Salmonella enterica is a gram-negative enteric bacterium that comprises about 2500 serovars 1. While some have a restricted host range (e.g. the serovars Typhi and Pullorum are restricted to humans and chickens, respectively), most of the S. enterica serovars can infect a broad range of warm-blooded animals and humans. S. enterica infects its hosts by the oral route and primarily causes two types of disease: a gastroenteritis characterized by the development of bacteria in the intestinal tract 2, and typhoid fever that results from the invasion of the systemic compartment 3. Typhoid fever is a serious health issue in developing countries 4 but is rare in the Western world. In contrast, Salmonella gastroenteritis is an important concern worldwide. Food products, including poultry, pork, egg, and milk constitutes an important source of Salmonella infection in humans 5. Salmonella enterica serovar Typhimurium (will be referred to hereafter as Typhimurium) is a broad host range serovar that infects humans, cattle, mice, and chickens, and is one of the major causes of food-borne human salmonellosis 67.

Typhimurium remains an important concern to the poultry industry 8 causing a systemic infection in newly hatched chicks, often resulting in death 9. In older birds infection by Typhimurium leads to an asymptomatic carriage state with colonization of the digestive tract and continuous shedding 1011. These healthy carrier birds constitute a risk of contamination of newly hatched chickens, as well as the food chain leading to both important economic losses and potential harm to human consumers.

The pathogenesis of Salmonella has been extensively studied in the mouse 12. In susceptible mice, Salmonella causes an acute systemic disease with limited intestinal manifestations 13. Recently, a model of Salmonella enterocolitis has been developed in streptomycin-treated mice 14. Studies using these mice and other animal models of Salmonella diseases have yielded substantial data about the molecular players involved at different levels. The Salmonella pathogenicity islands (SPIs) 1 and 2 are two major virulence determinants of S. enterica. They encode type III secretion systems (T3SS) that form syringe-like organelles on the surface of gram-negative bacteria and enable the injection of effector proteins directly into the cytosol of eukaryotic cells 1516. These effectors ultimately manipulate the cellular functions of the infected host and facilitate the progression of the infection. SPI1 and SPI2 play several roles in different organs within the host. SPI1 primarily promotes the invasion of non-phagocytic intestinal epithelial cells and the initiation of the inflammatory responses in the intestines 1718. It is also involved in the survival and persistence of Salmonella in the systemic compartment of the host 192021. The first characterized role of SPI2 was its ability to promote Salmonella survival and multiplication in phagocytic cells that constitute the main reservoirs for dissemination of the bacteria into systemic organs 16. SPI2 also plays an important role in the intestinal phase of Salmonella infection in mice 172223.

The regulation of SPI1 and SPI2 gene expression involves numerous transcriptional regulators located both inside and outside these pathogenicity islands. The regulation of SPI1 is particularly complex. SPI1 encodes for the five regulators HilA, HilC, HilD, InvF, and SprB (Figure 1). The first four of which are involved in regulatory pathways that lead to the activation of SPI1 genes and of genes encoding T3SS effectors located outside SPI1. In contrast to SPI1 the regulation of SPI2 genes is simpler with the SsrAB two-component system being the only transcriptional regulator encoded within SPI2 that activates the expression of SPI2 genes and of genes encoding T3SS effectors located outside SPI2. Interestingly, SPI1 regulators can regulate SPI2 genes. These include HilA that binds and represses the promoter of ssaH 24, and HilD that binds and activates the promoter of the ssrAB operon 25. In contrast, SsrAB has never been shown to act on the expression of SPI1 genes.

Few studies have investigated the role of SPI1 and SPI2 during the infection of chickens. In studies using Typhimurium, two approaches have provided data about the roles of SPI1 and SPI2. The first approach compared colonization in chickens by infecting with single strains and enumerating colonies from internal organs. Porter and Curtiss 26 found that mutations in structural genes of the SPI1 T3SS resulted in a reduction of the colonization of the intestines in day-old chickens. Jones et al. 27 generated strains with deletions of spaS and ssaU, genes that encode structural proteins of the SPI1 and SPI2 T3SS respectively, and compared their ability to colonize the cecum and liver in one-day and one-week old chickens to that of wild type. They concluded that both SPI1 and SPI2 play major roles in both the intestinal and the systemic compartments, with SPI2 contributing more than SPI1 in both compartments. The second approach screened random transposon libraries for reduced recovery from the chicken gastrointestinal tract through cloacal swabbing. Turner et al. 28 analyzed a library of 2,800 mutants for intestinal colonization in chickens. Among the mutants that showed reduced intestinal colonization they found one in which the SPI1 gene sipC was inactivated. No mutations in SPI2 genes were identified in this screen. Morgan et al. 29 screened a library of 1,045 mutants in chickens and found two mutations in SPI1 genes and one in a SPI2 gene that led to a reduction in colonization ability. The SPI1 mutants were unable to be recovered from 50% or 100% of the day old birds tested, while the single SPI2 gene was unable to be recovered in only 33%. In this study fourteen strains with mutations in SPI1 and fifteen strains with mutations in SPI2 did not show any defect in colonization. The authors of these two studies concluded that SPI1 and SPI2 play a marginal role in the colonization of chicken intestines by Typhimurium.

To gain better insight in the role of these important virulence factors we have taken a different approach. First, we performed mixed infections in which the strains that are being compared (the wild type and a mutant, or two different mutants) are co-administered. This approach more directly addresses the contribution of SPI1 and SPI2 by decreasing the animal to animal variations inherent in such studies and giving us the ability to test the fitness of two mutants directly against each other. Second, we used one-week-old chicks that are known to be resistant to acute infection by Typhimurium 1011 allowing us to follow the effect of the studied mutations over a relatively long period of time. Third, we used mutants in which the entire SPI1 and/or the entire structural operon of SPI2 are deleted (Figure 1). This inactivates all the genes involved in both SPI1 and SPI2 T3SS apparatus synthesis and prevents the action of SPI1 regulators on SPI2 gene expression. Using this approach, we compared the colonization of the wild-type to that of each of the mutants.

We report here that SPI1 contributes to the colonization of both the cecum and spleen of the chicken. In contrast, SPI2 contributes to colonization of the spleen but not the cecum and, in the absence of SPI1, inhibits cecal colonization. Additionally, we show that the contribution of SPI1 in the spleen is greater than that of SPI2. These results differ from those observed during the infection of mice by Typhimurium, where SPI2 plays a major role during systemic colonization.

To assess the roles of SPI1 and SPI2 in the colonization of the gut and internal organs of the chicken, we used a mixed infection approach 30. We orally infected one-week old chickens with mixtures of two strains. Each strain carried different antibiotic resistance markers providing a simple means of identification. At days three, seven, and fourteen post-infection, groups of chickens were euthanized. The spleen and a sample of cecum from each bird were recovered, processed and plated for enumeration of colonies as described in the Methods section. The ratio of the two strains recovered from each organ was determined and compared to the input ratio to determine the competitive index (CI, ratio of the two strains from an organ divided by the ratio in the suspension used for the infection).

SPI1 and SPI2 are important virulence determinants of S. enterica serovars that have been extensively studied in several animal models. Few studies have investigated the role of SPI1 and SPI2 in the colonization of the chicken by Typhimurium. These studies have analyzed the colonization of different organs in chickens infected with a wild type strain or with mutants of SPI1 or SPI2 in which a single T3SS structural gene was inactivated. To gain better insight in the roles played by SPI1 and SPI2 in the chicken we used an approach that combined mixed infections, large deletions in SPI1 and SPI2, and the tracking of infections for fourteen days. We found that SPI1 contributes to colonization of both the cecum and the spleen in chickens. In contrast, SPI2 plays a role in the colonization of the spleen, but not of the cecum. Furthermore, we show for the first time to our knowledge, that SPI1 plays a more important role than SPI2 in colonization of the chicken spleen by Typhimurium. We cannot exclude the possibility that the phenotypes conferred by the mutations we constructed resulted from pleiotropic effects given that we deleted the entire SPI1 and the majority of SPI2 T3SS genes, and that SPI1 and SPI2 regulators have been shown to act on the expression of genes located outside these pathogenicity islands whose functions are unrelated to T3SS 2432. This has been reported previously in mice where the deletion of the entire SPI1 had a different effect than a single gene deletion 33. However, it seems unlikely as other studies have yielded results that are consistent with some of our findings. For instance, two studies that screened transposon mutant libraries of Typhimurium for reduced colonization of the chicken gastrointestinal tract either found mutations in SPI1 but not in SPI2 28 or that SPI1 mutations had greater influence 29. Despite the fact that cecal swabbing was used to recover strains in these two studies, which may fail to catch low level colonization, both studies still identified SPI1 as important in intestinal colonization. Cecal colonization was also reported to decrease substantially after the deletion of SPI1 T3SS components 26. Additionally, a study with S. enterica serovar Enteritidis, which displays an infection pattern similar to Typhimurium, showed that deletion of the ssrA gene, encoding the sensor component of the SsrAB two-component system that is the major regulator of the SPI2 gene expression, did not affect the colonization of the chicken digestive tract 34. All together these results suggest that Typhimurium relies less on SPI2 than on SPI1 for colonization of the intestinal track in one-week-old chicks.

In contrast, Jones et al. 27 analyzed the contribution of SPI1 and SPI2 to the colonization of chickens by Typhimurium through the deletion of a single T3SS structural gene in each. They concluded that the SPI2 T3SS was required for systemic infection and played a significant role in the colonization of the gastrointestinal tract, while the SPI1 T3SS was involved in both compartments without being essential 27.

There are several important differences between that study and ours. First, Jones et al. used derivatives of the Typhimurium F98 strain 9 while we used derivatives of the UK-1 strain 36. While both have been well characterized for virulence and persistence in chickens, their mean lethal dose (LD₅₀) in day of hatch chicks differ by two orders of magnitude with F98 at 5 × 10⁵ cfu 35 and UK-1 at approximately 2 × 10³ 36. Second, they studied mutants in which a single structural T3SS gene was inactivated while in our mutants the entire SPI1 and all the SPI2 T3SS structural genes were deleted. Third, they determined the level of colonization of the chicken by calculating the bacterial density (number of colony forming unit per gram) in the organs after administration of single strains while we infected the chickens with mixtures of the two strains being compared and determined the competitive index. These differences may account for the differences in the results.

Through the use of different combinations of mutants used for infection we have observed that strains harboring the Δspi2 mutation have a modest advantage in the colonization of the chicken cecum, and therefore SPI2 may act to repress some factor needed for cecal colonization. However, this observation was only statistically significant when SPI1 was absent both in the strain that harbored the Δspi2 mutation and the competing strain (Figure 5A). We have come to this conclusion based on the above observation in addition to the fact that while the Δspi1 is out-competed by the wild type (Figure 2A), the double mutant Δspi1 Δspi2 is not (Figure 4A). We do not know the basis of this disadvantage conferred by the presence of SPI2 in the colonization of chicken cecum by Typhimurium. One explanation is that genes deleted from SPI2 may normally act to repress some factor needed for the colonization of the cecum but in their absence this factor is not repressed, thus increasing invasion. An alternative explanation may be that the phenotype conferred by the Δspi2 mutation in not decreasing intestinal colonization results from the absence of SPI1 regulators, such as HilD, that are known to regulate SPI2 genes, including the SsrAB central regulator. Additional investigations are needed to test these hypotheses.

In contrast to what we have observed in chickens, SPI2 is the major contributor for spleen colonization in BALB/c mice. The infection by Typhimurium in these two animal models leads to different outcomes. In mice, Typhimurium causes an acute systemic infection, frequently resulting in death, while in one-week or older chickens, the infection leads to heavy colonization of the intestinal track and asymptomatic carriage. It is interesting to note that in animal models where Salmonella infection results in acute systemic disease, SPI2 is a major player in the systemic infection. These include the infection of mice by Typhimurium 12, and the systemic disease in chickens infected by serovars Pullorum 37 and Gallinarum 38. In contrast, in animals where infection results in healthy carriage, such as in chickens, SPI2 plays a minor role in the persistence of the bacteria in the systemic compartment. This is demonstrated in the present study, and has been reported for Typhimurium in pigs 39, and for serovar Enteritidis in chicken 40. This difference in contribution of SPI2 in these two situations indicates that SPI2 is an important factor of Salmonella host specificity.

We have taken a mixed infection approach to study the role of SPI1 and SPI2 in the colonization of the chicken by Typhimurium. We confirmed the contribution of SPI1 to the colonization of both the cecum and the spleen, and showed that SPI2 is involved in the colonization of the spleen but not of the cecum and, may have a negative effect on cecal colonization. Additionally, we show that SPI1 plays a greater role than SPI2 in the colonization of the spleen in chickens. In contrast, SPI2 is more important than SPI1 for systemic colonization in mice. The approach we used in this study constitutes a sensitive assay that provided new insights into the role of SPI1 and SPI2 during infection.

RCIII provided the idea for this study. YD designed the experiments and constructed the mutants. YD, KA, and MM performed the animal experiments. YD wrote the manuscript.

RCIII and KA revised the manuscript. All authors read and approved the final manuscript.