activity or the expression profile of the bipA-lacZfusion in response to modulating signals, compared to that ofthe bipA-lacZ fusion in wild-type B. bronchiseptica strain RB50(Fig. 2 and 5A and B) (16). Note that the transcriptionalactivities of the different promoter derivatives in RKD100 weremeasured in the absence of modulators—conditions that resultin high levels of expression from the bipA promoter (16). De-letion of the upstream binding site IR1 resulted in almostcomplete inhibition of the bipA promoter activity (Fig. 2,�IR1). These results suggest that IR1 is essential for expres-sion and activation of the bipA promoter. Deletion of thedownstream binding sites IR2 and IR3 led to a greater than20-fold and 4-fold stimulation in �-galactosidase activity, re-spectively (Fig. 2, �IR2 and �IR3), suggesting that IR2 andIR3 are required for repression of transcription from the bipApromoter. Transcriptional activity of the different bipA pro-moter derivatives in this context was clearly BvgA dependent,since growth of the different strains in presence of MgSO4, aknown modulator of BvgA activity, caused reduction of the�-galactosidase activity to levels comparable to that for a strainlacking the reporter fusion (Fig. 2). Thus, the above-describedresults strongly suggest that while IR1 is essential for expres-sion of the bipA promoter, IR2 and IR3 are involved in re-pression from the promoter.

Introduction into RKD100 of promoter derivatives lackingboth IR2 and IR3 resulted either in a failure to generatecointegrants or in lac fusions that were highly unstable in theabsence of MgSO4 (Fig. 2, �IR2-3). To circumvent this prob-lem, IR2 and IR3 were deleted from the chromosome of RB50using allelic exchange (RKD101) (Fig. 3A). The level of bipAtranscript was then measured using primer extension and RT-PCR assays.

Deletion of the downstream BvgA binding sites IR2 and IR3from the B. bronchiseptica chromosome results in a hyperacti-vated promoter. Primer extension assays with RNA preparedfrom RKD101 produced a considerably shorter fragment (Fig.3B), which corresponds to the transcription initiation site ofthe bipA promoter obtained when RNA was isolated fromwild-type B. bronchiseptica strain RB50 or an intermediatephase-locked strain, RB53i (Bvgi) (Fig. 3B). The bipA-specifictranscript from RKD101 is smaller because of a deletion of 110nucleotides from RKD101 of the transcribed region (Fig. 3A).The presence of more than one band for the bipA-specifictranscript from RKD101 (Fig 3B) can be explained by theprevious identification of an additional start site two nucleo-tides downstream of the �1 nucleotide by rapid amplificationof cDNA ends PCR analysis (16). As determined by primerextension assays, the amount of bipA-specific transcript fromRKD101 was approximately 17-fold higher than the corre-sponding transcript from RB53i (Bvgi). As shown in Fig. 3C,the intensity of the product from RB50 (Bvg�) was consider-ably lower than that of the transcript from RB53i (Bvgi). Thisresult is in agreement with previous measurements of the rel-ative levels of bipA promoter activity in Bvg� and Bvgi strainsusing a number of independent assays (16).

In order to confirm the results of primer extension analysis,the levels of bipA transcript in different strain backgroundswere determined using RT-PCR (Fig. 3D). The level of the



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

bipA transcript from RKD101 was higher than that from RB50and RB53i. As a control for RNA levels, PCR was performedwith primers specific to the Bvg-independent gene, recA. Nosignificant variation in the expression of recA was observed inthe different strain backgrounds (data not shown). Therefore,these results show that deletion of the two downstream bindingsites IR2 and IR3 leads to very high levels of transcription fromthe bipA promoter.

Binding of BvgA to different promoter derivatives. To de-termine the effect of binding sequence alterations on the in-teraction of BvgA with the bipA promoter, the DNase I foot-prints of several mutant promoters similar to those used in thein vivo transcriptional assays were compared with those of thewild-type promoter.

Wild type. Similar to previous results (16), both BvgA andBvgA-P protected a region from about �55 to about �75,which includes IR1 (Fig. 4A). The adjacent region from �55 to�35, encompassing the half-site HS2, displayed greater BvgAoccupancy with increasingly higher concentrations of nonphos-phorylated BvgA and significantly greater protection withBvgA-P. Higher concentrations of BvgA-P resulted in the pro-tection of additional regions spanning from �10 to �85 andfrom �80 to �105 (Fig. 4A, lanes 6 to 10). This region includesthe downstream BvgA binding sites IR2 and IR3 and theupstream half-site HS1.

�IR1. Deletion of the upstream binding site IR1 led to areduction in the binding of BvgA-P to a region downstream

from �45 to �30, encompassing the binding site HS2 (Fig. 4Aand B, compare the intensities of the bands denoted by aster-isks). A higher concentration of BvgA-P was required for pro-tection in this region, and even with the highest concentrationof BvgA-P used, the protection observed was less completethan with the wild-type promoter (Fig. 4A and B, comparelanes 10). Note that in contrast to BvgA-P, nonphosphorylatedBvgA seems to bind more effectively to the region �45 to �30(Fig. 4B, lanes 3 to 5). The significance of this observation isnot clear. No apparent reduction in binding affinity or theextent of the DNase I footprint of BvgA for the region from�5 to �85, encompassing the binding sites IR2 and IR3, wasobserved as a result of the deletion of IR1. Thus, while efficientbinding to regions surrounding HS2 requires the presence ofIR1, BvgA occupancy of IR2 and IR3 occur independent ofIR1.

�IR2 and �IR3. Footprint analysis of the �IR2 fragmentindicated no apparent binding of BvgA-P to the regions con-tiguous to IR2 (Fig. 4C, lanes 6 to 10). Even at the highestBvgA-P concentration used, there was no observable protec-tion in the region from �5 to �17 and that from �59 to �85(Fig. 4A and C, compare lanes 10). Thus, these results clearlydemonstrate that IR2 is required for the binding of BvgA-P tothe adjacent regions, including IR3. In contrast to deletion ofIR2, deletion of IR3 caused no significant changes in the bind-ing affinity or the protection pattern of BvgA-P. However, theregion from �105 to �95, containing the binding site HS1, was

FIG. 2. Effect of deletions and substitutions in the bipA promoter region on the transcriptional activity. A schematic of the bipA promoterregion with the different BvgA binding site is shown at the top. Immediately below is the genetic structure of RKD100. Regions deleted aredepicted by gaps. Below the RKD100 diagram on the left are shown the lac fusion constructs and, on the right, the levels of transcriptional activitieswith standard deviations when the strains were grown under nonmodulating conditions or in the presence of 20 mM MgSO4 (�MgSO4). Alltranscriptional fusions were cloned into the bipA promoter assay vector pRD590 and integrated into the chromosome of the strain RKD100 at thebipA locus. N.D., not determined. Footnote 1, no conintegrants were obtained. Footnote 2, the corresponding lac fusions were highly unstableunder nonmodulating conditions.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

protected at a lower concentration of nonphosphorylatedBvgA than with the wild-type promoter (compare lanes 2 to 5,Fig. 4A and D). Note that as shown earlier (Fig. 4B), a similarenhancement in the binding affinity of BvgA to regions sur-rounding HS2 also occurred as a result of deletion of IR1.

Alterations in the bipA promoter region alter the expressionprofile of the bipA gene in response to modulators. It wasreported previously that the expression of bipA was maximalwhen wild-type B. bronchiseptica bacteria were grown undersemimodulating conditions (16). To compare the expressionpattern of the bipA gene in response to modulating signals withthat of different B. bronchiseptica genes, the �-galactosidaseactivities of different gene fusions in the wild-type strain RB50were measured in the presence of various concentrations ofnicotinic acid, a chemical modulator of bvg activity. The �-ga-lactosidase activities of lacZ fusions to prn, the gene encoding

the outer membrane protein pertactin, and frl, the motilitymaster regulatory locus that is repressed by BvgAS, were mea-sured. Expression of bipA was low in the absence of nicotinicacid, maximal at semimodulating concentrations, and nearbackground levels at high concentrations of the modulator(Fig. 5A). In contrast, prn-lacZ was expressed at very highlevels at concentrations of modulator from 0 to 0.4 mM and ata very low level at concentrations higher than 0.4 mM. Anexpression profile similar to that of prn was also seen for fha,encoding the major Bordetella adhesin, filamentous hemagglu-tinin (data not shown). The Bvg�-phase-specific gene frl wasexpressed in a manner reciprocal to that of prn and fha, withvery little expression in the absence of the modulating signaland maximal expression at high concentrations of modulator(Fig. 5A).

To determine the effect of deletions in the promoter region

FIG. 3. Measurement of the levels of the bipA transcript as a result of deletion of IR2-3. (A). Structure of the IR2-3 deletion in strain RKD101is diagrammed. The primer used for primer extension and DNA sequencing is BipAExt, as indicated. (B). Primer extension assays were performedwith RNA from wild-type strain RB50 (Bvg�), the phase-locked strains RB53i (Bvgi) and RB54 (Bvg�), and RKD101. The bipA-specific transcriptfrom RB50 and the phase-locked strains are marked with �1 (wt), whereas the corresponding transcript from RKD 101 is marked with �1(�IR2-3). G, A, T, and C represent the nucleotide sequencing ladder. The DNA template for sequencing was the plasmid pRD578. Reactionmixtures were electrophoresed on a 6% polyacrylamide gel and exposed to X-ray film for 24 h. (C) A longer exposure (8 days) of the samepolyacrylamide gel as in panel B. Relative intensities of the various transcripts were quantitated by densitometric analysis using a PersonalDensitometer SI and Image-Quant Software Program (Molecular Dynamics Inc., Sunnyvale, Calif.). (D). RT-PCR analysis was used to quantitatethe levels of bipA from different B. bronchiseptica strains. Total RNA from different strains was reverse transcribed using random hexamers andused as a template for PCR. g denotes the genomic DNA as a PCR template. Genomic DNA was isolated from wild-type B. bronchiseptica strainRB50. The volume of the PCR loaded in the RKD101 lane is half of the volumes for the Bvg�, Bvgi, and Bvg� lanes. The kilobase ladder (L) fromGibco BRL was used to determine the sizes of the PCR products.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

FIG. 4. DNase I protection analyses. 32P-labeled variants of the bipA promoter region were incubated with unphosphorylated BvgA (BvgA,lanes 2 to 5) or BvgA phosphorylated in vitro with acetyl phosphate (BvgA-P, lanes 6 to 10). Lanes 1 to 5 contain 0, 0.05, 0.2, 0.4, and 0.8 �Mconcentrations of BvgA, whereas lanes 6 to 10 contain 0.05, 0.1, 0.2, 0.4, and 0.8 �M concentrations of BvgA-P, respectively. (A) Wt, the wild-typeDNA fragment from �120 to �197 of the bipA promoter; (B) �IR1, bipA promoter fragment from �60 to �271; (C) �IR2, same as Wt fragmentexcept the region from �17 to �59, comprising the IR2 site of the bipA promoter, was deleted; (D) �IR3, �120 to �66 region of the bipA promoterlacking IR3. The nucleotide positions of the different DNA bands relative to the transcription initiation site are indicated at the left of each panel,and regions corresponding to HS1, IR1, IR2, and IR3 are indicated on the right of each panel, respectively. Asterisks denote the regions that showreduced affinity of BvgA-P binding as a result of deletion of IR1. �, absence of BvgA or BvgA-P.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

of bipA on the expression profile in response to modulators,the individual bipA-lacZ fusions integrated in the bipA pro-moter deletion strain RKD100 were tested at various concen-trations of nicotinic acid. Measurement of the transcriptionalactivity of the wild-type promoter fusion in RKD100 resultedin an expression profile similar to that observed with the wild-type strain RB50 (compare Fig. 5A and B). Deletion of IR2and IR3 resulted in quite different profiles of expression. Incontrast to the wild-type promoter derivative, lac fusions to the�IR2 or �IR3 promoter derivatives were highly expressed atnicotinic acid concentrations ranging from 0 to 0.8 mM. Re-markably, as a result of deletion of IR2, both the relativeactivity and the expression profile of bipA, a Bvg-intermediate-phase-specific gene, were altered to resemble those of theBvg�-phase-specific genes prn (compare Fig. 5A and B) andfha (data not shown). These observations suggest that thebinding sites IR2 and IR3 play a critical role in determining the

Bvg-intermediate-phase-specific pattern of the bipA gene inresponse to modulators.

DISCUSSION

The bipA gene represented the first identified example of aclass of genes in Bordetella that are maximally expressed undersemimodulating conditions (16, 45). Expression of bipA in re-sponse to modulating signals is unusual in that this gene ismost highly expressed at a point along the Bvg-regulatory con-tinuum where the activity and the level of BvgA-P are pre-dicted to be intermediate. It was hypothesized that the com-plicated expression pattern is the result of the differentialconcentration-dependent occupancy and the spatial location ofthe multiple BvgA binding sites relative to the start of tran-scription of the bipA promoter (16). The data presented here,from the genetic analyses of different BvgA binding sites, pro-vide the first experimental evidence supporting the differentialroles of BvgA binding sites in mediating the expression patternof bipA.

Deletion of the upstream binding site IR1 results in analmost complete abolition of the transcriptional activity of thebipA promoter, suggesting that IR1 is essential for activation ofthe bipA promoter. Analogous to the primary BvgA bindingsites of fha, cyaA, ptx, and prn promoters (10–12, 23–25, 49),IR1 is present upstream of the bipA promoter region. Similarto IR1, the primary binding sites for fha, cyaA, ptx, and prnhave been shown to be critical for activation of the individualpromoters (8, 12, 24–25). Similar locations of the primaryBvgA binding sites suggest the existence of a conserved mech-anism of transcriptional activation of these promoters.

The results from in vivo transcriptional analyses and quan-titation of bipA-specific transcripts reported in this study sug-gest that the binding sites IR2 and IR3 act in concert toefficiently repress transcription from the bipA promoter. IR2and IR3 are located in the ”exclusive zone of repression,” aterm coined by Gralla and Collado-Vides for the region down-stream of �30 (21). Binding of a regulatory protein to thisregion almost always interferes with the functioning of RNApolymerase (21). It is reasonable to speculate that phosphory-lation-induced oligomerization of BvgA may have a multitudeof outcomes for the bipA promoter. The binding of BvgA atvarious binding sites could lead to the formation of higher-order complexes that may either compete with RNA polymer-ase for binding or make the RNA polymerase binding regioninaccessible. One possible mechanism that could lead to theefficient formation of higher-order complexes involves DNAlooping or bending of the intervening DNA (33, 41). Interest-ingly, the two sites (IR1 and IR2) are separated by 105 bp(distance counted from centers of symmetry), similar to thedistance (113 bp) that separates the operators OE and OI ofthe two gal promoters (1, 5). When GalR binds to the opera-tors OE and OI, the two GalR dimers associate to form a DNAloop of the intervening DNA, leading to repression of the twogal promoters (1). A similar mechanism of repression can beenvisioned for the bipA promoter.

It is apparent from the mapping of the BvgA binding sites ona number of Bvg-regulated promoters that there is a great dealof variability in the spacing of the recognition sequences. Thehalf-sites comprising the BvgA binding sites of the fha and prn

FIG. 5. Effects of various concentrations of nicotinic acid on thetranscriptional activity of different promoter fusions. (A) Expressionprofiles of bipA-lacZ (bipA), prn-lacZ (prn), and frl-lacZ (frl) fusions.The different promoters were cloned in the vector pEGZ and inte-grated at the corresponding loci in the chromosome of the wild-type B.bronchiseptica strain RB50. (B) The wild-type bipA promoter and thedifferent promoter variants were cloned in the vector pRD590 andintegrated at the bipA locus in strain RKD100. The levels of transcrip-tional activities (103 �-galactosidase units) with standard deviationbars are shown. Note that the x axis is not drawn to scale.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

promoters are directly joined, whereas these are separated by2 and 10 bp in the cyaA and ptx promoters (11–12, 24). Similarto those of fha and prn, the two half-sites of IR1 are in tandem.In contrast, half-sites of IR2 and IR3 are separated by 27 bp(2.5 turns) and 37 bp (3.5 turns of the DNA helix), respec-tively (Fig. 1) (16). The large physical separation between thetwo half-sites poses an interesting question as to how BvgAmolecules that are bound to two opposite sides of the helixinteract to bridge the large gap separating the two half-sites inIR2 and IR3. A clear understanding of BvgA interaction withits recognition sequences at the bipA promoter will requireadditional detailed experimental analyses.

Studies of the interaction of BvgA with a number of Bvg-regulated promoters have contributed to a model where BvgAinitially binds to a relatively high-affinity primary binding site,which then serves as a starting point for progressive binding ofBvgA molecules to additional weak binding sites presentdownstream (9, 11–12). The data presented here support thismodel. Analyses of BvgA binding with different mutant bipApromoters revealed the presence of two core binding sites, IR1and IR2. Occupancy of BvgA by IR1 and IR2 is the primarydeterminant for the resultant binding of BvgA to the neigh-boring regions. Deletion of IR1 led to a decrease in the bindingaffinity of BvgA-P to HS2. In this respect, the interaction ofBvgA-P with the bipA promoter is similar to that with the fhapromoter. For the fha promoter, it has been observed thatBvgA-P binds more efficiently to the downstream secondaryregion in the presence of an intact primary site that is similarto IR1 (9). Similar to IR1, the binding site IR2 is essential forthe BvgA occupancy of contiguous regions including the down-stream binding site IR3. In each of these cases, phosphoryla-tion-induced binding of BvgA at the primary site contributes tosecondary site occupancy.

What is the molecular basis of differential gene expression ofbipA in response to modulating signals in Bordetella? It hasbeen proposed that depending on BvgA-P levels inside the cell,BvgA may either preferentially occupy the sites upstream ofthe promoter (including IR1), leading to maximal activation oftranscription in the Bvgi phase, or in addition it may bind thedownstream sites (IR2 and IR3), resulting in partial repressionin the Bvg� phase (16). Although intracellular levels ofBvgA-P have not been measured, results from �-galactosidaseassays of the different promoter constructs in response to var-ious concentrations of modulators provide indirect evidencefor such a mechanism. In particular, the repression of wild-typebipA promoter activity in the absence or presence of low con-centrations of modulators (Bvg�-phase conditions) was re-lieved as a result of either individually or collectively deletingIR2 or IR3. Thus, it seems likely that binding of IR2 and IR3by BvgA results in repression of transcription from the bipApromoter under Bvg�-phase conditions. The more than two-fold increase above the maximal expression levels of the wild-type promoter as a consequence of IR2 deletion suggests thatthere is a very tight control on the expression levels of bipApromoter even under Bvgi-phase conditions. Thus, althoughthe expression of wild-type bipA promoter in RKD100 at a 0.4mM concentration of nicotinic acid is maximal, this in factmight still be a partially repressed promoter. In summary, itcan be hypothesized that occupation of a combination of low-affinity repression sites (IR2 and IR3) and high-affinity activa-

tion sites (IR1) at the bipA promoter leads to a constellation ofprotein-DNA and protein-protein interaction that ultimatelyresults in the complex regulatory profile displayed by bipA.Note that the bipA promoter has a long untranslated region(Fig. 1). (16), raising the likelihood that mechanisms at theposttranscriptional level could also control bipA expression.

bipA is the first gene in Bordetella that has been shown toundergo such variegated changes in its expression pattern. Thedifferential regulation of bipA gene expression observed invitro as a result of quantitative differences in the concentra-tions of modulators reflects the ability of a precise and highlysensitive signal transduction system that is responsive to vari-ous microenvironments encountered during the Bordetella in-fectious cycle in mammalian hosts. Deciphering the complex-ities of this regulation would unravel new information thatwould ultimately lead to a better understanding of the mech-anisms employed by organisms of the Bordetella species tosurvive within mammalian hosts.

ACKNOWLEDGMENTS

I thank Jeff F. Miller for guidance and support. I am grateful toPurnima Dubey, Bill Hendrickson, and Tapan Misra for critical read-ing of the manuscript. I thank Linda Kenney for helpful suggestions. Iextend my thanks to Mingshun Liu and Peggy Cotter for providingstrains and plasmids. I am also grateful to the two anonymous review-ers whose detailed critiques of the manuscript led to significant im-provement.

This study was supported by grants to Jeff F. Miller. R.D. wassupported by microbial pathogenesis training grant T32A107323. Partsof this study were performed in collaboration with Purnima Dubey.

REFERENCES

1. Adhya, S., M. Geanacopoulos, D. E. Lewis, S. Roy, and T. Aki. 1998. Tran-scription regulation by repressosome and by RNA polymerase contact. ColdSpring Harbor Symp. Quant. Biol. 63:1–9.

2. Akerley, B. J., D. M. Monack, S. Falkow, and J. F. Miller. 1992. The bvgASlocus negatively controls motility and synthesis of flagella in Bordetella bron-chiseptica. J. Bacteriol. 174:980–990.

3. Akerley, B. J., and J. F. Miller. 1993. Flagellin gene transcription in Borde-tella bronchiseptica is regulated by the BvgAS virulence control system. J.Bacteriol. 175:3468–3479.

4. Akerley, B. J., P. A. Cotter, and J. F. Miller. 1995. Ectopic expression of theflagellar regulon alters development of the Bordetella-host interaction. Cell80:611–620.

5. Aki, T., and S. Adhya. 1997. Repressor induced site-specific binding of HUfor transcriptional regulation. EMBO J. 16:3666–3674.

6. Arico, B., V. Scarlato, D. M. Monack, S. Falkow, and R. Rappuoli. 1991.Structural and genetic analysis of the bvg locus in Bordetella species. Mol.Microbiol. 5:2481–2491.

7. Beattie, D., T. R. Shahin, and J. J. Mekalanos. 1992. A vir-repressed gene ofBordetella pertussis is required for virulence. Infect. Immun. 60:571–577.

8. Boucher, P. E., M. S. Yang, D. M. Schmidt, and S. Stibitz. 2001. Mutationalanalyses of the high-affinity BvgA binding site in the fha promoter of Bor-detella pertussis. Mol. Microbiol. 40:991–999.

9. Boucher, P. E., M. S. Yang, D. M. Schmidt, and S. Stibitz. 2001. Genetic andbiochemical analyses of BvgA interaction with the secondary binding regionof the fha promoter of Bordetella pertussis. J. Bacteriol. 183:536–544.

10. Boucher, P. E., F. D. Menozzi, and C. Locht. 1994. The modular architectureof bacterial response regulators. Insights into the activation mechanism ofthe BvgA transactivator of Bordetella pertussis. J. Mol. Biol. 241:363–377.

11. Boucher, P. E., and S. Stibitz. 1995. Synergistic binding of RNA polymeraseand BvgA phosphate to the pertussis toxin promoter of Bordetella pertussis.J. Bacteriol. 177:6486–6491.

12. Boucher, P. E., K. Murakami, A. Ishihama, and S. Stibitz. 1997. Nature ofDNA binding and RNA polymerase interaction of the Bordetella pertussisBvgA transcriptional activator at the fha promoter. J. Bacteriol. 179:1755–1763.

13. Cotter, P. A., and J. F. Miller. 1994. BvgAS-mediated signal transduction:analysis of phase-locked regulatory mutants of Bordetella bronchiseptica in arabbit model. Infect. Immun. 62:3381–3390.

14. Cotter, P. A., and J. F. Miller. 1997. A mutation in the Bordetella bronchi-septica bvgS gene results in reduced virulence and increased resistance to



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

starvation, and identifies a new class of Bvg-regulated antigens. Mol. Micro-biol. 24:671–685.

15. Deora, R., and T. K. Misra. 1996. Characterization of the primary sigmafactor of Staphylococcus aureus. J. Biol. Chem. 271:21828–21834.

16. Deora, R., H. J. Bootsma, J. F. Miller, and P. A. Cotter. 2001. Diversity in theBordetella virulence regulon: transcriptional control of a Bvg-intermediatephase gene. Mol. Microbiol. 40:669–683.

17. Edwards, R. A., L. H. Keller, and D. M. Schifferli. 1998. Improved allelicexchange vectors and their use to analyze 987P fimbria gene expression.Gene 207:149–157.

18. Fernandez, R. C., and A. A. Weiss. 1994. Cloning and sequencing of aBordetella pertussis serum resistance locus. Infect. Immun. 62:4727–4738.

19. Figurski, D., and D. Helinski. 1979. Replication of an origin-containingderivative of plasmid RK2 dependent on a plasmid function provided intrans. Proc. Natl. Acad. Sci. USA 76:1648–1652.

20. Finn, T. M., and L. A. Stevens. 1995. Tracheal colonization factor: a Bor-detella pertussis secreted virulence determinant. Mol. Microbiol. 16:625–634.

21. Gralla, J. D., and J. Collado-Vides. 1996. Organization and function oftranscription regulatory elements, p. 1232–1245. In F. C. Neidhardt (ed.),Escherichia coli and Salmonella. ASM Press, Washington, D.C.

22. Hoch, J. A. 2000. Two-component and phosphorelay signal transduction.Curr. Opin. Microbiol. 3:165–170.

23. Karimova, G., J. Bellalou, and A. Ullmann. 1996. Phosphorylation-depen-dent binding of BvgA to the upstream region of the cyaA gene of Bordetellapertussis. Mol. Microbiol. 20:489–496.

24. Karimova, G., and A. Ullmann. 1997. Characterization of DNA binding sitesfor the BvgA protein of Bordetella pertussis. J. Bacteriol. 179:3790–3792.

25. Kinnear, S. M., P. E. Boucher, S. Stibitz, and N. H. Carbonetti. 1999.Analysis of BvgA activation of the pertactin gene promoter in Bordetellapertussis. J. Bacteriol. 181:5234–5241.

26. Knapp, S., and J. J. Mekalanos. 1988. Two trans-acting regulatory genes (virand mod) control antigenic modulation in Bordetella pertussis. J. Bacteriol.170:5059–5066.

27. Lacey, B. W. 1960. Antigenic modulation of Bordetella pertussis. J. Hyg.58:57–93.

28. Ladant, D., and A. Ullmann. 2000. Bordatella pertussis adenylate cyclase: atoxin with multiple talents. Trends Microbiol. 7:172–176.

29. Loomis, W. F., G. Shaulski, and N. Wang. 1997. Histidine kinases in signaltransduction pathways of eukaryotes. J. Cell Sci. 110:1141–1145.

30. Marques, R. R., and N. H. Carbonetti. 1997. Genetic analysis of pertussistoxin promoter activation in Bordetella pertussis. Mol. Microbiol. 24:1215–1225.

31. Martinez de Tejada, G., P. A. Cotter, U. Heininger, A. Camilli, B. J. Akerley,J. J. Mekalanos, and J. F. Miller. 1998. Neither the Bvg� phase nor the vrg6locus of Bordetella pertussis is required for respiratory infection in mice.Infect. Immun. 66:2762–2768.

32. Martinez de Tejada, G., J. F. Miller, and P. A. Cotter. 1996. Comparativeanalysis of the virulence control systems of Bordetella pertussis and Bordetellabronchiseptica. Mol. Microbiol. 22:895–908.

33. Matthews, K. S. 1992. DNA looping. Microbiol. Rev. 56:123–136.34. Mattoo, S., A. K. Foreman-Wykert, P. A. Cotter, and J. F. Miller. 2001.

Mechanisms of Bordetella pathogenesis. Front. Biosci. 6:E168–E186.35. Merkel, T. J., and S. Stibitz. 1995. Identification of a locus required for the

regulation of bvg-repressed genes in Bordetella pertussis. J. Bacteriol. 177:2727–2736.

36. Miller, J. F., C. R. Roy, and S. Falkow. 1989. Analysis of Bordetella pertussisvirulence gene regulation by use of transcriptional fusions in Escherichia coli.J. Bacteriol. 171:6345–6348.

37. Perraud, A. L., V. Weiss, and R. Gross. 1999. Signalling pathways in two-component phosphorelay systems. Trends Microbiol. 7:115–120.

38. Roy, C. R., J. F. Miller, and S. Falkow. 1990. Autogenous regulation of theBordetella pertussis bvgABC operon. Proc. Natl. Acad. Sci. USA 87:3763–3767.

39. Roy, C. R., and S. Falkow. 1991. Identification of Bordetella pertussis regu-latory sequences required for transcriptional activation of the fhaB gene andautoregulation of the bvgAS operon. J. Bacteriol. 173:2385–2392.

40. Scarlato, V., A. Prugnola, B. Arico, and R. Rappuoli. 1990. Positive tran-scriptional feedback at the bvg locus controls expression of virulence factorsin Bordetella pertussis. Proc. Natl. Acad. Sci. USA 87:6753–6757.

41. Schleif, R. 1992. DNA looping. Annu. Rev. Biochem. 61:199–223.42. Stainer, D. W., and M. J. Scholte. 1970. A simple chemically defined medium

for the production of phase I Bordetella pertussis. J. Gen. Microbiol. 63:211–220.

43. Steffen, P., S. Goyard, and A. Ullmann. 1996. Phosphorylated BvgA is suf-ficient for transcriptional activation of virulence-regulated genes in Borde-tella pertussis. EMBO J. 15:102–109.

44. Stenson, T. H., and M. S. Peppler. 1995. Identification of two bvg-repressedsurface proteins of Bordetella pertussis. Infect. Immun. 63:3780–3789.

45. Stockbauer, K. E., B. Fuchslocher, J. F. Miller, and P. A. Cotter. 2001.Identification and characterization of BipA, a Bordetella Bvg-intermediatephase protein. Mol. Microbiol. 39:65–78.

46. Uhl, M. A., and J. F. Miller. 1994. Autophosphorylation and phosphotransferin the Bordetella pertussis BvgAS signal transduction cascade. Proc. Natl.Acad. Sci. USA 91:1163–1167.

47. Uhl, M. A., and J. F. Miller. 1996. Integration of multiple domains in atwo-component sensor protein: the Bordetella pertussis BvgAS phosphorelay.EMBO J. 15:1028–1036.

48. Yuk, M. H., E. T. Harvill, and J. F. Miller. 1998. The BvgAS virulencecontrol system regulates type III secretion in Bordetella bronchiseptica. Mol.Microbiol. 28:945–959.

49. Zu, T., R. Manetti, R. Rappuoli, and V. Scarlato. 1996. Differential bindingof BvgA to two classes of virulence genes of Bordetella pertussis directspromoter selectivity by RNA polymerase. Mol. Microbiol. 21:557–565.



http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

Documents

Differential Regulation of the Bordetella bipA Gene: Distinct Roles