Multiple Fimbrial Adhesins Are Required for Full Virulence of Salmonella typhimurium in Mice

Bacterial strains, bacteriophages, and recombinant DNA techniques.

Bacteria were grown overnight in Luria-Bertani broth at 37°C. Antibiotics, when required, were incorporated into the medium at the following concentrations: naladixic acid, 50 mg/liter; kanamycin, 60 mg/liter; chloramphenicol, 30 mg/liter; and carbenicillin, 100 mg/liter. Analytical-grade chemicals were purchased from Sigma (St. Louis, Mo.) or Boehringer Mannheim (Indianapolis, Ind.). AJB3 is a fully mouse virulent naladixic acid-resistant derivative of S. typhimurium SR-11 (3). SR-11 derivatives carrying a pefC::Tet^r allele (AJB9) or a deletion of the fim operon (AJB4) have been described previously (3, 4). Bacteriophage KB1int or P22HTint was used to transduce a pefC::Tet^r or lpfC::Kan^r mutation from S. typhimurium AJB7 (3) or AJB1 (5), respectively, into the desired SR-11 background. Recombinant DNA techniques and Southern hybridizations were performed by using standard protocols (1).

A 927-bp fragment internal to agfB was amplified from χ4252 (wild-type SR-11 [16]) with primers 5′-CTGACAGATGTTGCACTGCTGTG-3′ and 5′-TTCGCCCGATTATTTCCTCC-3′. This PCR product was cloned into the EcoRV site of pBluescript SK to yield plasmid pAV326. The agfB allele was inactivated upon insertion of a chloramphenicol resistance gene (a 1.2-kb SmaI fragment from pCMXX [6]) into a unique NruI site (nucleotide 466). This plasmid was digested with SacI and KpnI, and a 2.2-kb fragment was cloned into suicide vector pGP704 (19). The resulting plasmid (pAV328) was transformed into Escherichia coli S17λpir (15) and conjugated into S. typhimurium AJB3 (wild type) and AJB12 (Δfim lpfC pefC). A double cross-over was obtained by homologous recombination. Chloramphenicol-resistant, carbenicillin-sensitive (loss of vector pGP704) exconjugants were screened for and named AWM394 (agfB) and AWM401 (Δfim lpfC pefC agfB).

DNA probes specific for fim, lpf, pef, and agf were used as probes for Southern hybridization. In brief, a SphI fragment of pISF101 (7) and a SacI-KpnI fragment of pMS1054 (2) served as probes to detect fim- and lpf-specific loci, respectively. A 520-bp fragment internal to pefA was amplified by PCR with primers 5′-GGGAATTCTTGCTTCCATTATTGCACTGGG-3′ and 5′-TCTGTCGACGGGGGATTATTTGTAAGCCACT-3′ and cloned into the EcoRV site of pBluescript (21) to give rise to plasmid pAV323. The EcoRI- and ClaI-restricted insert of pAV323 was labeled and used as a pef-specific probe. A SacI-KpnI fragment of pAV326 was used to generate an agf-specific probe. Restriction enzyme-digested chromosomal DNA was separated on an agarose gel and transferred onto a positively charged membrane (Boehringer Mannheim). The predicted sizes of hybridizing fragments were as follows. A fim-specific probe detected a 13.7-kb fragment in SphI-restricted chromosomal DNA of fim⁺ strains (AJB3, AJB5, AJB9, AJB11, AWM394, and AWM400) and 10.5- and 3.1-kb fragments in SphI-restricted chromosomal DNA of fim mutants (AJB4, AJB6, AJB12, and AWM401). An lpf-specific probe detected a 3.7-kb fragment in PstI-restricted chromosomal DNA of lpfC⁺ strains (AJB3, AJB4, AJB9, and AWM394) and 2.8- and 1.7-kb fragments in PstI-restricted chromosomal DNA of lpfC mutants (AJB5, AJB6, AJB11, AJB12, AWM400, and AWM401). A pef-specific probe detected a 3.6-kb fragment in EcoRI- and HindIII-restricted chromosomal DNA of pefC⁺ strains (AJB3, AJB4, AJB5, AJB6, and AWM394) and a 2.8-kb fragment in EcoRI- and HindIII-restricted chromosomal DNA of pefC mutants (AJB9, AJB11, AJB12, AWM400, and AWM401). An agf-specific probe detected a 1.8-kb fragment in EcoRI- and SalI-restricted chromosomal DNA of agfB⁺ strains (AJB3, AJB4, AJB5, AJB6, AJB9, AJB11, and AJB12) and a 3.0-kb fragment in EcoRI- and SalI-restricted chromosomal DNA of agfB mutants (AWM394, AWM400, and AWM401). Detection was performed by using the Renaissance random primer fluorescein dUTP labeling and detection system from DuPont NEN (Boston, Mass.).

Mouse experiments.

Six- to eight-week-old female BALB/c mice (Jackson Laboratories, Bar Harbor, Maine) were used throughout this study. To determine the (two-step) LD₅₀, a series of 10-fold dilutions of overnight cultures in a 0.2-ml volume were injected intragastrically into groups of four mice. The LD₅₀s were calculated 28 days postinfection by the method of Reed and Muench (20). For course of infection studies, approximately 10⁸ bacteria were administered to groups of four mice by intragastric injection. Five days postinfection, the animals were sacrificed, after which internal organs (Peyer’s patches, villous intestinal tissues, mesenteric lymph nodes, spleens, and livers) were collected and homogenized in 5 ml of phosphate-buffered saline (PBS) by using a stomacher (Tekmar, Cincinnati, Ohio). To test the ability of a particular strain to colonize the intestinal lumen, fecal pellets were collected at days 1, 3, and 5 postinfection and homogenized in 5 ml of PBS. A 10-fold dilution series was plated on Luria-Bertani agar plates containing the appropriate antibiotics to determine the number of CFU. Results are reported in CFU per organ or per gram of feces (single strain infections) or as percentages of the total number of bacteria recovered (mixed infections). A paired t test was used to calculate statistical differences between arithmetic means.

Electron microscopy.

Bacterial strains were grown as 3-ml static broth cultures to promote expression of fimbrial structures. Subsequently, 15 μl of bacterial suspension was pipetted onto a Formvar-coated grid (Ted Pella Inc., Redding, Calif.). Bacteria were allowed to adhere for 2 min and then were fixed for 1 min with 1.5% glutaraldehyde in sodium cacodylate buffer (100 mM, pH 7.4). The grids were rinsed twice with water and negatively stained with 0.75% (wt/vol) uranyl acetate (pH 6.4) for 1 min. The grids were drained and subjected to microscopic studies.

Construction of S. typhimurium fimbrial mutants.

Mutations in three fimbrial operons, fim, lpf, and pef, have been reported previously (3, 5, 16) and were used to construct a set of isogenic S. typhimurium mutants that carried deletions of and/or insertions in essential fimbrial biosynthesis genes (Table 1). The lpfC::Kan^r allele of S. typhimurium ATCC 14028 derivative AJB1 (5) was transduced into SR-11 derivatives AJB3 (wild type) and AJB4 (Δfim) (3), yielding strains AJB5 (lpfC) and AJB6 (Δfim lpfC), respectively. The pefC::Tet^r allele of strain AJB7 (3) was transduced into AJB3 and AJB6 to give rise to strains AJB9 (pefC) and AJB12 (Δfim lpfC pefC), respectively (Table 1). The lpfC::Kan^r allele of AJB1 was then transduced into AJB9 to give rise to AJB11 (lpfC pefC) (Table 1). All mutants were confirmed by Southern blot analysis with the appropriate DNA probes (Fig. 1).

Thin aggregative fimbriae, which are encoded by the S. typhimurium agf operon, are assembled by an export machinery that is distinct from the chaperone- and usher-dependent transport systems of type 1 fimbriae, PE fimbriae, or LP fimbriae. Curli, encoded by the csg operon in E. coli, is the prototypic member of this novel pilus assembly class. Recent evidence by Hammar et al. suggests that CsgB, a membrane-associated nucleator protein, is required for the assembly of curli fimbriae on the bacterial cell surface (13). It was therefore decided to inactivate agfB, the csgB homolog in S. typhimurium (9). An agfB allele (carried on plasmid pAV328) was inactivated by insertion of a 1.2-kb chloramphenicol resistance cassette and introduced into strains AJB3 (wild type), AJB11 (lpC pefC), and AJB12 (Δfim lpfC pefC). Double cross-over events were obtained by homologous recombination, and the resulting strains were designated AWM394 (agfB), AWM400 (lpfC pefC agfB), and AWM401 (Δfim lpfC pefC agfB), respectively (Table 1). All three mutants were confirmed by Southern blot analysis with an agfB-specific DNA probe (Fig. 1).

Synergistic effect of mutations in fimbrial operons on mouse virulence.

LD₅₀ studies were conducted to investigate the effect of mutations in fimbrial operons on mouse virulence (Table 2) (20). Strains carrying mutations in a single fimbrial operon were either more virulent (AJB4, Δfim) or less than fivefold attenuated (AJB5, lpfC; AJB9, pefC; and AWM394, agfB) in comparison with the wild type (AJB3). Strain AJB12 (Δfim lpfC pefC) also exhibited slightly increased virulence, suggesting that the phenotype of a fim deletion mutant is dominant over the attenuating effect of mutations in lpf and pef. Interestingly, AWM400 (lpfC pefC agfB) is more strongly attenuated (>29-fold) than AWM401 (Δfim lpfC pefC agfB) (see below and Table 2). We and others have observed that all fim mutants tested had a slight increase in virulence compared to that of the wild type (16) (Table 2 and our unpublished results). These data suggest a dominant phenotype for mutant fim alleles.

AWM401 (Δfim lpfC pefC agfB), a quadruple fimbrial mutant, was more strongly attenuated (26-fold) than AJB12 (Δfim lpfC pefC) or any of the strains carrying a single fimbrial mutation. This result suggested an additive attenuating effect of these mutations on the ability of S. typhimurium to cause murine typhoid (Table 2). Furthermore, the increased virulence of strain AJB12 (Δfim lpfC pefC) compared to that of AWM401 (Δfim lpfC pefC agfB) supports the idea that the insertional inactivation of agfB is one of the mutations responsible for the strong attenuation of the quadruple mutant.

The quadruple mutant has a reduced ability to colonize liver, spleen, and intestine.

To investigate at which step during the infection process AWM401 (Δfim lpfC pefC agfB) is impaired, course of infection studies were conducted. Since our mouse virulence data (Table 2) suggested that a mutation in agfB in combination with a mutation in at least one other fimbrial operon is responsible for the attenuation of AWM401 (Δfim lpfC pefC agfB), strain AWM394 (agfB) was included in these studies. Groups of four mice were orally infected with 10⁸ CFU, and bacteria were recovered from Peyer’s patches, mesenteric lymph nodes, and spleens on days 3 (data not shown) and 5 postinfection (Fig. 2A). In addition, bacteria were recovered from the feces on days 1, 3, and 5 postinfection to monitor intestinal colonization (Fig. 2B). Compared to the wild type (AJB3), reduced numbers of both AWM401 (Δfim lpfC pefC agfB) and AWM394 (agfB) were recovered from internal organs and fecal pellets. However, these differences proved not to be statistically significant (P > 0.05). As bacterial numbers recovered from individual animals may vary greatly during infection, small differences between wild type and mutant may go undetected. In order to control for the variability between experimental animals, mixed infections with AJB3 (wild type), AWM394 (agfB), and AWM401 (Δfim lpfC pefC agfB) were performed, permitting a direct comparison between wild type and mutants. A group of four mice was orally infected with 10⁸ CFU of a mixture containing approximately equal amounts of AJB3 (wild type), AWM394 (agfB), and AWM401 (Δfim lpfC pefC agfB). On day 5 postinfection, CFU in internal organs (Peyer’s patches, villous intestinal tissues, mesenteric lymphs, spleens, and livers) were determined. In addition, bacteria were recovered from the feces up to 5 days postinfection to monitor intestinal colonization (Fig. 3). Both AWM394 (agfB) and AWM401 (Δfim lpfC pefC agfB) were able to compete with the wild type (AJB3) for colonization of Peyer’s patches and villous intestinal tissues in the terminal ileum. Interestingly, increased numbers of both the quadruple mutant (AWM401) and the agfB mutant (AWM394) were recovered from the mesenteric lymph nodes compared to that of the wild type (AJB3). These differences were not statistically significant (P > 0.05). However, both AWM394 (agfB) and AWM401 (Δfim lpfC pefC agfB) were outcompeted by the wild type (AJB3) for colonization of the liver and spleen (P < 0.05 and P < 0.01, respectively). In addition, AWM401 (Δfim lpfC pefC agfB) failed to compete with the wild type for colonization of the intestine, as suggested by recovery of significantly reduced numbers of AWM401 from fecal pellets (P < 0.05). These results provide evidence that fimbrial adhesins act synergistically during colonization of the mouse intestinal tract.

Identification of new fimbrial structures.

Although we have demonstrated that fimbrial adhesins of S. typhimurium play an important role during infection (Table 2), AWM401 (Δfim lpfC pefC agfB) was still able to cause a lethal systemic illness in mice when administered at higher doses. These data suggest that AWM401 (Δfim lpfC pefC agfB) may express yet other factors for intestinal attachment. To investigate this possibility, we examined strain AWM401 (Δfim lpfC pefC agfB) by electron microscopy for the presence of fimbriae. Interestingly, this mutant (AWM401) expressed thus far uncharacterized fimbrial structures (Fig. 4A), which could easily be distinguished from flagellar filaments required for cell motility (Fig. 4). Flagellar filaments varied in length from 5 to 10 μm, with a diameter of approximately 20 nm (18). Fimbriae could be distinguished from flagella by means of morphology and diameter (typically between 2 and 8 nm [17]). These data provide direct evidence for the expression of at least one yet uncharacterized fimbrial structure in S. typhimurium which may contribute to the redundancy of virulence factors involved in colonization of the intestinal mucosa.