In the current study we used chemotaxis and flagella mutants of the host adapted serovar S. Dublin and corresponding mutants of the broad host range serovar S. Typhimurium to study possible serovar differences in the importance of these genes for host pathogen interaction. The studies were based on defined mutants in one strain of each serovar, and we cannot rule out that there may be strain differences within serovar.
The constitutively tumbling cheB S. Dublin mutant, but not the constitutively smooth swimming cheA mutant, was negatively affected in invasion of epithelial cells. Since cheA has previously been shown to be important for S. Typhimurium cell invasion , which we also observed in our studies, S. Typhimurium and S. Dublin apparently differ with respect to the role of cheA in epithelial cell invasion. Lack of flagella (fliC mutation) caused reduced adhesion, which is in accordance with previously reported results for the effect of fliC/fljB mutation in S. Typhimurium  and our observations on the role of flagella in this serotype.
It has previously been reported that it is the flagella and not motility, which are important for cell adhesion and invasion , but it is currently unknown how precisely flagella influence this in a motility independent way, at least in cell culture experiments. Since we used centrifugation to maximize cell contact, it is also unlikely that our results were caused by reduced motility, which would lead to a reduction in number of contacts between bacteria and cells. Flagella in S. Typhimurium are expressed inside epithelial cells and can be demonstrated in infected cultured HeLa cells . During in vivo invasion, the stimulation of TLR-5 by flagellin and the following pro-inflammatory response may be important. However, invasion by S. Typhimurium in cell culture experiments happens within 15 minutes , and it is unlikely to be influenced by secretion of stimulating factors. A more likely explanation is down-regulation of SPI1 in flagella mutants, as suggested by Kim et al.. This down regulation can be caused by several regulatory systems, which control both flagella and virulence gene expression [24, 25].
Whether motility per se can be ruled out as important for invasion is still controversial since motility was shown to be essential for invasion of S. Typhi into cultured epithelial cells . A recent study with S. Typhimurium also suggests a requirement for motility in infection of epithelial cells. The invading population was demonstrated to consist of two populations. Some cells were only infected with few bacteria, which did not multiply to any great extent. These bacteria showed down-regulation of SPI-1 and fliC transcription. A fraction of approximately 10% of cells, however, was infected with bacteria that were motile, expressed invasion genes, possessed flagella, and multiplied at high rate. A speculation is that these cells may be ready to re-enter the lumen of the intestine to re-infect other cells . Whether a similar picture can be seen for S. Dublin remains to be investigated.
Similar to invasion into epithelial cells, mutation of chemotaxis and flagella genes caused reduced uptake by macrophage cells. The reason for this is unknown. The flagella and chemotaxis genes are down regulated once S. Typhimurium is inside a macrophage , probably to prolong the time the bacterium can stay inside the macrophage protected from neutrophil killing in the extracellular environment . The intracellular down regulation is controlled by the gene ydiV, which prevents transcription of the flagellin promoter . It is currently unknown how S. Dublin regulates it flagella expression in response to macrophage uptake. Despite the down regulation, flagella of S. Typhimurium are important for the outcome of the systemic phase of an infection, since lack of flagella leads to a decrease in the percentage of CD14+ and CD54+ cells resulting in a reduction of uptake of soluble antigens by these cells and fewer bacteria accumulating intracellular [29, 30]. Flagellin induces I-κBα degradation and subsequent NF-κB nuclear translocation, and induction of nitric oxide synthase [31–33]. This induces rapid de novo synthesis of tumour necrosis factor alpha (TNF-α), interferon gamma (IFN-γ), interleukin-1β (IL-1β) followed by IL-6 and IL-10, which is typical for a systemic inflammatory response.
Lack of flagella was found to allow net growth inside the macrophages over a 48 hours period, while wild type and chemotaxis mutant strains were reduced in numbers. The SPI-1 encoded type three-secretion system and flagella are important for rapid host cell death by pyroptosis seen after cell infection with S. Typhimurium . In the present investigation, lack of flagella caused reduced extracellular levels of lactate dehydrogenase, the intracellular enzyme used as an indicator of macrophage cell death, and this reduced killing can be the reason for the net growth observed with flagella-less mutants. The present investigation does not allow us to conclude which underlying mechanism that was responsible for the reduced cell death when flagella were absent. Wild type S. Typhimurium was significantly more cytotoxic than S. Dublin. When S. Dublin expressed S. Typhimurium fliC, the cytotoxicity increased above S. Typhimurium levels. This indicates that fliC is important for the level of cytotoxicity, however, the complemented strain used to show this had a higher number of flagella than the wild type strain, and we cannot rule out that this causes the increase in cytotoxicity. The plasmid used for complementation was based on pMF3, which has previously been used to complement knock out phenotypes in S. Typhimurium without adverse effects . More detailed studies are needed to demonstrate how these serotype differences relate to differences in the flagella sequence.
Significant cytokine production is generally assumed to require phagocytosis of the bacteria . This corresponds to uptake in our assays, and as pointed out by Winther et al. knock out mutants are not well suited to distinguish between lack-of-stimulation and lack-of-internalization responses. The flagella mutant of S. Typhimurium caused a reduced IL-6 cytokine production, but it also showed reduced uptake. We therefore included a control experiment where a 10 times higher challenge dose of the flagella mutant was used. The high challenge dose did not increase the IL-6 production, indicating that the lack of response was most likely not related to invasion levels. In support of this conclusion, the fliC and cheB mutants of S. Dublin also showed significantly reduced invasion, but absence of these genes in S. Dublin did not influence cytokine production. This result point to a fundamental difference between S. Dublin and S. Typhimurium in the way the flagella stimulates the host response, and calls for more detailed studies on structural functional relations in the signalling to the host.
The S. Dublin fliC mutant with S. Typhimurium provided in trans induced a lower response than the wild type strain. This result was surprising. Its phenotype is similar to a motA mutation, i.e. structurally the flagella appears normal, but they do not move. Naturally occurring motA mutants of S. Enteritidis stimulated transcriptional pro-inflammatory responses in Caco-2 cells , and there is no obvious reason why the complemented S. Dublin strain should not do the same. In cell culture experiment, a motA mutant of S. Typhimurium was non-invasive , which differs from the phenotype of our complemented mutant, and further studies are needed to clarify this observation.
Lack of stimulation of IL-6 expression has previously been seen with the host-specific serovar S. Gallinarum in a comparison to S. Typhimurium and S. Enteritidis after infection of a primary chicken cell line . No control was included in that study for the fact that S. Gallinarum contrary to S. Typhimurium and S. Enteritidis lacks flagella. Our results indicate that lack of IL-6 induction may be a general feature of host adapted/ host specific serotypes. Host specific serovar has been speculated to perform stealth like infection (i.e. down regulates several host responses) in comparison to the ubiquitous serovars . The lower cytotoxicity and lack of IL-6 responses support this assumption. In contrast to the role in IL-6 induction, none of the mutants differed significantly from the wild type strains in induction of oxidative responses. This result suggested that flagellin was not important for induction of the oxidative response.
Results on the role of flagella and chemotaxis genes in Salmonella host pathogen interaction have been contradictory (compare  and  with ), and we purposely looked for a sensitive assay to show subtle differences between strains. Co-infection assays have been shown to be more sensitive than assays where strains are tested individually . Using this assay, we found that flagella significantly influenced the number of bacteria that could be isolated from the spleen 4–5 days post oral infection of mice with S. Dublin, but not with S. Typhimurium. Chemotaxis genes were found to be dispensable in this assay, as previously reported for S. Typhimurium . Animal welfare regulations dictated us to scarify mice when they were severely affected by infection, and this prevented us from using one single end-point of infection. Potentially, this may have influenced the competitive indexes for S. Typhimurium, since this serovar propagated at different speed at systemic sites depending on the presence of flagella genes (see below). However, all mice were killed within a 24 hours period, and we do not believe that this significantly influenced our results.
Like cheA mutation, mutation of cheR confers a constitutively smooth swimming phenotype. We have not included this gene in our investigation, and we cannot rule out that it may have a different role in host pathogen interaction than cheA. We have performed preliminary testing of an S. Dublin cheR mutant and found that it corresponds to cheA with respect to phenotypes in cell assays and oral challenge of mice (unpublished), however, we do not have S. Typhimurium results to compare it to.
Flagella have been found to be important for the outcome of oral infection with S. Typhimurium in streptomycin treated mice, which is a model for studies of the entero-pahtogenesis of Salmonella. In this model flagella are essential for initiation of inflammation, creating an environment in which Salmonella prevails over the normal flora, and in this model, chemotaxis genes were also essential for the outcome of infection. Cattle are the natural host for S. Dublin, and in addition to differences caused by the choice of animal model, studies have shown that virulence factors may differ depending on the host . This must be taken into account when concluding on the current results. The changes in virulence observed when flagella were removed were relatively modest. We have previously demonstrated that flagella do not play an important role during extra animal survival , and we believe that selection for stable maintenance of the flagella apparatus must happen during the interaction with the host. It thus appears that these small differences are enough to provide the selective force.
It has previously been reported that a flagella mutant of S. Typhimurium is hyper virulent following intraperitoneal challenge of mice  and we confirmed this result. In contrast, the S. Dublin flagella mutant was not different from the wild type strain after intraperitoneal challenge. In conjunction with the results of IL-6 induction and cytotoxicity, this indicates that flagella are most important for S. Dublin in the initial invasion phase in the intestine, while it plays a minor role during the systemic phase. We suggest that a likely explanation for the contradicting results on the role of flagella in virulence of S. Typhimurium is that the results depends very much on the time point where bacterial load is measured. At early time points, lack of flagella causes a lower invasion, but at later time points, this is balanced by a higher ability to grow in the systemic phase.