Analysis of the Lateral Flagellar Gene System of Aeromonas hydrophila AH-3

JOURNAL OF BACTERIOLOGY, Feb. 2006, p. 852–862 Vol. 188, No. 30021-9193/06/$08.00�0 doi:10.1128/JB.188.3.852–862.2006Copyright © 2006, American Society for Microbiology. All Rights Reserved.

Analysis of the Lateral Flagellar Gene Systemof Aeromonas hydrophila AH-3

Rocıo Canals,1 Maria Altarriba,1 Silvia Vilches,1 Gavin Horsburgh,2Jonathan G. Shaw,2 Juan M. Tomas,1* and Susana Merino1

Departamento de Microbiologıa, Facultad de Biologıa, Universidad de Barcelona, Diagonal 645,08071 Barcelona, Spain,1 and Division of Molecular and Genetic Medicine, University of

Sheffield Medical School, Sheffield S10 2RX, United Kingdom2

Received 25 July 2005/Accepted 21 October 2005

Mesophilic Aeromonas strains express a polar flagellum in all culture conditions, and certain strains producelateral flagella on semisolid media or on surfaces. Although Aeromonas lateral flagella have been described asa colonization factor, little is known about their organization and expression. Here we characterized thecomplete lateral flagellar gene cluster of Aeromonas hydrophila AH-3 containing 38 genes, 9 of which (lafA-U)have been reported previously. Among the flgLL and lafA structural genes we found a modification accessoryfactor gene (maf-5) that is involved in formation of lateral flagella; this is the first time that such a gene hasbeen described for lateral flagellar gene systems. All Aeromonas lateral flagellar genes were located in a uniquechromosomal region, in contrast to Vibrio parahaemolyticus, in which the analogous genes are distributed in twodifferent chromosomal regions. In A. hydrophila mutations in flhAL, lafK, fliJL, flgNL, flgEL, and maf-5 resultedin a loss of lateral flagella and reductions in adherence and biofilm formation, but they did not affect polarflagellum synthesis. Furthermore, we also cloned and sequenced the A. hydrophila AH-3 alternative sigmafactor �54 (rpoN); mutation of this factor suggested that it is involved in expression of both types of flagella.

Mesophilic Aeromonas strains are ubiquitous waterbornebacteria and pathogens of reptiles, amphibians, and fish (4).They can be isolated as part of the fecal flora of a wide varietyof other animals, including some animals consumed by hu-mans, such pigs, cows, sheep, and poultry. In humans, Aero-monas hydrophila strains belonging to hybridization group 1(HG1) and HG3, Aeromonas veronii biovar sobria (HG8/HG10), and Aeromonas caviae (HG4) have been associatedwith gastrointestinal and extraintestinal diseases, such aswound infections, and less commonly with septicemias of im-munocompromised patients (14). The pathogenicity of meso-philic aeromonads has been linked to a number of differentdeterminants, such as toxins, proteases, outer membrane pro-teins (28), lipopolysaccharide (23), and flagella (23, 32).

Mesophilic Aeromonas strains usually have a single polarunsheathed flagellum in all culture conditions, but it is knownthat 50% to 60% of the mesophilic aeromonads most com-monly associated with diarrhea (18) also have many un-sheathed peritrichous lateral flagella when they are grown inviscous environments or on surfaces (38). Different workershave shown that lateral flagella increase bacterial adherenceand are required for swarming motility and biofilm formation(10, 19).

The expression of two distinct flagellar systems is relativelyuncommon, although it has been observed in Vibrio parahae-molyticus (19), Azospirillum brasilense (26), Rhodospirillum cen-tenum (15), Helicobacter mustelae (29), and Plesiomonas shigel-loides (13). V. parahaemolyticus is the best-studied organism,

and it has two distinct flagellar systems. Recently, an Esche-richia coli O42 lateral flagellar gene cluster (Flag-2) has beendescribed (33), and the presence of Flag-2-like gene clusters inYersinia pestis, Yersinia pseudotuberculosis, and Chromobacte-rium violaceum suggests that the coexistence of two flagellarsystems in the same species is more common than previouslysuspected (33). The polar flagellum (Fla) of V. parahaemolyti-cus requires around 60 genes distributed in five clusters onchromosome I for biogenesis and assembly (41), whereas lat-eral flagella (Laf) are coded for by 38 different genes distrib-uted in two clusters on chromosome II (39). We have previ-ously described lateral flagellar regions containing 9 genes inA. hydrophila and 10 genes in A. caviae (10). These regionsencode one flagellin (LafA) in A. hydrophila and two flagellins(LafA1 and LafA2) in A. caviae, a HAP-2 protein (LafB), aprotein with an unknown function (LafX), a flagellin chaper-one (LafC), some regulatory proteins (LafE, LafF, and LafS),and the motor proteins (LafT and LafU). Characterization ofmutants and nucleotide and N-terminal sequencing demon-strated that the lateral flagellins were distinct from their polarflagellin counterparts (10, 32). Mutations of the lateral flagel-lar genes, such as lafB, lafS, or lafA, did not affect polar fla-gellum synthesis and vice versa.

Little is known about the organization and expression oflateral flagella in mesophilic Aeromonas strains. Some geneshave been described previously, but many others are requiredfor the expression of flagella. In this work we employed trans-poson mutagenesis and complementation of homologous mu-tants to isolate the A. hydrophila AH-3 chromosomal regionsinvolved in expression of lateral flagella. Furthermore, we in-vestigated the distribution of the genes in the mesophilic Aero-monas species, characterized several Aeromonas strains withdefined mutations in different lateral flagellar genes, and stud-

* Corresponding author. Mailing address: Departamento Microbio-logıa, Facultad Biologıa, Universidad Barcelona, Diagonal 645, 08071Barcelona, Spain. Phone: 34-93-4021486. Fax: 34- 93-4039047. E-mail:[email protected].

852

ied the motility of these strains, the presence of both types offlagella, the adherence to HEp-2 cells, and the ability to formbiofilms.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions. The bacterial strains andplasmids used in this study are listed in Table 1. E. coli and Klebsiella pneumoniaestrains were grown on Luria-Bertani (LB) Miller broth and on LB Miller agar at37°C; Aeromonas strains were grown either in tryptic soy broth (TSB) or ontryptic soy agar (TSA) at 30°C; and Vibrio cholerae strains were grown either inLB broth or on LB agar with 2 mM glutamine at 37°C. When required, ampicillin(50 �g/ml), kanamycin (50 �g/ml), chloramphenicol (25 �g/ml), rifampin (100�g/ml), spectinomycin (50 �g/ml), and tetracycline (20 �g/ml) were added to thedifferent media.

Motility assays (swarming and swimming). Freshly grown bacterial colonieswere transferred with sterile toothpicks into the center of swarm agar (1%tryptone, 0.5% NaCl, 0.6% agar). Agar plates containing LB medium with 0.3%agar and 2 mM glutamine were used to measure V. cholerae motility. The plateswere incubated face up for 16 to 24 h at 30°C, and motility was assessed byexamining the migration of bacteria through the agar from the center toward theperiphery of the plate. In addition, swimming motility was assessed by lightmicroscopy.

TEM. For transmission electron microscopy (TEM) bacterial suspensionswere placed on Formvar-coated grids and negative stained with a 2% solution of

uranyl acetate (pH 4.1). The preparations were observed with a Hitachi 600transmission electron microscope.

Mini-Tn5Km-1 mutagenesis. Conjugal transfer of transposition element mini-Tn5Km-1 from E. coli S17-1�pirKm-1 (8) to A. hydrophila AH-405 (rifampin-resistant AH-3) was carried out in a conjugal drop incubated for 6 h at 30°C witha ratio of S17-1�pirKm-1 to AH-405 to HB101(pRK2073) (helper plasmid) of1:5:1. Serial dilutions of the mating mixture were plated on TSA supplementedwith rifampin and kanamycin in order to select mutants.

DNA techniques. DNA manipulations were carried out essentially as describedpreviously (35). DNA restriction endonucleases, T4 DNA ligase, the E. coli DNApolymerase Klenow fragment, and alkaline phosphatase were used as recom-mended by the suppliers. PCR was performed using Taq DNA polymerase(Invitrogen) with a Perkin-Elmer Gene Amplifier PCR System 2400 thermalcycler. Amplification of 4,000 bp was performed using High Fidelity PlatinumTaq DNA polymerase (Invitrogen) as recommended by the supplier. Colonyhybridization was carried out by colony transfer onto positive nylon membranes(Roche), and then preparations were lysed according to the manufacturer’sinstructions. Probe labeling with digoxigenin, hybridization, and detection (Am-ersham) were carried out as recommended by the suppliers.

Cloning of DNA flanking mini-Tn5Km-1 insertions. Chromosomal DNA ofmini-Tn5Km-1 mutants was digested with EcoRI, PstI and EcoRV, purified,ligated into the vector pBCSK (Stratagene), and introduced into E. coli XL1-Blue. Recombinant plasmids containing the transposon with flanking insertionswere selected in LB agar plates supplemented with kanamycin and chloramphen-icol. The mini-Tn5Km-1 flanking sequences were obtained by using specific

TABLE 1. Bacterial strains and plasmids used in this study

Strain or plasmid Genotype and/or phenotypea Reference or source

A. hydrophila strainsAH-3 Wild type, serogroup O:34 22AH-405 AH-3, spontaneous Rifr 2AH-5501 AH-405, flhAL::mini-Tn5Km-1 This studyAH-5502 AH-405, rpoN::Kmr This studyAH-5503 AH-405, lafK::Kmr This studyAH-5504 AH-405, flhAL::Kmr This studyAH-5505 AH-405, fliIL::Kmr This studyAH-5506 AH-405, flgNL::Kmr This studyAH-5507 AH-405, flgEL::Kmr This studyAH-5508 AH-405, maf-5::Kmr This study

E. coli strainsDH5� F�endA hdsR17(rK

�1mK�) supE44 thi-1 recA1 gyr-A96 �80lac 12

XL1-Blue endA1 recA1 hsdR17 supE44 thi-1 gyrA96 relA1 lac (F� proAB lacIZ�M15 Tn10) StratageneS17-1�pir thi thr1 leu tonA lacY supE recA::RP4–2 (Tc::Mu) Kmr �pir with mini-Tn5Kml 8MC1061�pir thi thr1 leu6 proA2 his4 argE2 lacY1 galK2 ara14 xy15 supE44 �pir 34

K. pneumoniae 52145 O1:K2 27V. cholerae KKV56 �rpoN1 �lacZ 20Plasmids

pGEMT Cloning vector, Apr PromegapBCSK Cloning vector with lacZ gene, Cmr StratagenepLA2917 Cosmid vector, Tcr Kmr 1pACYC184 Plasmid vector, Cmr Tcr 35pRK2073 Helper plasmid, Spr 34pFS100 pGP704 suicide plasmid, pir dependent, Kmr 34pDM4 Suicide plasmid, pir dependent with sacAB genes, oriR6K, Cmr 23COS-FLG pLA2917 with AH-3 polar flagellar region 1 (flgA-LP), Tcr 2pLA-FLIL1 pLA2917 with AH-3 fliML-flhAL genes, Tcr This studypLA-FLIL2 pLA2917 with AH-3 lafK-flgNL genes, Tcr This studypLA-FLGL pLA2917 with AH-3 flgB-flgLL genes, Tcr This studypLA-MAFL pLA2917 with AH-3 maf-5 gene, Tcr This studypLA-RPON pLA2917 with AH-3 rpoN gene, Tcr This studypACYC-RPON pACYC184 with AH-3 rpoN gene, Tcr This studypFS-FLHAL pFS100 with an internal fragment of AH-3 flhAL gene, Kmr This studypFS-FLIJL pFS100 with an internal fragment of AH-3 fliJL gene, Kmr This studypFS-FLGNL PSF100 with an internal fragment of AH-3 flgNL gene, Kmr This studypFS-MAFL pFS100 with an internal fragment of AH-3 maf-5 gene, Kmr This studypFS-RPON pFS100 with an internal fragment of AH-3 rpoN gene, Kmr This studypDM-LAFK pDM4 with AH-3 lafK::Km, Cmr Kmr This studypDM-FLGEL pDM4 with AH-3 flgEL::Km, Cmr Kmr This study

a Tcr, tetracycline resistant; Kmr, kanamycin resistant; Apr, ampicillin resistant; Rifr, rifampin resistant; Cmr, chloramphenicol resistant; Spr, spectinomycin resistant.

VOL. 188, 2006 LATERAL FLAGELLAR GENE SYSTEM OF A. HYDROPHILA 853

primers for the I and O ends of mini-Tn5Km-1 (5�-AGATCTGATCAAGAGACAG-3� and 5�-ACTTGTGTATAAGAGTCAG-3�, respectively), as well asprimers M13for and T3.

Nucleotide sequencing and computer sequence analysis. Plasmid DNA forsequencing was isolated by using a QIAGEN plasmid purification kit (QIAGEN,Inc. Ltd.) as recommended by the supplier. In some cases, inverse PCR was usedto amplify a chromosomal DNA fragment that was not present in the A. hy-drophila library for sequencing. Briefly, 2 �g of A. hydrophila AH-3 chromosomalDNA was digested with appropriate restriction endonucleases, cleaned, andsubjected to overnight ligation. The ligation reaction mixture was phenol andchloroform extracted and resuspended in water, and 100 to 200 ng of ligatedDNA was then subjected to inverse PCR with specific primers. PCR productswere visualized in an agarose gel, and amplified fragments were recovered forDNA sequencing. Double-stranded DNA sequencing was performed by usingthe Sanger dideoxy chain termination method (36) with an ABI Prism dyeterminator cycle sequencing kit (Perkin-Elmer). Custom-designed primers usedfor DNA sequencing were purchased from Amersham Biosciences.

The DNA sequence was translated in all six frames, and all open readingframes (ORFs) that were more than 100 bp long were inspected. Deduced aminoacid sequences were compared with sequences in the GenBank, EMBL, andSwiss-Prot databases by using the BLASTX, BLASTP, or PSI-BLAST networkservice at the National Center for Biotechnology Information (3). A proteinfamily profile was determined using the Pfam Protein Family Database at theSanger Center (5). Possible terminator sequences were determined by using theTerminator program from the Genetics Computer Group package (GeneticsComputer Group, Madison, Wis.). Other online sequence analysis services werealso used.

RT-PCR. Total RNA was isolated from A. hydrophila AH-3 grown in solid agar(TSA) by using the Trizol reagent (Invitrogen). To ensure that RNA was devoidof contaminating DNA, the preparation was treated with RNase-free DNase I(amplification grade; Invitrogen). The isolated RNA was used as a template inreverse transcription PCRs (RT-PCRs), utilizing the Thermoscript RT-PCRsystem (Invitrogen) according to the manufacturer’s instructions. A PCR withoutreverse transcriptase was also performed to confirm the absence of contaminat-ing DNA in the RNA samples. RT-PCR amplifications were performed at leasttwice with total-RNA preparations obtained from a minimum of two indepen-dent extractions. The RT-PCR and PCR products were analyzed by agarose gelelectrophoresis.

Construction of defined insertion mutants. To obtain single defined insertionmutants with mutations in the flhAL, fliJL, flgNL, maf-5, and rpoN genes, we useda method based on suicide plasmid pFS100 (34). Briefly, an internal fragment ofthe selected gene was amplified by PCR, ligated into pGEM-T Easy (Promega),and transformed into E. coli XL1-Blue. The DNA insert was recovered by EcoRIrestriction digestion and was ligated into EcoRI-digested and phosphatase-treated pFS100. The ligation preparation was transformed into E. coli MC1061(�pir) with selection for kanamycin resistance (Kmr). Triparental mating with themobilizing strain HB101/pRK2073 was used to transfer the recombinant plasmid

into A. hydrophila AH-405 rifampin-resistant (Rifr) strains to obtain definedinsertion mutants, with selection for Rifr and Kmr.

To obtain mutants with mutations in lafK and flgEL, the genes were amplifiedby PCR, ligated into the vector pGEM-T Easy (Promega), and transformed intoE. coli XL1-Blue. The Tn5-derived kanamycin resistance cartridge (nptll) frompUC4-KIXX was inserted into each of the genes. This cartridge contains anoutward-reading promoter that ensures expression of downstream genes when itis inserted in the correct orientation (6); however, such an insertion alters theregulation of the genes. The SmaI-digested cassette was inserted into a restric-tion site internal to each gene, and the presence of a single HindIII site in theSmaI-digested cassette allowed its orientation to be determined. Constructscontaining the mutated genes were ligated to suicide vector pDM4 (25) andelectroporated into E. coli MC1061 (�pir), which was plated on chloramphenicolplates at 30°C. Plasmids with mutated genes were transferred into rifampin-resistant A. hydrophila AH-405 by triparental mating using E. coli MC1061 (�pir)containing the insertion constructs and the mobilizing strain HB101/pRK2073.Transconjugants were selected on plates containing chloramphenicol, kanamy-cin, and rifampin. PCR analysis confirmed that the vector had integrated cor-rectly into the chromosomal DNA. To complete the allelic exchange, the inte-grated suicide plasmid was forced to recombine out of the chromosome byadding 5% sucrose to the agar plates. The pDM4 vector contains sacB, whichproduces an enzyme that converts sucrose into a product that is toxic to gram-negative bacteria. Transconjugants surviving on plates with 5% sucrose that wererifampin resistant, kanamycin resistant (Kmr), and chloramphenicol sensitive(Cms) were chosen and confirmed by PCR.

Plasmid construction. For complementation studies A. hydrophila AH-3 DNAfragments with fliML-flhAL and lafK-flgAL clusters were PCR amplified usingprimers FLIMF1 (5�-GCTCTAGATGCAACAGAGAGCAAACCG-3�) andVIRG (5�-GCTCTAGAGATTGGGAATGGATTGG-3�) and primers FHIA(5�-GCTCTAGAAGTTATTGGGACACTGGG-3�) and LFGNF1 (5�-GCTCTAGAGCTGCGGGTCAAGCAAC-3�), respectively. Amplified fragments thatwere 7,584 bp and 7,822 bp long, respectively, were XbaI digested, blunt endedwith Klenow DNA polymerase (the XbaI sites are underlined in the primers),and ligated into pLA2917 (1). The wild-type DNA fragment with the flgB to flgLL

cluster and the maf-5 gene were amplified using primers LFGBF (5�-GAAGATCTGTGCATTCAGCCAGATAG-3�) and LFGLR (5�-GAAGATCTGATCCAGCCTTGAAACCAC-3�) and primers LFGFIN (5�-GAAGATCTCTTAAACGTCTGGAGCAGC-3�) and LAFAB (5�-GAAGATCTGGAGAAAATTGAGCCGGAG-3�), respectively. Amplified fragments of 9,724 bp and 3,212 bp,respectively, were BglII digested (BglII sites are underlined on the primers)ligated separately into pLA2917 (1). Attempts to introduce the lateral flagellarplasmids into E. coli DH5� by transformation were unsuccessful. However, wewere able to introduce these plasmids into the nonflagellated organism K. pneu-moniae strain 52145 by electroporation, resulting in plasmids pLA-FLIL1 (fliML-flhAL), pLA-FLIL2 (lafK-flgAL), pLA-FLGL (flgBL-flgLL), and pLA-MAFL(maf-5) (Fig. 1). A plasmid containing only the rpoN gene from A. hydrophilaAH-3 was obtained by PCR amplification of genomic DNA using oligonucleo-

FIG. 1. Genetic organization of the A. hydrophila AH-3 lateral flagellar gene region. ORFs and their directions of transcription are indicatedby large arrows; the designations of the ORFs are based on the designations of homologues in other bacterial species. Bent arrows indicate thelocations of putative promoter sequences. Lollipop symbols indicate the approximate positions of the putative rho-independent transcriptionalterminators. The cross-hatched large arrows indicate lateral flagellar genes described previously (10). The five clusters determined by RT-PCRsin the sequenced region (small arrows) and the genes contained in different plasmids are also indicated.

854 CANALS ET AL. J. BACTERIOL.

tides 5�-TGTCTTGATCACCGACCAC-3� and 5�-GCTTGTCCAGCAGGGTATC-3� to generate a 1,927-bp band. The amplified band was ligated into pGEM-TEasy (Promega). The DNA insert was recovered by EcoRI restriction digestionand was ligated into EcoRI phosphatase-treated pACYC184 (34) to generateplasmid pACYC-RPON. This vector was electroporated into E. coli DH5�, andrecombinant plasmids surviving on tetracycline plates were selected on the basisof chloramphenicol sensitivity.

Whole-cell protein preparation and immunoblotting. Whole-cell proteinswere obtained from Aeromonas strains grown at 30°C. Equivalent numbers ofcells were harvested by centrifugation, and each cell pellet was suspended in 50to 200 �l of sodium dodecyl sulfate-polyacrylamide gel electrophoresis loadingbuffer and boiled for 5 min. Following sodium dodecyl sulfate-polyacrylamide gelelectrophoresis and transfer to nitrocellulose membranes, the membranes wereblocked with bovine serum albumin (3 mg/ml) and probed with either polyclonalrabbit anti-polar flagellin or anti-lateral flagellin antibodies (1:1,000) that wereobtained previously (10). The unbound antibody was removed by three washes inphosphate-buffered saline (PBS), and a goat anti-rabbit horseradish peroxidase-conjugated secondary antibody (1:1,000) was added. The unbound secondaryantibody was removed by three washes in PBS. The bound conjugate was thendetected by addition of 2 ml of 0.5% 4-chloro-1-naphthol (Sigma) prepared inmethanol and diluted in 8 ml PBS containing 50 �l of 30% H2O2.

Assay of adherence to HEp-2 cells. An HEp-2 tissue culture was maintained asdescribed by Thornley et al. (40). The adherence assay was performed by usinga slight modification of the method described by Carrello et al. (7). Bacteria weregrown statically in brain heart infusion broth at 37°C, harvested by gentle cen-trifugation (1,600 g, 5 min), and resuspended in PBS (pH 7.2) at a concen-tration of approximately 107 CFU/ml (A600, 0.07). The HEp-2 monolayer wasinfected with 1 ml of the bacterial suspension for 90 min at 37°C in 5% CO2.Following infection, the nonadherent bacteria were removed from the monolayerby three washes with PBS. The remaining adherent bacteria and the monolayerwere then fixed in 100% methanol for 5 min. Methanol was removed by washingwith PBS, and the HEp-2 cells with adherent bacteria were stained for 45 min in10% (vol/vol) Giemsa stain (BDH) prepared in Giemsa buffer. The coverslipswere air dried, mounted, and viewed by oil immersion using a light microscope.Twenty HEp-2 cells/coverslip were randomly chosen, and the number of bacteriaadhering/HEp-2 cell was recorded. Assays were carried out in duplicate ortriplicate.

Biofilm formation. A quantitative biofilm formation experiment was per-formed in a microtiter plate as described previously (31), with minor modifica-tions. Briefly, bacteria were grown on TSA, and several colonies were gentlyresuspended in TSB (with or without the appropriate antibiotic); 100-�l aliquotswere place in a microtiter plate (polystyrene) and incubated 48 h at 30°C withoutshaking. After the bacterial cultures were poured out, the plate was washedextensively with water, fixed with 2.5% glutaraldehyde, washed once with water,and stained with a 0.4% crystal violet solution. After solubilization of the crystalviolet with ethanol-acetone (80:20, vol/vol) the absorbance at 570 nm was deter-mined.

Statistical analysis. The differences in adherence to HEp-2 cells or biofilmformation in vitro between the wild-type and mutant strains were analyzed by thet test, using the Microsoft Excel software.

Nucleotide sequence accession numbers. The nucleotide sequences deter-mined in this study have been deposited in the GenBank/EMBL database underaccession numbers DQ124694 and DQ124695.

RESULTS

Isolation and characterization of A. hydrophila AH-3 mu-tants with a reduced swarming phenotype. In order to find theA. hydrophila AH-3 lateral flagellar genetic regions, we per-formed mini-Tn5Km-1 mutagenesis using A. hydrophila AH-405 (rifampin-resistant AH-3) as the recipient strain, andtransconjugants were screened for an altered or reducedswarming phenotype on swarm agar. Of 7,500 transconjugantsanalyzed, 34 transposon insertion mutants exhibited a repro-ducible reduction in swarming. The swimming motility of thesemutants was subsequently analyzed by light microscopy, andthe mutants which did not exhibit wild-type swimming motilitywere discarded. The remaining 12 transposon insertion mu-tants with a highly reduced swarming phenotype and wild-type

swimming motility were analyzed by transmission electron mi-croscopy after growth in swarm agar and in liquid media anddivided into two groups on the basis of their ability to producelateral flagella. The members of the first group were able toproduce polar flagella but not lateral flagella (nine mutants),and the members of the second group were able to produceboth types of flagella, although they had a reduced swarmingphenotype (three mutants).

As no EcoRV restriction sites were present in the transpo-son, the two groups of mutants were analyzed for the presenceof the transposon by Southern hybridization of EcoRV chro-mosomal DNA digests. A single band was detected in everymutant, indicating that each mutant had a single copy of themini-transposon in its genome (data not shown).

Sequence analysis of genes interrupted by mini-Tn5Km-1.The DNA flanking the transposon was isolated and cloned intopBCSK (see Materials and Methods). Nucleotide sequencingof the cloned fragments from most of the transposon insertionmutants in the first group (six of nine mutants) revealed ORFsthat exhibited high levels of homology to different structuraland regulatory A. hydrophila AH-3 lateral flagellar genes (lafAto lafS) reported previously by our laboratory (10). However,the amino acid sequence predicted from DNA flanking themini-transposon of mutant AH-5501 exhibited homology toFlhAL of the V. parahaemolyticus lateral flagellar export appa-ratus (38). The amino acid sequence predicted from DNAflanking the mini-transposon in the other two insertion mu-tants exhibited extremely low levels of homology to possiblechaperones, which recently have been identified as ORF14 andORF15 (Fig. 1). Therefore, we decided to focus on mutantAH-5501. Nucleotide sequencing of members of the secondmutant group revealed ORFs that exhibited high levels ofhomology to the A. hydrophila AH-3 proton flagellar motorgenes lafT and lafU reported previously by our laboratory (10).

Organization of the A. hydrophila AH-3 lateral flagellar clus-ter. Transposon flanking sequences from the AH-5501 mutantwere used to synthesize an internal probe of the flhAL-likegene and to screen a previously constructed genomic library ofA. hydrophila AH-3 (28). We screened 12,000 recombinantclones, and no positive recombinant was identified. In order toanalyze the A. hydrophila lateral flagellar loci, progressive in-verse PCR with specific oligonucleotides (see Materials andMethods) was used to amplify and sequence this chromosomalregion that was not present in the A. hydrophila AH-3 library.

Sequence analysis of amplified fragments revealed a28,000-bp gene cluster upstream of the A. hydrophila AH-3lafA lateral flagellin gene (10); this region contained 29 com-plete ORFs, most of which were related to the V. parahaemo-lyticus lateral flagellar gene system (39) (Fig. 1). ORF1 (fliML)to ORF14 (fliJL) and lafA-U (downstream of ORF29) (Fig. 1)were similar to region 2 of the V. parahaemolyticus lateralflagellar gene system, with no putative motYL motor gene,whereas ORF15 (flgNL) to ORF28 (flgLL) were related toregion 1 of the V. parahaemolyticus lateral flagellar gene system(Fig. 1 and 2). ORF1 (fliML) to ORF14 (fliJL) and ORF18(flgBL) to ORF29 (maf-5) were transcribed in the same direc-tion, and ORF17 (flgAL) to ORF15 (flgNL) were transcribed inthe opposite direction (Fig. 1). Upstream of ORF1 and tran-scribed in the same direction was a truncated ORF that en-coded a homologue of the putative methyltransferase YarL,

VOL. 188, 2006 LATERAL FLAGELLAR GENE SYSTEM OF A. HYDROPHILA 855

which appears not to be involved either in motility or in flagel-lar biosynthesis; a strong stem-loop termination sequence, AAAATCCCGGCACCTTCTGGTGCCGGGATTTT, was lo-cated 469 bp downstream of its TAA stop codon. ORF7(flhAL) was separated from ORF8 (lafK) by 423 bp; ORF28(flgLL) was separated from ORF29 (maf-5) by 1,073 bp; andORF29 was separated from A. hydrophila AH-3 lafA (9) by 459bp. The other ORFs overlapped or were located one behindthe other with intergenic regions less than 80 bp long. PutativeShine-Dalgarno sequences were found upstream of each ORF.Data summarizing the locations of the 29 complete ORFs areshown in Table 2. Sequence analysis in silico revealed possibletranscriptional terminator rho-independent sequences down-stream of ORF7 (flhAL), ORF14 (fliJL), ORF28 (flgLL), andORF29 (maf-5). Putative �54 promoter sequences were foundupstream of ORF1 (fliML), ORF17 (flgAL), ORF18 (flgBL),and ORF29 (maf-5); a putative �28 promoter sequence waslocated upstream of ORF16 (flgML); and a different putativepromoter sequence was found upstream of ORF8 (lafK) (Fig.1). RT-PCR using specific primers and total RNA from A.hydrophila AH-3 grown on solid agar (TSA) showed that therewas amplification (data not shown) between ORF1 to ORF7(fliML to flhAL), ORF8 to ORF14 (lafK to fliJL), ORF17 toORF15 (flgAL to flgNL), and ORF18 to ORF28 (flgBL to flgLL)and ORF29 (maf-5) (Fig. 1). However, no amplification (datanot shown) was obtained with pairs of oligonucleotides fromORF7 (flhAL) and ORF8 (lafK), ORF28 (flgLL) and ORF29(maf-5), and ORF29 and lafA, confirming that at least sixclusters are present in the A. hydrophila AH-3 lateral flagellargene region. The nomenclature of the A. hydrophila lateralflagellar genes is consistent with that of the V. parahaemolyticusgenes (39), which are designated with reference to the genes

encoding the polar flagella, so that fliML is the lateral flagellargene equivalent of the polar flagellar gene fliM (Fig. 1).

The characteristics of the individual proteins and their pro-tein homologues were analyzed using the BLASTP program(3) of the National Center for Biotechnology Information da-tabase and are shown in Table 2. The first 13 A. hydrophilalateral flagellar proteins (FliM, FliN, FliP, FliQ, FliRL, FlhB,FlhAL, LafK, FliE, FliF, FliG, FliH, and FliIL homologues)exhibited 30 to 65% amino acid identity with the export andassembly (FliP, FliQ, FliR, FliH, FliIL, FlhB, and FlhAL),switch (FliM, FliN, and FliGL), regulatory (LafK), basal body(FliEL), and M-ring (FliFL) orthologous proteins in V. para-haemolyticus lateral flagellar system region 2 (39). In contrast,BLASTP analysis of the ORF14 protein (FliJ homologue)showed that its predicted amino acid sequence did not matchthe sequence of the V. parahaemolyticus orthologous proteinbut did exhibit a low level of identity with a potential non-specific chaperone involved in flagellar export in Pseudo-monas putida, FliJ (32% identity). ORF15 (FlgN gene homo-logue) encoded a protein that exhibited a low level of identitywith hypothetical proteins of Photobacterium profundum(PBPRA0027) and Yersinia pestis (YP3459) (26 and 22% iden-tity, respectively), as determined using BLASTP, although PSI-BLAST analysis did reveal identity with the lateral flagellarchaperone LfgN of E. coli O42 (32) and V. parahaemolitycus(39) (31 and 28% identity, respectively), which is required forfilament assembly. The predicted amino acid sequences en-coded by the next 13 ORFs, ORF16 to ORF28 (FlgML, FlgAL,FlgB, FlgC, FlgD, FlgE, FlgF, FlgG, FlgH, FlgI, FlgJ, FlgK,and FlgLL), exhibited 26 to 66% identity to the anti-�28 factor(FlgML), P and L-ring formation (FlgA, FlgH, and FlgIL), androd and hook formation (FlgB, FlgC, FlgD, FlgE, FlgF, FlgG,

FIG. 2. Comparative diagrams of the A. hydrophila AH-3 lateral flagellar gene region and lateral flagellar gene regions of V. parahaemolyticusand E. coli O42. Arrows with the same pattern represent homologous genes in the bacteria.

856 CANALS ET AL. J. BACTERIOL.

FlgJ, FlgK, and FlgLL) proteins of V. parahaemolyticus lateralflagellar system region 1 (39). The last protein (encoded byORF29) contained a protein family domain (DUF115) be-tween amino acids 194 and 347 and exhibited homology withproteins of Shewanella oneidensis (SO3259), Clostridium aceto-butylicum (CAC2196), and Campylobacter jejuni (Cj1337) (36,27, and 26% identity, respectively). A recent study suggestedthat CAC2196 of C. acetobutylicum and Cj1337 of C. jejuni aremembers of the maf (motility accessory factor) gene familyinvolved in flagellin modification and phase variation (16).

Construction of lateral flagellar gene defined insertion mu-tants and complementation studies. To determine the role ofsome of the identified genes in motility and flagellar biosyn-thesis, amplified internal fragments of flhAL, fliJL, flgNL, andmaf-5 were ligated to pFS100, creating pFS-FLHAL, pFS-FLIJL, pFS-FLGNL, and pFS-MAFL, respectively. The plas-mids were independently introduced into AH-405 by conjuga-tion. Defined insertion mutants with mutations in flhAL (AH-5504), lafK (AH-5503), fliJL (AH-5505), flgNL (AH-5506),flgEL (AH-5507), and maf-5 (AH-5508) were obtained. ThelafK and flgEL genes were mutated using the Km cassette toavoid polar effects. Correct construction of all mutants wasverified by Southern blot hybridization (data not shown). Mu-tants were subjected to swarming assays in semisolid agarplates and swimming assays with light microscopy in liquidmedia. All mutants exhibited reduced swarming motility withmotility zones that were approximately one-half those of thewild type, but they exhibited wild-type levels of swimmingmotility (Table 3), like mini-Tn5 mutant AH-5501. To confirm

that the alterations in swarming motility were in fact caused bya loss of the lateral flagella, an examination of both flagellartypes (polar and lateral) in mutants grown on semisolid andliquid media was carried out by TEM. All mutants exhibitedpolar flagella and were unable to produce lateral flagella onsemisolid media (Table 3). These results suggest that the A.hydrophila fliML-flhAL, lafK-fliJL, flgAL-flgNL, and flgBL-flgLL

TABLE 2. Characteristics of the A. hydrophila AH-3 lateral flagellar gene region

ORF Nucleotideposition

Protein size(amino acids)

Mol wt(103) pI Predicted function Homologous gene % Identity/

% similarity

1 1181–2054 291 31.9 4.94 Switch fliML of Vibrio parahaemolyticus 44/652 2064–2439 125 13.5 4.29 Switch fliNL of Vibrio parahaemolyticus 65/773 2441–3173 244 27 5.13 Export/assembly fliPL of Vibrio parahaemolyticus 63/784 3165–3444 93 10 4.5 Export/assembly fliQL of Vibrio parahaemolyticus 65/785 3443–4223 260 28.2 6.02 Export/assembly fliRL of Vibrio parahaemolyticus 51/696 4222–5356 378 42.7 9.99 Export/assembly flhBL of Vibrio parahaemolyticus 41/607 5362–7435 691 75.3 5.42 Export/assembly flhAL of Vibrio parahaemolyticus 57/738 7860–9234 458 50.7 6.73 Regulation lafK of Vibrio parahaemolyticus 50/689 9248–9572 108 11.8 4.69 Basal body component fliEL of Vibrio parahaemolyticus 50/7110 9585–11323 580 63.9 5.03 M-ring fliFL of Vibrio parahaemolyticus 42/6111 11322–12348 342 38.3 4.91 Switch fliGL of Vibrio parahaemolyticus 46/7012 12363–13053 230 25.3 4.72 Export/assembly fliHL of Vibrio parahaemolyticus 30/5113 13048–14377 443 48 5.71 Export ATP synthase fliIL of Vibrio parahaemolyticus 61/7914 14389–14803 138 16.2 9.51 Export/assembly fliJ of Pseudomonas putida 32/50a

15 15252–14813 147 16.6 6.41 Chaperone PBPRA0027 of Photobacterium profundum 26/3516 15524–15253 91 10.1 6.55 Anti-�28 lfgML of Vibrio parahaemolyticus 38/6117 16413–15605 270 30.4 9.39 P-ring assembly flgA of Vibrio parahaemolyticus 26/4918 16444–16831 129 13.9 5.53 Rod lfgBL of Vibrio parahaemolyticus 53/6819 16836–17259 141 15.3 4.62 Rod lfgCL of Vibrio parahaemolyticus 57/7520 17265–17958 231 24.6 4.69 Rod lfgDL of Vibrio parahaemolyticus 37/5521 17979–19170 397 41.7 5.15 Hook lfgEL of Vibrio parahaemolyticus 54/7022 19183–19915 244 25.9 5.1 Rod lfgFL of Vibrio parahaemolyticus 47/6723 19953–20778 275 29.4 4.75 Rod lfgGL of Vibrio parahaemolyticus 66/7724 20783–21449 222 24.5 9.23 L-ring lfgHL of Vibrio parahaemolyticus 60/7425 21409–22564 385 40.9 9.14 P-ring lfgIL of Vibrio parahaemolyticus 57/7326 22575–22971 132 14.7 5.02 Peptidoglycan hydrolase lfgJL of Vibrio parahaemolyticus 44/6327 22979–24290 437 46.8 5.04 HAP1 lfgKL of Vibrio parahaemolyticus 37/6528 24303–25200 299 32.7 4.87 HAP3 flgL of Vibrio parahaemolyticus 33/5729 26276–27560 428 48.2 6.37 Motility accesory factor SO3259 of Shewanella oneidensis 36/53

a Levels of identity and similarity obtained using PSI-BLAST.

TABLE 3. Phenotypes of the A. hydrophila AH-3 mutants obtained

StrainMotility Flagellation

typec Gene defectinterval

Swarminga Swimmingb Lateral Polar

AH-5502 � � � � rpoNAH-5503 � � � � lafKAH-5504 � � � � flhALAH-5505 � � � � fliJLAH-5506 � � � � flgNLAH-5507 � � � � flgELAH-5508 � � � � maf-5

a �, nonmotile in swarm semisolid plates; �, reduced expansion in swarmsemisolid plates; �, expansion equivalent to that of the wild-type strain. Thereduced expansion in swarm semisolid plates may have been due to the presenceof the constitutively expressed polar flagellum in mutants AH-5503 to AH-5508.

b �, nonmotile in swim semisolid plates and liquid media; �, motility equiv-alent to that of the wild-type strain in swimming plates or liquid media.

c �, absence of flagella; �, flagella equivalent to these of the wildtype. Thepresence of polar flagella was estimated by using swimming plates, light micros-copy (motility), electron microscopy, and Western blotting with specific anti-serum against polar flagellin. The presence of lateral flagella was estimated byusing swarming plates, electron microscopy, and Western blotting with specificantiserum against lateral flagellin.

VOL. 188, 2006 LATERAL FLAGELLAR GENE SYSTEM OF A. HYDROPHILA 857

clusters, as well as the maf-5 gene, are involved in biosynthesisof lateral flagella and not in the formation of polar flagella(Fig. 3).

Wild-type copies of the mutated genes were amplified inorder to complement in trans the motility phenotype andflagellar biosynthetic defects caused by the transposon and thedefined insertions, as described in Materials and Methods.Plasmids pLA-FLIL1, pLA-FLIL2, pLA-FLGL, and pLA-MAFL from K. pneumoniae strain 51245 were introduced sep-arately into AH-5501, AH-5503, AH-5504, AH-5505,AH-5506, AH-5507, and AH-5508 by mating. The AH-5501and AH-5504 mutants were able to swarm on plates and pro-duce lateral flagella (as determined by TEM) when plasmidpLA-FLIL1 was introduced into them. Plasmid pLA-FLIL2was able to fully complement AH-5503, AH-5505, and AH-5506. Plasmids pLA-FLGL and pLA-MAFL were able to com-plement motility and the flagellar phenotype in the AH-5507and AH-5508 mutants, respectively (Fig. 3). No complemen-tation was observed when the plasmid vector (pLA21917)alone was introduced into the mutants.

Additionally, when the COS-FLG plasmid (2) carrying theA. hydrophila AH-3 flg polar flagellar loci (flgA, flgM, flgNP, andflgB to flgLP) was introduced into AH-5506 and AH-5507, itwas unable to rescue motility or lateral flagellar expression.Although the flg polar flagellar genes have functions homolo-gous to those of their lateral flagellar counterparts, they couldnot perform their functions in lateral flagellar biogenesis.

Cloning of the A. hydrophila AH-3 alternative sigma factor�54 (rpoN). In order to clone the A. hydrophila alternativesigma factor �54 (rpoN), the genomic library of A. hydrophilaAH-3 (28) was transferred by mating into rifampin-resistant V.cholerae rpoN in-frame deletion mutant strain KKV56 (21).Transconjugants were selected for rifampin and tetracyclineresistance and inoculated into LB agar (0.3% agar) platessupplemented with 2 mM glutamine and tetracycline. Thecomplemented colonies which spread on the plates were iso-lated, and plasmid pLA-RPON was recovered. Sequence anal-ysis of pLA-RPON revealed a complete 1,440-bp ORF whichencoded a protein consisting of 479 amino acids and having apredicted molecular mass of 53.8 kDa. Moreover, we foundtwo putative promoter sequences approximately 111 bp and156 bp upstream of the start codon and a putative rho-inde-pendent sequence 179 bp downstream of the stop codon. Asearch of the protein database showed that the deduced aminoacid sequence exhibited high levels of homology (55 to 57%identity; 68 to 70% similarity) with the alternative �54 sigmafactors from different Enterobacteriaceae, such as Salmonellaenterica, E. coli, and Shigella flexneri, as well as from Photobac-terium profundum and different Vibrio species.

Construction of rpoN defined insertion mutants andcomplementation studies. To study the role of rpoN in theregulation of formation of Aeromonas flagella, an amplifiedinternal fragment of the rpoN gene was ligated into pFS100(pFS-RPON), and the suicide recombinant plasmid was intro-duced into AH-405 by conjugation. A defined insertion mutantwith a mutation in rpoN (AH-5502) was obtained, whose cor-rect construction was verified by Southern blot hybridization(data not shown). The AH-5502 mutant was tested for swarm-ing in semisolid plates and for swimming with light microscopyin liquid media (Table 3). The mutant was absolutely unable to

FIG. 3. (A) Transmission electron microscopy of A. hydrophilaAH-3 (wild type), mutant AH-5508, and mutant AH-5508 comple-mented with plasmid pLA-MAFL (panels 1, 2, and 3, respectively).Bacteria were grown at 30°C in solid medium (TSA). The AH-5508mutant was also grown at 30°C in liquid medium (TSB) (panel 4).Bacteria were gently placed onto Formvar-coated copper grids andnegatively stained using 2% uranyl acetate. Bars 0.5 �m. (B) West-ern blot analysis with anti-polar flagellin polyclonal antibodies (1:1,000) of whole-cell preparations of A. hydrophila AH-3 (wild-type),mutant AH-5508, and mutant AH-5508 complemented with plasmidpLA-MAFL(lanes 1, 2, and 3, respectively) grown at 30°C in liquidmedium (TSB). (C) Western blot analysis with anti-lateral flagellinpolyclonal antibodies (1:1,000) of whole-cell preparations of the strainsused in the experiment shown in panel B grown at 30°C in solidmedium (TSA).

858 CANALS ET AL. J. BACTERIOL.

swarm or swim. TEM assays after growth in semisolid or liquidmedia revealed that AH-5502 lacked both types of flagella(Fig. 4A), in contrast to the wild type. Specific immunoblots ofAH-5502 (rpoN) whole cells after growth in solid or liquidmedia using lateral flagellin- or polar flagellin-specific antibod-ies showed that the mutant was unable to produce polar orlateral flagellins (Fig. 4B). These results suggest that inactiva-tion of A. hydrophila rpoN eliminated formation of lateral andpolar flagella completely.

Complementation studies of the AH-5502 mutant with plas-mid pACYC-RPON containing the A. hydrophila AH-3 rpoNgene alone (see Materials and Methods) showed that there wascomplete recovery of motility (swimming and swarming) andthat polar and lateral flagella were present, as determined byTEM (Fig. 4). V. cholerae mutant strain KKV56, when com-plemented with plasmid pACYC-RPON, was able to spread onVibrio motility plates as fast as the wild type, and the formationof polar flagella was confirmed by TEM. No complementationwas observed when only the vector pACYC184 was introducedinto the mutants.

Distribution of the lateral flagellar genes and the rpoN genein mesophilic Aeromonas strains. The distribution of lateralflagellar genes in mesophilic Aeromonas strains was analyzedby dot blot hybridization experiments with total genomic DNAusing independent PCR probes. The distribution of lateralflagella was screened using five separate PCR probes for fliQL-flhBL, lafK-fliFL, flgA-NL, flgD-EL, and maf-5.

The percentages of strains positive for lateral flagellar geneswere 70% for fish isolates (7/10 strains), 62% for clinical strains(31/50 strains), and 55% for strains isolated from foods (22/40strains). Only five of the reference strains for hybridizationgroups, corresponding to groups 4, 5a, 6, 9, and 12, exhibited apositive hybridization reaction. Of the reference strains usedfor O serotyping (44 strains), 47% showed positive hybridiza-tion reactions, and these strains corresponded to serotypes 2, 3,7, 8, 9,10, 11, 12, 14, 16, 17, 19, 25, 26, 28, 30, 38, 39, 40, 42, and44. The representative Aeromonas strains A. caviae Sch3N andA. veronii biovar sobria AH-1 hybridized positively to all lateralflagellar probes. All the strains that were positive for lateralflagella as determined by dot blot hybridization were able toswarm on plates (data not shown). An internal rpoN probehybridized to the chromosomal DNA of all mesophilic Aero-monas strains tested.

Adhesion to HEp-2 cells and biofilm formation. We exam-ined the adhesion of the wild type and lateral flagellar andrpoN mutants to cultured monolayers of HEp-2 cells. Differ-ences in adherence were calculated by determining the averagenumbers of bacteria adhering to HEp-2 cells (Fig. 5A.). Also,we compared the abilities of the strains to form biofilms inmicrotiter plates (Fig. 5B). For the A. hydrophila wild-typestrain, AH-3, 18.3 (18.3 � 1.7) bacteria adhered per HEp-2cell, and this strains could form biofilms (optical density at 570nm, 1.3 [1.3 � 0.1]). For all the defined lateral flagellar mutantstrains, which were polar flagellum positive but lateral flagel-lum negative, there was an approximately 80% reduction inadhesion to HEp-2 cells and a 62% reduction in the ability toform biofilms compared to the wild-type strain. Similar obser-vations were obtained with the A. hydrophila lafB (AH-1982)and lafS (AH-1983) mutant strains described previously (10).For the AH-5502 (rpoN) mutant, which lacked both types offlagella, there was a dramatic reduction (89%) in adhesion toHEp-2 cells compared to the wild-type strain, as well as areduction in the ability to form biofilms (75%). When lateralflagellar mutant strain AH-5504 (flhAL) was complementedwith plasmid pLA-FLIL1, when AH-5503 (lafK), AH-5505(fliJL), and AH-5506 (flgNL) were complemented with plasmidpLA-FLIL2, and when AH-5507 (flgEL) and AH-5508 (maf-5)were complemented with plasmids pLA-FLGL and pLA-MAFL, respectively, adhesion and biofilm formation values

FIG. 4. (A) Transmission electron microscopy of A. hydrophilaAH-3 (wild-type), mutant AH-5502 (rpoN), and mutant AH-5502 com-plemented with plasmid pACYC-RPON (panels 1, 2, and 3, respec-tively) grown at 30°C in liquid medium (TSB) and the same strainsgrown at 30°C on solid medium (TSA) (panels 4, 5, and 6, respec-tively). Bacteria were gently placed onto Formvar-coated copper gridsand negatively stained using a 2% solution of uranyl acetate. Bars 0.5 �m. (B) Western blot analysis with anti-polar flagellin polyclonalantibodies (1:1,000) of whole-cell preparations of A. hydrophila AH-3(wild-type), mutant AH-5502 (rpoN), and mutant AH-5502 comple-mented with plasmid pACYC-RPON (lanes 1, 2, and 3, respectively)grown at 30°C in liquid medium (TSB). (C) Western blot analysis withanti-lateral flagellin polyclonal antibodies (1:1,000) of whole-cell prep-arations of the strains used in the experiment shown in panel B grownat 30°C in solid medium (TSA).

VOL. 188, 2006 LATERAL FLAGELLAR GENE SYSTEM OF A. HYDROPHILA 859

similar to those obtained for the wild type were obtained.Complementation of the AH-5502 (rpoN) mutant with plasmidpACY-RPON restored wild-type levels of adhesion to HEp-2cells and biofilm formation.

DISCUSSION

Certain mesophilic Aeromonas strains are able to producetwo different types of flagella depending on the growth me-dium; a single polar flagellum is expressed under all laboratoryconditions, whereas lateral flagella are produced only on solidand semisolid media. Although previous work demonstratedthat A. hydrophila AH-3 and A. caviae Sch3N produce lateralflagella, only a few genes involved in the formation of lateralflagella have been described previously (10). The isolation ofA. hydrophila AH-3 transposon mutants with polar flagella andwithout lateral flagella, followed by subsequent cloning and

gene walking, allowed us to genetically characterize 29 genes(Table 2) upstream of lafA (10). Most of these sequencedgenes encoded protein orthologues of V. parahaemolyticus lat-eral flagellar proteins or other flagellar proteins. In addition,we found downstream of flgLL a modification gene (maf-5)that has not been described previously in lateral flagellar sys-tems. Mutations in the export-assembly protein FlhAL, the�54-dependent transcriptional regulator LafK, the chaperonesFliJL and FlgNL, the hook protein FlgEL, and the modificationaccessory factor protein Maf-5 all eliminated formation of lat-eral flagella but did not result in the loss of polar flagella.These mutants also exhibited 80% reductions in adhesion toHEp-2 cells and 62% reductions in the ability to form biofilms(Fig. 5). The wild-type phenotype was restored by introductionof the pLA-FLIL1 plasmid into the flhAL mutant, by introduc-tion of pLA-FLIL2 into the lafK, fliJL, and flgNL mutants, byintroduction of pLA-FLGL into the flgEL mutant, and by in-troduction of pLA-MAFL into the maf-5 mutant. It is impor-tant to point out that we were unable to clone in E. coli theregions of the Aeromonas lateral flagellar gene cluster in plas-mids pLA-FLIL1, pLA-FLIL2, and pLA-GLGL, but this situ-ation was overcome by cloning them in a nonflagellated bac-terium, K. pneumoniae strain 52145. Transfer of the COS-FLGplasmid (2), carrying the A. hydrophila AH-3 flg polar flagellarloci (flgA, flgM, flgN, flgB-flgL), into flgEL and flgNL lateralflagellar mutants did not induce expression of lateral flagella.This suggests that although the genes encode similar proteins,they are specific for each specific type of flagella.

The A. hydrophila lateral flagellar gene system contains 38genes in a single chromosomal region (Fig. 2), similar to therecently described Flag-2 locus of E. coli O42 (33). This isdifferent from the situation in V. parahaemolyticus, whose lat-eral flagellar genes are distributed in two different chromo-somal regions (39). A. hydrophila RT-PCR assays showed thatmost of the switch and export-assembly genes are dividedamong two clusters transcribed in the same direction; the firstpredicted cluster contains fliM, fliN, fliP, fliQ, fliRL, flhB, andflhAL, and the second cluster contains the regulatory gene lafK,as well as fliE, fliF, fliG, fliH, fliI, and fliJL. Moreover, we foundupstream of fliML a putative �54 promoter sequence and down-stream of flhAL a putative terminator sequence. The regulatorygene lafK starts 423 bp upstream of flhAL, and downstream offliJL we found a putative terminator sequence. The two clustersare arranged and transcribed differently in V. parahaemolyticus,in which they are transcribed in opposite directions and con-tain the gene motYL (39). Similar to Flag-2 of E. coli O42 (33)and the lateral flagellar gene system of V. parahaemolyticus, theA. hydrophila lateral flagellar gene clusters do not contain theexport-assembly gene fliOL, which is typically found in otherflagellar gene systems. The role of FliO is poorly understood,even in S. enterica serovar Typhimurium and E. coli (37). TheA. hydrophila lateral flagellar gene clusters (flgA, flgM, flgNL,and flgB to flgLL) exhibit the same distribution and direction oftranscription as the V. parahameolyticus cluster in lateral flagel-lar region 1 (39). Interestingly, we found between flgLL andlafA (10) a modification accessory factor gene (maf-5) that istranscribed independently in the same direction as the flgB-flgLL cluster. The maf-5 gene-encoded protein is homologousto the modification accessory factor (Maf) proteins found inHelicobacter pylori, C. acetobutylicum, and C. jejuni. In all these

FIG. 5. (A) Adhesion of A. hydrophila AH-3 and lateral flagellarmutants to Hep-2 cells, expressed as the mean number of adherentbacteria per Hep-2 cell. (B) Biofilm formation abilities of A. hydrophilaAH-3 and lateral flagellar mutants. The optical density at 570 nm(O.D.570) was used to quantify the crystal violet retained by the bio-films on the microtiter plates after staining. Bar 1, AH-3 (wild type);bars 2 to 8, AH-5504 (flhAL), AH-5503 (lafK), AH-5505 (fliJL), AH-5506 (flgNL), AH-5507 (flgEL), AH-5508 (maf-5), and AH-5502(rpoN), respectively; bar 9, AH-5504 with plasmid pLA-FLIL1; bars 10to 12, AH-5503, AH-5505, and AH-5506 with plasmid pLA-FLIL2,respectively; bar 13, AH-5507 with plasmid pLA-FLGL; bar 14, AH-5508 with plasmid pLA-MAFL; bar 15, AH-5502 with plasmid pA-CYC-RPON. The averages of three independent experiments (eachexperiment performed in duplicate) are shown. The error bars indicatestandard deviations.

860 CANALS ET AL. J. BACTERIOL.

bacteria, genes encoding Maf proteins are linked to flagellarbiosynthesis genes and/or genes involved in sugar biosynthesisand transport (11, 16). Some reports have stated that mafgenes contain homopolymeric tracts that result in phase vari-ation via a slipped-strand mispairing mechanism. The fact thatinsertional mutation of the A. hydrophila AH-3 maf-5 geneeliminated only expression of lateral flagella (Fig. 3) and thefact that the A. hydrophila AH-3 lateral flagellins, like the A.caviae Sch3N lateral flagellins, are glycosylated (10) suggestthat the Maf-5 protein maybe involved in specific posttransla-tional glycosylation of lateral flagella but not polar flagella.Two possible roles could be attributed to this gene: a differencein the last step of polysaccharide synthesis (common for polarand lateral flagella) (unpublished data) linked to only lateralflagella or the specific glycosyl transferase able to link thepolysaccharide to the lateral flagellin. If the gene has the latterrole, we should be able to find another gene able to transfer thepolysaccharide to the polar flagellin.

Different reports have associated the alternative �54 factorrpoN to the formation of polar flagella in different bacteria,such as Pseudomonas aeruginosa, V. cholerae (20), V. parahae-molyticus (39), Vibrio anguillarum (30), and Vibrio alginolyticus(17). The lateral flagellar gene clusters of A. hydrophila, and V.parahaemolyticus (39) are transcribed from promoters recog-nized by two different, characteristic alternative sigma factors,a specific �28 factor that is encoded by the lafS gene (10) anda �54 factor that has not been described previously in Aeromo-nas. The A. hydrophila RpoN protein has the three typicaldomains of �54 factors, the activator interaction domain be-tween amino acids 3 and 51, the core binding domain betweenamino acids 106 and 306, and the DNA binding domain be-tween amino acids 318 and 477, as well as �54 factor familysignature motif 1 (PMVLNDIAEAVEMHESTISRV) and sig-nature motif 2 (RRTIAKY) between amino acids 366 and 386and amino acids 457 and 463, respectively. Defined insertioninto the A. hydrophila rpoN gene resulted in a loss of motility(swarming and swimming), the absence of both types of flagella(lateral and polar), and elimination of polar and lateral flagel-lin expression (Fig. 4). The wild-type phenotypes were restoredwhen plasmid pACYC-RPON was transferred into the mutantstrain. In summary, these results indicated that �54 is essentialfor transcription of both polar and lateral flagellar gene sys-tems, even though both systems have specific �28 factors. Thisis the first demonstration that RpoN regulates biosynthesis ofAeromonas lateral flagella.

By comparing the results obtained with mutants that areunable to produce both polar and lateral flagella (rpoN mu-tant), mutants that are able to produce polar flagella but notlateral flagella (24; this study), and mutants that are able toproduce lateral flagella but not polar flagella (2), we concludedthat both types of flagella contribute to HEp-2 cell adhesionand biofilm formation in A. hydrophila AH-3. Only the polarflagella in V. parahaemolyticus appear to be involved in thesepathogenic features (9).

ACKNOWLEDGMENTS

This work was supported by grants from Plan Nacional de I�D(Ministerio de Ciencia y Tecnologıa, Spain), from Generalitat de Cata-lunya, and from the Wellcome Trust. R.C., M.A., and S.V. were sup-ported by fellowships from the Universidad de Barcelona, Ministerio

de Ciencia y Tecnologıa (Spain), and Generalitat de Catalunya, re-spectively.

We thank Maite Polo for her technical assistance and K. E. Klose forproviding strain KKV56.

REFERENCES

1. Allen, L. N., and R. S. Hanson. 1985. Construction of broad-host-rangecosmid cloning vector: identification of genes necessary for growth of Methy-lobacterium organophilum on methanol. J. Bacteriol. 161:955–962.

2. Altarriba, A., S. Merino, R. Gavın, R. Canals, A. Rabaan, J. G. Shaw, andJ. M. Tomas. 2003. A polar flagella operon (flg) of Aeromonas hydrophilacontains genes required for lateral flagella expression. Microb. Pathog. 34:249–259.

3. Altschul, F. S., T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zang, W. Miller,and J. Lipman. 1997. Gapped BLAST and PSI-BLAST: a new generation ofprotein database search programs. Nucleic Acids Res. 25:3389–3402.

4. Austin, B., and C. Adams. 1996. Fish pathogens, p. 197–243. In B. Austin, M.Altwegg, P. J. Gosling, and S. W. Joseph (ed.), The genus Aeromonas. JohnWiley and Sons, New York, N.Y.

5. Bateman, A., E. Birney, L. Cerruti, R. Durbin, L. Etwiller, S. R. Eddy, S.Griffiths-Jones, K. L. Howe, M. Marshall, and E. L. L. Sonnhammer. 2002.The pfam protein families database. Nucleic Acids Res. 30:276–280.

6. Bott, M., M. Meyer, and P. Dimroth. 1995. Regulation of anaerobic citratemetabolism in Klebsiella pneumoniae. Mol. Microbiol. 18:533–546.

7. Carrello, A., K. A. Silburn, J. R. Budden, and B. J. Chang. 1988. Adhesionof clinical and environmental Aeromonas isolates to Hep-2 cells. J. Med.Microbiol. 26:19–27.

8. De Lorenzo, V., M. Herrero, U. Jakubzik, and K. N. Timmis. 1990. Mini-Tn5transposon derivatives for insertion mutagenesis, promoter probing, andchromosomal insertion of cloned DNA in gram-negative eubacteria. J. Bac-teriol. 172:6568–6572.

9. Enos-Berlage, J. L., Z. T. Guvener, C. E. Keena, and L. L. McCarter. 2005.Genetic determinants of biofilm development of opaque and translucentVibrio parahaemolyticus. Mol. Microbiol. 55:1160–1182.

10. Gavın, R., A. A. Rabaan, S. Merino, J. M. Tomas, I. Gryllos, and J. G. Shaw.2002. Lateral flagella of Aeromonas species are essential for epithelial celladherence and biofilm formation. Mol. Microbiol. 43:383–397.

11. Guerry, P., P. Doig, R. A. Alm, D. H. Burr, N. Kinsella, and T. J. Trust. 1996.Identification and characterization of genes required for post-translationalmodification of Campylobacter coli VC167 flagellin. Mol. Microbiol. 19:369–378.

12. Hanahan, D. 1983. Studies on transformation of Escherichia coli with plas-mids. J. Mol. Biol. 166:557–580.

13. Inoue, K., Y. Kosako, K. Suzuky, and T. Shimada. 1991. Peritrichous flag-ellation in Plesiomonas shigelloides strains. Jpn. J. Med. Sci. Biol. 44:141–146.

14. Janda, J. M., and S. L. Abbott. 1998. Evolving concepts regarding the genusAeromonas: an expanding panorama of species, disease presentation, andunanswered questions. Clin. Infect. Dis. 27:332–344.

15. Jiang, Z. Y., and C. E. Bauer. 1997. Analysis of a chemotaxis operon fromRhodospirillum centenum. J. Bacteriol. 179:5712–5719.

16. Karlinshev, A. V., D. Linton, N. A. Gregson, and B. W. Wren. 2002. A novelparalogous gene family involved in phase-variable flagella-mediated motilityin Campylobacter jejuni. Microbiology 148:473–480.

17. Kawagishi, I., M. Nakada, N. Nishioka, and M. Homma. 1997. Cloning of aVibrio alginolyticus rpoN gene that is required for polar flagella formation. J.Bacteriol. 179:6851–6854.

18. Kirov, S. M., B. C. Tassel, A. B. T. Semmler, L. A. O’Donovan, A. A. Rabaan,and J. G. Shaw. 2002. Lateral flagella and swarming motility in Aeromonasspecies. J. Bacteriol. 184:547–555.

19. Kirov, S. M. 2003. Bacteria that express lateral flagella enable dissection ofthe multifunctional roles of flagella in pathogenesis. FEMS Microbiol. Lett.224:151–159.

20. Klose, K. E., and J. J. Mekalanos. 1998. Distinct roles of an alternative sigmafactor during both free-swimming and colonizing phases of the Vibrio chol-erae pathogenic cycle. Mol. Microbiol. 28:501–520.

21. McCarter, L. L., and M. Silverman. 1990. Surface-induced swarmer celldifferentiation of Vibrio parahaemolyticus. Mol. Microbiol. 4:1057–1062.

22. Merino, S., S. Camprubi, and J. M. Tomas. 1991. The role of lipopolysac-charide in complement-killing of Aeromonas hydrophila strains of serotypeO:34. J. Gen. Microbiol. 137:1583–1590.

23. Merino, S., X. Rubires, A. Aguilar, and J. M. Tomas. 1996. The O:34-antigenlipopolysaccharide as a adhesin in Aeromonas hydrophila. FEMS Microbiol.Lett. 139:97–101.

24. Merino, S., X. Rubires, A. Aguilar, and J. M. Tomas. 1997. The role offlagella and motility in the adherence and invasion to fish cell lines byAeromonas hydrophila serogroup O:34 strains. FEMS Microbiol. Lett. 151:213–217.

25. Milton, D. L., R. O’Toole, P. Horstedt, and H. Wolf-Watz. 1996. Flagellin Ais essential for the virulence of Vibrio anguillarum. J. Bacteriol. 178:1310–1319.

26. Moens, S., K. Michiels, V. Keijer, F. Van Leuven, and J. Vanderleyden. 1995.

VOL. 188, 2006 LATERAL FLAGELLAR GENE SYSTEM OF A. HYDROPHILA 861

Cloning, sequencing, and phenotypic analysis of laf1, encoding the flagellinof the lateral flagella of Azospirillum brasilense Sp7. J. Bacteriol. 177:5419–5426.

27. Nassif, X., J. M. Fournier, J. Arondel, and P. J. Sansonetti. 1989. Mucoidphenotype of Klebsiella pneumoniae is a plasmid-encoded virulence factorInfect. Immun. 57:546–552.

28. Nogueras, M. M., S. Merino, A. Aguilar, V. J. Benedi, and J. M. Tomas. 2000.Cloning, sequencing, and role in serum susceptibility of porin II from me-sophilic Aeromonas hydrophila. Infect. Immun. 68:1849–1854.

29. O’Rouke, J., A. Lee, and J. G. Fox. 1992. An ultrastructural study of Heli-cobacter mustelae and evidence of a specific association with gastric mucosa.J. Med. Microbiol. 36:420–427.

30. O’Toole, R., D. L. Milton, P. Horstedt, and H. Wolf-Watz. 1997. RpoN of thefish pathogen Vibrio (Listonella) anguillarum is essential for flagellum pro-duction and virulence by the water-borne but not intraperitoneal route ofinoculation. Microbiology 143:3849–3859.

31. Pratt, L. A., and R. Kotler. 1998. Genetic analysis of Escherichia coli biofilmformation: role of flagella, motility, chemotaxis and type I pili. Mol. Micro-biol. 30:285–293.

32. Rabaan, A. A., I. Gryllos, J. M. Tomas, and J. G. Shaw. 2001. Motility andthe polar flagellum are required for Aeromonas caviae adherence to Hep-2cells. Infect. Immun. 69:4257–4267.

33. Ren, C.-P., S. A. Beatson, J. Parkhill, and M. J. Pallen. 2005. The Flag-2

locus, an ancestral gene cluster, is potentially associated with a novel flagellarsystem from Escherichia coli. J. Bacteriol. 187:1430–1440.

34. Rubires, X., F. Saigi, N. Pique, N. Climent, S. Merino, S. Albertı, J. M.Tomas, and M. Reque. 1997. A gene (wbbL) from Serratia marcescens N28b(O4) complements the rfb-50 mutation of Escherichia coli K-12 derivatives. J.Bacteriol. 179:7581–7586.

35. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: alaboratory manual, 2nd ed. Cold Spring Harbor Laboratory, Cold SpringHarbor, N.Y.

36. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencing withchain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463–5467.

37. Schoenhals, G. J., M. Kihara, and R. M. MacNab. 1998. Translation of theflagellar gene fliO of Salmonella typhimurium from putative tandem starts. J.Bacteriol. 180:2936–2942.

38. Shimada, T., R. Sakazaki, and K. Suzuki. 1985. Peritrichous flagella inmesophilic strains of Aeromonas. Jpn. J. Med. Sci. Biol. 38:141–145.

39. Stewart, B. J., and L. McCarter. 2003. Lateral flagellar gene system of Vibrioparahaemolyticus. J. Bacteriol. 185:4508–4518.

40. Thornley, J. P., J. G. Shaw, I. Gryllos, and A. Eley. 1996. Adherence ofAeromonas caviae to human cell lines Hep-2 and Caco-2. J. Med. Microbiol.45:445–451.

41. Yun-Kyeong, K., and L. McCarter. 2000. Analysis of the polar flagellar genesystem of Vibrio parahaemolyticus. J. Bacteriol. 182:3693–3704.

862 CANALS ET AL. J. BACTERIOL.