Salmonellosis is one of the important causes of food-borne diseases throughout the world 1. Salmonella enterica contains more than 2,300 serotypes among which Typhimurium is an important causative agent of gastroenteritis. Adhesion of bacteria to the host epithelial cells is a prerequisite step in establishing infection. Specific adhesion requires the interaction of specialized complementary molecules in a ligand-receptor interaction between bacterial surfaces and host tissues 2. Proteinaceous hair-like structures called fimbriae on the surface of bacteria have been implicated in such an event 3. Many members of the family Enterobacteriaceae including Salmonella produce type 1 fimbriae, the most commonly found type of fimbrial appendages 4. Type 1 fimbriae comprise a family of rod-shaped organelles which are 7 nm in diameter and 0.2–2.0 μm long 5. Type 1 fimbriae adhere to different cell types including erythrocytes, leukocytes, intestinal cells, respiratory cells, protozoa, yeast, fungal hyphae, and plant root hairs 6. Several studies indicated that type 1 fimbriae also contribute to virulence 789. For example, type 1 fimbriae expressing S. Typhimurium caused persistent infection in swine 7. Type 1 fimbriae may also modulate the bacterial tropism to the gut of the host 8, and type 1-fimbriated Salmonella were more virulent than the fim-minus ones 9. The fact that more than 80% of Salmonella isolates produce this fimbrial type suggests that type 1 fimbriae play an important role at some stage in the life cycle of bacteria 10.

Phenotypic variation of the expression of type 1 fimbriae in S. Typhimurium was first characterized by Old et al. 1112. They described a biphasic growth pattern associated with the outgrowth of fimbriate bacteria incubated in static, liquid broth culture. Briefly, strongly type 1 fimbriate phase S. Typhimurium cells were isolated following serial passage every 48 h in static broth culture. Non-fimbriate phase bacteria were obtained by growth on solid media.

Current data indicate that the phenotypic expression of type 1 fimbriae in S. Typhimurium is the result of the interaction and cooperation of several genes in the fim gene cluster 131415161718. The FimZ transcriptional factor activates the fimA expression by binding to the fimA promoter, and in addition, FimZ also positively regulates its own transcription 13. FimZ requires another co-activator FimY to activate fimA expression, however, no evidence has revealed any interaction of FimY and the fimA promoter 19. Interestingly, fimY gene possesses five rarely used arginine codon which is recognized by an arginine tRNA gene fimU. A S. Typhimurium fimU mutant did not produce type 1 fimbriae and the fimY translation was inhibited 1617. fimU modulates the fimY expression in the translational level. A fimW deleted strain, on the contrary, overproduced type 1 fimbriae as compared to the parental strain. This negative regulation was proposed to be mediated through an inhibitory effect on the FimZ protein by a FimW-FimZ protein-protein interaction 15. Other than gene products within the fim gene cluster, McFarland et al. recently reported that a knockout mutation in lrp, encoding the leucine-responsive regulatory protein (LRP), was non-fimbriated 20. They demonstrated that LRP positively regulated type 1 fimbrial expression by binding to the promoter region of fimZ 20. Other gene products that may play a role in type 1 fimbrial expression were seldom characterized. Our long-term goal in the laboratory is focused on finding any genetic determinants, other than those previously described within the fim gene cluster, which may also play a role in the phenotypic expression of type 1 fimbriae in S. Typhimurium. Initially we were interested in characterizing the genetic elements that may be associated with the ability of S. Typhimurium to switch type 1 fimbrial expressions in response to solid agar and static broth culture conditions. We used transposon mutagenesis to generate a library for S. Typhimurium and screened mutants that exhibited different type 1 fimbrial phenotypes than the parental strain between solid agar and static broth culture conditions. The flanking DNA fragments of the transposon insertion sites in the mutants of interest were cloned and sequenced. Herein, we report on the genetic determinants in S. Typhimurium that may play a role in the regulation of type 1 fimbrial expression between solid agar and static broth culture conditions.

It is not surprising that many bacterial fimbriae have the ability to switch expression from a fimbriate to a non-fimbriate phenotype. Since bacterial population may face unpredictable milieu during their life cycle, having cells in both fimbriate and non-fimbriate phases offers the bacterial population an advantage of surviving in the diverse situations it will encounter 34. The mechanisms that control the fimbrial regulation have been described in detail for only few fimbrial types. For example, the regulation in E. coli type 1 fimbriae is primarily controlled by inversion of a 314-bp segment of DNA (fim switch region) upstream of the fimA gene coding for the major fimbrial subunit 35. Accessory proteins including the integration host factor (IHF), LRP and the histone-like nucleoid-structuring (H-NS) protein, also bind to the fim switch region to re-orientate the DNA into a configuration favourable for inversion 363738. Interestingly, the E. coli K-12 strain also possesses the sfm gene cluster (sfmACDHFZU), which is the orthologue of S. Typhimurium fim gene cluster 39. Nevertheless not much characterization has been performed on sfm. The control system for type 1 fimbriae in S. Typhimurium differs from that of E. coli. No DNA inversion event has been detected and several fim gene products have been implicated as regulators of type 1 fimbrial expression 13141516171840. Previous studies did not reveal that global regulator like IHF or H-NS play a role in S. Typhimurium type 1 fimbrial expression. In order to explore other elements outside the fim gene cluster that might also participate in type 1 fimbrial expression, we constructed a transposon library of S. Typhimurium. From the mutants that no longer exhibited the same type 1 fimbrial phenotypes as the parental strain whether grown on solid agar and in static broth cultures, 12 groups were classified according to the putative functions of the genes that were interrupted by the transposons. Not surprisingly, several mutants that had defects in the fim genes did not produce type1 fimbriae in either condition. The mutants possessing transposons in fimA, fimC, fimD, fimH, fimZ, and fimY all exhibited the predicted phenotypes. These fim genes are required for type 1 fimbrial biogenesis or regulation, and disruption of these genes causes the bacteria to be non-fimbriate in either condition. Interestingly, the present study identified 3 mutants that had defective genes for other fimbrial systems, and they produced type 1 fimbriae in both culture conditions. K23 had a transposon inserted in stbC that encodes the fimbrial usher protein for Stb fimbriae 41. K70 had insertion in fimbrial chaperone protein for long polar fimbriae (Lpf). Those two gene products are involved in fimbrial biogenesis. How disruption of the outer membrane usher gene, stbC, of the Stb fimbrial system affects another fimbrial system poses an interesting research topic. We did try to complement the stbC mutant with a recombinant plasmid possessing the stbC coding sequence. However, the stbC mutant exhibited resistance to a battery of antibiotics, including ampicillin and chloramphenicol those frequently used in molecular cloning. This unexpected finding impeded us to transforming available plasmids into this strain for the time being. PefI identified in K61 is the homolog of PapI in E. coli. PefI negatively regulates the production of PefA, the major fimbrial subunit of plasmid-encoded fimbriae (Pef) 42. Cross-talk between different fimbrial systems within the same microorganism has been demonstrated previously. For example, minor fimbrin of type 1 and F1C fimbriae of E. coli can be reciprocally exchanged to generate hybrid fimbriae with changed receptor specificities 43. The regulator PapB of pyelonephritis-associated pili (Pap) prevents inversion of the fim switch controlling the expression of type 1 fimbriae in the same E. coli strain 44. Recently Nuccio et al. also demonstrated that the presence of type 1 fimbrial biosynthesis genes interfered with the expression of plasmid-encoded fimbriae in S. Typhimurium 45. Although additional research is required to elucidate the interaction of different fimbrial systems, our results did stress the observation that different types of fimbriae may coordinate for their own benefit.

A global regulator, LRP, has been demonstrated to play a role in the regulation of both type 1 fimbriae and P fimbriae of E. coli and Pef of S. Typhimurium. Recently McFarland et al. demonstrated that an lrp knockout S. Typhimurium strain did not produce type 1 fimbriae and Lrp is a positive regulator for fim gene expression by interacting with the fimZ promoter region 20. FimZ is a positive regulator for fimA 14. Their data suggested that a positive contribution of fim gene expression by Lrp/fimZ promoter interaction may be due to the displacement of H-NS, a repressor for fimZ 2046. The K46 mutant with the transposon inserted into lrp did not produce type 1 fimbriae in either culture condition. Our result concurred with their finding. The role that lrp plays in terms of type 1 fimbrial regulation in S. Typhimurium becomes apparent.

Several S. Typhimurium strains had transposons inserted in the genes that encode enzymes. The K3 strain had a defect in cafA that encodes a ribonuclease named RNase G in E. coli. Umitsuki et al. reported that RNase G was involved in the in vivo degradation of adhE mRNA which encodes a fermentative alcohol dehydrogenase, AdhE 47. A mutation at cafA resulted in accumulation of adhE mRNA and prolonged the half-life of its mRNA. Interestingly, the K16 strain had a mutation in the gene that encodes alcohol dehydrogenase. The K16 strain constitutively expressed type 1 fimbriae. Overexpression of AdhE inhibited type 1 fimbrial expression, while the absence of AdhE caused type 1 fimbrial expression on solid agar or in static broth. How an enzyme involved in metabolism influences fimbrial expression is not clear, but previous studies demonstrated that type 1 fimbrial expression was linked to the nutrient status of the bacterial encounter. Broth culture supplemented with glucose favours Shigella flexneri remaining in a non-fimbriate state 6. Morgenroth and Duguid described a S. Typhimurium biotype called the FIRN strain; it is non-type 1 fimbriate, non-inositol-fermenting, and non-rhamnose-fermenting 48. The genes controlling fimbrial expression may be linked to the genes involved in carbohydrate metabolism. Interestingly, similar to the K16 strain, a variety of mutants in the present study having transposons inserted in the genes encoding "house-keeping" enzymes all constitutively exhibited type 1 fimbriae. Barak et al. used transposon mutagenesis to identify genetic determinants required for S. Newport to attach to plant tissue 49. They also found that alcohol dehydrogenase, hydrolase, bacteriophage protein, and several membrane associated proteins, including transporter protein, similar to our findings, were the essential components for this Salmonella serotype to adhere to plant tissue 49. The ability of Salmonella to colonize plant tissue and modulate type 1 fimbrial expression may share a common regulatory network.

A complementation test allowed us to examine if providing the coding sequence of the gene carried in a plasmid back to the mutant stain defective in such a gene would switch the phenotype of type 1 fimbrial expression to that originally observed. A comprehensive complementation test is not within the scope of the present study; however, we did test some strains. The complemented K3 cafA (pCafA), K43 miaA (pMiaA), and K73 nusA (pNusA) strains did not exhibit the type 1 fimbrial phenotype as the parental strain. Polar effect caused by the transposon could be one of the reasons. Another possibility could be due to the gene dosage effect conferred by the cloning vector. The K5 res (pRes) strain did repress the type 1 fimbrial production on solid agar as the parental strain LB5010. However, in static broth culture, the complemented strain did not exhibit type 1 fimbriae. The res gene, encoding for the restriction subunit of type III restriction-modification (R-M) system of S. Typhimurium, locates upstream of the mod gene which encodes the methyltransferase subunit of R-M system. Since Res subunit requires Mod to form a complex to recognize and cleave DNA 50, failure of mod expression caused by polar effect could account for the partially complemented phenotype of the K5res (pRes) strain. NusA is a transcription termination factor and under some circumstance it may also function as a transcription elongation factor 51. The ability of NusA to enhance RNA polymerase pausing would result in tight coupling of transcription and translation, thus interferes with the Rho-mRNA interactions and consequently blocks termination of transcription 52. The transposon inserted in nusA in K73 strain could cause transcription/translation coupling defect in a variety of genes including those that may affect fim expression. The pNusA transformed in K73 strain may not produce precise stoichiometric ration of NusA, leading K73 nusA (pNusA) strain remained unchanged as the K73 nusA strain. MiaA is an isopentenylpyrophosphate transferase that is required for tRNA modification. Ericson and Björk demonstrated that a miaA mutant in S. Typhimurium LT2 which was deficient in 2-methylthio-N6-(4-hydroxyisopentenyl) adenosine (ms²io⁶A37) in its tRNA induced pleiotropic effects on bacterial physiology such as reduced growth rate and polypeptide chain elongation rate 53. Non-fimbriate phenotype exhibited by K43 miaA strain in both culture conditions could be the result of such pleiotropic effects. The K48 ubiB mutant transformed with the pUbiB plasmid restored the type 1 fimbrial phenotype as the parental strain LB5010. The K48 ubiB (pUbiB) cultured on agar did not agglutinate yeast cells. Although signals of both LB5010 and K48 ubiB (pUbiB) from agar could be detected by densitometry in RT-PCR assay, we argued that such fimA amount would not be enough for scoring positive in yeast agglutination test. Our complementation assay did prove that the ubiB mutant could exhibit the same phenotype as LB5010 when it harboured a plasmid containing the coding sequence of ubiB. The ubiB in S. Typhimurium could encode a NAD(P)H-flavin-reductase, whose homologue in E. coli has been proved to be required for the first monooxygenase step in ubiquinone biosynthesis 54. How a gene involving in biosynthesis of ubiquinone, an essential component of electron transport chain influences type 1 fimbrial expression is currently unknown. However, Leonard et al. reported that UbiB protein contains motifs found in eukaryotic-type protein kinase 55. It was speculated that phosphorylation may be required for ubiquinone biosynthesis-activating proteins 55. Possibly UbiB would also phosphorylate another substrate protein that affects fimA expression. The function of ubiB in S. Typhimurium and the relationship of ubiB and fim expression demands more careful characterization. Our data clearly showed that an ubiB mutant produced fimA at the condition that either favours or inhibits type 1 fimbrial expression and complementing this mutant with a plasmid possessing ubiB coding sequence restored it to the phenotype similar to the parental strain.

The present study revealed that the genetic determinants other than the fim genes could have the potential to affect type 1 fimbrial expression in response to different standard laboratory culture conditions. However, a battery of genes identified by transposon mutagenesis still requires detail characterization to validate its linkage to fimbrial expression. The phenotype caused by polar effect is one of the drawbacks in this kind of study. Validation of the genes of interest that are associated with type 1 fimbrial expression in S. Typhimurium is presently under investigation in our laboratory.

The ability to switch between fimbrial expressions in bacteria may be beneficial for survival. Fimbriate bacteria may facilitate colonizing host cells at certain stages of infection, while non-fimbriate bacteria may also have their own role to play, for example, as in avoiding recognition by phagocytes, or allowing the bacteria to shed and colonize another site on the host. Compared to E. coli, Salmonella has a higher survival rate in the external environment, which promotes transmission to a new host 56. Persistent infection of Salmonella in the environment is enhanced by adhesion and biofilm formation, both of which involve fimbrial appendages 21. How S. Typhimurium fluctuates between type 1 fimbriate phase and non-fimbriate phase in vivo remains to be uncovered. Nonetheless the present study reveals that several genetic determinants may be associated with the ability of S. Typhimurium to modulate type 1 fimbrial expression between solid agar and static broth culture conditions. These two laboratory culture conditions may to some extent mimic the environmental milieu that Salmonella would encounter. We report here that other gene products besides those in the fim gene cluster are also required for type 1 fimbrial expression. How each gene influences type 1 fimbrial expression is an interesting research topic which warrants further investigation.

Y-CC drafted the manuscript. K-CW participated in the sequence alignment and molecular cloning experiments. Y-TC, C-HY, S-CM, and C-CF were responsible for the library screening, molecular cloning, and hybridization experiments. L-HC carried out the complementation tests. K-SY conceived and coordinated the study, and helped to draft the manuscript. All authors read and approved the final manuscript.