Proc. Natl. Acad. Sci. USAVol. 86, pp. 6671-6675, September 1989Genetics

Sequences required for expression of Bordetella pertussis virulencefactors share homology with prokaryotic signaltransduction proteins

(vir/bvgA, bvgB, and bvgC/two-component regulatory systems/virulence regulation)

BEATRICE ARic6*, JEFF F. MILLERt, CRAIG Royt, ScoTT STIBITZt, DENISE MONACKt, STANLEY FALKOWt,RoY GRoss*, AND RINo RAPPUOLI*§*Sclavo Research Center, Via Fiorentina 1, 53100 Siena, Italy; tDepartment of Microbiology and Immunology, Stanford University, Stanford, CA 94305;and tCenter for Biologics Evaluation and Research, Food and Drug Administration, Bethesda, MD 20892

Contributed by Stanley Falkow, May 9, 1989

ABSTRACT The bvg locus of Bordetella pertussis is re-quired for coordinate regulation of several factors associatedwith virulence. The control system is modulated by variousenvironmental signals, including low temperature, MgSO4,and nicotinic acid. The nucleotide sequence of the bvg regionhas been determined and three open reading frames, bvgA,bvgB, and bvgC, are present. Twelve-base-pair linker insertionmutations in any of these open reading frames result in a Bvg-phenotype. The predicted protein products of bvgA and bvgCshare homology with a family of prokaryotic regulatory pro-teins that respond to environmental stimuli and are membersof two-component sensory transduction systems. We propose amodel in which BvgB and the N-terminal portion of BvgC arelocalized in the periplasm. Environmental signals are recog-nized, transduced to the cytoplasmic portion of BvgC, and thentransmitted to BvgA, a positive regulator of transcription.

The host-microbe interaction that occurs during infectionand disease is a dynamic process in which the infectingorganism encounters a variety of external conditions. A needfor efficient adaptation to changing environments has re-sulted in the evolution of specialized systems that control theexpression of bacterial virulence (1).

Bordetella pertussis is a human pathogen that causeswhooping cough. This organism produces a number of factorsthat appear to be involved in pathogenesis (2). Filamentoushemagglutinin (3) and pili (4) probably play a role in adher-ence of the bacteria to ciliated epithelial cells in the upperrespiratory tract. Other products, such as tracheal cytotoxin(5), dermonecrotic toxin (6), hemolysin (6, 7), adenylatecyclase toxin (8), and pertussis toxin (7, 9), may be involvedin local damage, evasion of host defenses, and systemicdisease. With the exception oftracheal cytotoxin, expressionof these virulence factors is coordinately regulated andrequires the products of the bvg (flordetella yvirulence gene)locus, which was initially designated vir (10-12). Positivecontrol of the pertussis toxin operon by bvg has been shownto occur at the level of transcription (13).

Inactivation of the bvg locus by transposon mutagenesisresults in a loss of virulence (7). Alternative virulence statesof B. pertussis also occur by two mechanisms, phase varia-tion and phenotypic modulation. Phase variation (11) is dueto a low-frequency reversible alteration in the bvg locus that,for the Tohama III strain, has recently been shown to involvea frameshift mutation in bvg (14). In contrast, phenotypicmodulation is a reversible response to environmental condi-tions. Growth at low temperature (280C vs. 370C), or in thepresence of MgSO4 or nicotinic acid, results in a lack of

expression of bvg-regulated genes and an avirulent pheno-type (15, 16).

In this paper we report the complete nucleotide sequenceof the B. pertussis bvg region. The predicted products of twobvg loci are homologous to a family of regulatory proteinsthat transmit sensory signals using a conserved two-component motif (1, 17). A model for the role of the bvgproducts in sensory transduction and gene regulation is alsopresented.

MATERIALS AND METHODSBacterial Strains. B. pertussis strains BP165 (Bvg'; ref. 9),

BP536 (Bvg'; ref. 10), and the TnS insertion derivativesBP347 (virl::TnS; ref. 10) and BP359 (vir2::TnS; ref. 11) havebeen described. BP370 is a virulent phase derivative ofBP326(11). Escherichia coli strain SmlO (18) was used for IncP-mediated conjugative transfer from E. coli to B. pertussis asdescribed below.

Chloramphenicol Acetyltransferase (CAT) Fusions. A pro-moterless CAT gene (19) was isolated as a BamHI fragmentand inserted into the Bgl II site of bvgC in both orientations.CAT fusions were returned to the chromosome of BP370 byhomologous recombination (20). The orientation of the fu-sions within the B. pertussis chromosome was confirmed bySouthern analysis. The resulting CAT fusion strains CF1470and CF1770 carry the CAT gene in the orientations shown inFig. 1. CAT assays were performed as described (9).

Nucleotide Sequence Determination. The nucleotide se-quence of the 5995-base-pair (bp) EcoRI-Xho I fragmentshown in Fig. 1 was determined by the standard dideoxychain-termination method (21). The sequence of both strandswas obtained by using synthetic oligonucleotide primers.

Linker Insertion Mutagenesis. Twelve-base-pair linker in-sertions were introduced into the bvg region at the locationsshown in Fig. 1. Insertions at Sal I, Pst I, or Mlu I sitesconsisted of the sequence 5'-GACCTGCAGGTC-3', 5'-GCTCGAGCTGCA-3', or 5'-CGCGTCTCGAGA-3', respec-tively. Mutations were constructed in E. coli and transferredto B. pertussis by shuttle mutagenesis (20). The structures ofthe bvg loci in the resulting strains were verified by Southernblot analysis.

Sequence Analysis. The Sequence Analysis Software Pack-age distributed by the University of Wisconsin GeneticsComputer Group (22), running on a VAX/VMS computer,PC Gene (Intelligenetics), and Genepro (Riverside Scientific,Seattle) were used for protein and nucleic acid sequenceanalysis, database searching, and homology assessment.

Abbreviations: CAT, chloramphenicol acetyltransferase; ORF, openreading frame.ITo whom reprint requests should be addressed.

6671

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Proc. Natl. Acad. Sci. USA 86 (1989)

RESULTSB. pertussis bvg Locus. The bvg locus was initially identified

by transposon TnS insertions in the chromosomes of B.pertussis strains BP347 and BP359 (10, 11). The bvg locus hasbeen cloned (12) and a restriction map is shown in Fig. 1. Toinvestigate the transcriptional activity of bvg, a promoterlessCAT gene was inserted in both orientations at the unique BglII site within the bvg sequences of strain BP370. The orien-tations and insertion point of the bvg-CAT fusions carried byCF1470 and CF1770 were confirmed by Southern blot analysisand are shown in Fig. 1. The Bgi II site in bvg lies between thesites of TnS insertion in BP347 and BP359. Insertion of theCAT cassette at this site disrupts the bvg locus, CF1470 andCF1770 have a Bvg- phenotype.CAT activity was detected only in CF1770, in which the

CAT gene is present in the left to right orientation as depictedin Fig. 1 (data not shown). Introduction of a plasmid carryingthe wild-type bvg locus into these strains results in a Bvg+phenotype and does not alter the relative pattern of CATexpression. We conclude from these results that the bvglocus is transcribed in Bvg+ and Bvg- strains and that thedirection of transcription is from left to right as shown inFig. 1.

Nucleotide Sequence ofthe bvg Locus. The boundaries ofthebvg locus are contained within a 6-kb DNA fragment delim-ited by the EcoRI-Xho I restriction sites shown in Fig. 1. InB. pertussis (10, 11) and E. coli (12), Bvg- TnS insertions mapwithin this region. In addition, this fragment is sufficient fortranscriptional activation ofanflJaB-lacZYA operon fusion inE. coli in a manner that responds to environmental signalsthat modulate virulence expression by B. pertussis (J.F.M.,C.R., and S.F., unpublished data).We have determined the complete nucleotide sequence of

the 5995-bp EcoRI-Xho I fragment containing the bvg locusand have deposited this sequence in the EMBL/GenBankdata base (accession no. M25401). The overall G+C contentof the sequence is 65%, a value similar to that of other B.pertussis sequences (4, 8, 23). As summarized in Fig. 1, threetandem ORFs with codon usage patterns typical of B. per-tussis genes are detected. These ORFs, which may form anoperon, are designated bvgA (Bordetella virulence gene A),bvgB, and bvgC. Putative ribosomal binding sites foFeach ofthese loci are located 7-14 nucleotides upstream from AUGsequences. The proposed direction of bvg expression corre-sponds to the direction of transcription detected with theCAT fusions described above.bvgA, bvgB, and bvgC are predicted to encode proteins of

209, 275, and 936 amino acids, with molecular weights of

BP359(n)

E M M P_LII

+-

CF1470 CF1770SCATSCPl

S PS S. P Bll EI I I inl l l L

+' +T^s

23,000, 30,000, and 102,000, respectively. Characteristic fea-tures of a prokaryotic signal peptide are found at the Nterminus of BvgB. Two basic amino acids at positions 6 and10 are followed by 23 uncharged, predominantly hydrophobicresidues. This pattern is similar to that present at the Ntermini of the E. coli Tsr, Tar, and Tap chemoreceptors (24).BvgC contains two hydrophobic segments. The first, whichmay also have a role in export, contains a lysine residue atposition 8 followed by a relatively short 13-amino acidhydrophobic core. In addition, a sequence of 22 hydrophobicresidues beginning at position 240 is predicted to form atransmembrane helix [transfer free energy - +44 kcal/mol (1kcal = 4.18 kJ); ref. 25] and is followed by a highly basicstretch of9 amino acids. This putative transmembrane regionis similar to those found in several bacterial proteins that spanthe inner membrane, such as the Agrobacterium VirA protein(26) and the M13 procoat protein (27).Although sequences upstream of bvgA are unusually rich

in A+T content, putative bacterial promoters (28) could notbe found. In addition, structures resembling constitutiveterminators (Rho independent; ref. 29) do. not appear to bepresent. Sequences downstream of bvgC show a low codingprobability in the left to right orientation. The probability ofcoding in this region is high in the opposite strand; however,no obvious ORFs have been identified. We have determinedthe sites of transposon TnS insertion in the Bvg- B. pertussismutants BP347 and BP359 in relation to the bvg ORFs andfound that these insertions map in bvgC and bvgA, respec-tively (Fig. 1).BvgA and BvgC Belong to a Family of Regulatory Proteins

Involved in Signal Transduction. The predicted amino acidsequences of BvgA, BvgB, and BvgC were compared tosequence entries in the National Biomedical Research Foun-dation database, and to sequences translated from the Gen-Bank database. Localized regions of homology were foundbetween BvgA, BvgC, and members of a family of bacterialproteins that mediate signal transduction using a conserved"two-component" motif (Fig. 1; refs. 17 and 30). In thesesystems a "sensor" protein directly or indirectly perceivessensory stimuli and transmits a signal to a "regulator"protein that functions to control transcription or some othercellular function. Sensor proteins share a homologous C-terminal domain of =250 amino acids that has been termedthe "transmitter," and regulator proteins typically share anN-terminal region ofabout 115 residues, which comprises the"receiver" (30).The central portion of BvgC (residues 423-655) shares

homology with the transmitter domains of EnvZ (31), FixL

BP347(Tn5)

kb i

S S P E E XI I I I-

+bvgA bvgB

RECEIVER

bvgC

TRANSMITTER RECEIVER

FIG. 1. Restriction map and features of the bvg locus. The 5995-bp EcoRI-Xho I fragment that contains the BP165 bvg locus is shown. Therestriction map is derived from the nucleotide sequence. Arrows denote the position of transposon Tn5 insertions in strains BP359 and BP347.The location and orientation of the CAT fusions in CF1470 and CF1770 are also indicated. bvgA, bvgB, and bvgC are open reading frames (ORFs)predicted from the sequence. Transmitter and receiver indicate the relative locations ofhomologous sequences in the predicted protein products.Solid portions ofthe arrows indicate putative N-terminal signal sequences in BvgB and BvgC and the predicted transmembrane domain of BvgC.The location and orientation of the gene encoding filamentous hemagglutinin (ffiaB) are also shown. Positions of 12-bp linker insertion mutationsat Mlu I, Sal I, and Pst I sites in the bvg region are indicated by upward vertical arrows. +, Insertions resulting in a Bvg+ phenotype; -, insertionsgiving a Bvg- phenotype. E, EcoRI; M, Mlu I; P, Pst I; S, Sal I; BII, Bgl II; X, Xho I. kb, Kilobase.

thaB4=...m

6672 Genetics: Arico' et al.

Genetics: Arico et al.

(32), VirA (26), NtrB (33), PhoR (34), PhoM (35), and othertransmitter proteins (Fig. 2). The similarity to FixL, a factorrequired for positive control of nitrogen fixation by Rhizo-bium (32), is observed over the entire length of FixL. Inaddition, regions of similarity with VirA, a sensor proteininvolved in Agrobacterium virulence control (26), can befound extending from the transmitter to the N terminus ofBvgC. FixL and VirA contain predicted transmembranedomains similar to the one identified in BvgC, and VirA hasbeen shown to be located in the Agrobacterium cytoplasmicmembrane (26).BvgA and BvgC also contain receiver domains (Fig. 3).

These are located at the N terminus of BvgA (residues 1-115)and, surprisingly, at the C terminus of BvgC (residues672-789). The homology between BvgA, FixJ (32), and UhpA(40) extends over the entire length of these proteins, whereasmost other similarities are confined to N-terminal sequences.FixJ is a transcriptional activator that acts in conjunctionwith FixL (see above), and UhpA positively regulates expres-sion of the sugar-phosphate transporter gene uhpT in E. coli(40). Many proteins that contain receiver modules are tran-scriptional activators that fall into different categories ac-cording to their C-terminal sequences (17). BvgA appears tobelong to a group that previously has been shown to containFixJ and UhpA (32). A measure of the significance of thisrelationship was obtained by using the RDF program ofLipman and Pearson (41). When BvgA was compared withrandomly permuted versions of the FixJ or UhpA sequences,the optimized alignments scored 14.7 (FixJ) and 16.5 (UhpA)standard deviations above the mean for 20 randomized com-

Proc. Natl. Acad. Sci. USA 86 (1989) 6673

parisons. Although FixJ and UhpA have been suggested tocontain helix-turn-helix DNA binding motifs (32, 40), we findno such evidence of this motif in BvgA, FixJ, or UhpA usingthe refined method for identification of helix-turn-helix se-quences recently described by Brennan and Matthews (42).BvgA, BvgB, and BvgC Are Required for the Expression of

Virulence Factors. Previous identification of the bvg locus inB. pertussis has relied on transposon TnS insertions that arecapable of causing polar effects on distal transcription (10,11). To determine if each of the three bvg ORFs is requiredfor expression of virulence factors, 12-bp linker insertionmutations were introduced into the Sal I, Pst I, and Mlu Isites shown in Fig. 1. Since these mutations result in anin-frame insertion of 4 amino acids, polar effects would notbe generated. Insertions at the Mlu I site in bvgA, the Sal Isite in bvgB, and several Pst I and Sal I sites in bvgC causea Bvg- phenotype as shown by the loss of hemolytic activity.Mutations at the Mlu I site preceding bvgA, or the Sal I andPst I sites located downstream from bvgC, had no effect onhemolytic activity. These results indicate that bvgA, bvgB,and bvgC are required for expression of virulence factors byB. pertussis.

DISCUSSIONThe bvg regulon of B. pertussis includes several unlinkedgenes and operons that encode virulence factors. Expressionof these loci, and their response to environmental signals,require the products of the bvg locus in trans (1, 10, 12, 43).An additional set of genes that are repressed by bvg and

BvgC 5 PVVKVAVLNLFAPFTLFRTDEQFGGISAAVLQLLQLRTGLDE'EIIGVD.TVEELIAKLRSGEADMAGALFVNSAR.ESF..LSFSRPYVRNGMVIVTRQDPDAPVDADHLDGR... .TV 115:A: :: A :A At A*A: :: :: ::: :A:A:: A* **Ai ** : : :: :A*A : : ::

VirA 32 AVLAIGPIKNEKSIEAILTELQSIDVDCALLQRNVLRA ..HAGLLRNYRPLMVPLGRVRSSIANLQQ. LFKK .ARVEDVGELSELLARLKSSINT.....TDAAVASFGAQNVVFESL 141

FixL MIRAEPIGEGLLLFSFIPAILWALIGGRNP. ILFAAGLSLVAAVSH: A : : : : AA *A: A A : : *

BvqC ALVRNSAAIPLLQRRYPQ ....AKVVTADNPSEAMLMVANGQADAWQTQISASYYVNRYFAGKLRIASALDLPPAEIALATIRGQTELMSILNKALYSISNDE. LAS. IISRHRGSDGA 2A *AA : AAA :: : AA : :3::: As:A : : :A : : : A: * : ::: A : AA

VirA ATFNQS. ISSLLRTSDSRDLNAAKV. . PELGYL. MLQFS.FRPNTELALQITQSLDQLQMSTNADKVAIQEVVRNGRVILGVLPRLNErVKLVQ. ASGTFEDITKKLQRAYLEADSLARV

FixL

BvqC

VirA

FixL

BvqC

VirA

EDvZ

NtrB

PhoR

PhoM

FixL

BvgC

VirA

EnvZ

NtrB

PhoR

PhoM

FixL

BvqC

VirA

EnvZ

NtrB

PhoR

PhoM

QQIS. SAWGPSWELLVFG.SAVLLIVALGEVLEAARRAIDRTE1VVRARDAHLRSIML'VPDATVVSATDGTIVSFNAAAVRQFGYAEEEVIGQNLR.... ILMPEPYRHDGYLQR: : A: *: A AA:: A A A:*A t :A:: *A:: A :A:: A A: AA : :A: A A A : A A

DPRTWYAYRNEIYLLIGLGLLSALLFLSWIVYLRRQ IRQRRARLND LEFMVIWTPNPIYVRDKEKMECRDYMLGVTDAVLGKTIPEANVVG PAREELLTA :* :A AA A ,* : :: *A: : s : A: :: 3

.......VEQRVRTFLGAVSVF'FCFGIVILVHKLRRRTl)RLARRLDF-EEVIKKIGVCFMlSTErKQSLKSSAEAALGTIENFFEANQCVLGLVNVTElIAETFSASAPPPSHNERIRKI

YMATGEKRIIGIDRWSGQRKDGS ..TFPMKLAVGEMRGE 'GFIRDLTEREESAARLEQIQAELARLARL ..... ................... NEDGEMASTLAHELNQPLSA* A A A :A : A A :*:::: : A AA A3: A: : : A :A: :**: A: A

VA.M1EI.HIPYGDSLGELKGIIGGWIDITERAELLREADAAN.............. ATMSHEIRTPHIAA ::A :A:* A A : s ::*AA

VSLVQADBISIF-RDYPARKASCFNEAPRALVAF'KVSDRLVAVFCL-RDPVQPASSEVQIKCAAGCVVHYVIRCQTQRDILERRLKHAERLEAVGTLAGGIAHEFNILGV:*: :A: A*

... MAGVSHDLRTPLTR: A: :*::: AA:

... VRGLAHEIKNPLGG: :**AA:: A

. .. FANVSHELRTPLTV: :**A: s**s

... VYALTHELKSPLAA

IANYSHGCTRLLRDMDD. AVATRIREALEEVASQSLRAGQIIKHLRFTKGEIE APMIRKLVEESAALALVGSREQGVRTFEY . LPGAEDVLVD ........ RIQVQQVLINLMRA AA* A2: *A A2 A A A : :* AA : t k A: At :A IA AA :: AA** AA

IIGKfLEA. .LL MPRSQADASLLGILIK~GFLPRARLEARV'GAQGEVKDVVD D.....PLRI4CQVLSNLVGAt A A At : : A AAS* : :AA:: * :: AAA**: A* AA

ILGYA. MAQNILHRR.TYARHYIDRINAESNRARLIIDQILALSRRRERTARPFNLSALVREIAPSLRVALPSEVEVDFNIQ ... ............ SAQMIVEGNPLEIEQILMNLCXA :*A:: A : **: : AA* AS : A:: A :: :AA:s : A AAA : *A:I... RLATEMMSEQDGY... LAESINKDIEECNAIIEQFIDYLRTGQE .. MPMEADLNAVLGEVIAAESGYEREIETALY............... PGSIEVKMHPLSIKRAVANVA: A IS: t* A A AAA : : A A t:SA : ::: ::*A A2 A A A* : A:

L.... RGAAQLLSKALP. DPSLLEYT. KVIIEQADRLRNLVDRLLGPQL .. PGTRVTESIHVAE...... RVVTLVSMELPDNVRLI..... RDYDPSLPELAHDPDQIEQVLLNIVRA A: : :A :AA*: A: AA AA*t : A : AAA : A:3: :: : : *:

L.... QGYLDffIEQPL. Ei.VREALHIRETQRMEGLVKQLLTLS .. XIEAPTLILLNEKVWPM. RVVEREAQTLSQKKQTF .... TF. EIDNGLKVSGNEDQLRSAISNLVYAA* : AA A A I : A AA: :AA AA* * :At* S *S tI *AAA t A: :**AA AA*A: 3: :A A: S**:

I.... RGAAEILREGPPPE. WA. RE'TDILTQNARMQALVETLLRQA .. RLENRQEVVLTA. VDVAALFRRVSEARTVQLADEKITLHV. .TPTEV.NVAAEPALLEQ ... ALGNLLD

NAIEAM... . RHVDRRELTIRTHPADPGEVAVVVEIXG= IPEEVAGQLFKPFV ....... TTK.ASGMGIGLSISKRIVEAHGGEKtVSKNEAGGATFRFT .. LPAYLDERIVAND#AAA A:: A A AAA AA:A AAAAAA^ A3A A2AAAAA3A:*:*3 AA : : : :*A A: :: : A

NAIKFT.... TEGQVLAVTARPDGDAAHQFSVSDrCISEADQRQLFKPFSQVGGSAEAGP.APGTGLGL8I!RRLVELTLVMRSAPGVGTVSVDLRLTM.VEKSVQAAP.AA :2 : : A A AAA* : : :AA : :AAAAAAA A A:A :AAAA : A A A:A: :

NAAESV .. .YRSFVHKHVLANGTIPAGDYILLSTEDNGGGIGQAALPHIIFEPFTR........AQCGGTGLGLSTVHGHVSAMAGFVDVISTVGRGTRFMIYLP.TSAX.KPVNS...AAA A:A :2:: :A:AAAAA: A*:AA A :AAAAAA: A A A ::: :: A :AA: :: :

NAARY.G. . NG.NI.K ..... VSSGTEPNRANFQVEDDGPGIAPEQRKHLFQPFVRGD. SA ... RTISGTGLGLAIVQRIVDNHNGMLELGTSERGGLSIRANLPVPVTRAQGTTKEG#AA A A : : A :*AAA*AA*** AA A A A A :AAAAAA: : :A*: A :A :: A :AA:NALQALGPEGGLRTRTAFQLTLHGERYRLAARIDVWNGPGIPPHLQDTLFYPMVSG. .R.EGLGLSIANLIDQHSGKIEFTSWP.GHTEFSVYLPIRK*AA: A I A:: :*A AAAAAAAAAlA AAA*AAAA3A t : 32::: A A AA :A A

NAVN.. HPEGTHITVRWQRVPHGAESVENGPGIAPEHIPRLTERFYRVDKARSRQT ..GGSGLGLAIVKHAVNHHESRLNIESTVGKGTRFSFVIPERLIAKNSD#AA:: AAA : :AA*A A A:AA: ::A: AAAA* * * : A:* AAAA A 2 3**** A A : A :

NAID. . FTPESGCITLSAEVDQEIVTLKVLDGSGIPDYALSRIFERFYSLPRANG.Q ... KSSGLGLAPVSEVARLFNGEVTLRNVQE3GVLASLRL. HlffT#

46

228

254

159

347

368

252

435

487

251

147

221

273

361

544

588

346

247

327

377

464

655

704

450

349

431

474

FIG. 2. Amino acid sequence comparisons between BvgC and members of bacterial sensory transduction systems that contain transmitterdomains. Identical amino acids (*) and similar amino acids (:) between the adjacent sequences are indicated. The groups of similar amino acidsare C; S, T, P, A, G; N, D, Q, E; M, I, L, V; F, Y, W. Proposed transmembrane domains and signal sequences are indicated by dashed lines.References for sequences are given in the Results.

6674 Genetics: Arico et al. Proc. Natl. Acad. Sci. USA 86 (1989)

BvgA MYNKLIIMKPVLRFAVRVLMEIFIGET. DNGIDGLKIARE.KIPNLVVLDIGIPKLDGLEVIARLQSLGLP .......LRVLVLTGQW=LF CA *:** :* : s: :AA A : : A A*: : : A*: :*: *:*s: AA : : :*:* : A : s*A *: *

FixJ KTDYTVHIVDUEEVXSLAf.GFAVKMhI.QQSAEAFLAFAPDVR.NGVLVTDLRMPNGVLTDLLRULGDLK.NIPS.IV..... ISA E IKPFA*A ::** ** : A :A:::: *: ** * * * : *: A**:A**:** :A: :*::: * :: :*: *** *: *

UhpA MITVALIDDHLIVRSGFAQLLGLEPDLQWAEFGSGREALAGLP. GRGVQVCICDISMPDISGLELLSQLPKG. MA ... ...... TIMSVHDSPALVEQALNAGARGFLSKRC* : :*** * * : : ** AA* : : *A* ** *** A* : : * : : * A*A

BvgC 672 . . RVLVVIDWPNLMLLRQQLDYLGQRVIA. .ADSGEAALALIRH.AFDVVITDCNMPGISGYELARRIRAAE&APGYGRTRCILFGFTASAQ MDEAQACRAAGQMCLFKPI: :A:*A :AA: : A* * **A : *:*I * :**: .* *: :A:: : *A :: :* * : * A*

SpoOF MMNEKILIVDDQYGIRILLNEVFNKEGYQTF .. QAANGLQALDIVT. KERPDLVLLDMKIPGMDGIEILXRMKVIDENIR. VIIMTAYGELDMIQESKELGALThFAKPF**A:A :: :: :: A:: :A: A: *:AAA:A:A:: *:::**: :A : :: **: A**

CheY MADKELKFLVVDDFSTRRIVRNLLKIEGFNNV ..E.A DGVDAINKLQAGGYGFVISDNMPNMDGTr'rL-LTIRADGAMSA. LPVLMVTAEAKKEDIIAAAQAGASGYVVKPF: : * ****A :* :: * * ** A : :* * : :: * :* *A* : : :A::: : :***A : : * : : A* *A: A*

OupR MQENYKNLVVDDIRALLERYLTEQGFQ .VRSVANAEQMDRLL.TRESFH. LMVLDLMLPGEDGLSICRRLRSQSNPMP ....... IIMVTAXGEEVDRIVGLEIGADDYIPKPFA AAA* :* :: : A A A : :: AA: :* AAA :A AAA : :A:AA AA : :AA:AA *:: :**

CheB MSKIRVLSVMSALMRQIMTEIINSHSDMVATAPDPLVARDLIKFN . PDVLTLDVEMPRMDGLLEKLMRLR. PMP .....VVtvSSLTGKGSEV.TLRPLELGAIDFVTKPQA:A A:A A :*::: ::: : A :A AA:: AA :A A: :A: :A A :A: : A AA: :A A :A:AA AA A: :AAA

PhoB MARRILVVEDEAPIREVCFVLEQNGFQPVEA.EDYDSAVNQLN. . EPPDLILLDI4LPGGSGIQFIKHLKRE. . SMTRD.. . IPVVMLTARGEEEDRVRGLETGADDYITKPF: :**:A*** *: : : * :A:: ** : : * A :*:::* *A* A: : * ** :::: **:* :* *:: *** AAAAAAAAAA

Dye MQTPHILIVEVrRNTLKSiF EYVM.GAEMHQIL .. SEYDINLVIMDINLPGKNGLLLAREL.REQAN . V. MFTGRDKEVDKILGLEIGADDYITKPFA: :: : :A: A A:: : :: A A A*:::: :AA :* :AA*: A AA :: AA: :A ::: :AA :* AAA

SpoOA MEXIKVCVADDNRELVSLLSEYIEGQEDMEVIGVAYNGQE LSLF.XEKDPDVLVLDIIMPHLDGLAVLERL .RESDLKKQP .... .NVIMLTAFGQEDVTIAVDLGASYFILKPFA:: : **: :* : A : A:A A A AAAA: AA AA :**AA:A : :: **:I:*A :: A : A : *A :: AAA

NtrC MQRGIAWIVDDDSSIRMV)LERAI.TGAGLSCTTFES. ...................GNEVLDALT .KTPDVLLSDIRMPGDLALQI KQRHP. MLPVIIMTAHSDLDAAVSAYQQGAFDYLPXPF

BvgA NLHEVINAAKAVMAGYTYFPSTT:A :A

FixJ EDTVIIEAIERASEHL .......:AA:

UhpA SPDELIAAVHTVATGG......: A A

BvgC GVDALRQRLN... 79(: At :A:1

SpoOF DIDEIRDAVK... il1

CheY TAATLEKLN...: A : :

OupR NPRELLARIR...

CheB LGIREGMLAY ...

PhoB SPKELVARIK...A:AA A

Dye PRELTIRAR..: :*A : A

SpoOA DIIENLVGHIR...

A::: At: :NtrC DIDEAVALVD...

rLSEWRMGDKRKSDSTLISVLSMRMTVLQ . LLA. QGMSIKDIADSMFLSIOVEVSTYIRLLQKL LIDLVELA#A: :* : :AA:*AA AA ::A A: :A** : :A :*A :: A : A *s*AA: g :

.VAAEADVDDRANDIRARLQTLSERERQVLSAWA. . GLPNKSIAYDLDISPRTVEVHRANVMAKMKAKSLPHLVRMALAGGFGPS#:A ::A: AAAAA :* A:: A AA :A :A*:AA AAAA*:* A: A A A

CYLTPDIAIKLASGRODPLT[REROVAEK. LA. OGMAVKEIAAELGLSPKHVHRANLMEKLGVSNDVELARRMFGN#

L016

121

116

119

117

115

120

116

FIG. 3. Amino acid sequence comparisons between BvgA, BvgC, and members of bacterial sensory transduction systems that containreceiver domains. Similarities and identities are as described in the legend to Fig. 3. References for members of two-component systems thatare not described in the text are as follows: SpoOF (36), SpoOA (37), CheY and CheB (38), and Dye (39).

derepressed by modulation signals has recently been identi-fied (43); however, their functions are unknown.Our analysis ofthe nucleotide sequence has identified three

ORFs, designated bvgA, bvgB, and bvgC, which residewithin the bvg region. In B. pertussis, transposon insertionsin bvgA or bvgC disrupt bvg function (10), and spontaneousphase variation in the TohamaIII strain results from theinsertion (and deletion) of a cytosine residue at position 4129in bvgC (amino acid 800; ref. 14). In this report we show thatinsertion of a promotorless CAT gene within bvgC eliminatesbvg activity and that nonpolar linker insertion mutations inbvgA, bvgB, and bvgC result in a Bvg- phenotype in B.pertussis. The bvgC-CAT fusions also confirm that transcrip-tion occurs from left to right as shown in Fig. 1.Sequences within the EcoRI-Xho I fragment (Fig. 1) are

sufficient for trans activation and modulation ofanjhaB-IacZtranscriptional fusion in E. coli (J.F.M., C.R., and S.F.,unpublished data). This suggests that bvg-specific compo-nents involved in signal transduction are present within theseboundaries. Sequences downstream from bvgC do not showany obvious ORFs, and linker insertions within this regionhad no effect on bvg activity. Knapp and Mekalanos (43) haverecently described the isolation of mutations in the bvg regionthat constitutively express virulence factors in the presenceof modulating conditions. The nature and location of thesemutations have not yet been reported.Amino acid sequence similarities suggest that BvgA and

BvgC belong to a large family of bacterial regulatory proteinsthat use conserved transmitter and receiver domains totransduce environmental signals. A growing body of evi-dence indicates that the mechanism of signal transductioninvolves autophosphorylation of transmitter proteins, fol-lowed by phosphorylation of regulator proteins that containreceiver domains (44-47). In relation to this "two-com-ponent motif," the bvg locus is interesting in several re-spects. Three components, one of which (BvgB) does notcontain a transmitter or receiver domain, are needed forregulation. Three linked loci (uhpABC) are also responsible

for exogenous induction of sugar-phosphate transport in E.coli (48), and the products of these genes share functionalsimilarities with the proposed Bvg proteins. Another unusualaspect of the B. pertussis system is the presence of atransmitter and a receiver domain in the same polypeptidesequence (see below).A model describing the proposed cellular locations, func-

tions, and interactions of the bvg products is shown in Fig. 4.It is suggested that BvgB and the N-terminal domain ofBvgC(residues 1-240) are localized in the periplasm. Residues60-276 of BvgB share striking similarities with the periplas-mic region of BvgC, indicating that these proteins mayinteract with a common factor, or with each other. According

p BvgC

CM

c IBvgA

(inactive)

. < Transmittermodule

Receivermodule

BvgA(active)

FIG. 4. Model for the regulation of virulence factors in B.pertussis. Details are presented in Discussion. P, CM, and C desig-nate periplasm, cytoplasmic membrane, and cytoplasm, respec-tively.

106

106

103

780

106

111

106

109

107

105

110

106

209

204

196

19

2.07

BvgB

Proc. Natl. Acad. Sci. USA 86 (1989) 6675

to our model, the transmembrane signal in BvgC (residues240-261) anchors this protein in the cytoplasmic membrane,and the C-terminal transmitter and receiver modules arelocated in the cytoplasm along with BvgA.

Several properties of the bvgA sequence suggest that itencodes a transcriptional activator. The presence of a re-ceiver domain and extensive homology with FixJ and UhpAsupport this contention. We have recently observed that thejhaB promoter is activated in E. coli by hyperexpression ofBvgA alone (C.R., J.F.M., and S.F., unpublished data),although both BvgB and BvgC are normally required forBvgA activity in both E. coli and B. pertussis.

In the absence of modulating signals, we propose thatBvgC activates BvgA. It is likely that this activation resultsfrom phosphorylation ofBvgA by a mechanism involving thetransmitter domain of BvgC. BvgB could exert its effect bydirectly interacting with BvgC, or by inactivating an uniden-tified inhibitor of BvgC. Although the modulating effect oftemperature could act at any level, the inhibitory effects ofsignals like MgSO4 or nicotinic acid may occur in the peri-plasm through an interaction with BvgB and/or BvgC.The role ofthe additional receiver module at the C terminus

of BvgC is unknown; however, it may be involved in mod-ulating the proposed kinase activity of the transmitter do-main. The frameshift mutation described above, which re-sults in a Bvg- phenotype, lies 10 residues downstream fromthis receiver (14). Deletion of sequences that encode the final26 amino acids of BvgC also disrupts bvg activity (J.F.M.,C.R., and S.F., unpublished data). A role for the C terminusof BvgC in DNA binding can also not be ruled out.

We thank D. Relman for advice, critical reading of the manuscript,and transatlantic communications and S. Ricci for oligonucleotidesynthesis. This work was supported by Damon Runyon-WalterWinchell Postdoctoral Fellowship DRG-911 (to J.F.M.), by a Na-tional Institutes of Health Public Health Service grant and a NationalInstitutes of Health National Research Service Award (to S.S.), andby National Institutes of Health Grant A129345 (to S.F.). B.A., R.G.,and R.R. were supported by a grant from ENI (Ente NazionaleIdrocarburi).

1. Miller, J. M., Mekalanos, J. J. & Falkow, S. (1989) Science243, 916-922.

2. Weiss, A. A. & Hewlett, E. L. (1986) Annu. Rev. Microbiol.40, 661-686.

3. Urisu, A., Cowell, J. L. & Manclark, C. R. (1986) Infect.Immun. 52, 695-701.

4. Livey, I., Duggleby, C. J. & Robinson, A. (1987) Mol. Micro-biol. 1, 203-209.

5. Goldman, W. E., Klapper, D. G. & Baseman, J. B. (1982)Infect. Immun. 36, 782-794.

6. Wardlaw, A. C. & Parton, R. (1983) Pharmacol. Ther. 19, 1-53.7. Weiss, A. A., Hewlett, E. L., Myers, G. A. & Falkow, S.

(1984) J. Infect. Dis. 150, 219-222.8. Glaser, P., Landant, D., Sezer, O., Pichot, F., Ullmann, A. &

Danchin, A. (1988) Mol. Microbiol. 2, 19-30.9. Nicosia, A. & Rappuoli, R. (1987) J. Bacteriol. 169, 2843-2846.

10. Weiss, A. A., Hewlett, E. K., Myers, G. A. & Falkow, S.(1983) Infect. Immun. 42, 33-41.

11. Weiss, A. A. & Falkow, S. (1984) Infect. Immun. 43, 263-269.12. Stibitz, S., Weiss, A. A. & Falkow, S. (1988) J. Bacteriol. 170,

2904-2913.13. Gross, R. & Rappuoli, R. (1988) Proc. Natl. Acad. Sci. USA 85,

3913-3917.

14. Stibitz, S., Aaronson, W., Monack, D. & Falkow, S. (1989)Nature (London) 338, 266-269.

15. Lacey, B. W. (1960) J. Hyg. 58, 57-93.16. Idigbe, 0. R., Parton, R. & Wardlaw, A. C. (1981) J. Med.

Microbiol. 14, 409-418.17. Ronson, C. W., Nixon, B. T. & Ausubel, F. M. (1987) Cell 49,

579-581.18. Simon, R., Priefer, U. & Puhler, A. (1983) Biotechnology 1,

784-791.19. Close, T. J. & Rodriguez, R. L. (1982) Gene 20, 305-316.20. Stibitz, S., Black, W. & Falkow, S. (1986) Gene 50, 133-140.21. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl.

Acad. Sci. USA 74, 5463-5467.22. Devereux, J., Haeberli, P. & Smithies, 0. (1984) Nucleic Acids

Res. 12, 387-395.23. Relman, D. A., Domenighini, M., Tuomanan, E., Rappuoli, R.

& Falkow, S. (1989) Proc. Natl. Acad. Sci. USA 86,2637-2641.24. Krikos, A., Mutoh, N., Boyd, A. & Simon, M. I. (1983) Cell 33,

615-622.25. Engelman, D. M., Steitz, T. A. & Goldman, A. (1986) Annu.

Rev. Biophys. Chem. 15, 321-353.26. Leroux, B., Yanofsky, F. M., Winans, C. S., Ward, E. J.,

Ziegler, F. S. & Nester, E. W. (1987) EMBO J. 6, 849-856.27. Kuhn, A., Wickner, W. & Kreil, G. (1986) Nature (London)

322, 335-339.28. Harley, C. B. & Reynolds, R. P. (1987) Nucleic Acids Res. 15,

2343-2361.29. Brendel, V., Hamm, G. H. & Trifonov, E. N. (1986) J. Biomol.

Struct. Dyn. 3, 705-723.30. Kofoid, E. C. & Parkinson, J. S. (1988) Proc. Natl. Acad. Sci.

USA 85, 4981-4985.31. Mizuno, T., Wurtzel, T. & Inouye, M. (1982) J. Biol. Chem.

257, 13692-13698.32. David, M., Daveran, M., Batut, J., Dedieu, A., Domergue, O.,

Ghai, J., Hertig, C., Boistard, P. & Kahn, D. (1988) Cell 54,671-683.

33. MacFarlane, A. A. & Merrik, M. (1985) Nucleic Acids Res. 13,7591-7606.

34. Makino, K., Shinagawa, H. H., Amemura, M. & Nakata, A.(1986) J. Mol. Biol. 192, 549-556.

35. Amemura, M., Makino, K., Shinagawa, H. & Nakata, A. (1986)J. Bacteriol. 168, 294-302.

36. Trach, K. A., Chapman, J. W., Piggot, P. J. & Hoch, J. A.(1985) Proc. Natl. Acad. Sci. USA 82, 7260-7264.

37. Ferrari, F. A., Trach, K., LeCoq, D., Spence, J., Ferrari, E. &Hoch, J. A. (1985) Proc. Natl. Acad. Sci. USA 82, 2647-2651.

38. Stock, A., Koshland, D. E. & Stock, J. (1985) Proc. Natl.Acad. Sci. USA 82, 7989-7993.

39. Drury, L. S. & Buxton, R. S. (1985) J. Biol. Chem. 260,4236-4242.

40. Friedrich, M. J. & Kadner, R. J. (1987) J. Bacteriol. 169,3556-3563.

41. Lipman. D. J. & Pearson, W. R. (1985) Science 227, 1435-1441.

42. Brennan, R. G. & Matthews, B. W. (1989) J. Biol. Chem. 264,1903-1906.

43. Knapp, S. & Mekalanos, J. (1988) J. Bacteriol. 170, 5059-5066.44. Ninfa, A. J., Ninfa, E. G., Lupas, A. N., Stock, A., Ma-

gasanik, B. & Stock, J. (1988) Proc. Natl. Acad. Sci. USA 85,5492-54%.

45. Hess, J. F., Oosawa, K., Kaplan, N. & Simon, M. I. (1988)Cell 53, 79-87.

46. Igo, M. M. & Silhavy, T. J. (1988) J. Bacteriol. 170, 5971-5973.47. Keener, J. & Kustu, S. (1988) Proc. Natl. Acad. Sci. USA 85,

4976-4980.48. Weston, L. A. & Kadner, R. J. (1988) J. Bacteriol. 170, 3375-

3383.

Genetics: Arico' et al.