Regulation of Capsule Synthesis and Cell Motility in Salmonella enterica by the Essential Gene igaA

Corresponding author: Josep Casadesús, Facultad de Biología, Universidad de Sevilla, Avenida Reina Mercedes 6, Sevilla 41012, Spain., casadesus{at}us.es (E-mail)

Communicating editor: B. L. BASSLER

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mutants of Salmonella enterica carrying the igaA1 allele, selected as able to overgrow within fibroblast cells in culture, are mucoid and show reduced motility. Mucoidy is caused by derepression of wca genes (necessary for capsule synthesis); these genes are regulated by the RcsC/YojN/RcsB phosphorelay system and by the RcsA coregulator. The induction of wca expression in an igaA1 mutant is suppressed by mutations in rcsA and rcsC. Reduced motility is caused by lowered expression of the flagellar master operon, flhDC, and is suppressed by mutations in rcsB or rcsC, suggesting that mutations in the igaA gene reduce motility by activating the RcsB/C system. A null igaA allele can be maintained only in an igaA⁺/igaA merodiploid, indicating that igaA is an essential gene. Lethality is suppressed by mutations in rcsB, rcsC, and yojN, but not in rcsA, suggesting that the viability defect of an igaA null mutant is mediated by the RcsB/RcsC system, independently of RcsA (and therefore of the wca genes). Because all the defects associated with igaA mutations are suppressed by mutations that block the RcsB/RcsC system, we propose a functional interaction between the igaA gene product and either the Rcs regulatory network or one of its regulated products.

A screen for mutants able to overgrow within a fibroblast cell line identified functions of Salmonella enterica involved in growth restraint within nonphagocytic eukaryotic cells (CANO et al. 2001 ). One of such mutants carried a point mutation in a hitherto unknown locus, located near mcrA at centisome 75. This novel locus was called igaA (intracellular growth attenuator; EMBL accession no. AJ301649). The same study showed that IgaA^- mutants of S. enterica are mucoid and avirulent in the murine typhoid model (CANO et al. 2001 ).

Analysis in silico of the igaA DNA sequence indicated the existence of an open reading frame (ORF) homologous to the yrfF ORF of Escherichia coli. The yrfF and igaA ORFs are located at an equivalent position on the chromosomal gene maps of E. coli and and S. enterica serovar Typhimurium (BLATTNER et al. 1997 ; MCCLELLAND et al. 2001 ). A BLAST databank search (ALTSCHUL et al. 1997 ) indicated that the igaA and yrfF loci are in turn homologous to the umoB gene of Proteus mirabilis (DUFOUR et al. 1998 ). The umoB gene was identified as a multicopy suppressor of the swarming defect associated with flgN mutations and proved to be required for both swarming and motility (DUFOUR et al. 1998 ). Current evidence suggests that umoB encodes a membrane protein that regulates genes involved in flagellar synthesis (DUFOUR et al. 1998 ).

Below we show that the igaA gene of S. enterica is essential under laboratory conditions. We also provide evidence that the igaA gene product of S. enterica is a pleiotropic regulator that exerts positive control on the flagellar master operon flhDC and negative control on the colanic acid cluster wca. Genetic analysis also unveils a functional relationship between the igaA gene product and the two-component regulatory system RcsB-RcsC. Originally described as a regulator of capsule synthesis in E. coli (GOTTESMAN et al. 1985 ), the RcsB-RcsC system has been shown to participate in a variety of cellular processes, which include cell division control in E. coli (CARBALLES et al. 1999 ), regulation of Vi antigen synthesis in Salmonella typhi (VIRLOGEUX et al. 1996 ), synthesis of flagellin and invasion proteins in S. typhi (ARRICAU et al. 1998 ), expression of the E. coli tolQRA operon (CLAVEL et al. 1996 ), synthesis of the E. coli outer membrane protein OsmC (DAVALOS-GARCIA et al. 2001 ), resistance to chlorpromazine-induced stress in E. coli (CONTER et al. 2002 ), and production of exopolysaccharide by Erwinia amylovora (BERESWILL and GEIDER 1997 ). The RcsB-RcsC system is made of two main components, the sensor protein RcsC and the transcriptional activator RcsB (BRILL et al. 1988 ), but includes also a second transcriptional activator, RcsA (GOTTESMAN et al. 1985 ); a recently described ancillary protein, YojN (TAKEDA et al. 2001 ); and additional sensors that transmit external signals to the membrane-bound RcsC sensor (CLARKE et al. 1997 ; KELLEY and GEORGOPOULOS 1997 ; CHEN et al. 2001 ). We show that mutations affecting any of the rcsB, rcsC, and yojN genes suppress the defects associated with igaA mutations (lethality, mucoidy, and reduced motility). Mucoidy is also suppressed by rcsA mutations. These observations provide genetic evidence that IgaA might be part of a novel signaling pathway involving the RcsB-RcsC system.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Bacterial strains, bacteriophages, and strain construction:
All the S. enterica strains listed in Table 1 belong to serovar Typhimurium. Unless indicated otherwise, the strains derive from the mouse-virulent strain SL1344 (HOISETH and STOCKER 1981 ). Transductional crosses using phage P22 HT 105/1 int201 (SCHMIEGER 1972 ; G. ROBERTS, unpublished results), henceforth called P22 HT, were used for strain construction operations involving chromosomal markers and for transfer of plasmids among Salmonella strains. The transduction protocol was as described by MALOY 1990 . To obtain phage-free isolates, transductants were purified by streaking on green plates (CHAN et al. 1972 ). Phage sensitivity was tested by cross-streaking with the clear-plaque mutant P22 H5. The alleles flhDC5213::MudA and fliA6002::MudA were originally harbored by strains KK1107 and KK1108, respectively (KUTSUKAKE and IINO 1994 ). Both strains derive from LT2 and were obtained from Kazuhiro Kutsukake (Okayama University, Okayama, Japan). MudA is a Mud1 derivative with conditional transposition (HUGHES and ROTH 1984 ). The duplication DUP4102[fliC5050*MudA*flhDC5213], originally carried on a LT2-derived strain (TH4313), was a gift from Kelly T. Hughes (University of Washington, Seattle). Strain MST1753, used for transposon replacement, carries an F-prime containing a MudP element (YOUDERIAN et al. 1988 ); the strain was obtained from Stanley R. Maloy (University of Illinois, Urbana, IL). The allele rcsB70::Tn10dCm was obtained from Eduardo Groisman (Washington University School of Medicine, St. Louis). This allele was originally carried by a 14028s derivative (strain EG12711). Allele numbers for S. enterica mutants described in this study were obtained from the Salmonella Genetic Stock Center (University of Calgary, Calgary, Alberta, Canada). The E. coli recipient of plasmids was strain DH5 (YANISCH-PERRON et al. 1985 ). The E. coli host for -dependent suicide plasmids was S17 pir (SIMON et al. 1983 ).

Media and chemicals:
The E medium of VOGEL and BONNER 1956 was used as the standard minimal medium. The rich medium was Luria broth (LB), prepared according to MALOY 1990 . Carbon sources were either 0.2% glucose or 0.2% arabinose. Solid media contained agar at 1.5% final concentration. Auxotrophic requirements and antibiotics were used at the final concentrations described by MALOY 1990 . As an exception, viability assays involving plasmid pNG1166 were carried out on plates containing a lower concentration of ampicillin (30 µg/ml), as recommended by GUZMAN et al. 1995 . D-Mecillinam, a gift from Juan A. Ayala (Centro de Biología Molecular, CSIC, Cantoblanco, Spain), was used as described below. Green plates were prepared according to CHAN et al. 1972 , except that methyl blue (Sigma, St. Louis) substituted for aniline blue. Motility assays were carried out in LB prepared without yeast extract (GILLEN and HUGHES 1991 ). Solid motility medium contained agar at 0.25% final concentration.

Surveys of d-mecillinam resistance:
Minimal inhibitory concentrations (MIC) of D-mecillinam were determined in solid media (LB and E) using late exponential cultures of the strains to be tested, prepared in Luria broth. The final concentrations of D-mecillinam ranged from 0.05 µg/ml to 1 µg/ml. Levels of mecillinam resistance were also assessed with filter paper discs soaked in a solution of mecillinam (40 µg/ml). The discs were placed on LB or E plates, previously spread with 10⁵ colony-forming units of the strain to be tested. After 18–24 hr incubation at 37°, the diameters of the inhibition halos were measured.

Plasmids:
Plasmid pBAD18 (Ap^r) is a member of the pBAD series of vectors, designed for the study of essential genes (GUZMAN et al. 1995 ). Plasmid pUC4-KIXX (Km^r Ap^r), a product of Pharmacia Biotech Europe (Sant Cugat del Vallès, Spain), carries the Tn5 kanamycin-resistance gene flanked by restriction sites that facilitate cloning. Plasmid pGEM-3Z (Ap^r) is a cloning vector from Promega (Madison, WI). Plasmid pIZ994, constructed in our laboratory, is a pGEM-3Z derivative that carries the igaA locus of S. enterica. Plasmid pIZ998 (Ap^r Km^r) is a pIZ994 derivative in which the igaA locus has been interrupted with the KIXX cassette from pUC4-KIXX, to generate the igaA2::KIXX allele. Plasmid pGP704 (Ap^r) is a pBR322 derivative that carries the origin of replication of R6K and the mob region of RP4 (MILLER and MEKALANOS 1988 ). Plasmid pIZ1551 (Ap^r) is a pGP704 derivative constructed in our laboratory and carries the igaA2::KIXX allele from pIZ998 cloned at the EcoRI site of pGP704.

Construction of plasmid pNG1166:
Genomic DNA from strain SL1344 was PCR amplified using primers from both sides of the igaA gene: 5'-TCT GTG GTA CCA CGC CTG ACA GAC-3' and 5'-CAA TAT CTA GAT GCA TGG GGA ACT GC-3', which introduce KpnI and XbaI sites (underlined), respectively. The amplified DNA fragment was digested with KpnI and XbaI to permit oriented cloning of the igaA gene on pBAD18 (GUZMAN et al. 1995 ). Ligation mixtures were used to transform E. coli DH5, selecting Ap^r transformants on LB-ampicillin plates. One of the transformants was the source of plasmid pNG1166, which carries the igaA gene under the control of the arabinose-dependent P_BAD promoter (GUZMAN et al. 1995 ).

Mutagenesis with MudJ:
We employed the cis-complementation procedure of HUGHES and ROTH 1988 , in which a defective MudJ element is cotransduced with a Mud1 element that transiently provides transposition functions. Mud1 is the specialized transducing phage Mud1(Ap Lac cts62) constructed by CASADABAN and COHEN 1979 . MudI1734[KmLac] (CASTILHO et al. 1984 ) is a transposition-deficient Mu derivative that generates operon fusions upon insertion; the element was renamed MudJ by HUGHES and ROTH 1988 .

Mutagenesis with Tn10dTc:
A lysate grown on strain TT10423 was used to transduce the strain to be mutagenized. The latter carried plasmid pNK2280 (KLECKNER et al. 1991 ) to permit complementation of the defective Tn10dTc element by ATS transposase. Transducing mixtures were made directly on LB plates containing tetracycline and were incubated 24 hr at 37°. For the preparation of insertion pools, plates containing Tc^r transductants were replica printed to LB-EGTA. Several thousands of colonies were then harvested and used to inoculate a single, mixed culture in LB-EGTA. When saturated, the culture was harvested by centrifugation and washed three to five times with LB. An aliquot was then used for standard lysis with P22 HT.

Replacement of the MudJ element by MudP:
A lysate grown on strain MST1753 was used for transposon replacement by homologous recombination. In each substitution, a MudJ element is replaced by a MudP element upon recombination between the ends of the elements. P22 HT transductions selecting the incoming marker (Cm^r) were carried out. Cm^r transductants were then scored for loss of the resident marker (Km^r).

ß-Galactosidase assays:
Levels of ß-galactosidase activity were assayed as described by MILLER 1972 , using the CHCl₃-sodium dodecyl sulfate permeabilization procedure.

Bacterial transformation:
Transformation of E. coli DH5 with plasmids followed the procedure of INOUE et al. 1990 . Transformation of S. enterica was according to LEDERBERG and COHEN 1974 . The Salmonella host for plasmids was strain TR5878. Plasmids transformed into TR5878 were transferred to suitable recipients by transduction with P22 HT.

Matings:
Transfer of pIZ1551 from E. coli S17 pir to S. enterica was assayed using exponential cultures of both the donor and the recipient. To obtain higher cell concentrations, cultures were harvested by centrifugation and concentrated 10- to 100-fold. Donor and recipient aliquots of 100 µl were mixed on an LB plate and incubated 8 hr at 37° before replica printing to selective plates.

DNA amplification with the polymerase chain reaction:
Amplification reactions were carried out in a Perkin Elmer GeneAmp PCR System 2400 (Perkin Elmer Cetus, Foster City, CA). The final volume of all reactions was 50 µl, and the final concentration of MgCl₂ was 1 mM. Reagents were used at the following concentrations: dNPTs, 200 µM; primers, 1 µM; and Taq polymerase, 1 unit per reaction. The thermal program included the following steps: (i) initial denaturation, 10 min at 94°; (ii) 35 cycles of denaturation (94°, 30 sec), annealing (55°, 30 sec), and extension (72°, 1 min); and (iii) final incubation at 72° for 10 min, to complete extension.

DNA sequencing:
Chromosomal DNA was prepared from 1.5-ml overnight cultures in LB. Cells were harvested by centrifugation and resuspended in 1.5 ml of 10 mM Tris-HCl, pH 8.0, 25 mM EDTA, pH 8.0. A volume of 0.55 ml of lysozyme solution (10 mg/ml in 10 mM Tris-HCl, pH 8.0, 1 mM EDTA, pH 8.0) was added. The mixture was incubated for 20 min at 37°. Proteinase K (100 µg/ml) was then added, and the preparation was incubated for 1 hr at 55°. After three to four extractions with phenol and chloroform-isoamyl alcohol (24:1), DNA was precipitated with ammonium acetate and absolute ethanol, and finally suspended in 10 mM Tris-HCl, pH 8.0, 1 mM EDTA, pH 8.0. DNA was further sheared with a 23G needle and cleaned in a Sephadex G50 column. The primer used for sequencing of the boundaries of Tn10 insertions was 5'-CTA ATG ACA AGA TGT GT-3' (WAY and KLECKNER 1984 ). For MudJ insertions, the primer was 5'-CGA ATA ATC CAA TGT CCT CC-3' (TORREBLANCA et al. 1999 ), henceforth called MuL. Sequencing was performed as described elsewhere (CANO et al. 2001 ).

Antibodies:
Rabbit anti-FtsZ polyclonal antibody was a gift from Miguel Vicente (Centro Nacional de Biotecnología, CSIC, Madrid).

Rabbit anti-IgaA polyclonal antibody was raised against a recombinant N-IgaA-6xHis protein, which was produced and purified as follows: a 707-bp PCR fragment (nucleotide positions 64–771 from the putative translation start of igaA; EMBL accession no. AJ272210) was PCR amplified with the primers 5'-TCC GGG CCA TGG CCA GAC GAG GG-3' and 5'-AAT TTC CTC GAG CGC GCT TTT CGA-3', which introduce NcoI and XhoI sites (underlined), respectively. The resulting fragment was digested with NcoI and XhoI and inserted in-frame upstream from the His tag sequence in the expression vector pET21d+ (Novagen, Madison, WI). The resulting plasmid, pETN1-2, was verified by sequencing the insert from both junctions. This plasmid was then used to transform E. coli BL21 (DE3; Novagen). The resulting strain was grown at 37° with shaking until early exponential phase, and expression of IgaA6xHis was induced by addition of 0.1 mM isopropyl thiogalactoside (IPTG). After 3 hr of induction, bacteria were harvested by centrifugation, suspended in column-binding buffer (5 mM imidazole, 500 mM NaCl, and 5 mM HEPES, pH 7.9), and disrupted with a French press. Bacterial debris was removed by centrifugation and protein was purified from the supernatant by cobalt-affinity chromatography according to the manufacturer's instructions. Following elution with 150 mM imidazole, purified IgaA6xHis was concentrated to the desired volume using centriplus YM-10 (Millipore, Bedford, MA). Polyclonal antibodies were obtained by standard immunization of rabbits with the N-IgaA-6xHis protein.

SDS-PAGE and Western blot analysis:
Proteins from bacterial extracts were prepared as described by KANIGA et al. 1995 and analyzed by SDS-PAGE in Tris/Tricine buffer by using 10% acrylamide gels. Blots were probed with rabbit affinity-purified anti-FtsZ (1:200) or rabbit anti-IgaA (1:5000) polyclonal antibodies. Goat anti-rabbit HRP-conjugated secondary antibody (Bio-Rad Life Science, El Prat de Llobregat, Spain) was used as secondary antibody to detect specific proteins by the enhanced chemiluminescence assay.

Shift to IgaA deprivation conditions:
An overnight culture of strain SV4578 [igaA2::KIXX/pNG1166 (igaA⁺)] grown in LB medium supplemented with 0.2% arabinose and ampicillin (100 µg/ml) was diluted 1:100 in the same medium. After 2 hr of incubation at 37°, bacteria were washed twice in PBS, suspended in LB medium adjusted to an OD₆₀₀ of 0.05, and transferred to two separate flasks, to which glucose (2%) or arabinose (0.2%) was added. To maintain bacteria in continuous exponential growing conditions, cultures were refreshed every hour. Samples were taken every hour for SDS-PAGE and Western blot analysis.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Suppression of the mucoid phenotype of an igaA1 mutant by MudJ insertions:
The igaA1 mutation causes mucoidy, which is best observed in colonies grown in rich medium. To identify genes involved in the mucoid phenotype associated with the igaA1 mutation, an igaA1 strain was mutagenized with MudJ; nonmucoid colonies were then sought among the the Km^r transductants. The strain used, SV4254, carried a Tn10 insertion (zhf-6311::Tn10dTc) cotransducible with the igaA1 mutation (CANO et al. 2001 ). Nonmucoid Km^r transductants were subjected to the following tests:

1. For reconstruction analysis the candidates were lysed with P22 HT, and the lysates were used to transduce SV4254, selecting kanamycin resistance. A 100% linkage between the Km^r marker and the nonmucoid phenotype indicated the existence of a suppressor mutation. Only six independent isolates passed this test.

2. The same lysates were used to transduce the wild-type strain, selecting tetracycline resistance. If true suppression had occurred, the igaA1 mutation would persist on the chromosome, and its phenotype would reappear in the absence of suppressors. The six candidates passed this test: in all cases, 35% of the Tc^r transductants were mucoid, indicating that the igaA1 mutation had been cotransduced with the insertion zhf-6311::Tn10dTc.

3. Cotransduction between the zhf-6311::Tn10dTc insertion and the MudJ element was not detected in any of the suppressor-carrying isolates, indicating that the suppressor mutations were not linked to the igaA locus.

To identify the loci where the MudJ element had inserted, DNA sequencing was performed directly on the chromosome, using the MuL primer (TORREBLANCA et al. 1999 ). The MudJ insertions that suppressed mucoidy in the igaA1 background were found in the gmm, manB, wcaK, and wcaF genes, which are part of the wca cluster for colanic acid synthesis (STEVENSON et al. 1996 ), and in rcsA and rcsC, which are part of the Rcs network that controls capsule synthesis (GOTTESMAN and STOUT 1991 ; GOTTESMAN 1995 ). Interestingly, all the mutations that suppressed the mucoid phenotype also reduced the intracellular proliferation rate in NRK fibroblast cell cultures (data not shown), suggesting that capsule overproduction is necessary for the igaA1 mutant to overgrow inside eukaryotic cells.

A recent study describes S. enterica mutants that display a mucoid phenotype that is suppressed by rcsC mutations (COSTA and ANTON 2001 ). The map position of the locus affected, mucM, is similar to that of igaA (CANO et al. 2001 ; COSTA and ANTON 2001 ), thus raising the possibility that the igaA1 allele and mucM alleles could all be ascribed to a single class. Because mucM mutants were isolated by resistance to D-mecillinam (COSTA and ANTON 2001 ), we carried out tests of D-mecillinam resistance in strains SL1344 (igaA⁺) and SV4450 (igaA1). Both strains exhibited MIC values around 0.8 µg/ml in LB agar and 0.6 µg/ml in E plates. The diameters of growth inhibition halos were also similar for both strains. This absence of differences suggests that the igaA1 allele is unrelated to the mucM alleles described by COSTA and ANTON 2001 . Further differences between igaA1 and mucM alleles will be described below.

The igaA1 mutation increases wca expression:
In the gmm::MudJ insertion described above as a suppressor of the mucoid phenotype of the igaA1 mutant (henceforth called gmm-21::MudJ), the MudJ element had inserted in the appropriate orientation to generate a lac fusion: The insertion was Lac^- in an IgaA⁺ background and Lac⁺ in the presence of the igaA1 mutation. Analysis of ß-galactosidase activity confirmed that the igaA1 mutation caused an increase in the expression of the gmm-21::lac fusion (Table 2). For these experiments, cultures were prepared in LB, where the mucoidy caused by the igaA mutation is higher. In the wild type, expression of gmm was low, as previously described for E. coli wca genes (GOTTESMAN 1995 ). An additional observation was that the presence of an rcsB mutation suppressed derepression of gmm in the igaA1 background (Table 2). This result suggests that derepression of colanic acid synthesis by the igaA1 mutation is exerted via the Rcs regulatory system. This view is supported by the observation, described above, that rcsA and rcsC mutations suppress the mucoid phenotype of an igaA1 mutant.

The igaA1 mutation causes a defect in motility:
The DNA sequence relatedness between the igaA ORF and the umoB gene of P. mirabilis (DUFOUR et al. 1998 ) raised the possibility that igaA1 mutants might be impaired in motility. Hence, we compared the motility of isogenic IgaA⁺ and IgaA^- strains. Cultures were prepared in motility medium and incubated at 37° with shaking. At the stage of midexponential growth, a sterile toothpick was soaked in the culture and used to inoculate a motility agar plate. The diameters of the bacterial growth halos were calculated every hour. Halo formation by the igaA1 mutant proceeded at a speed about half that of the wild type: 3.2 ± 0.2 mm/hr vs. 7.5 ± 0.9 mm/hr, respectively. The doubling time of the igaA1 mutant in liquid motility medium was identical to that of the wild type (data not shown). Therefore, delayed halo formation associated with the igaA1 mutation was interpreted as a motility defect.

The igaA gene regulates the expression of the flagellar master operon flhDC:
The motility defect of UmoB^- mutants of P. mirabilis is known to be caused by their inability to activate the flhDC operon (DUFOUR et al. 1998 ). In enteric bacteria, as in P. mirabilis, the flhDC operon encodes transcription factors necessary to trigger the genetic cascade that controls flagellar synthesis (GYGI et al. 1995 ; MACNAB 1996 ; FURNESS et al. 1997 ; CLARET and HUGHES 2000 ). Hence, a defect in flhDC expression causes reduced motility (MACNAB 1996 ). The homology between igaA and umoB led us to investigate whether the igaA1 mutation affected the expression of flagellar genes, using transcriptional lac fusions in flhDC and fliA. Evidence that flhDC undergoes autogenous positive regulation (KUTSUKAKE 1997 ) led us also to compare flhDC expression in FlhDC⁺ and FlhDC^- backgrounds. The latter experiments were performed in a merodiploid strain (SV4325), which permits analysis of the flhDC5213::lac fusion in the presence of a wild-type flhDC operon. The results shown in Fig 1 can be summarized as follows:

1. Expression of flhDC decreased sixfold in an igaA1 background. The difference was only threefold when flhDC expression was measured in a FlhDC^-/FlhDC⁺ merodiploid.

   
   

   View larger version (29K): [in this window] [in a new window] [Download PPT slide]   Figure 1. Effect of the igaA1 mutation on the expression of flhDC::lac and fliA::lac fusions. The following strains were used, from top to bottom: SV4614, SV4615, SV4324, SV4325, SV4630, and SV4631. The data shown are the means and standard deviations of at least three independent experiments.

2. As expected, the igaA1 mutation also reduced expression of an fliA::lac fusion, used as representative of class II genes, which require FlhDC activation (MACNAB 1996 ).

The correlation between reduced motility and lowered flhDC expression in an igaA1 background is in agreement with the report that other mutations that lower flhDC expression (e.g., crp, cya, and hns) also decrease motility (KUTSUKAKE 1997 ).

MudJ insertions in rcsB or rcsC, but not in rcsA, suppress the defect in flhDC expression caused by the mutation igaA1:
To investigate whether reduced expression of the flhDC operon in an igaA1 background was suppressed by suppressors of mucoidy, we constructed a set of strains that carried an flhDC::lac fusion and the igaA1 mutation in combination with wca or rcs mutations. Aside from the rcsA and the rcsC alleles described above, a collection rcsB allele (rcsB70::Tn10dCm) was also used. As shown in Fig 2, rcsB and rcsC mutations suppressed the flhDC expression defect in an igaA1 background. None of the wca insertions exerted any effect on flhDC expression. An rcsA mutation failed also to restore flhDC expression in an igaA1 mutant, suggesting that IgaA might regulate flhDC via RcsB/RcsC and that RcsA is not required. It must be noted that other examples of loci regulated by RcsB/RcsC, but not by RcsA, are found in the literature (GOTTESMAN 1995 ; CLAVEL et al. 1996 ; VIRLOGEUX et al. 1996 ; DAVALOS-GARCIA et al. 2001 ).

View larger version (57K): [in this window] [in a new window] [Download PPT slide]   Figure 2. Effects of rcs and wca mutations on flhDC operon expression in an igaA1 background. The strains used were, from top to bottom, SV4614, SV4615, SV4616, SV4623, SV4625, SV4617, SV4618, SV4620, SV4619, SV4622, and SV4624. The data shown are the means and standard deviations of at least three independent experiments.

The igaA1 mutation is recessive:
Strain 4215 carries a G A transition in the putative igaA ORF (CANO et al. 2001 ). To examine whether this mutation was dominant or recessive, the igaA1 mutation was introduced into a merodiploid strain carrying a chromosomal duplication of the argA-cysG region, which encompasses centisomes 64–75 (CAMACHO and CASADESUS 2001 ). For this purpose, strain SV4321 was transduced with a P22 HT lysate grown on SV4254. Tc^r Cm^r transductants were selected. Because the zhf-6311::Tn10dTc insertion is 35% linked to the igaA1 mutation (CANO et al. 2001 ), dominance can be expected to result in 35% of mucoid transductants. A screen of >2000 independent transductants did not reveal any mucoid colony, suggesting that the igaA1 mutation was recessive.

Segregation analysis (ANDERSON and ROTH 1977 ) confirmed the recessivity of the igaA1 mutation. Twenty-four Tc^r Cm^r transductants were grown in LB without antibiotics. Cultures were diluted and spread on LB agar. Single colonies were then replica printed to LB agar with tetracycline. Eight of the 24 transductants segregated mucoid colonies. All the mucoid colonies were Cm^s, thereby confirming that the nonmucoid merodiploids carried the igaA1 mutation. The igaA1 mutation is therefore recessive and, in principle, may cause loss of function.

The igaA gene of S. enterica is essential:
The igaA1 mutation is a base substitution that putatively changes a histidine moiety to arginine (CANO et al. 2001 ); thus, a priori, the igaA1 allele could be null or leaky. To compare the properties of igaA1 with those of a definite null allele, the igaA gene was knocked out in vitro. The wild-type igaA allele carried on plasmid pIZ998 was interrupted at its single BamHI site with the KIXX (Km^r) cassette from pUC4-KIXX. Attempts to replace the chromosomal igaA gene with this construct failed (data not shown), thereby raising the possibility that igaA was an essential gene. To test this hypothesis, a suicide plasmid containing the igaA2::KIXX mutation (pIZ1551) was constructed. Plasmid pIZ1551 is a pGP704 derivative (MILLER and MEKALANOS 1988 ) and was stably maintained in E. coli S17 pir (SIMON et al. 1983 ). Plasmid pIZ1551 was then transferred by conjugation to strain SV4321, which carries a MudP-held duplication of the cysG-argA chromosomal region (CAMACHO and CASADESUS 2001 ). Transconjugants were selected on LB plates supplemented with kanamycin (to select plasmid transfer) and chloramphenicol (to counterselect the donor). Because the plasmid cannot replicate in a strain devoid of -protein, Km^r transconjugants can be formed only by either plasmid integration in the chromosome or homologous recombination between one resident igaA gene and the incoming igaA2::KIXX allele. The latter event yields Km^r Ap^s transconjugants, while plasmid integration generates Km^r Ap^r transconjugants. An ampicillin-sensitive igaA⁺/igaA2::KIXX merodiploid (SV4322) was obtained by this procedure. To confirm the presence of both alleles, DNA from SV4322 was PCR amplified using igaA-flanking primers. Two amplification fragments of 4.2 and 5.8 kb were obtained (data not shown). The size of the igaA locus is 4.2 kb and that of the KIXX cassette is 1.6 kb: hence the merodiploid contained both an intact igaA locus and a knockout allele.

When the duplication carried by strain SV4322 was allowed to segregate in chloramphenicol-free medium, all Cm^s colonies were kanamycin sensitive. The absence of haploid segregants carrying the igaA2::KIXX allele supported the view that a null igaA allele might be lethal. This view received further support from experiments in which a P22 HT lysate grown on strain SV4322 (igaA⁺/igaA2::KIXX) was used to transduce SL1344 (igaA⁺, haploid) and SV4321 (igaA⁺/igaA⁺, merodiploid). Kanamycin-resistant transductants appeared at frequencies 1000-fold higher in SV4321 than in SL1344: around 10^-6 and 10^-9 per plaque-forming unit (PFU), respectively. Altogether, these observations suggest that the igaA locus of S. enterica is essential, at least under the conditions assayed. A corollary is that the original igaA1 mutation may be leaky. Interestingly, the essential condition of igaA in S. enterica is not shared by the related gene umoB of P. mirabilis, where insertion mutations have been described (DUFOUR et al. 1998 ).

Because igaA is the promoter-proximal gene within a putative operon that may contain four genes (yrfF, yrfG, yrfH, and yrfI), we investigated whether the lethality of the igaA2::KIXX mutation was caused by a polar effect on downstream genes. For this purpose, a KIXX cassette was introduced in the second ORF of the putative operon, yrfG (data not shown). The yrfG::KIXX mutant was viable, as indicated by the observation that strains SL1344 (igaA⁺, haploid) and SV4321 (igaA⁺/igaA⁺, merodiploid) could be transduced at similar frequencies (10^-6/PFU) with a P22 lysate carrying the yrfG::KIXX construct. We thus concluded that the lethality of the igaA2::KIXX mutation was solely due to igaA inactivation. Furthermore, the yrfG::KIXX mutant was nonmucoid and unable to derepress the flhDC5213::lac fusion (data not shown), indicating that mucoidy and flhDC derepression were also igaA-associated traits.

Finally, we tested whether the lethality caused by a null igaA mutation could be complemented by a wild-type igaA allele carried on a plasmid. For this purpose, plasmid pNG1166 was transduced to SL1344. The resulting strain was then transduced with a P22 HT lysate grown on strain SV4322 (igaA⁺/igaA2::KIXX), selecting kanamycin resistance on LB plates containing 0.2% arabinose. Km^r transductants appeared at frequencies 10^-6/PFU, suggesting that the igaA2::KIXX mutation did not impair viability in the presence of the igaA⁺ allele carried on pNG1166. To confirm that the igaA2::KIXX construct had recombined with the recipient chromosome (and not with plasmid pNG1166), a putatively complemented isolate was lysed with P22 HT. The lysate was used to transduce SL1344, selecting Ap^r. Transductants were obtained at a frequency of 10^-5/PFU, and all (100/100) were Km^s. If the same lysate was used to select kanamycin resistance, transductants were extremely rare (10^-9/PFU) and were Ap^s. Transduction of kanamycin resistance to the merodiploid strain SV4321 was, however, successful. These experiments confirmed that the donor carried both an intact plasmid pNG1166 and the chromosomal igaA2::KIXX construct. This strain was propagated as SV4578.

Occurrence of complementation was confirmed in plate tests, shown in Fig 3. Strain SV4578 (igaA2::KIXX/ pNG1166) was able to grow on arabinose-containing plates, but not on glucose. In contrast, IgaA⁺ and IgaA^- RcsC^- strains (SV4577 and SV4629) grew well on both media. These experiments confirm that lack of igaA expression alone prevents growth of S. enterica in standard laboratory conditions.

View larger version (87K): [in this window] [in a new window] [Download PPT slide]   Figure 3. Growth of IgaA+, IgaA-, and IgaA- RcsC- strains on LB-glucose-ampicillin (left) and LB-arabinose-ampicillin (right). The strains used were SV4577 (igaA+/pNG1166), SV4629 (igaA2::KIXX rcsC52::MudP/pNG1166), and SV4578 (igaA2::KIXX/pNG1166). In the latter, expression of the plasmid-borne igaA gene is conditional, driven by the arabinose-dependent PBAD promoter. The assays were carried out as described by GUZMAN et al. 1995 , and the concentration of ampicillin was 30 mg/liter.

MudJ insertions in rcsB or rcsC, but not in rcsA, suppress the lethality caused by a null igaA allele:
We have described above that rcsA and rcsC mutations suppress mucoidy and reduced motility caused by the viable allele igaA1. To investigate whether mutations affecting the Rcs regulatory system also suppressed the lethality caused by the null allele igaA2::KIXX, transductional crosses were performed. The recipients were SV4379 (RcsA^-), SV4380 (RcsB^-), and SV4406 (RcsC^-). The donor was an igaA⁺/igaA2::KIXX merodiploid (SV4322). Kanamycin-resistant transductants were selected upon P22 HT transduction. Their frequencies (per PFU) were 10^-6 in the RcsB^- and RcsC^- recipients and <10^-9 in the RcsA^- recipient. Hence, lack of RcsB or RcsC functions does suppress the lethality of an igaA null mutation, but lack of RcsA does not. This suppressor pattern is identical to that found for flhDC expression in an igaA1 background (see above).

Search for Tn10 insertions that suppress the lethality caused by a null igaA allele:
Mutagenesis with Tn10dTc was carried out in strain SL1344; five pools of Tn10dTc insertions, each containing some 10,000 independent isolates, were then prepared. The pools were grown in LB + EGTA until midexponential phase and used as recipients in P22 HT-mediated transductional crosses. The donor was the merodiploid strain SV4322 (igaA⁺/igaA2::KIXX). Km^r transductants were selected on LB-kanamycin plates. Each cross yielded 5–30 Km^r transductants. Segregation analysis after nonselective growth (in LB without kanamycin) allowed us to discard igaA⁺/igaA2::KIXX merodiploids that segregated Km^s colonies. Stable Km^r isolates carried putative Tn10dTc insertions that suppressed the lethality associated with the igaA2::KIXX mutation. These isolates were then subjected to reconstruction analysis:

1. Fifty independent candidates were lysed with P22 HT, and the resulting lysates were used to transduce SL1344, selecting tetracycline resistance.

2. Tc^r transductants obtained in these crosses were transduced with an SV4322 lysate, selecting kanamycin resistance. As a control, the merodiploid strain SV4321 (igaA⁺/igaA⁺) was also transduced with the same lysate. Whenever Km^r transductants were obtained at similar frequencies in SV4231 and in the putative suppressor-carrying isolates, evidence existed that the igaA2::KIXX mutation was viable in the genetic background of the latter. In such cases, the frequency of transductants was 10^-6/PFU. Otherwise, transduction occurred at frequencies 10^-9/PFU; these rare transductants were presumed to carry suppressor mutations of spontaneous origin and were discarded.

These experiments provided us with six independent suppressor-carrying isolates in which suppression of lethality appeared to be associated with a Tn10dTc insertion. One boundary of each Tn10dTc insertion was then sequenced using a Tn10-derived primer (WAY and KLECKNER 1984 ). Two of the insertions mapped in rscC, thereby confirming that lack of RcsC suppresses the lethality of an igaA null mutation. The four remaining Tn10dTc insertions mapped in yojN, a locus located between ompC and rcsB on the Salmonella chromosome (MCCLELLAND et al. 2001 ). Mutations in yojN also suppressed the mucoidy associated with igaA mutations and the low expression of flhDC (data not shown). Suppression by yojN mutations strengthens the evidence that IgaA interacts with the RcsB-RcsC regulatory system, since YojN has been shown to participate in the phosphorelay signaling pathway RcsC RcsB (TAKEDA et al. 2001 ).

The lethality caused by a null igaA allele does not involve changes in the level of the cell division protein FtsZ:
Because IgaA appears to interact with the RcsB-RcsC regulatory system and the latter has been shown to regulate positively the cell division genes ftsA and ftsZ (CARBALLES et al. 1999 ), a conceivable explanation for the lethality associated with null igaA alleles might be cell division arrest, as a consequence of overproduction of these cell division proteins. In E. coli, the key protein that regulates the frequency and the timing of cell division is FtsZ (WARD and LUTKENHAUS 1985 ). On these grounds, we investigated whether a decrease in IgaA synthesis altered the level of FtsZ. For this purpose, we carried out shift-down experiments in which strain SV4578 (igaA2::KIXX/pNG1166) was transferred from arabinose to glucose medium. The amounts of IgaA and FtsZ in cell extracts obtained at different times after arabinose depletion were analyzed by Western blotting using anti-IgaA and anti-FtsZ polyclonal antibodies. The results were clear-cut: While a swift decrease in IgaA was observed, the level of FtsZ remained unchanged (Fig 4). These data suggest that the viability defect of igaA mutants is not caused by FtsZ overproduction and reveal a further difference between igaA and mucM alleles, since the latter increase transcription from the ftsA1 promoter at the ftsQAZ operon (COSTA and ANTON 2001 ). Further evidence that the proposed functional interaction between IgaA and the RcsB-RcsC system does not involve changes in the level of FtsZ was provided by the following observations: (i) The levels of FtsZ remained largely unchanged in strains carrying igaA mutations, alone or combined with rcsA, rcsB, and rcsC alleles (Fig 5) and (ii) strains carrying the igaA1 mutation do not form minicells (not shown).

View larger version (35K): [in this window] [in a new window] [Download PPT slide]   Figure 4. Levels of IgaA and FtsZ in protein extracts from strain SV4578 (igaA2::KIXX/pNG1166) upon transfer to glucose medium (left) or to fresh arabinose medium (right). Samples were taken every hour for SDS-PAGE and Western blot analysis. The time of sampling (in hours) is indicated on the top of each sample; zero designates the time of transfer to fresh media.

View larger version (36K): [in this window] [in a new window] [Download PPT slide]   Figure 5. Levels of IgaA and FtsZ in protein extracts from SL1344 (wildtype, lane 1), SV4215 (igaA1, lane 2), SV4379 (rcsA, lane 3), SV4406 (rcsB, lane 4), SV4380 (rcsC, lane 5), SV4343 (igaA1 rcsA, lane 6), SV4402 (igaA1 rcsB, lane 7), SV4404 (igaA1 rcsC, lane 8), SV4345 (igaA2::KIXX rcsC, lane 9), and SV4443 (igaA2::KIXX rcsB, lane 10). The extracts were prepared from bacteria growing exponentially in LB.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mutants of S. enterica carrying the igaA1 allele show a mucoid phenotype caused by derepression of wca genes. The induction of wca gene expression in an igaA1 mutant is suppressed by mutations in rcsA and rcsC. The igaA1 mutation also causes lowered expression of the flagellar master operon, flhDC, which results in lowered motility. The latter defect is suppressed by mutations in rcsB or rcsC. Altogether, these observations support a model in which loss of IgaA causes activation of the RcsB/C system, leading both to derepression of wca genes and to repression of the flhDC operon. However, a noteworthy observation is that, in an igaA1 background, activation of colanic acid synthesis requires RcsA while repression of flagellar formation is RcsA independent.

The opposite effects of the igaA gene product on colanic acid synthesis and flagellar formation are reminiscent of the inverse relationship between capsule synthesis and flagellar biogenesis described in several bacterial species: (i) Biofilm-forming E. coli repress flagellar genes and induce capsular gene expression (PRIGENT-COMBARET et al. 1999 ); (ii) production of alginate and synthesis of flagella by Pseudomonas aeruginosa show inverse and mutually exclusive regulation (GARRETT et al. 1999 ); and (iii) absence of flagella is a signal to induce capsular polysaccharide synthesis in Vibrio cholerae (WATNICK et al. 2001 ). One may thus hypothesize that formation of flagella and capsule synthesis could be also mutually exclusive in Salmonella and that the igaA gene product might coordinate them, perhaps in response to environmental signals. This coordination would be exerted via the RcsB/C regulatory system.

The idea that IgaA might be a sensor protein is speculative at this stage. However, analysis in silico of the predicted IgaA amino acid sequence unveils the presence of putative phosphorelay motifs similar to those found in two-component regulatory systems (data not shown). Several potential transmembrane domains (HOFMANN and STOFFEL 1993 ) are also detected. Immunodetection upon subcellular fractionation has confirmed that the IgaA protein is membrane bound (data not shown). With these features in mind, it is tempting to speculate that IgaA might be part of a novel pathway of signal transduction through the RcsB-RcsC system. A membrane protein, DjlA, that interacts with RcsB/C and participates in wca activation has been described (CLARKE et al. 1997 ; KELLEY and GEORGOPOULOS 1997 ), and additional transmitters appear to exist (CHEN et al. 2001 ). IgaA may thus be a tentative addition to the list of sensors converging in the Rcs system. An alternative possibility is that the functional interaction between IgaA and the RcsB/C system might be indirect. In E. coli, mutants affected in the mdoH gene (EBEL et al. 1997 ) show several defects similar to those of igaA1 mutants of S. enterica (e.g., mucoidy suppressed by mutations in rcsA, rcsB, or rcsC, and impaired motility). The mdoH gene product is involved in the synthesis of membrane-derived oligosaccharides (MDOs; KENNEDY 1996 ). A model proposes that periplasmic levels of MDOs act to signal RcsC to activate capsule synthesis (EBEL et al. 1997 ). Mutations in genes for lipopolysaccharide synthesis also activate RcsB/C (PARKER et al. 1992 ). In an analogous fashion, the igaA gene product might affect the RcsB-RcsC system through specific components of the cell envelope.

Null igaA mutations are lethal and can be maintained only in a merodiploid carrying a wild-type igaA allele, either chromosomal or plasmid borne. However, null IgaA^- mutants are viable in the presence of rcsB, rcsC, or yojN mutations, indicating again the existence of a functional relationship between IgaA and the RcsB-RcsC system. A tentative explanation is that, in the absence of IgaA, activation of the RcsB-RcsC system might lead to derepression of a gene or operon whose overexpression is lethal. An obvious candidate was the ftsQZA operon, which is known to be positively regulated by the RcsB/C system (CARBALLES et al. 1999 ; COSTA and ANTON 2001 ). However, we provide evidence that the viability defect of null igaA mutants does not involve FtsZ-mediated cell division arrest.

The viability of the igaA1 allele admits a simple (albeit at this stage unproved) explanation. The mutation causes a single amino acid substitution and is recessive. Hence, the igaA1 allele may be leaky, and residual function may permit cell survival. However, the view that only point igaA mutations may be tolerated is countered by observations made in E. coli, where a partial deletion in the igaA homolog, yrfF, appears to be viable and causes mucoidy (MEBERG et al. 2001 ).

	FOOTNOTES

¹ Present address: Diabetes Center, Department of Medicine, University of California, San Francisco, 513 Parnassus Ave., San Francisco, CA 94143-0540.

	ACKNOWLEDGMENTS

We are grateful to John Roth, Kazuhiro Kutsukake, Kelly Hughes, Stan Maloy, and Eduardo Groisman for providing strains; to Nuria Gómez-López for DNA sequencing reactions and construction of plasmid pNG1166; and to Francisco Ramos for critical reading of the manuscript. This work was supported by grants from the Spanish Ministry of Science (BIO2001-0232-CO2) and the European Union (QLK2-1999-00310). A.T. is recipient of a Ph.D. fellowship from the Comunidad de Madrid.

Manuscript received April 29, 2002; Accepted for publication September 3, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALTSCHUL, S. F., T. L. MADDEN, A. A. SCHAFFER, J. ZHANG, and Z. ZHANG et al., 1997  Gapped BLAST and PSI-BLAST: a new generation of database search programs. Nucleic Acids Res. 25:3389-3402.[Abstract/Free Full Text]

ANDERSON, R. P. and J. R. ROTH, 1977  Tandem genetic duplications in phage and bacteria. Annu. Rev. Microbiol. 31:473-505.[Medline]

ARRICAU, N., D. HERMANT, H. WAXIN, C. ECOBICHON, and P. S. DUFFEY et al., 1998  The RcsB-RcsC regulatory system of Salmonella typhi differentially modulates the expression of invasion proteins, flagellin and Vi antigen in response to osmolarity. Mol. Microbiol. 29:835-850.[Medline]

BERESWILL, S. and K. GEIDER, 1997  Characterization of the rcsB gene from Erwinia amylovora and its influence on exopolysaccharide synthesis and virulence of the fire blight pathogen. J. Bacteriol. 179:1354-1361.[Abstract/Free Full Text]

BLATTNER, F., G. PLUNKERT, III, C. A. BLOCH, N. T. PERNA, and V. BURLAND et al., 1997  The complete genome sequence of Escherichia coli K-12. Science 277:1432-1434.[Free Full Text]

BRILL, J. A., C. QUINLAN-WALSHE, and S. GOTTESMAN, 1988  Fine-structure mapping and identification of two regulators of capsule synthesis in Escherichia coli K-12. J. Bacteriol. 170:2599-2611.[Abstract/Free Full Text]

CAMACHO, E. M. and J. CASADESÚS, 2001  Genetic mapping by duplication segregation in Salmonella enterica.. Genetics 157:491-502.[Abstract/Free Full Text]

CANO, D. A., M. MARTINEZ-MOYA, M. G. PUCCIARELLI, E. A. GROISMAN, and J. CASADESÚS et al., 2001  Salmonella enterica serovar Enterica response involved in attenuation of pathogen intracellular proliferation. Infect. Immun. 69:6463-6474.[Abstract/Free Full Text]

CARBALLÈS, F., C. BERTRAND, J. P. BOUCHÉ, and K. CAM, 1999  Regulation of Escherichia coli cell division genes ftsA and ftsZ by the two-component system rcsC-rcsB.. Mol. Microbiol. 34:442-450.[Medline]

CASADABAN, M. J. and S. N. COHEN, 1979  Lactose genes fused to exogenous promoters in one step using Mu-lac bacteriophage: in vivo probe for transcriptional control sequences. Proc. Natl. Acad. Sci. USA 76:4530-4533.[Abstract/Free Full Text]

CASTILHO, B. A., P. OLFSON, and M. J. CASADABAN, 1984  Plasmid insertion mutagenesis and lac gene fusion with Mini-Mu bacteriophage transposons. J. Bacteriol. 158:488-495.[Abstract/Free Full Text]

CHAN, R. K., D. BOTSTEIN, T. WATANABE, and Y. OGATA, 1972  Specialized transduction of tetracycline by phage P22 in Salmonella typhimurium. II. Properties of a high frequency transducing lysate. Virology 50:883-898.[Medline]

CHEN, M. H., S. TAKEDA, H. YAMADA, Y. ISHII, and T. YAMASHINO et al., 2001  Characterization of the RcsC YojN RcsC phosphorelay signalling pathway involved in capsular synthesis in Escherichia coli.. Biosci. Biotechnol. Biochem. 65:2364-2367.[Medline]

CLARET, L. and C. HUGHES, 2000  Functions of the subunits in the FlhD(2)C(2) transcriptional master regulator of bacterial flagellum biogenesis and swarming. J. Mol. Biol. 303:467-478.[Medline]

CLARKE, D. J., I. B. HOLLAND, and A. JACQ, 1997  Point mutations in the transmembrane domain of DjlA, a membrane-linked DnaJ-like protein, abolish its function in promoting colanic acid production via the Rcs signal transduction pathway. Mol. Microbiol. 25:933-944.[Medline]

CLAVEL, T., J. C. LAZZARONI, A. VIANNEY, and R. PORTALIER, 1996  Expression of the tolQRA genes of Escherichia coli K-12 is controlled by the RcsC sensor protein involved in capsule synthesis. Mol. Microbiol. 19:19-25.[Medline]

CONTER, A., R. STURNY, C. GUTIERREZ, and K. CAM, 2002  The RcsCB His-Asp phosphorelay system is essential to overcome chlorpromazine-induced stress in Escherichia coli.. J. Bacteriol. 184:2850-2853.[Abstract/Free Full Text]

COSTA, C. S. and D. N. ANTÓN, 2001  Role of the ftsA1p promoter in the resistance of mucoid mutants of Salmonella enterica to mecillinam: characterization of a new type of mucoid mutant. FEMS Microbiol. Lett. 200:201-205.[Medline]

DAVALOS-GARCIA, M., A. CONTER, I. TOESCA, C. GUTIERREZ, and K. CAM, 2001  Regulation of osmC gene expression by the two-component system rcsB-rcsC in Escherichia coli.. J. Bacteriol. 183:5870-5876.[Abstract/Free Full Text]

DUFOUR, A., R. B. FURNESS, and C. HUGHES, 1998  Novel genes that upregulate the Proteus mirabilis flhDC master operon controlling flagellar biogenesis and swarming. Mol. Microbiol. 29:741-751.[Medline]

EBEL, W., G. J. VAUGHN, H. K. PETERS, III, and J. E. TREMPY, 1997  Inactivation of mdoH leads to increased expression of colanic acid capsular polysaccharide in Escherichia coli.. J. Bacteriol. 179:6858-6861.[Abstract/Free Full Text]

FURNESS, R. B., G. M. FRASER, N. A. HAY, and C. HUGHES, 1997  Negative feedback from a Proteus class II flagellum export defect to the flhDC master operon controlling cell division and flagellum assembly. J. Bacteriol. 179:5585-5588.[Abstract/Free Full Text]

GARRETT, E. S., D. PERLEGAS, and D. J. WOZNIAK, 1999  Negative control of flagellum synthesis in Pseudomonas aeruginosa is modulated by the alternative sigma factor AlgT (AlgU). J. Bacteriol. 181:7401-7404.[Abstract/Free Full Text]

GILLEN, K. L. and K. T. HUGHES, 1991  Negative regulatory loci coupling flagellar synthesis to flagellar assembly in Salmonella typhimurium.. J. Bacteriol. 173:2301-2310.[Abstract/Free Full Text]

GOTTESMAN, S., 1995 Regulation of capsule synthesis: modification of the two-component paradigm by an accessory unstable regulator, pp 253–262 in Two-Component Signal Transduction, edited by J. A. HOCH and T. A. SILHAVY. American Society for Microbiology, Washington, DC.

GOTTESMAN, S. and V. STOUT, 1991  Regulation of capsular polysaccharide synthesis in Escherichia coli K12. Mol. Microbiol. 5:1599-1606.[Medline]

GOTTESMAN, S., P. TRISLER, and A. TORRES-CABASSA, 1985  Regulation of capsular polysaccharide synthesis in Escherichia coli K-12: identification of three regulatory genes. J. Bacteriol. 162:1111-1119.[Abstract/Free Full Text]

GUZMAN, L. M., D. BELIN, M. J. CARSON, and J. BECKWITH, 1995  Tight regulation, modulation, and high level expression by vectors containing the arabinose P_BAD promoter. J. Bacteriol. 177:4121-4130.[Abstract/Free Full Text]

GYGI, D., M. J. BAILEY, C. ALLISON, and C. HUGHES, 1995  Requirement for FlhA in flagella assembly and swarm-cell differentiation in Proteus mirabilis.. Mol. Microbiol. 15:761-769.[Medline]

HOFMANN, K. and W. STOFFEL, 1993  TMbase—a database of membrane spanning protein fragments. Biol. Chem. Hoppe-Seyler 347:166.

HOISETH, S. K. and B. A. STOCKER, 1981  Aromatic-dependent Salmonella enterica are non-virulent and effective as live vaccines. Nature 291:238-239.[Medline]

HUGHES, K. T. and J. R. ROTH, 1984  Conditionally transposition-defective derivative of Mu d1(Amp Lac). J. Bacteriol. 159:130-137.[Abstract/Free Full Text]

HUGHES, K. T. and J. R. ROTH, 1988  Transitory cis complementation: a method to provide transposition functions to defective transposons. Genetics 109:263-282.[Abstract/Free Full Text]

INOUE, H., H. NOJIMA, and H. OKAYAMA, 1990  High efficiency transformation of Escherichia coli with plasmids. Gene 96:23-28.[Medline]

KANIGA, K., S. TUCKER, D. TROLLINGER, and J. E. GALAN, 1995  Homologs of the Shigella IpaB and IpaC invasins are required for Salmonella typhimurium entry into cultured epithelial cells. J. Bacteriol. 177:3965-3971.[Abstract/Free Full Text]

KELLEY, W. L. and C. GEORGOPOULOS, 1997  Positive control of the two component RcsB/C signal transduction network by DjlA: a member of the DnaJ family of molecular chaperones in Escherichia coli.. Mol. Microbiol. 25:913-931.[Medline]

KENNEDY, E. P., 1996 Membrane-derived oligosaccharides (periplasmic ß-D-glucans) of Escherichia coli, pp. 1064–1071 in Escherichia coli and Salmonella: Cellular and Molecular Biology, edited by F. C. NEIDHARDT, R. CURTISS, III, J. L. INGRAHAM, E. C. C. LIN, K. B. LOW, B. MAGASANIK, W. S. REZNIKOFF, M. RILEY, M. SCHAECHTER and H. E. UMBARGER. American Society for Microbiology, Washington, DC.

KLECKNER, N., J. BENDER, and S. GOTTESMAN, 1991  Uses of transposons with emphasis on Tn10.. Methods Enzymol. 204:139-180.[Medline]

KUTSUKAKE, K., 1997  Autogenous and global control of the flagellar master operon, flhD, in Salmonella typhimurium.. Mol. Gen. Genet. 254:440-448.[Medline]

KUTSUKAKE, K. and T. IINO, 1994  Role of the FliA-FlgM regulatory system on the transcriptional control of the flagellar regulon and flagellar formation in Salmonella typhimurium.. J. Bacteriol. 176:3598-3605.[Abstract/Free Full Text]

LEDERBERG, E. M. and S. N. COHEN, 1974  Transformation of Salmonella typhimurium by plasmid deoxyribonucleic acid. J. Bacteriol. 119:1072-1074.[Abstract/Free Full Text]

MACNAB, R. M., 1996 Flagella and motility, pp. 123–145 in Escherichia coli and Salmonella: Cellular and Molecular Biology, edited by F. C. NEIDHARDT, R. CURTISS, III, J. L. INGRAHAM, E. C. C. LIN, K. B. LOW, B. MAGASANIK, W. S. REZNIKOFF, M. RILEY, M. SCHAECHTER and H. E. UMBARGER. American Society for Microbiology, Washington, DC.

MALOY, S. R., 1990 Experimental Techniques in Bacterial Genetics. Jones & Bartlett, Boston.

MCCLELLAND, M., K. E. SANDERSON, J. SPIETH, S. W. CLIFTON, and P. LATREILLE et al., 2001  Complete genome sequence of Salmonella enterica serovar Enterica. Nature 413:852-856.[Medline]

MEBERG, B. M., F. C. SAILER, D. E. NELSON, and K. D. YOUNG, 2001  Reconstruction of Escherichia coli mcrA (PBP 1a) mutants lacking multiple combinations of penicillin binding proteins. J. Bacteriol. 183:6148-6149.[Abstract/Free Full Text]

MILLER, J. H., 1972 Experiments in Molecular Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

MILLER, V. I. and J. J. MEKALANOS, 1988  A novel suicide vector and its use in construction of insertion mutations: osmoregulation of outer membrane proteins and virulence determinants in Vibrio cholerae requires toxR.. J. Bacteriol. 170:2575-2583.[Abstract/Free Full Text]

PARKER, C. T., A. W. KLOSER, C. A. SCHNAITMAN, M. A. STEIN, and S. GOTTESMAN et al., 1992  Role of the rfaG and rfaP genes in determining the lipopolysaccharide core structure and cell surface properties of Escherichia coli K-12. J. Bacteriol. 144:2525-2538.

PRIGENT-COMBARET, C., O. VIDAL, C. DOREL, and P. LEJEUNE, 1999  Abiotic surface sensing and biofilm-dependent regulation of gene expression in Escherichia coli.. J. Bacteriol. 181:5993-6002.[Abstract/Free Full Text]

SCHMIEGER, H., 1972  Phage P22 mutants with increased or decreased transducing abilities. Mol. Gen. Genet. 119:75-88.[Medline]

SIMON, R., U. PRIEFER, and A. PÜHLER, 1983  A broad host range mobilization system for in vivo genetic engineering: transposon mutagenesis in gram-negative bacteria. Methods Enzymol. 118:640-659.

STEVENSON, G., K. ADRIANOPOULOS, M. HOBBS, and P. R. REEVES, 1996  Organization of the Escherichia coli K-12 gene cluster responsible for production of the extracellular polysaccharide colanic acid. J. B