INFECTION AND IMMUNITY, Aug. 1994, p. 3289-3298 Vol. 62, No. 80019-9567/94/$04.00+0Copyright ©D 1994, American Society for Microbiology

The Vibrio cholerae acfB Colonization Determinant Encodes anInner Membrane Protein That Is Related to a Family

of Signal-Transducing ProteinsKEITH D. EVERISS, KEITH J. HUGHES, MICHAEL E. KOVACH, AND KENNETH M. PETERSON*

Department of Microbiology and Immunology, Louisiana State UniversityMedical Center, Shreveport, Louisiana 71130

Received 14 March 1994/Returned for modification 29 April 1994/Accepted 21 May 1994

Vibrio chokrae accessory colonization factor genes (acfA, B, C, and D) are required for efficient intestinalcolonization. Expression ofacfgenes is under the control of a regulatory cascade that also directs the synthesisof cholera toxin and proteins involved in the biogenesis of the toxin-coregulated pilus. The gene for acfB wascloned by using an acfB::TnphoA fusion junction to probe a V. cholerae 0395 bacteriophage lambda library.DNA sequence analysis revealed that aciB is predicted to encode a 626-amino-acid protein related to the V.cholerae HlyB and TcpI proteins. These three Vibrio proteins have amino acid sequence similarity in a regionhighly conserved among bacterial methyl-accepting chemotaxis proteins. Analysis of the predicted AcfB aminoacid sequence suggests that this colonization determinant possesses a membrane topology and domainorganization similar to those of methyl-accepting chemotaxis proteins. Heterologous expression of aciB inEscherichia coli generates four polypeptide species with apparent molecular masses of 34, 35, 74, and 75 kDa.The 74- and 75-kDa proteins appear to represent modified forms of the full-length AcfB protein. The 34- and35-kDa polypeptide species most likely correspond to a C-terminal 274-amino-acid polypeptide that resultsfrom internal translation initiation of acfB mRNA. Localization studies with AcfB-PhoA hybrid proteinsindicate that AcfB resides in the V. cholerae inner membrane. V. cholerae acfB::TnphoA mutants display analtered motility phenotype in semisolid agar. The relationship between AcfB and Vibrio motility and the aminoacid similarities between AcfB and chemotaxis signal-transducing proteins suggest that AcfB may interact withthe V. cholerae chemotaxis machinery. The data presented in this report provide preliminary evidence that acpiencodes an environmental sensor/signal-transducing protein involved in V. cholerae colonization.

Asiatic cholera is a severe, potentially fatal diarrheal diseasecaused by the gram-negative bacterium Vibrio cholerae. Vcholerae remains an important human pathogen responsiblefor significant morbidity and mortality throughout the worldbecause of the limitations of water treatment facilities and thelack of effective prophylactic vaccines. Following the ingestionof contaminated food or water, V cholerae organisms colonizethe intestinal epithelial tissues of the human small bowel via acomplex and ill-defined series of events. To successfully colo-nize the small bowel, the vibrios must penetrate the protectivemucous gel and attach to and colonize the brush borders of theintestinal mucosa. The voluminous diarrhea associated withcholera is correlated primarily with the activity of choleraenterotoxin (reviewed in reference 10).V cholerae coordinately regulates the expression of a num-

ber of virulence and colonization determinants through acomplex and incompletely defined regulatory cascade (8, 9)involving at least two transcriptional regulators, ToxR (31, 32)and ToxT (19, 33). Two virulence factors known to be con-trolled by the ToxR/ToxT regulatory cascade are cholera toxin(28, 30, 31, 32) and toxin-coregulated pilus (TCP) (42, 46).ToxR mediates transcriptional activation of the cholera toxinoperon (cttAB) by binding to the nucleotide sequence lITlCGAT, which is repeated up to seven times in the promoterregion of the cholera toxin operon (32). Assembly of TCP,which is essential for colonization of the suckling mouse (46)and humans (18) by V cholerae, requires ToxR and additional

* Corresponding author. Mailing address: Department of Microbi-ology and Immunology, LSU Medical Center, 1501 Kings Highway,Shreveport, LA 71130. Phone: (318)-675-5753. Fax: 318-675-5764.

factors to control the expression of 15 genes necessary forefficient pilus biogenesis (34, 36).

In addition to controlling the synthesis of cholera toxin andTCP biogenesis, the ToxRiToxT regulatory cascade also coor-dinates the expression of the V cholerae accessory colonizationfactor (ACF) gene cluster (36). Disruption of any of the fouridentified acf genes (acfA, B, C, and D) by TnphoA reduces theability of V cholerae 0395 to colonize the intestines of sucklingmice (36). V. cholerae acf::TnphoA mutant strains exhibit anapproximately 10-fold decrease in colonization and a 10-foldincrease in the 50% lethal dose relative to the parental strain,V. cholerae 0395. The defect in colonization observed for Vcholerae acf mutant strains is not as severe as that caused bymutations that abolish TCP synthesis, leading to the proposalthat the acf gene cluster is responsible for the production of anACF. Presently, the molecular basis for reduced colonizationand virulence of V. cholerae acf mutants is unknown. Previousstudies in our laboratory have determined that the ACF andTCP gene clusters are physically linked (11), suggesting thatthe acf genes may encode a protein(s) that functions togetherwith TCP to promote vibrio colonization of the small bowel.To provide insight into the possible role of the V cholerae

acf gene products in vibrio intestinal colonization, we havefocused on molecular characterization of the ACF gene clus-ter. In this study, we report the isolation, characterization, andexpression of the acfB structural gene and the subcellularlocalization of AcfB-PhoA fusion proteins. The data presentedbelow suggest that the AcfB colonization determinant belongsto a family of proteins involved in environmental sensing andsignal transduction.

3289

on July 12, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

3290 EVERISS ET AL.

TABLE 1. Bacterial strains, plasmids, and bacteriophages used in this study

Strain, plasmid or phage Relevant characteristics Source or reference

Vibrio cholerae0395 Sm Str' 46KP3.51 0395 acfB::TnphoA Strr Kanr 38KP2.22 0395 acfB::TnphoA Strr Kanr 38KP1.17 0395 toxR12 Strr Ampr 38

Escherichia coliER1451 F' traD36 proAB+ lacIq AlacZM15IendA4 gyrA96 thi-1 hsdR2 (or R17) supE44 New England Biolabs

relAl A(lac-proAB) mcrBl mcrAKW251 F- supE44 galK2 galT22 metBI hsdR2 mcrBI mcrA (argA481::TnlO) recD1014 Promega

Tetr

PlasmidspBluescript II KS+ lacZ' Ampr StratagenepUC18 lacZ' Ampr Bethesda Research LaboratoriespB53.2 pBluescript II KS+ containing a 2.2-kbp EcoRI fragment from KP3.51 11pACFB pBluescript II KS+ containing a 2.0-kbp AvaIl fragment from 0395 This studypACFB-2 pBluescript II KS+ containing a 1.9-kbp NsiI-Avall fragment from pACFB This studypACFB-3 pBluescript II KS+ containing a 1.4-kbp MspI-Avall fragment from pACFB This studypA316 pBluescript II KS+ containing the 3' 1.0 kbp of acfB generated by Bal 31 This study

digestion of pACFB-3pA394 pBluescript II KS+ containing the 3' 0.8 kbp of acfB generated by Bal 31 This study

digestion of pACFB-3pACFB-4 pBluescript II KS+ containing a 0.4-kbp BglII-EcoRV fragment from pACFB This studypACFB-5 pBluescript II KS+ containing 0.5-kbp BglII-Avall fragment from pACFB This studypSPB-1 pUC18 containing a 5.0-kbp SalI-Pacl fragment from LambdaGEM-11-B-7 This studypB2-1 pBluescript II KS+ containing a 0.9-kbp Clal-PacI fragment from pSPB-1 This study

BacteriophagesLambdaGEM-11 Cloning vector PromegaLambdaGEM-11-B-7 LambdaGEM-1 1 containing a 15-kbp partial Sau3A fragment from 0395 This study

MATERIALS AND METHODSBacterial strains, plasmids, media, and reagents. The Esch-

erichia coli and V cholerae strains and plasmids used in thisstudy are listed in Table 1. V. cholerae and E. coli strains weremaintained at -70°C in Luria-Bertani medium (LB) contain-ing 25% (vol/vol) glycerol. E. coli strains were grown at 37°C inLB or M9 medium supplemented with amino acids, glucose,and thiamine. V cholerae were grown either in LB at pH 6.5and 30°C or in LB at pH 8.4 and 37°C as described previously(32). Antibiotics were used at the following concentrations:ampicillin, 100 ,ig/ml (LB) or 25 ,ug/ml (M9); kanamycin, 45,ug/ml; and rifampin, 200 ,ig/ml. EXPRE35S35S protein label-ing reagent ([35S]methionine/cysteine) was purchased fromNew England Nuclear, Boston, Mass. Isotopes for radiolabel-ing nucleic acids, [ot-32P]dATP and [a-35S]dATP, were pur-chased from Amersham, Arlington Heights, Ill. Unless other-wise indicated, all chemical reagents were obtained fromSigma Chemical Co., St. Louis, Mo.Genomic library construction and identification of acfB-

containing clones. Total DNA from V cholerae Ogawa 395(0395) was subjected to partial Sau3A digestion by enzymedilution as described by Maniatis et al. (26). A sample of eachreaction mix was separated on a 0.6% agarose gel and visual-ized with ethidium bromide staining. Samples enriched forfragments in the 10- to 20-kb range were pooled, extracted withphenol, precipitated with ethanol, and ligated into BamHI-digested LambdaGEM-1 1 vector DNA (Promega Corp., Mad-ison, Wis.). The ligated products were packaged into lambdaphage particles with an in vitro packaging kit (Promega Corp.)and amplified by using E. coli KW251 grown in the presence of0.4% maltose and 1 mM MgSO4. Recombinant phage plaquematerial was transferred to nitrocellulose filters and screened

by hybridization (26) with random-primed, [_-32P]dATP-la-beled DNA restriction fragments from the cloned acfB::TnphoA fusion junction (11).DNA sequence analysis. Restriction fragments from pACFB

were ligated into pBluescript II KS+. Plasmids carrying theappropriate inserts as determined by restriction enzyme anal-ysis were subjected to single- and double-stranded DNAsequencing using both strands as templates by the dideoxyoli-gonucleotide method (39) with the Sequenase version 2.0 kit(U.S. Biochemical Co., Cleveland, Ohio) with synthesizedsequencing primers. DNA sequence analysis was performedwith the MacVector (Eastman Kodak Co., New Haven, Conn.)and PCGENE (IntelliGenetics, Inc., Mountain View, Calif.)software packages.

Expression of aciB in E. coli. T7 analysis was performed asoutlined previously (44) except that expression was induced byinfection with M13mpl8 phage encoding T7 RNA polymerase(M13/T7) under the control of a Tac promoter (Invitrogen,San Diego, Calif.). E. coli strains containing the expressionplasmids pACFB, pA316, and pA394 were grown on M9 agarplates to ensure retention of the F' episome prior to expressionstudies. Cultures were grown to saturation in LB containingampicillin. Samples were diluted 1:100 with LB containingampicillin, grown to the mid-log stage, and harvested bycentrifugation. The cells were washed and resuspended in M9medium containing all common amino acids except methio-nine and cysteine at 0.01% and cultured at 37°C for 1 to 2 h.M13/T7 phage (at a multiplicity of infection of approximately10) and isopropyl-,B-D-thiogalactopyranoside (IPTG) (to a finalconcentration of 2 mM) were added simultaneously to induceexpression of the T7 RNA polymerase gene. After 30 min,rifampin was added to 200 ,ug/ml, and incubation was contin-

INFECT. IMMUN.

on July 12, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

1VIBRIO CHOLERAE acfB COLONIZATION DETERMINANT 3291

ued for 45 min prior to a 5-min pulse with 10 p.Ci of[35S]methionine/cysteine (EXPRE35S35S protein labeling re-agent; New England Nuclear Research Products, Boston,Mass.). Cells were harvested by centrifugation and resus-pended in sodium dodecyl sulfate-polyacrylamide gel electro-phoresis (SDS-PAGE) solubilization buffer (60 mM Tris-Cl[pH 6.8], 1% mercaptoethanol, 8 M urea, 10% glycerol, 3%SDS, 0.01% bromophenol blue). Solubilized cell extracts wereboiled for 10 min prior to SDS-PAGE. Gels were treated withEnlightning (New England Nuclear Research Products), dried,and exposed to Kodak X-Omat film.

Nested 5' deletions within acJtB were generated by Bal 31exonuclease digestion of pACFB-3 from the BamHI site in thevector polylinker. Following incubation with Bal 31 exonucle-ase, the samples were end repaired with the Klenow fragmentof DNA polymerase (26), and the insert was liberated bydigestion with HindIII. The reaction mixtures were separatedby agarose gel electrophoresis, and the insert DNA wasisolated by using GeneClean II (Bio 101, La Jolla, Calif.),ligated with EcoRV- and Hindlll-digested pBluescript KS+,and transformed into E. coli ER1451. Plasmid clones wereanalyzed by agarose gel electrophoresis and DNA sequenceanalysis by standard procedures (26).

Subcellular fractionation of V. chokrae KP2.22. V. choleraeKP2.22 cells were fractionated by the method of Kelley andParker (21) with some modifications. Briefly, V. choleraeKP2.22 cells grown in LB, pH 6.5, at 30°C were harvested bycentrifugation, washed with 0.01 M HEPES (N-2-hydroxyeth-ylpiperazine-N'-2-ethanesulfonic acid) buffer, pH 7.4, andstored at -70°C. Frozen pellets were thawed on ice andsuspended in 0.1 original culture volume of HEPES buffer, pH7.4, containing 18% (wt/wt) sucrose. DNase (3 p.g/ml) wasadded to the cell suspension, and the reaction mix wasincubated for 30 min at 4°C. The cell suspension was thenpassed through a French pressure cell at 20,000 lb/in2. Undis-rupted cells were removed by centrifugation at 5,000 x g for 20min at 4°C. The cell lysate was diluted with an equal volume ofHEPES buffer and layered onto sucrose step gradients con-taining 1 ml of 55% (wt/wt) sucrose and 4 ml of 14% (wt/wt)sucrose in HEPES buffer. The gradients were centrifuged for 2h, at 33,000 rpm in an SW41 rotor at 4°C. The top 3 ml fromthe tubes was collected and incubated at 23°C for 30 min priorto centrifugation at 200,000 x g for 1 h to yield the solublefraction (cytoplasm and periplasm). The membrane fractionwas collected and washed in 3 volumes of HEPES buffer andcentrifuged at 200,000 x g for 1 h on a 55% (wt/wt) sucrosecushion to yield the total envelope fraction. Inner and outermembrane fractions were isolated by isopycnic sucrose densitygradient fractionation (40). Briefly, samples (approximately 5mg) of protein were layered onto a sucrose step gradientconsisting of 1 ml of 55%, 2.5 ml of 50%, 2.5 ml of 45%, 2.5 mlof 40%, and 2.5 ml of 3.5% (wt/wt) sucrose in HEPES bufferand centrifuged for 24 to 30 h in an SW41 rotor at 33,000 rpm.Fractions were collected dropwise from the bottom of thecentrifuge tubes. Membrane-containing fractions were identi-fied by quantitating the protein concentration of each fractionwith the bicinchoninic acid reagent as recommended by Pierce(Rockford, 111.). The outer membrane and soluble fractionswere less than 15% contaminated by inner membranes, asdetermined by succinate dehydrogenase assays (35). Inner andouter membrane fractions were also isolated by N-lauroylsarcosine extraction of total membrane fractions (12). Approx-imately 10 mg of total envelope protein in 0.5 ml of 0.01 MHEPES, pH 7.4, was added to 0.5 ml of 1% N-lauroyl sarcosinein the same buffer. Samples were incubated at 37°C for 1 h withoccasional vortexing. Insoluble material (outer membrane)

was collected by centrifugation at 200,000 x g for 1 h at 4°C.Samples were examined by SDS-PAGE and stained withCoomassie brilliant blue to confirm separation of inner andouter membranes.

Alkaline phosphatase activity. Enzyme activity was deter-mined by adding samples of the subcellular fractions in 20 p.l of0.01 M HEPES buffer (approximately 5 p.g of protein), pH 7.4,for each sample to reaction mixtures containing 160 p.l of 1 MTris-Cl (pH 8.0) and 20 .1I of 0.4% p-nitrophenyl phosphate.Reaction mixes were incubated at 37°C for 20 min, after which20 p.l of 1 M K2PO4 was added to stop further color develop-ment. The A421) was determined and referenced against theA570. Alkaline phosphatase activity is expressed as [(A410 -

A570)/time] x 1,000.Swarm plate assays. LB agar (0.3%) plates, pH 6.5 or 8.4,

were inoculated by stab with the appropriate strains andincubated at 37 or 27°C for 12 or 24 h, respectively.

Nucleotide sequence accession number. The acJB nucleotidesequence has been deposited in the GenBank database underaccession number L25660.

RESULTS

Cloning of the gene encoding AcfB (acJB). The acfB genewas initially identified by the TnphoA insertion KP3.51 (36),which lies within acJB and creates a gene fusion between acfBand phoA marked by the adjacent Tn5-encoded Kmr gene. Wehave previously reported the isolation and nucleotide sequenceof the acfB::TnphoA fusion junction from V. cholerae KP3.51(11). To isolate the intact acfB gene, a V. cholerae genomiclibrary was constructed in bacteriophage LambdaGEM-1 1.Recombinant bacteriophage carrying acfB sequences wereidentified by in situ hybridization of phage plaques with a0.66-kbp ClaI fragment from pB53.2 (Table 1) as an acfB-specific probe. A single recombinant bacteriophage, the Lamb-daGEM-11 clone designated LambdaGEM-11-B-7, was iso-lated and characterized by restriction and Southern blotanalyses (data not shown). A 1.9-kb Avall fragment fromLambdaGEM-11-B-7 was inserted into the EcoRV site ofpBluescript II KS+, and the recombinant plasmid was desig-nated pACFB (Table 1, Fig. 1).DNA sequence of acJB. To characterize and define acfB,

restriction fragments of LambdaGEM-11-B-7 DNA were in-serted into pBluescript II KS+ and sequenced by the dideoxy-oligonucleotide method (39). The nucleotide sequence foracfB was determined for both strands. The DNA sequencedata were in complete agreement with DNA sequence infor-mation obtained previously from a cloned acfB::TnphoA fusion(KP3.15) (11). The acfB nucleotide sequence and deducedprotein sequence are illustrated in Fig. 2. The sequence datareveal that acJ3 consists of a 1,878-nucleotide open readingframe that is preceded by a characteristic Shine-Dalgarnoelement and encodes a 626-amino-acid-residue protein with a

predicted molecular mass of 69,319 Da and isoelectric point of4.36. As noted previously (11), the ATG initiation codon ofacfB is located 9 bp downstream from the end of tcpJ (20). Theterminating TAA codon precedes the start of acfC by 14 bp(Fig. 2). The proximity of acfB to tcpJ and acfC and the absenceof characteristic terminating structures in the correspondingmRNA suggest that tcpJ, acfBl, and acfC are encoded on a

polycistronic message.Two regions within the predicted AcfB polypeptide exhibit

properties characteristic of membrane-spanning segments. Thefirst is located near the N terminus of the protein, spans aminoacid residues 7 to 25, and represents the leader peptide of thisprotein. This region lacks a characteristic signal peptidase I

VOL. 62, 1994

on July 12, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

3292 EVERISS ET AL.

AH B P

II IP B RV B RV RV RV

l1 I Il I I Iac r fY|laYc I or I orfX I fA I I D |o

4 - - *~~~~~~~~~~~~2.

B 3.51 2.22

RV* A* N PV P

.½\1 I I IB RV C A RV

I II1/PACFB I'-

pA316pT7

pA394

pT7

A

acfB

I

I

PT7 A

500 basepairs

FIG. 1. (A) Physical organization of the ACF gene cluster. Restriction sites are indicated above the line. Dark bars at the 5' ends of genesindicate the presence of a leader peptide, as determined by PhoA fusion protein activity and nucleotide sequencing. Arrows indicate the directionof transcription. (B) T7 RNA polymerase/promoter constructs. The V. cholerae DNA in each plasmid is designated by a thick line, and vectorsequences (pBluescriptIl KS+) are denoted by thin lines. Deleted sequences are noted by a A below a stippled line. TnphoA insertion sites areindicated by the numbers 3.51 and 2.22. T7 polymerase promoters are noted by an arrow labeled pT7. Restriction sites: RV, EcoRV; H, HindIII;B, BglII; X, XhoI; C, ClaI; A, AvaII; N, NsiI; P, Pacl; PV, PvuI. Asterisks denote restriction sites that have been disrupted.

cleavage site (47), suggesting that it is not removed from thenascent AcfB polypeptide. A second contiguous stretch ofhydrophobic amino acids is located between amino acids 274and 295. Both regions are predicted (22, 23, 37) to representmembrane-spanning segments. The two identified TnphoAinsertions in acfB (Fig. 1) map to the region bounded by thesehydrophobic stretches. These data, combined with the require-ment that alkaline phosphatase (PhoA) be extracytoplasmic inorder to express enzymatic activity, suggest that the firsthydrophobic region of AcfB represents a secretory leaderpeptide, which, in the absence of signal peptide removal,anchors AcfB in the inner membrane. The following 250 aminoacids bounded by the second hydrophobic stretch would belocalized to the periplasmic space. The absence of a thirdcontiguous stretch of hydrophobic amino acids suggests thatthe remaining portion of AcfB corresponds to a cytoplasmicdomain. The results of limited TnphoA and TnlacZ mutagen-esis of acfB support this prediction. Enzymatically activeAcfB-PhoA fusions occur only when PhoA is fused to theputative AcfB periplasmic domain. Conversely, enzymaticallyactive AcfB-LacZ hybrid proteins are produced only whenLacZ is fused to the predicted cytoplasmic domain of AcfB(data not shown).Amino acid sequence similarity to prokayotic signaling

proteins. Computer-aided searches of the SWISS-Protein(SWISS-PROT) and National Center for Biotechnology Infor-mation (NCBI) databases revealed similarity between AcfBand prokaryotic methyl-accepting chemotaxis proteins (MCPs).MCPs exist as homodimeric integral inner membrane proteinspossessing two membrane-spanning regions that organize the

proteins into distinct periplasmic ligand-binding and cytoplas-mic signaling domains (6, 15). The most significant amino acidsequence similarity between AcfB and the MCPs is restrictedto the carboxyl half of AcfB in a region of the MCPs referredto as the highly conserved domain (HCD) (1) (Fig. 3). TheHCD of MCPs is contained within the cytoplasmic portion ofthese proteins and is believed to interact directly with CheW,a modulator of the chemotaxis signaling pathway (24). Recentstudies have shown that closely related HCD sequences arepresent in other bacterial proteins with primary functions notthought to be involved in chemotaxis (3, 7, 14). These includeTcpI of V cholerae (Fig. 4) and PilJ ofPseudomonas aeruginosa(7), two proteins involved in type 4 pilus biogenesis (7, 14). Anadditional V cholerae HCD-containing protein, HlyB (Fig. 4),is required for efficient secretion of the hemolysin HlyAdeterminant (3).AcfB exhibits 24, 27, and 20% amino acid sequence identity

with TcpI, HlyB, and the E. coli serine transducer Tsr, respec-tively. The overall relatedness between AcfB and these pro-teins is 39% (TcpI), 44% (HlyB), and 38% (Tsr). Sequenceidentity to the AcfB HCD is 75% for TcpI, 67% for HlyB, and65% for Tsr (Fig. 3 and 4). Flanking the HCD of bacterialchemosensors are two regions, designated Kl and Rl, whichbecome methylated and demethylated at the second residue ofGlu-Glu pairs in response to environmental stimuli (38). Threeof the four methylated glutamate-glutamate or glutamine-glutamate (deamidation yields Glu-Glu pairs) pairs within theTsr chemosensor (38) are not conserved in AcfB (Fig. 3). It isnot presently known whether AcfB is methylated.The greatest variability in the reported MCP amino acid

. ___j- - -

INFEC'T. IMMUN.

I

I

on July 12, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

VIBRIO CHOLERAE acfB COLONIZATION DETERMINANT 3293

tcpJ> acfB31 GGACCAAGCATTATTATCTCTC=ATT7rrrATCTTTAATlACMTGAAGT""CTATCAAGTTACTAATIG

G P S I I I S P V I V P F S I R L N M K P S I K L L M

91 ATFTACA¶CAATTATAATTACTATTTCATCTATCCTTACTACTTrCAGATGCATGGACCAATGAATATCTTGATCGTACGATAAAAI F T S I I I T I S S I L T Y F Q M H G T N E Y L D R T I K

181 GATACTATTTCTGAAACGCTAGACITCTrGGAAGATAAAATAAATCTFTCrATTGAGTCCAAGGTCGATGTGTCAAAATCTGTTATCGACD T I S E T L D S L E D K I N L S I E S K V D V S K S V I D

271 ATTTATGAAAArACTGGAGGGGGAATCCCCTCTCAGTAAATTTAAGGATGTTGATGTCAATCATCTTTCAAATGTrrTCAATTGTTCI Y E N L L E G E S P L S K F K D V D V N H L S N V F Q L F

361G Y A D E R T G E I I T N D P N F K V P T G F D P R T R S W

451 TATTTAAATGCAAAAAAATTGAATACCTTTAGTTTOTCTGAGCCGTATGTCGATCTCATTACAGAGAAGTTAATGGTGACAACAAGTGCAY L N A K K L N T F S L S E P Y V D L I T E K L M V T T S A

541P I Y N K N D L T G V I FF D I P L D D V Q E L I K S Y N P

631 TTTGATGCCGGCACGATTTTTATCGTTGATAACAGTGIAATAATTTTCGGTAATAAAAATGATATATCGGGTAAAAATTTATTTGGAP D A G T I F I V D N S G K I I F G N K N D I S G K N L F G

721 GACTTTGATAGCTTCCCTCTATCGGTAAGTGAGTCAAAAACGAAAGATAAAAATGGGGTAAATTACGATGTATTCATCAAAATGTCAGACD F D S F P L S V S E S K T K D K N G V N Y D V F I K M S D

811 TTTGGCGATGGAATCTGTTTCTATTATCGACCATGATAAAGCACGCTCTGATATTATCACGTTGAGAAACAATAGTATATTTACTGCCF G D W N L V S I I D H D K A R S D I I T L R N N S I F T A

901 GTCATTTTAGCAAGTGTTrTCTTTGCTATCTTGTTGI TACTATGCGGTTAATGCTGAAGCCATrCCATCAATTAACCGATGCAATGGTA

991

V I L A S V F F A I L L F T M R L M L K P L H Q L T D A M V

=================IPA316==========AATATTTCATCGGGCAGTYCTGWITCTTACGGTTCGTATCCCGAATAGTACGGATCAAGAGITTCAAAAATAATAAATTCCTTTAACATTN I S S G S A D L T V R I P N S T D Q E F S K I I N S F N I

1081 TTTTGTAATCTGCAATCTATTGTGTCO9S aAAAAT GAACTCAGAAAAAATAAATTGCATAACTACTGAAACACAAGAGCTAGTCF V G N L Q S I V S E V K M N S E K I N C I T T E T Q E L V

I- PA3941171 GAAGTTTCCAATAATAGCGTCATGGATCAGTATCGTGAACTTGATATGCTAGCTAGCTCAATGAATGAGATGGTGGCAACATCTAATCAG

E V C N N S V M D Q Y R E L D M L A S S M N E M V A T S N Q

1261 ATTGCACAAATTACATCCGAAGCATCAGAGATCACATCTAAGATTAACGGTCAAGTGAATGAAGGAGTCGGTGCTGTTTCTTCTGTTACCI A Q I T S E A S E I T S K I N G Q V N E G V G A V S S V T

1351 GAAAGTGTAGGTAATTTAGTTGAGAAGTPGGATAAGACGAAATCTG'rATCCAAGATCTPAATCGCCAAACTCAAAATATAGATGTTATTE S V G N L V E K L D K T K S V I Q D L N R Q T Q N I D V I

1441 TTGAAGGCTATTAATGATATTTGACCAAACAAACCTTC?TGCTCTTAATGCTGCAATTGAGGCTGCACGAGCTGGGGAAAATGGACGAL K A I N D I A D Q T N L L A L N A A I E A A R A G E N G R

1531G F A V V A D E V R S L A I K T Q E S T K N I G S I I H I L

1621 CAAGAAAACTCTTTGCTATCCGTTCATGTTATGGATGAAAGT"TTAATATTGCCTCTGAAACTATGACTATATCAGCTGATTCAAAACAAQ E N S L L S V H V M D E S F N I A S E T M T I S A D S K Q

1711 TGTTTAGACAATATTAGTCAATCAGTAATACAAATTGTTGATATCACAAATCAAGTAGCTACAGCTGCTTACGAACAAAGTCACGTATCAC L D N I S Q S V I Q I V D I T N Q V A T A A Y E Q S H V S

1801 GAGGAAATAAACAGCAACTCTATTTCTATAAAAGATAAAGCAGATACCTTATCAAG1TGGGTAATAAGATTTCTCAACAGGCTTATTCTE E I N S N S I S I K D K A D T L S S L G N K I S Q Q A Y S

aofC1891 CAAAAAGCACTGATTGGTCATCAAGATGATrTAATCAGTAAGTTTATCATTTAATATl:&TOaATTTCATGAAAAGTAAGAATCGATmls

Q K A L I G H Q D D L I S K F I I M K S K N R F L

1981

9

39

69

99

129

159

189

219

249

279

309

339

369

399

429

459

489

519

549

579

609

626

L L I S L L S F S T S V F A D V N L Y G

FIG. 2. Nucleotide sequence of the V. cholerae acfB gene. Numbers on the left indicate nucleotide positions. Numbers on the right indicateamino acid positions within AcfB. Putative Shine-Dalgarno sequences are underlined. Gene designations are indicated above the nucleotidesequence. Predicted membrane-spanning regions of AcfB are denoted by a double underline. The deletions in plasmids pA316 and pA394 aredesignated by arrows.

sequences is present in the receptor ligand-binding periplasmicdomains. Only limited amino acid sequence similarity was

detected between AcfB and the periplasmic domains of theMCPs or other HCD-containing proteins (data not shown).Similarity searches comparing the first 300 amino acid residuesof AcfB with those of proteins in the NCBI and SWISS-PROTdatabases failed to identify significant similarities. Limitedregions within the putative periplasmic domain of AcfB were

found to be similar to sequences within TcpI (Fig. 4). Aminoacid residues 154 to 207 in AcfB and 149 to 195 in TcpI show

20% identity and 59% overall similarity. This region fails toalign with any sequences in HlyB (Fig. 4).

Heterologous expression of the V. cholerae acfB gene in E.coli. The T7 promoter provided by the pBluescript II KS+plasmid derivative pACFB-1 (Fig. 1) was used to examineAcfB synthesis in E. coli. The T7 promoter was induced byinfection with recombinant M13/T7 phage and addition ofIPTG. Expression of pACFB-1 resulted in the synthesis of fourpolypeptides. Two protein doublets with apparent molecularmasses of 74 and 75 kDa (74/75-kDa doublet) and of 34 and 35

VOL. 62, 1994

on July 12, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

3294 EVERISS ET AL.

AcfB - MKFSIKLLMIFTSIIITISSILTYFQMHGTNEYLDRTIRDTISETLDSLEDKINLII .I^ .

I I^Tsr - MLKRIKIV---TSLLLVLAVFGLLQLTSGGLFFNALKNDKENFTV---------

TM1AcfB - SIESKVDVSKSVIDIYENLLEGESPLSKFKDVDVNHLSNVFQLFGYADERTGEII

1~ . A.1 A 11 A

Tsr - -LQTIRQQQSTLNGSWVALLQTRNTLNR---------- AGIRYMMDQNNIGSGST

-55

-42

-110

-86

AcfB - TNDPNFKVPTGFDPRTRSWYLNAKKLNTFSLSEPYVDLITEKLMVTTSAPIYNKN -165^^ I 1^^1 "I

Tsr - VAELMESASISLKQAEKNW------------- ADYEALPRDPRQSTAAAAEIKRN -128

AcfB- DLTGVIFFDIPLDDVQELIKSYNPFDAGTIFIVDNSGKIIFGNKNDISGKNLFGD -220III I III

.

Tsr - -------YDIYHNALAELI--------- LLGAGKINEFFDQPTQGYQDGFEKQ -166

AcfB - FDSFPLSVSESKTKDKNGVNYDVFIKMSDFGD--WNLVSIIDHDKARSDIITLRN.1. 11 . 11 ^

Tsr - YVAYM-------EQNDRLHDIAVSDNNASYSQAMWILVGV--------------

AcfB - NSIFTAVILASVFFAILLFTMRLMLKPLHQLTDAMVNISSGSADLTVRIPNSTDQ^^^li^ ^ I I^ ^I 11

Tsr - ----- MIVVLAVIFAVWFGIKASLVAPMNRLIDSIRHIAGG--DLVKPIEVDGSN

TM2AcfB - EFSKIINSFNIFVGNLQSIVSEVKMNSEKINCITTETQELVEVCNNSVMDQYREL

^ I I. -

Tsr - EMGQLAESLRHMQGELMRTVGDVRNGANAIYSGASEIATGNNDLSSRTEQQAASL

AcfB -

Tsr

AcfB -

Tsr

DMLASSMNEMVATSNQIAQITSEASEITSKINGQVNEGVGAVSSVTESVGNLVEK^ 1^11 111^IIII ^

EETAASMEQLTATVKQNAENARQASHLALSASETAQRGGKVVDNVVQTM-* Kl *

LDKTKSVIQDLNRQTQNIDVILKAINDIADQTNLLALNAAIEAARAGENGRGFAV1^ ^1 1^I 11 11

--------RDISTSSQKIADIISVIDGIAFQTNILALNAAVEAARAGEQGRGFAVHCD

AcfB - VADEVRSLAIKTQESTKNI------------- GSIIHILQENSLLSVHVMDESF11 III 11 - ^I^^

Tsr - VAGEVRNLAQRSAQAAREIKSLIEDSVGKVDVGSTLVESAGETMAEIVSAVTRVT

-273

-199

-328

-247

-383

-302

-438

-351

-493

-398

-534

-453

AcfB - NIASETMTISADSKQCLDNISQSVIQIVDITNQVATAAYEQSHVSEEINSNSISI -589^ ^ ^l ^1^1 1^ . " ^ .

Tsr - DIMGEIASASDEQSRGIDQVGLAVAEMDRVTQQNAALVEESAAAAAAIEEQASRL -508* Rl *

-SSLGNKISQQAYSQKALIGHQDDLISKFII -626^l . ..

-551TEAVAVFRIQQQQRETSAVVKTVTPAAPRKMAVADSEENWETF

FIG. 3. Sequence alignment of V cholerae AcfB with the Tsr MCPof E. coli. Identical residues are marked by vertical lines, and con-served residues are marked by carets. Dashes indicate gaps inserted inthe sequences to improve the alignments. The underlined residueslabeled TM1 and TM2 correspond to membrane-spanning regions.The HCD of Tsr is underlined and labeled. The Kl and Rl domains of

Tsr are underlined and labeled. Sites of methylation are noted withasterisks.

kDa (34/35-kDa doublet) were observed (Fig. 5). The 74/75-kDa doublet most likely represents modified versions of the

full-length AcfB (626 amino acids). SDS-PAGE analysis of the

methylated and demethylated forms of MCPs shows that

methylation results in minor shifts in electrophoretic migration(27, 38); thus, it is possible that the 74- and 75-kDa proteinsobserved represent methylated species of AcfB.The 34/35-kDa doublet could be the product of a proteolytic

event that would cleave AcfB into essentially equivalently sizedportions. Alternatively, these two protein species may repre-sent the products of internal translation initiation within theacfB transcript. Preliminary evidence for the second possibilitywas provided by expression studies with pACFB-3, which lacksthe N-terminal coding region of acfB (Table 1). The produc-tion of the 34- and 35-kDa proteins upon expression of thetruncated acfB gene in pACFB-3 (data not shown) indicatedthat these proteins result from internal translation initiation.To localize the site of initiation, deletions were introduced intothe 5' end of the acfB gene present in pACFB-3. T7 promoter/polymerase-directed expression of truncated versions of acJB(Fig. 1) in which the coding sequences for the first 316 (pA316)and first 394 (pA394) amino acids have been deleted localizedthe initiation site to within 234 nucleotides. Within this region,there are six potential translation initiation codons (ATGs andGTGs) that could yield significant proteins. Only one of thepotential initiation sites is preceded by a suitable Shine-Dalgarno element (43) (Fig. 2), making it the most likely site of

translation initiation. Translation initiating at this positionwould yield a protein consisting of the C-terminal 274 aminoacids of AcfB, with a predicted molecular mass of 29 kDa.Interestingly, the 34/35-kDa doublet appears to be present ingreater quantities than the 74/75-kDa doublet. This may reflectmore efficient recognition of the internal ribosome-binding siteby the E. coli translation machinery. The 34/35-kDa doubletdetected in these assays probably results from the samemechanism that alters the migration of full-length AcfB. Thisidea is consistent with the fact that the Ki and Rl methylationsites are still present in the pA316 construct.

Localization of AcfB-PhoA fusion proteins in V. choleraeKP2.22. Western blot (immunoblot) analysis of V choleraeKP2.22 (acfB::TnphoA) and KP3.51 (acJB::TnphoA) with alka-line phosphatase-reactive antiserum revealed that strainKP2.22 synthesized a stable AcfB-PhoA hybrid protein (datanot shown) and was therefore chosen for localization studies.Enzymatic analysis of V cholerae subcellular fractions indi-cated that the alkaline phosphatase activity associated with theAcfB-PhoA fusion was enriched in the total membrane frac-tion (inner and outer membranes) of strain KP2.22 (Fig. 6).Western blot analysis of the V. cholerae subcellular fractionsrevealed that the proteins detected by PhoA-specific antiserumin the soluble fraction (cytoplasm and periplasm) comigratedwith E. coli alkaline phosphatase (data not shown), indicatingthat the PhoA activity associated with the soluble fraction is aresult of fusion protein degradation. AcfB-PhoA fusion pro-teins present in the total membrane fraction were of the sizeexpected for the full-length fusion protein (data not shown).Further fractionation of the total envelope preparations byusing isopycnic sucrose density gradients or extraction withN-lauroyl sarcosine (12) revealed that the AcfB-PhoA hybridproteins were enriched in the cytoplasmic membrane fraction(Fig. 6). The localization of the KP2.22 AcfB-PhoA fusionprotein to the V. cholerae inner membrane fraction supportsour earlier contentions that the AcfB signal peptide is notcleaved and that the hydrophobic N-terminal hydrophobicamino acids serve as a membrane anchor to localize AcfB tothe inner membrane.AcfB affects V. cholerae swarm plate activity. The relation-

ship between AcfB and MCPs and the observation that V.cholerae cells grown under conditions in which the ToxRregulon is maximally expressed (LB, pH 6.5) display reducedswarming when inoculated into semisolid (0.3% agar) LB agarplates (13) prompted us to examine V. cholerae KP3.51(acJB::TnphoA), KP2.22 (acJ3::TnphoA), and KP1.17 (toxR12)(Table 1) for altered swarm plate phenotypes. The swarmingpatterns of four V. cholerae strains are depicted in Fig. 7. V.cholerae KP1.17 carrying an insertional disruption of the toxRstructural gene and 0395 derivatives with TnphoA insertions inacJB display increased swarming activity in semisolid agarcompared with the parental strain 0395. This effect is mostpronounced under in vitro culture conditions promoting syn-thesis of ToxR regulon determinants (LB, pH 6.5) (Fig. 7) andless evident under conditions resulting in reduced expressionof ToxR-activated genes (LB, pH 8.4) (data not shown). Theseresults suggest that the effect of toxR and acfB mutations is anincrease in swarming on the part of the mutant vibrioscompared with that of the wild-type 0395 cells. We do notbelieve that the increased swarming of acJB::TnphoA mutantsis the result of polar effects on downstream genes, sinceTnphoA insertions in acfC do not affect the swarming patternof V. cholerae (data not shown). Additional evidence support-ing the idea that AcfB affects vibrio swarming is provided byexperiments in which V. cholerae 0395 cells were transformedwith a pBR322 derivative that constitutively expresses acfB.

INFECT. IMMUN.

on July 12, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

VIBRIO CHOLERAE acfB COLONIZATION DETERMINANT 3295

T11M ---KFS IKLLIFTSII ITISSILTYFQMGTNEYLDRTIKDTISETLDSLEDKMIINKFSLKWMLAIAVAIPAIRLLFVAFTSLNTMSVMQAQSNSLYANTAAPMRAMMIaIISV-FLLLACI ITSAFTAFFYHSKLSDQTKSISSLSSQQAQERLQSYQD-* r* . .. . . .. . . .

AcfB INLSIESKVDVSKSVIDIYENLLEGESPLSKFKDVDVN-HLSNVFQLFGYADERTH1yB A----EATSRIPRMRVGI-DMMLLQETALKDAKGV----LKRV------EEARTTcpI -SLDFYKKLNIS-LSVAIANSL--RDKAVEELNAIALRIQENHGFIGVTFASLDG

*

GEIITNDPNFKVPTGFDPRTRSWYLNAKKLNTFSLSEPYVDLITEKLMVTTSAPIEDIPEMRQRMQVAVDSQVNP------ ELKEQARKLQARFEQMVREEL-------TMFTDIGT---LDWNAKTLRRDWFVKTVELGTKHYTAFDIDKTTGQHVLTIATPV

YNKNDLTGVI FFDIPLDDVQELIKSYNPFDAGTI FIVDNSGKI I FGN--------------------------- EPMLQA-------------------FAN-------YVGNDIVGSVALDIAGDQIASP------- NGSGMFIMMTDRNFNVFASDLTHSTLI

* ~~~~~~~~~*-

AcfB -KNDISGKNLFGDFDSFPLSVSESKTKDKNGVNYDVFIKMSDFGDWNLVS IIDHDHlyB -NDMTTAQNIYRD--KYAPTYGEMRKQANQIL--DTLLQQADRQNHASVESFEAGTcpI GKDLTKEKPLFKNL--VSGQYVTFSDADSHWFAVSQTEIDGENKLFTI IDIQQIV

T2AcfB KARSDI ITLRNNSI FTAVILASVFFAILLFTMRIMLKPLHQLTDAMVNI SSGSADHlyB RTKQMVI------IAAGLIIS--FITSLVIITNLR-SRVAYLKDRMSS-AAANLSTcpI QTYKRDIQLI IAGFSG----FSCVMLIGLYWVLSKELSGVRQIREWILSLSDG--Q

* *

LTVRIPNSTDQEFSKI INSFNI FVGNLQSIVSEVKMNSEKINCITTETQELVEVCLRTRLELDGNDELCDIGKSFNAFIDKVHHSIEEVAENSKELATMASSVSQRAHMTIKERRPIKFHNELDTIAQSLENLQFRLLDVVRNSHRTMNDLSIKQTDITYSIEGN

* * * *

NNSVMDQYRELDMLASSMNEMVATSNQIAQITSEASEITSKINGQVNEGVGAVSSQSNCASQRDRTVQVATAIHELGATVSEIASNAAMARDVANEATLHSGEGKKVVGETNNSQQELGLIEQVATATTQLSVTSFDVMQQAQSAELNAETAQKLIAESHDI IDS

* * * *

HCDVTESVGNLVEKLDKTKSVIQDLNRQTQNIDVILKAINDIADQTNLLALNAAIE.AAVQNRIQTLVNELDNATQVVSSLATQINGISSTLDTIRSISEQTNLLALNAAIEAAoexe!Ymw%m"nCTetvww^T rT^T 01> S<! nXGT C! entYTr TktTIM C! rnl^fPlT TAtWA A T P k

* * * * **************

RAGENGRGFAVVADEVRSLAIKTQESTKNIGSIIHILQENSLLSVHVMDESFNIARAGEQGRGFAWADEVRTLASRSAASTEBIQQVINRLQTESTRAVRAMKGRSQSRAGEQGRGFAVVADBVRSLAVKTQQSTIDIQGIILKLQEQSQIADQVMTRNVSLE

~****** **** ** ** * * ** * *

SETMTISADSKQCLDNI SQSVIQIVDITNQVATAAYEQSHVSEEINSNS ISIKDKDVVVEFSAKANQSLTEINSQIDQINDQNIQVATATEEQSTVVEDINRNVEDINQLHETQVANRALIASFNLI SDKVLEISNINSIVSTAANEQKIVTEDVAKQMEDIRYL

* t * ** ** * * *

525553

10694

104

161144156

208144204

262194257

317239307

372294362

427349417

482404472

537459527

592514582

ADTLSSLGNKISQQAYSQ1KLIGHQDDLISKFII---- 626

TTETSHVADELSRASASLQRLSSQLDKLVGSFEL---- 548VQENLSAMERTKQANQNISDLTTNLNDALSFFKIELTS 620*.. - - * * - * - * -

FIG. 4. Multiple alignment of AcfB, Tcpl, and HlyB. Residues conserved in all three sequences are designated by asterisks. Highly conservedresidues are designated by dots. The proposed membrane-spanning regions are underlined and labeled TM1 and TM2. The HCD is underlinedand labeled.

These cells display reduced swarming activity compared withV. cholerae 0395 transformed with pBR322 when culturedunder conditions that repress the synthesis of ToxR-regulatedproteins (data not shown).

DISCUSSIONColonization of the intestinal epithelial tissues lining the

small bowel is a crucial step in V cholerae pathogenesis. Weare characterizing a cluster of genes encoding proteins thatpromote efficient colonization of the small bowel by V. chol-erae. This report describes the isolation and characterization of

the acfB gene and the subcellular localization of AcfB-PhoAhybrid proteins. The nucleotide sequence of the acfB generevealed a 626-amino-acid reading frame (Fig. 2) which couldyield a polypeptide with a predicted molecular mass of 69 kDa.NCBI and SWISS-PROT database searches revealed signifi-cant similarity between the predicted acfB product and pro-karyotic MCPs. MCPs are inner membrane proteins thatmonitor the periplasmic concentrations of attractants andrepellents. Changes in the concentrations of these chemicalslead to changes in the bias of the flagellar motor. Upon bindingof the proper attractant, a signal is transferred to the cytoplas-

AcfBHlyBTcpI

AcfBHlyBTcpI

AcfBHlyBTcpl

AcfBHlyBTcpI

AcfBHlyBTcpI

AcfBHlyBTcpI

AcfBHlyBTcpI

AcfBHlyBTcpI

AcfBHlyBTcpI

VOL. 62, 1994

on July 12, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

3296 EVERISS ET AL.

1 2 3 4 kDa A

-97

-68

-43 B C

-29

FIG. 5. T7 RNA polymerase/promoter-directed expression of acfBin E. coli ER1451. Arrows on the left point to the individual polypep-tides. Numbers to the right indicate molecular masses of radiolabeledstandard proteins. Fluorographs were made of cell extracts of ER1451carrying the following plasmids and induced with IPTG and separatedby SDS-PAGE: lane 1, pAcfB; lane 2, pA316; lane 3, pA&394; and lane4, pBluescript II KS+.

mic signaling domain, resulting in a conformational changethat promotes the interaction of CheW with the chemosensor.CheW, in conjunction with CheA, an autokinase, transfersphosphoryl groups to CheB and CheY, leading to changes inthe methylation status of the MCP. Alterations in the methyl-ation status result in flagellar rotor switching and sensoryadaptation (1, 4, 5, 15-17). The change in flagellar rotorswitching regulates the frequency of tumbling and smoothswimming of the bacteria, enabling the cells to swim towardattractants and away from repellents. Four species of MCPshave been identified in E. coli, and each responds to a differentcollection of attractants and repellents (15). Extensive conser-vation of primary amino acid sequence is observed in thecytoplasmic signaling domain of MCPs, whereas little similarityis observed among the MCP receptor-binding domains. Theamino acid similarity between AcfB and the MCPs is restrictedto an HCD found in all bacterial MCPs. The proposed role ofthe HCD is to mediate MCP interactions with CheW (24).

50

40

30

20

10

0

WCL Soluble TM Inner Outer SS Si

FIG. 6. Subcellular localization of the V cholerae KP2.22 AcfB-PhoA hybrid protein. WCL, whole-cell lysate; Soluble, cytoplasm andperiplasm contents; TM, total membrane fraction; Inner, inner mem-branes; Outer, outer membranes; SS, N-lauroyl sarcosyl-soluble mem-branes; SI, N-lauroyl sarcosyl-insoluble membranes.

DFIG. 7. Motility plate analysis of V cholerae acfB::TnphoA mu-

tants. The bacterial strains used are: (A) V. cholerae 0395, (B) KP2.22(acfB::TrphoA), (C) KP1.17 (toxR12), and (D) KP3.51 (acJB::TrphoA).

AcfB is predicted to share a number of structural featureswith the MCPs, including the number and organization of thetransmembrane regions, the size of the cytoplasmic domains,and spatial conservation of the distance between the HCD andthe C terminus of the AcfB (Fig. 3). The Glu-Glu or Gln-Glupairs that are the substrate for the methyltransferase in E. coliand Salmonella typhimurium are not well maintained in AcfB.We are currently attempting to determine if AcfB is methyl-ated in V cholerae cells. Several lines of evidence suggest thatthe membrane topological organization of AcfB is also similarto that of the MCPs. Three methods (22, 23, 37) for predictingmembrane-spanning segments were used to identify two puta-tive transmembrane regions within AcfB. The first is locatednear the N terminus and spans residues 7 to 25. This regionlacks a characteristic signal peptidase cleavage site conformingto the -3, -1 rule proposed by von Heijne (47), suggestingthat it is not removed in the mature protein. This firsthydrophobic region may serve to anchor AcfB in the innermembrane of V cholerae, an idea consistent with the observa-tion that AcfB-PhoA hybrid proteins localize to the innermembrane (Fig. 6). A second contiguous stretch of hydropho-bic amino acids predicted to span the membrane is locatednear the middle of AcfB protein and encompasses amino acids275 to 294. The presence of two predicted membrane-spanningregions indicates that AcfB exhibits the same membranetopological organization established for MCPs. These predic-tions are supported by preliminary studies with TnphoA andTnlacZ as topology probes. Limited analysis of acfB::TnphoAand acfB::TnlacZ insertions indicates that AcfB possesses acytoplasmic domain and a periplasmic domain.Amino acid sequences with similarity to the HCD of MCPs

are also found in the V cholerae TcpI (14) and HlyB (3)proteins (Fig. 4). TcpI has been implicated as a negativeregulator of TcpA synthesis (45). The HCD is also present inPilJ, a protein required for type 4 pilus expression in P.aeruginosa (7). HlyB, a putative outer membrane protein, isrequired for efficient release of the V cholerae hemolysin HlyAfrom exponentially growing E. coli cells (3). E. coli cellsexpressing the V cholerae hlyA gene in the absence of afunctional hlyB product produce the hemolysin but do notsecrete HlyA into the medium (3). These observations suggest

INFECT. IMMUN.

2

on July 12, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

VIBRIO CHOLERAE acfB COLONIZATION DETERMINANT 3297

that HlyB may function posttranscriptionally to facilitate HlyAsecretion. AcfB and TcpI have regions of limited amino acidsimilarity in the periplasmic domain. The greatest degree ofsimilarity spans amino acids 154 to 208 of AcfB and is localizedto a region absent from the other known HCD-containing V.cholerae protein, HlyB (Fig. 4). The similarity exhibited be-tween AcfB and TcpI in the putative periplasmic domains mayrepresent a shared functional domain responsible for recog-nizing a common stimulus. Future studies aimed at generatingdirected mutations in these regions will be used to test thishypothesis. It may be noteworthy that a pH gradient exists atthe surface of the small bowel (25, 41) (i.e., the environmentbecomes more acidic as the epithelial surface is approached),since we know from previous studies that optimal choleratoxin, TCP, and ACF synthesis occurs at pH 6.5 (36, 46).Future studies examining the signals recognized by AcfB andTcpI may help to elucidate the signal transduction pathwayinvolved in modulating the synthesis and possibly the export ofV. cholerae colonization determinants.T7 promoter/polymerase-directed expression of acfB in E.

coli results in the production of four distinct protein species(Fig. 6). A doublet of proteins with apparent molecular massesof 74 and 75 kDa is observed by SDS-PAGE and fluorographyand may represent a modified form of the full-length acfBproduct. The two polypeptide species observed could beexplained by removal of the AcfB leader sequence, but theabsence of a characteristic signal peptidase recognition se-quence and the knowledge that AcfB-PhoA hybrid proteinslocalize to the inner membrane, presumably anchored by theleader peptide, argue against this possibility. MCPs exhibitmultiple forms on SDS-PAGE gels because of methyl modifi-cation. Methyl modification of AcfB may explain the presenceof the 74- and 75-kDa proteins observed in the E. coliexpression experiments. A second set of smaller doublet bandswith apparent molecular masses of 34 and 35 kDa was alsoobserved; these most likely result from internal translationinitiation in the acfB transcripts. Synthesis of the truncatedproducts by V. cholerae has not been determined. We arecurrently attempting to generate AcfB-specific antibodies toanalyze AcfB synthesis and possible methyl modification in Vcholerae cells. The absence of a hydrophobic leader peptide onthe truncated AcfB and unlikely translation initiation sites inaddition to the ones identified support the assertion that thetwo proteins in each doublet do not result from removal of asignal peptide from the proteins. Instead, the two formsprobably result from posttranslational modifications (methyl-esterification?) in regions shared by the full-length and trun-cated proteins.

Vibrios in which toxR has been disrupted exhibit increasedmotility in semisolid agar when incubated under conditionspromoting the expression of ToxR-activated genes, suggestingthat motility and virulence are oppositely regulated (13) (Fig.7). This observation suggests that the increased swarmingactivity associated with V cholerae toxR mutants is correlatedwith the expression of genes positively regulated by ToxR. Thealtered motility phenotype may be the result of a decreasedability of V. cholerae to orchestrate an effective chemotacticresponse when ToxR-regulated genes such as acfB (Fig. 7) andtcpI (14) are expressed. These results bring to light thepossibility that ToxR-controlled expression of V choleraecolonization determinants downregulates vibrio motility orchemotaxis and thus may enhance microcolony formation bylimiting the spread of vibrios from the anatomical site thatpromotes maximum ACF and TCP expression. The similarityof the proposed cytoplasmic domains of AcfB and TcpI tothose of proteins involved in bacterial chemotaxis may indicate

a mechanism by which vibrio factors (CheA and CheW ho-mologs) binding to the HCDs within AcfB and TcpI areprevented from interacting with the methyl-accepting chemo-taxis proteins that usually control vibrio chemotaxis. Thesequestration of factors required for the generation of thechemotactic response by AcfB and TcpI may result in abreakdown of the signaling pathway. Support for this modelcomes from experiments in which acfB was expressed in cellsunder conditions that repress ToxR-activated gene expression.Heterologous expression of acfB in these cells results indecreased swarming. A similar result is seen when OrfI fromTnl 721 (2), an HCD-containing protein of unknown function,is overexpressed in E. coli. The decreased swarming in semi-solid medium observed for these cells presumably occurs bythe diversion of chemotaxis proteins from the endogenousMCPs by Orfl. Future experiments aimed at determiningwhether AcfB undergoes methylation and demethylation, aswell as studies characterizing AcfB interactions with compo-nents of the vibrio chemotactic signaling pathway, will examinethese possibilities.The information presented in this report suggests that AcfB

represents an environmental sensory molecule that reduces theswarming activity (motility-chemotaxis) of V cholerae. We donot believe, however, that the sole function of AcfB in thecolonization process is the dampening of the V. choleraechemotactic response. This conclusion stems from preliminarystudies indicating that TcpI (V. cholerae HCD-containingToxR-regulated protein) functions as an environmental regu-lator of tcpA expression (45). Interestingly, TcpI also down-modulates V. cholerae motility plate activity when it is synthe-sized at high levels (14). By analogy, we believe that AcfB maybe involved in regulating the synthesis and/or export of one ormore of the V cholerae ACF proteins, i.e., we have preliminaryevidence that AcfC is secreted by V. cholerae cells into culturesupernatants. A better understanding of the role of AcfB incolonization will come from studies examining the expressionand structure-function of the other determinants encoded bythe ACF gene cluster.

ACKNOWLEDGMENTSThis work was supported by NIH grant AI28502 and an award from

the LSUMC-Shreveport Center for Excellence in Cancer Research,Treatment and Education.We thank V. DiRita for critical review of the manuscript and A.

Darzins, C. Gardel, J. J. Mekalanos, and R. K. Taylor for helpfuldiscussions.

REFERENCES1. Alley, M. R. K., J. R. Maddock, and L. Shapiro. 1992. Polar

localization of a bacterial chemoreceptor. Genes Dev. 6:825-836.2. Allmeier, H., B. Cresnar, M. Greck, and R. Schmitt. 1992.

Complete nucleotide sequence of Tnl721: gene organization anda novel gene product with features of a chemotaxis protein. Gene111:11-20.

3. Alm, R. A., and P. A. Manning. 1990. Characterization of the hlyBgene and its role in the production of the El Tor haemolysin ofVibrio cholerae 01. Mol. Microbiol. 4:413-425.

4. Bollinger, L., C. Park, S. Harayama, and G. L. Hazelbauer. 1984.Structure of the Trg protein: homologies with and differences fromother sensory transducers of Escherichia coli. Proc. Natl. Acad. Sci.USA 81:3287-3291.

5. Borkovich, K. A., N. Kaplan, J. F. Hess, and M. I. Simon. 1989.Transmembrane signal transduction in bacterial chemotaxis in-volves ligand-dependent activation of phosphate group transfer.Proc. Natl. Acad. Sci. USA 86:1208-1212.

6. Boyd, A., K. Kendall, and M. I. Simon. 1983. Structure of theserine chemoreceptor in Escherichia coli. Nature (London) 301:623-626.

VOL. 62, 1994

on July 12, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

3298 EVERISS ET AL.

7. Darzins, A. 1993. Characterization of a Pseudomonas aeruginosagene cluster involved in pilus biosynthesis and twitching motility:sequence similarity to the chemotaxis proteins of enterics and thegliding bacterium Myxococcus xanthus. Mol. Microbiol. 11:137-153.

8. DiRita, V. J. 1992. Coordinate control of virulence gene expressionby ToxR in Vibnio cholerae. Mol. Microbiol. 6:451-458.

9. DiRita, V. J., C. Parsot, G. Jander, and J. J. Mekalanos. 1991.Regulatory cascade controls virulence in Vibrio cholerae. Proc.Natl. Acad. Sci. USA 88:5403-5407.

10. Donowitz, M., and M. J. Welsh. 1987. Regulation of mammaliansmall intestine electrolyte secretion, p. 1351-1388. In L. R. John-son (ed.), Physiology of the gastrointestinal tract. Raven Press,New York.

11. Everiss, K. D., K. J. Hughes, and K. M. Peterson. The accessorycolonization factor and toxin-coregulated pilus gene clusters arephysically linked on the Vibrio cholerae chromosome. DNA Se-quence, in press.

12. Filip, C., G. Fletcher, J. L. Wulff, and C. F. Earhart. 1973.Solubilization of the cytoplasmic membrane of Escherichia coli bythe ionic detergent sodium-lauryl sarcosinate. J. Bacteriol. 115:717-722.

13. Gardel, C. L., and J. J. Mekalanos. 1994. Modus operandi ofVibrio cholerae: swim to arrive; stop to kill. The relationship amongchemotaxis, motility and virulence. J. Cell. Biochem. 18A:65.

14. Harkey, C., K. D. Everiss, and K. M. Peterson. 1994. The Vibriocholerae toxin-coregulated-pilus gene tcpI encodes a homolog ofmethyl-accepting chemotaxis proteins. Infect. Immun. 62:2669-2678.

15. Hazelbauer, G. L. 1992. Bacterial chemoreceptors. Curr. Opin.Struct. Biol. 2:505-510.

16. Hazelbauer, G. L., H. C. Berg, and P. Matsumura. 1993. Bacterialmotility and signal transduction. Cell 73:15-22.

17. Hazelbauer, G. L., and S. Harayama. 1983. Sensory transuction inbacterial chemotaxis. Int. Rev. Cytol. 81:33-70.

18. Herrington, D. A., R. H. Hall, G. Losonsky, J. J. Mekalanos, R. K.Taylor, and M. M. Levine. 1988. Toxin, toxin-coregulated pili andthe toxR regulon are essential for Vibrio cholerae pathogenesis inhumans. J. Exp. Med. 168:1487-1492.

19. Higgens, D. E., E. Nazareno, and V. J. DiRita. 1992. The virulencegene activator ToxT from Vibrio cholerae is a member of the AraCfamily of transcriptional activators. J. Bacteriol. 174:6974-6980.

20. Kaufman, M. R., J. M. Seyer, and R. K. Taylor. 1991. Processingof TCP pilin by TcpJ typifies a common step intrinsic to a newlyrecognized pathway of extracellular protein secretion by gram-negative bacteria. Genes Dev. 5:1834-1846.

21. Kelley, J. T., and C. D. Parker. 1981. Identification and prelimi-nary characterization of Vibrio cholerae outer membrane proteins.J. Bacteriol. 145:882-889.

22. Klein, P., M. Kanehisa, and C. DeLisi. 1985. The detection andclassification of membrane-spanning proteins. Biochim. Biophys.Acta 815:468-476.

23. Kyte, J., and R. F. Doolittle. 1982. A simple method for displayingthe hydropathic character of a protein. J. Mol. Biol. 157:105-132.

24. Liu, J., and J. S. Parkinson. 1991. Genetic evidence for interactionbetween the CheW and Tsr proteins during chemoreceptor sig-nalling by Escherichia coli. J. Bacteriol. 173:4941-4951.

25. Lucas, M. L., B. T. Cooper, F. H. Lei, I. T. Johnson, G. K. T.Holmes, J. A. Blair, and W. T. Cooke. 1978. Acid microclimate incoeliac and Crohn's disease: a model for folate malabsorption.Gut 19:735-742.

26. Maniatis, T., E. F. Fritsch, and J. Sambroolk 1982. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

27. McBride, M. J., T. Kohler, and D. R. Zusman. 1992. Methylation

of FrzCD, a methyl-accepting taxis protein of Myxococcus xanthus,is correlated with factors affecting cell behavior. J. Bacteriol.174:4246-4257.

28. Mekalanos, J. J. 1985. Cholera toxin: genetic analysis, regulation,and role in pathogenesis. Curr. Top. Microbiol. Immunol. 118:97-118.

29. Mekalanos, J. J. 1992. Environmental signals controlling expres-sion of virulence determinants in bacteria. J. Bacteriol. 174:1-7.

30. Miller, V. L., V. J. DiRita, and J. J. Mekalanos. 1989. Identifica-tion of toxS, a regulatory gene whose product enhances ToxR-mediated activation of the cholera toxin promoter. J. Bacteriol.171:1288-1293.

31. Miller, V. L., and J. J. Mekalanos. 1984. Synthesis of cholera toxinis positively regulated at the transcriptional level by toxR. Proc.Natl. Acad. Sci. USA 81:3471-3475.

32. Miller, V. L., R. K. Taylor, and J. J. Mekalanos. 1987. Choleratoxin transcriptional activator ToxR is a transmembrane DNAbinding protein. Cell 48:271-279.

33. Ogierman, M. A., and P. A. Manning. 1992. Homology of TcpN, aputative regulatory protein of Vibrio cholerae, to the AraC familyof transcriptional activators. Gene 116:93-97.

34. Ogierman, M. A., S. Zabihi, L. Mourtzios, and P. A. Manning.1993. Genetic organization and sequence of the promoter-distalregion of the tcp gene cluster of Vibrio cholerae. Gene 126:51-60.

35. Osborn, M. J., J. E. Gander, E. Parisi, and J. Carson. 1972.Mechanism of assembly of the outer membrane of Salmonellatyphimurium. J. Biol. Chem. 247:3962-3972.

36. Peterson, K. M., and J. J. Mekalanos. 1988. Characterization ofthe Vibrio cholerae ToxR regulon: identification of novel genesinvolved in intestinal colonization. Infect. Immun. 56:2822-2829.

37. Rao, M. J. K., and P. Argos. 1986. A conformational preferenceparameter to predict helices in integral membrane proteins.Biochim. Biophys. Acta 869:197-214.

38. Rice, M. S., and F. W. Dahlquist. 1991. Sites of deamidation andmethylation in Tsr, a bacterial chemotaxis sensory transducer. J.Biol. Chem. 266:9746-9753.

39. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencingwith chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA74:5463-5467.

40. Schnaitman, C. A. 1970. Examination of the protein compositionof the cell envelope of Eschenichia coli by polyacrylamide gelelectrophoresis. J. Bacteriol. 104:882-889.

41. Shaiu, Y., P. Fernandez, M. J. Jackson, and S. McMonagle. 1985.Mechanisms maintaining a low-pH microclimate in the intestine.Am. J. Physiol. 248:G608-G617.

42. Shaw, C., K. M. Peterson, J. J. Mekalanos, and R. K. Taylor. 1988.Genetic studies of Vibno cholerae TCP biogenesis, p. 5-17. In Y.Takeda and R. B. Sack (ed.), Advances in research on cholera andrelated diarrheas. KTK Science Publishers, Tokyo, Japan.

43. Shine, J., and L. Dalgarno. 1975. Determinant of cistron specificityin bacterial ribosomes. Nature (London) 254:34-38.

44. Tabor, S., and C. C. Richardson. 1985. A bacteriophage T7 RNApolymerase/promoter system for controlled expression of specificgenes. Proc. Natl. Acad. Sci. USA 82:1074-1078.

45. Taylor, R. K. 1989. Genetic studies of enterotoxin and otherpotential virulence factors of Vibrio cholerae, p. 309-329. In D. A.Hopwood and K. F. Chater (ed.), Genetics of bacterial diversity.Academic Press, London.

46. Taylor, R. K., V. L. Miller, D. B. Furlong, and J. J. Mekalanos.1987. The use ofphoA gene fusions to identify a pilus colonizationfactor coordinately regulated with cholera toxin. Proc. Natl. Acad.Sci. USA 84:2833-2837.

47. von Heijne, G. 1986. A new method for predicting signal sequencecleavage sites. Nucleic Acids Res. 14:4683-4690.

INFECT. IMMUN.

on July 12, 2020 by guesthttp://iai.asm

.org/D

ownloaded from