Molecular genetic basis for complex flagellar antigen expression in a triphasic serova
Molecular Genetic Basis for Complex Flagellar Antigen Expression in a TriphasicSerovarof SalmonellaNH Smith, and RK SelanderPNAS 1991;88;956-960doi:10.1073/pnas.88.3.956This information is current as of February 2007.www.pnas.org#otherarticlesThis article has been cited by other articles:E-mail Alertsat the top right corner of the article or click here.Receive free email alerts when new articles cite this article - sign up in theboxRights & Permissionswww.pnas.org/misc/rightperm.shtmlTo reproduce this article in part (figures, tables) or in entirety, see:Reprintswww.pnas.org/misc/reprints.shtmlTo order reprints, see:Notes:
Proc. Nati. Acad. Sci. USAVol. 88, pp. 956-960, February 1991MicrobiologyMolecular genetic basis for complex flagellar antigen expressioninatriphasic serovarofSalmonella(flagellin/phase variation/plasmids/recombination/polymerase chain reaction)NOEL H. SMITH AND ROBERT K. SELANDER*Institute of Molecular Evolutionary Genetics, Mueller Laboratory, PennsylvaniaState University, University Park, PA 16802Contributed by Robert K. Selander, October 22, 1990ABSTRACT StrainsofmostSalmoneUaserovarsproduceeither one (monophasic) or two (diphasic) antigenic forms offlagellin protein, but strains capable of expressing three ormore serologically distinct flagellins ("complex" serovars)have occasionally been reported. Amolecular genetic analysisof a triphasic strain of the normally diphasic serovar SalnoneUarubislaw revealed that it has three flgellin genes, includingthenormalfliC(phase 1)andfljB(phase2)chromosomalgenes encoding type r and type e,n,x flagellins, respectively,andathirdlocus-(hereindesignatedasflpA) that islocatedona large plasmid (pRKSO1) and codes fora type d flagellin. Thecodingsequenceoftheplasmid-bornegene issimilartothatofa phase 1 chromosomal gene, but the sequence of its promoterregion is homologous to that of a phase 2 chromosomal gene.The irreversible loss of the ability to express a type d flagellinthat occurs when the triphasic strain is grown in the presenceofdantiserum iscausedbydeletionofpartor alloftheflpAgene. Thus, the molecular basis for the unusual serologicalreactions of the triphasic strain of S. rubislaw and, by inference,othercomplexserovarsofSalmonela isexplained. Plasmidsof the type carried by the triphasic strain of S. rubislawprovide a mechanism for the generation of new serovarsthrough the lateral transfer and recombination of flagellingenes.Strains of the diphasic serovars of Salmonella produce twoserologically distinct types of flagellin, which is the majorcomponent of the bacterial flagellar filament. The alternateexpression of different flagellin proteins is known as phasevariation;cellsinphase 1expressthefliCflagellingeneandthose in phase 2 express the fljB flagellin gene and theassociated repressor (fljA) of the phase 1 gene (1). Bothflagellin genes are located on the chromosome (2), and phasetransition is controlled by a site-specific recombinationmechanism that periodically inverts the promoter region ofthephase2flagellinoperon. Antigenic variationofthephase1 and phase 2 flagellins, in combination with that of thecell-walllipopolysaccharide, isthebasisoftheKauffmann-
White serotyping scheme, which is universally employed toclassify and identify strains of Salmonella (3, 4).Isolates of the more than 2200 recognized Salmonellaserovars are normally either diphasic or monophasic (producingonlyone type offlagellin), but strains that can expressthree or more flagellins ("triphasic" or other "complex"serovars) have occasionally been reported (4). For example,the serovar Salmonella rubislaw is normally diphasic, withthe antigenic formula ll:r:e,n,x, which indicates a lipopolysaccharideoftype11,aphase 1flagellinoftype r,andaphase2 flagellin exhibiting three antigenic factors, e, n, and x.However, a strain recovered from a human in 1963 wasidentified as a triphasic form ofS. rubislawby McWhorter etThepublicationcostsofthis articleweredefrayed inpartbypagechargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.al. (5). The serological reactions of this strain, which in kindare typical of all triphasic serovars, are as follows: cells inphase 1 produce only antigenic typedflagellin, whereasthosein phase 2 express flagellins with four antigenic factors, di, e,n, and x. Hence, the antigenic formula is ll:d:d,e,n,x, whichinvolvesacombinationofphase 1.andphase2antigensnotrecognized in the Kauffmann-White scheme (4). When cellsin either phase 1 or phase 2 are grown in the presence of dantiserum, colonies are recovered that have irreversibly lostthe ability to produce type d flagellin and, instead, express anew phase 1 flagellin of an r antigenic type. The passagedcells are now serologically indistinguishable from normaldiphasic strains of S. rubislaw.Several hypotheses have been advanced to accountfor theoccurrence of triphasic strains of Salmonella. On the arbitraryassumption that the long-term evolution of the salmonellaehas involved progressive reduction in the number offlagellingenes(6), ithasbeensuggestedthatthetriphasicandothercomplex serovars are archetypal forms from which thediphasicand monophasic serovarsevolved (7). Alternatively,Douglas and Edwards (8) concluded that some triphasicserovars are merely diphasic strains that happen to expressflagellinsofthesamemajorantigentype inbothphases.Theoccurrence ofa third phase was interpreted as an example of"Rphase" variation, in which serological challenge inducesthe expression of new, presumably mutant antigens (9).We here report the resultsofamoleculargenetic analysisof the triphasic strain of S. rubislaw in which we havediscovered that the type d flagellin is encoded by a third
flagellin gene located on a plasmid and that the irreversiblelossofthis flagellin iscausedby deletionofpartor alloftheplasmid-borne gene.tMATERIALS ANDMETHODSBacterial Strains. The triphasic S. rubislaw strain (RKS5310;CDC2209-63)wasobtainedfromA. C.McWhorter,ofthe Centers for Disease Control, Atlanta, GA (5). Strains ofdiphasic S. rubislaw (RKS 938), Salmonella typhimurium(RKS 284), Salmonella muenchen (RKS 3121), and Salmonellatyphi (RKS 3333) were described by Smith et al. (10).Threederivative strainsofRKS5310thatcannotexpressadflagellin antigen were selected by growth in semi-solid agar(cystine tryptic agar; Difco) containing d antiserum. Plateswere inoculated with the triphasic S. rubislaw strain and,after overnight incubation, isolates were obtained from theleadingedge ofa50-mmzoneofgrowth extendingfrom thepoint of inoculation.Chromosomal DNA. Methods of preparation, restrictiondigestion, Southern transfer and hybridization, and polymerase chain reaction (PCR) amplification ofchromosomalDNAAbbreviation: PCR, polymerase chain reaction.*To whom reprint requests should be addressed.tThe sequence reported in this paper has been deposited in theGenBank data base (accession no. M58145).956
958 Microbiology: Smith and Selander Proc. Natl. Acad. Sci. USA 88 (1991)0 500 1000 1500IlAVIPlO> <P8 P1> P5> <P11 <P2P9>XXr ,FIG. 3. Primers,probes,andregionssequenced.Schematicrepresentationoftheplasmid-borneflpAgeneandflankingregions;thescalerepresentsthedistance,inbp,fromtheinitiationcodon(AUG)ofthegene.ThepositionsanddirectionsofPCRprimersaremarked,asarethe regions sequenced(filled boxes). Openboxes represent the position oftheuniversal flagellin probe (left)andthe dflagellin alleleprobe (right).The small deletion detected in the plasmid of strain RKS 5313 is indicated bythe hatched box.pRKS01 isconsistentwiththeobservationthatonlyonecopyofthisgene ispresentinthetriphasicS.rubislaw strain(Fig.1, lane 3). To confirm that the third flagellin gene is plasmidborne,pRKS01 DNA was excised from an agarose gel,purified,andsubjectedtoPCRamplificationwithprimers 1and 2 (Fig. 3), which have been used to amplify phase 1 andphase 2 flagellin genes from strains of various Salmonellaserovars (10, 11). A single fragment was amplified, anddigestion with Pst I revealed a restriction pattern identical tothatofthechromosomaldalleleofS. typhi(strainRKS3333)(10). The extracted plasmidDNAwas also subjected toPCRamplification with primers 2 and 7, a combination that wehave used to specifically amplify the fliC gene in strains ofseveralotherSalmonellaserovars(unpublished data). [Primer7wasderivedfrom the sequence ofthefliD gene, whichis adjacent tofliC in S. typhimurium (21).] But under variousreaction conditions withpRKS01DNAas template,wewereunable to amplify a flagellin sequence with this primer.In sum, the triphasic S. rubislaw strain has, in addition tothetwonormalchromosomalflagellingenes, athird, plasmidborne,typedflagellin gene. Because the plasmid-borne geneisnotassociatedwitheitherafliDgeneorahingene, itcannotbe identified as either afliC or afljB gene. Accordingly, wedesignated it asflpA (flagellin-plasmid located).PromoterSequenceRthefipATGene.The serologicalreactionsofthetriphasicserovarsofSalmonella
indicatethattheflpA gene is expressed independently of the chromosomalflagellin phase ofthe host cell. This implies that the sequenceof the promoter region of the flpA gene differs from thecorrespondingsequencesofthefliCandfljBgenes. Supportfor this conclusion comes from the absence ofa hin gene onthe plasmid and our failure to amplify the flpA gene withprimers derivedfrom thefliDgene.To obtain the sequenceofthepromoterregion oftheflpAgene,weusedan inversePCR(14), and from this sequencewe designed aPCRprimerlocated 181 bp from the initiation codon of the flpA gene(primer10,Fig. 3).Thisprimerwasused,incombinationwithd allele-specific primers, to PCR amplify and sequence thepromoter region of the flpA gene; the region sequencedextendsfrom 164bp infrontofthe initiationcodon to60bpinside the coding region (Fig. 3).To assess the extent ofsequence variation in the promoterregionofdallelesfrom various strains,weusedprimers 2and7 toPCRamplify and sequence 244bp directly in front ofthefliC(d) gene of strains of S. muenchen (RKS 3121) and S.typhi(RKS3333) (10).Thesegmentextendsfromthecodingregion ofthefliCgene to within 10bp ofthe coding region ofthefliDgeneandwas inbothcases identical in sequence tothe published sequence of this region of the S. typhimuriumchromosome (22, 23).In Fig. 4, the promoter sequence and promoter-proximalcoding sequence oftheflpA gene may be compared with thecorresponding sequencesofthefliCgeneofS. typhimurium(22,23)andthefljBgeneofSalmonellaadelaide (21).Atfiveofthe six sites thatarepolymorphic inthe first20codonsofthe these genes, flpA and fljB share nucleotides, and thisflpA gggggtacttcactttcctagttagtcccgtggagccttaacgaatgtgatcgaagccaatcc 63fliC ggaataatgatgcataaagcggctatttcgccgcctaagaaaaagatcgggggaagtgaaaaa 63fljB gacacagtattcacctggaaaggctttttaatcaaaatgttagatgtaagcaattacggacag 63flpAfliC ..gggt.cttca.t.tccta.ttag.cc..tg.agcctt..cg.atgt.atc....cc..tccflpAfljB .acacagta.tcacc.gga.aggct.tttaatc.aaa.gtta..tgtaagcaatta.gg.cagflpA aaaaattagtaaagtttatgtgctaaatgtcgataacctggatgacacaggtaagcctggcat 126fliC ttttctaaagttcgaaattcaggtgccgatacaagggttacggtgagaaaccgtgggcaacag 126fljB aaaaaatagtaaagtttatgcctcaactgtcgataacctggatgacacaggtaagcctggcat 126flpAfliC aaaaa.t.gtaaa.ttta.gt.c.aaatg.cg.taacc.ggatgacac.ggtaa.cctgg..t. cctc.flpAfljB ...a.. .. ...flpAfliCfljBflpAfliCflpAfljB
Microbiology: Smith and SelanderProc. Natl. Acad. Sci. USA 88 (1991) 959sequence homology continues for 101 bp in front ofthe genes.However, the 63-bp region farthest from the coding region offlpA shows little similarity with the corresponding region offljB (Fig. 4); the promoter region offlpA has less homologywith that offliC.In sum, the promoter region and the promoter-proximalcoding sequence of the plasmid-borne flagellin gene (flpA)has greater homology with a phase 2 gene (fljB) than with aphase 1 gene (fliC).Molecular Basis for Loss of the d Antigen. Three independentderivatives of the triphasic S. rubislaw strain thatmigrated through semi-solid agar containing d-specific flagellinantiserum were designated as strains RKS 5313, RKS5314, and RKS 5315. When extrachromosomal DNA preparedfrom these strains and the parental triphasic strain wasvisualized by agarose gel electrophoresis, no variation ineither plasmid content or size was apparent. The plasmidDNAwas then tested with the universal flagellin gene probe,to which the flpA-bearing plasmid (pRKS01) hybridizedstrongly, as did the plasmid of similar size from RKS 5313;but the plasmidDNAfromRKS 5314 andRKS 5315 did notanneal with this probe. The extrachromosomal DNA wasthen used as a template for PCR amplification with primersspecific for d flagellin alleles. Primers 5 and 11 (Fig. 3)amplified a 378-bpsegmentfrom the central region oftheflpAgene of pRKS01 but failed to amplify sequences from thethree derivative strains. Primers 10 and 11 (Fig. 3) amplifieda 1217-bp segment that included the promoter region andmuch of the flpA gene from pRKS01 but did not amplifysequences from the plasmid DNA of strains RKS 5314 andRKS 5315. However, a 950-bp sequence was amplified fromtheplasmid ofRKS 5313. These findings indicate thatmostoftheflpAgenewas deletedfrom the plasmidsofRKS5314andRKS5315 and that a261-bp segmentwas deletedfrom theplasmid ofRKS 5313.It has been reported thatatypejfliC allele canbe generatedby deletion of a 261-bp segment from the antigen-codingregion of the normal fliC (d) allele of S. typhi (24). Becaused antiserum does not cross-react with type j flagella, cellsexpressing type j flagellin are unaffected by d antiserum inmotility experiments. To confirm that the plasmid of derivativestrain RKS 5313 carries a j allele, we sequenced theappropriateregionoftheflpAgeneandcompared itwiththe
sequence ofthesame regionofpRKS01fromthe triphasicstrain. Single-stranded DNA was derived from segmentsamplifiedbyPCRprimers 10and 11, and the regionfrombasepair 457 to base pair 1027 (15) was sequenced (Fig. 3). ThissegmentofpRKS01was only slightlydivergent in sequencefromthatofthefliC(d) alleleofS.muenchenreportedbyWeiand Joys (15); we observed only two changes in 570 bp. (Thefull sequence of the flpA gene will be reported elsewhere.)When the same region of the plasmid from strain RKS 5313was sequenced on one strand, it was found to have a deletionof261bp(Fig. 3)andwasidenticalinsequencetothejallelederivative of the fliC (d) gene of S. typhi (24).DISCUSSIONWe have shown that a triphasic strain of S. rubislaw has aplasmid-borne flagellin gene (flpA) in addition to the normalphase 1 and phase 2 chromosomal flagellin genes (fliC andfljB). TheflpAgene hybridizes withaprobe specific forthedflagellin allele and has a central antigen-coding region thatis virtually identical in sequence to thedallele ofthefliCgeneofS.muenchen (15).AlthoughflpAencodesatypedflagellinand, therefore, represents an allele that is almost invariablyassociated with phase 1 flagellins (4), the promoter andpromoter-proximal coding sequences showgreaterhomologywith a phase 2 flagellin gene. We consider this homologysignificant because the four fliC alleles that have been sequenced (refs. 21 and 22; this study) show little or nonucleotide variation in this region. The sequence homologybetween the promoter regions of flpA and fljB does notextend to the hin region; indeed, a hin gene could not bedetected on these plasmids by Southern hybridization. Consequently,flpA has a mosaic structure and appears to consistofaphase 2 flagellinpromotercoupled toacentral, antigen-determining region derived from a phase 1 gene.Ourfindings explain at the molecular level the unusual, andhitherto puzzling, serological reactions of the triphasic S.rubislaw strain and, by inference, complex Salmonella serovarsin general. Because the sequence ofthe promoterofflpAis similar to that of a phase 2 gene, flpA is unlikely to besensitive to repression by the product ofthefljA gene, whichis cotranscribed with the phase 2 flagellin gene (22). For thetriphasic strainofS. rubislaw,wesuggest that cells inphase2 express the phase 2 chromosomal flagellin gene (fljB) andthe plasmid-borne gene (flpA). Simultaneous expression ofthese genes results in flagellar filaments composed of bothflagellin proteins (see ref. 25), and, thus, the antigenic reactionsofthesestrainsare d, e, n,and x. Incontrast,whenfljBand the repressor offliC are not being expressed, chromosomallyencoded phase 1 flagellin is absent and only a type d
flagellin encoded by the plasmid-borne gene is produced.This observation suggests thepresenceofarepressorofthefliCgenethat iscotranscribedwiththeflpAgene inamanneranalogous to the fljA gene of the phase 2 flagellin operon.Finally, the irreversible loss of the type d flagellin is explainedby the selection of cells with plasmids from whichpartor alloftheflpAgenehasbeendeleted.When thegeneis absent, cells express the normal diphasic S. rubislaw phase1 and 2 flagellins; but when the deletion is partial and merelyconvertsflpAfromadtoajallele, the cellsremaintriphasicand fail to express the chromosomal type r phase 1 flagellin.The observation that growth in d antiserum selects fordeletions rather than mutations that silence the flpA genesuggests that expression of flpA is not controlled by asite-specific inversion system (1).In related work, we have interpreted the occurrence ofthesamefliC alleles invery distantly related serovarsofSalmonellaas evidence that the lateral transfer and chromosomalintegrationofflagellingenesequences isamajorevolutionaryprocess generating new serovars (10, 16). We now suggestthat plasmids of the type we have identified are vehicles forthe lateral transfer of flagellin genes, and, significantly, wehave recently discovered that strains of several other, chromosomallyunrelated triphasic Salmonella serovars containplasmids that are similar in size to pRKS01 (see Fig. 2) andalso carryflpA (d)genes(unpublished data). TransferofthepRKS01plasmid, eitherbyconjugationorasaphage,wouldalter the array of antigenic flagellin types produced by therecipient cell and, thus, would generate a novel complexserovar. Moreover, because theirflagellingenes are availablefor homologous recombination with chromosomal flagellingenes, plasmids ofthis type also provide amechanismfor theevolution ofnewmonophasicand diphasic serovars.We thank A. C. McWhorter for providing the triphasic strain anddantiserumand E. Elsinghorst forsupplying strains bearing standardplasmids. This research was supported by Grant AI-22144 from theNational Institutes of Health.1.Macnab, R. M. (1987) in Escherichia coli and Salmonellatyphimurium:CellularandMolecularBiology, eds. Neidhardt,F. C., Ingraham, J. L., Low, K. B., Magasanik, B., Schaechter,M.&Umbarger,H. E.(Am.Soc.Microbiol.,Washington,DC), Vol. 1, pp. 732-759.2.Sanderson, K. E. & Roth, J. R. (1988) Microbiol. Rev. 52,485-532.3.LeMinor,L.(1984)inBergeysManualofSystematicBacte
960 Microbiology: Smith and Selanderriology,eds.Krieg,N.R.&Holt, J.G.(Williams&Wilkins,Baltimore), Vol. 1, pp. 427-458.4.Ewing, W. H. (1986) Edwards and Ewings Identification ofEnterobacteriaceae (Elsevier, New York).5.McWhorter,A.C.,Balland,M.M.&Freeman,B.0.(1964)J.Bacteriol. 87, 967.6.White, P. B. (1926) Further Studies ofthe Salmonella Group(His Majestys Stationery Office, London), Med. Res. CouncilSpec. Rep. Ser. no. 103.7.Edwards, P. R., Kauffmann, F. & Huey, C. R. (1957) ActaPathol. Microbiol. Scand. 41, 517-520.8.Douglas,G.W.&Edwards,P.R.(1962)J.Gen.Microbiol.29,367-372.9.Kauffmann, F. (1961) Die Bakteriologie der Salmonella-Gruppe (Munksgaard, Copenhagen).10.Smith,N.H.,Beltran,P.&Selander,R.K.(1990)J.Bacteriol.172, 2209-2216.11.Smith, N. H. & Selander, R. K. (1990) J. Bacteriol. 172,603-609.12.Higuchi, R. G. & Ochman, H. (1989) Nucleic Acids Res. 17,5865.13.Kado, C. I. & Liu, S.-T. (1981) J. Bacteriol. 145, 1365-1373.14.Ochman,H.,Gerber,A. S.&Hartl,D. L.(1988)Genetics120,621-623.15.Wei, L.-N. & Joys, T. M. (1985) J. Mol. Biol. 186, 791-803.Proc. Nad. Acad Sci. USA 88 (1991)16.Beltran, P., Musser, J. M., Helmuth, R., Farmer, J. J., III,Frerichs, W. M., Wachsmuth, I. K., Ferris, K., McWhorter,A. C., Wells, J. G., Cravioto, A. & Selander, R. K. (1988)Proc. Natl. Acad. Sci. USA 85, 7753-7757.17.Selander,R.K.&Smith,N.H.(1990)Rev.Med.Microbiol.1,219-228.18.Macrina,F.L.,Kopecko,D.J.,Jones,K.R.,Ayers,D.J.&McCowen, S. M. (1978) Plasmid 1, 417-420.19.Sanderson, K. E. & Stocker, B. A. D. (1987) in Escherichiacoli and Salmonella typhimurium: Cellular and MolecularBiology, eds. Neidhardt, F. C., Ingraham, J. L., Low, K. B.,Magasanik, B., Schaechter,M.&Umbarger,H. E.(Am. Soc.Microbiol., Washington, DC), Vol. 2, pp. 1220-1224.20.
Kopecko, D. J., Holcombe, J. & Formal, S. B. (1979) Infect.Immun. 24, 580-582.21.Szekely, E. & Simon, M. (1983) J. Bacteriol. 155, 74-81.22.Inoue,Y.H.,Kutsukake,K.,Iino,T.&Yamaguchi,S.(1989)Gene 85, 221-226.23.Joys, T. M. (1985) J. Biol. Chem. 260, 15758-15761.24.Frankel, G., Newton, S. M. C., Schoolnik, G. K. & Stocker,B.A.D. (1989)EMBOJ. 8,3149-3152.25.Kutsukake, K. & Iino, I. (1983) in Nucleic Acid Research:Future Developments, eds. Kizobuchi, K., Watanabe, I. &Watson, J. D. (Academic, New York), pp. 395-406.