A large panel of chicken cells are invaded in vivo by Salmonella Typhimurium even when depleted of all known invasion factors.

Poultry are the main source of human infection by Salmonella. As infected poultry are asymptomatic, identifying infected poultry farms is difficult, thus controlling animal infections is of primary importance. As cell tropism is known to govern disease, our aim was therefore to identify infected host-cell types in the organs of chicks known to be involved in Salmonella infection and investigate the role of the three known invasion factors in this process (T3SS-1, Rck and PagN). Chicks were inoculated with wild-type or isogenic fluorescent Salmonella Typhimurium mutants via the intracoelomic route. Our results show that liver, spleen, gall bladder and aortic vessels could be foci of infection, and that phagocytic and non-phagocytic cells, including immune, epithelial and endothelial cells, are invaded in vivo in each organ. Moreover, a mutant defective for the T3SS-1, Rck and PagN remained able to colonize organs like the wild-type strain and invaded non-phagocytic cells in each organ studied. As the infection of the gall bladder had not previously been described in chicks, invasion of gall bladder cells was confirmed by immunohistochemistry and infection was shown to last several weeks after inoculation. Altogether, for the first time these findings provide insights into cell tropism of Salmonella in relevant organs involved in Salmonella infection in chicks and also demonstrate that the known invasion factors are not required for entry into these cell types.

Murine AML12 hepatocytes allow Salmonella Typhimurium T3SS1-independent invasion and intracellular fate.

Numerous studies have demonstrated the key role of the Salmonella Pathogenicity Island 1-encoded type III secretion system (T3SS1) apparatus as well as its associated effectors in the invasion and intracellular fate of Salmonella in the host cell. Several T3SS1 effectors work together to control cytoskeleton networks and induce massive membrane ruffles, allowing pathogen internalization. Salmonella resides in a vacuole whose maturation requires that the activity of T3SS1 subverts early stages of cell signaling. Recently, we identified five cell lines in which Salmonella Typhimurium enters without using its three known invasion factors: T3SS1, Rck and PagN. The present study investigated the intracellular fate of Salmonella Typhimurium in one of these models, the murine hepatocyte cell line AML12. We demonstrated that both wild-type Salmonella and T3SS1-invalidated Salmonella followed a common pathway leading to the formation of a Salmonella containing vacuole (SCV) without classical recruitment of Rho-GTPases. Maturation of the SCV continued through an acidified phase that led to Salmonella multiplication as well as the formation of a tubular network resembling Salmonella induced filaments (SIF). The fact that in the murine AML12 hepatocyte, the T3SS1 mutant induced an intracellular fate resembling to the wild-type strain highlights the fact that Salmonella Typhimurium invasion and intracellular survival can be completely independent of T3SS1.

Multiplicity of Salmonella entry mechanisms, a new paradigm for Salmonella pathogenesis.

The Salmonella enterica species includes about 2600 diverse serotypes, most of which cause a wide range of food- and water-borne diseases ranging from self-limiting gastroenteritis to typhoid fever in both humans and animals. Moreover, some serotypes are restricted to a few animal species, whereas other serotypes are able to infect plants as well as cold- and warm-blooded animals. An essential feature of the pathogenicity of Salmonella is its capacity to cross a number of barriers requiring invasion of a large variety of phagocytic and nonphagocytic cells. The aim of this review is to describe the different entry pathways used by Salmonella serotypes to enter different nonphagocytic cell types. Until recently, it was accepted that Salmonella invasion of eukaryotic cells required only the type III secretion system (T3SS) encoded by the Salmonella pathogenicity island-1. However, recent evidence shows that Salmonella can cause infection in a T3SS-1-independent manner. Currently, two outer membrane proteins Rck and PagN have been clearly identified as Salmonella invasins. As Rck mediates a Zipper-like entry mechanism, Salmonella is therefore the first bacterium shown to be able to induce both Zipper and Trigger mechanisms to invade host cells. In addition to these known entry pathways, recent data have shown that unknown entry routes could be used according to the serotype, the host and the cell type considered, inducing either Zipper-like or Trigger-like entry processes. The new paradigm presented here should change our classic view of Salmonella pathogenicity. It could also modify our understanding of the mechanisms leading to the different Salmonella-induced diseases and to Salmonella-host specificity.

Emergency and therapeutic vaccination--is stimulating innate immunity an option?

There is increasing evidence that activation of innate immunity, in animals and man, by live vaccines, sub-unit vaccines or synthetic or non-synthetic stimulants can induce a profound and rapidly induced resistance to pathogens, including infectious agents that are unrelated to the stimulating antigen or agent. We review the evidence for this phenomenon and present the proposition that this approach might be used to stimulate immunity during the life of the animal when susceptibility to infection is high and when normal vaccination procedures may be inappropriate.

Use of a reduced set of single nucleotide polymorphisms for genetic evaluation of resistance to Salmonella carrier state in laying hens.

Salmonella propagation by apparently healthy chickens could be decreased by the selection and use of chicken lines that are more resistant to carrier state. Using a reduced set of markers, this study investigates, for the first time to the authors' knowledge, the feasibility of a genomic selection approach for resistance to carrier state in hen lines. In this study, commercial laying hen lines were divergently selected for resistance to Salmonella carrier state at 2 different ages: young chicks and adults at the peak of lay. A total of 600 birds were typed with 831 informative SNP markers and artificially infected with Salmonella Enteritidis. Phenotypes were collected 28 d (389 young animals) or 38 d (208 adults) after infection. Two types of variance component analyses, including SNP data or not, were performed and compared. The set of SNP used was efficient in capturing a large part of the genetic variation. Average accuracies from mixed model equations did not change between analyses, showing that using SNP data does not increase information in this data set. These results confirm that genomic selection for Salmonella carrier state resistance in laying hens is promising. Nevertheless, a denser SNP coverage of the genome on a greater number of animals is still needed to assess its feasibility and efficiency.

Characterization of isolates of Listeria monocytogenes from sludge using pulsed-field gel electrophoresis and virulence assays.

AIMS: To study the diversity and virulence of Listeria monocytogenes isolated from sludge. METHODS AND RESULTS: A total of 60 isolates of L. monocytogenes from sludge were characterized by serotyping, PFGE typing and using in vitro and in vivo virulence assays. The PFGE patterns were compared with those of food and human isolates to determine whether specific group clones are associated with environmental samples. The 60 isolates gave 44 different combined ApaI/AscI PFGE patterns. The PFGE patterns of most isolates were similar or very similar to those of epidemic isolates. The majority (93%) of isolates were found to be virulent by plaque-forming assay and by mouse virulence assay. CONCLUSIONS: Our findings suggest that L. monocytogenes strains found in non-sanitized sludge are virulent and represent a potential health hazard. Although no case of listeriosis related to sludge spread onto agricultural land has been reported, particular attention to this pathogen is needed. SIGNIFICANCE AND IMPACT OF THE STUDY: This is the first study dealing with the characterization of L. monocytogenes isolates from non-sanitized sludge samples by molecular typing methods and in vitro and in vivo virulence assays. Our findings provide relevant information for evaluating the health risks associated with spreading sludge onto agricultural land.

A Listeria monocytogenes strain is still virulent despite nonfunctional major virulence genes.

The low-virulence Listeria monocytogenes strains have been previously assigned to 4 phenotypic groups. This study aimed to characterize the A23 strain, which exhibits a pulsed-field gel electrophoresis profile specific to low-virulence strains. This strain has the same causal mutations as the group III strains and a supplementary mutation in the mpl gene, leading to the absence of internalin A expression and the presence of inactive internalin B, phosphatidyl-inositol phospholipase C, and phosphatidylcholine phospholipase C. Despite these mutations in major virulence genes, the A23 strain formed plaques in cell monolayers and contaminated 100% of inoculated mice, suggesting that it evolved from group III strains by acquiring new virulence genes.

QTL for resistance to Salmonella carrier state confirmed in both experimental and commercial chicken lines.

The ability of chickens to carry Salmonella without displaying disease symptoms is responsible for Salmonella propagation in poultry stocks and for subsequent human contamination through the consumption of contaminated eggs or meat. The selection of animals more resistant to carrier state might be a way to decrease the propagation of Salmonella in poultry stocks and its transmission to humans. Five QTL controlling variation for resistance to carrier state in a chicken F(2) progeny derived from the White Leghorn inbred lines N and 6(1) had been previously identified using a selective genotyping approach. Here, a second analysis on the whole progeny was performed, which led to the confirmation of two QTL on chromosomes 2 and 16. To assess the utility of these genomic regions for selection in commercial lines, we tested them together with other QTL identified in an [Nx6(1)] x N backcross progeny and with the candidate genes SLC11A1 and TLR4. We used a commercial line divergently selected for either low or high carrier-state resistance both in young chicks and in adult hens. In divergent chick lines, one QTL on chromosome 1 and one in the SLC11A1 region were significantly associated with carrier-state resistance variations; in divergent adult lines, one QTL located in the major histocompatibility complex on chromosome 16 and one in the SLC11A1 region were involved in these variations. Genetic studies conducted on experimental lines can therefore be of potential interest for marker-assisted selection in commercial lines.

Prevalence of low-virulence Listeria monocytogenes strains from different foods and environments.

Various studies have demonstrated variations in the levels of virulence of different L. monocytogenes strains. In our laboratory, a plaque-forming assay followed by subcutaneous footpad inoculation of mice enabled us to estimate the prevalence of the low-virulence strains. This value fell from 16.3% to 1.7% with bacteria collected before 1994 and after 1997 respectively. This could be related to the modification in 1997 of the reference method EN ISO 11 290-1 of Listeria detection which recommended the use of polymyxin-acriflavine-LiCl-ceftazidime-aesculin-mannitol (PALCAM) medium. The aim of this study was to determine whether the percentage of low-virulence strains detected has changed due to the modification of the detection method recommending the use of the ALOA medium. After analyzing 380 L. monocytogenes strains, no increase in the percentage of low-virulence strains could be detected. The prevalence reached only 2.6% (ten of the 380 strains tested). The low virulence of L. monocytogenes strains was not related to rare serotypes and was also observed in serotypes usually involved in human disease. Low-virulence strains were found in dairy, meat, ready-to-eat products and also in the environment, highlighting the absence of one specific source. These results are discussed in terms of detection methods and the definition of low virulence.

Poor detection of low-virulence field strains of L. monocytogenes is related to selective agents in selective media and is unrelated to PrfA.

We have previously shown a relationship between the virulence level of Listeria monocytogenes strains and their detection on PALCAM medium. To account for the fact that only 40% of low-virulence field strains of L. monocytogenes were detected on PALCAM medium compared to 92% on ALOA medium, the detection of virulent and low-virulence strains on decomposed selective ALOA and PALCAM media was compared. This showed that better detection of the strains was not explained by the growth factors added to the ALOA medium. On the other hand, the presence of acriflavine in the PALCAM medium partly explained the delay in detection of the low-virulence strains, while the presence of ceftazidime was related to growth inhibition. However, the effect of these two components was modified when they were combined in the PALCAM medium. As some of these low-virulence strains had an inactive PrfA (the transcriptional activator of the main virulence genes of L. monocytogenes), its role in the poor detection of these low-virulence strains was investigated. However, complementing these strains with the wild-type prfA gene or deleting the prfA gene from a virulent strain suggested that this poor detection was unrelated to PrfA, but was related to their higher susceptibility to the antimicrobial components in the selective media.

An integrated approach of genetic resistance to Salmonella carrier state in fowls: from genetics to genomics and modelling.

Increasing resistance to acute Salmonellosis (that is, contamination level shortly after infection) is not sufficient to reduce the risk for consumers to be contaminated by Salmonella. Indeed, animals may remain contaminated at a low level for weeks or months. Increased resistance to the Salmonella carrier state, i.e., animals' ability to clear bacteria, is needed; it involves measuring bacterial contamination several weeks after inoculation with a low dose. To study such resistance traits, three convergent approaches were used. A quantitative trait loci (QTL) study was performed, taking advantage of inbred lines differing in resistance. Several QTLs controlling resistance at a younger age were identified and are currently being confirmed in a new cross before finer mapping, using advanced intercross lines. These inbred lines are also presently being compared using functional genomics. In parallel, a selection experiment for increased or decreased resistance at a younger and a later age was undertaken. Besides providing genetic models differing in their levels of resistance, it underlined the importance of the choice of selection criterion, whether marker assisted or not. Indeed, genes controlling resistance are strongly dependant on age; selecting for resistance at a younger age might result in increased susceptibility at an older age. Finally, the results of this experiment were used in a model of the intra-flock propagation of Salmonella. It showed that introducing a proportion of resistant animals within a flock of susceptible hens could dramatically change the evolution of contamination. Moreover, it demonstrated the magnitude of synergy between selection and vaccination, which should enhance the interest of increased resistance. The results show that selection for increased resistance to the Salmonella carrier state may be efficient, providing that the appropriate criteria of selection are used.

Multiple point mutations in virulence genes explain the low virulence of Listeria monocytogenes field strains.

In order to understand the causes of the low virulence of Listeria monocytogenes field strains, five low-virulence strains were analysed. These five strains showed changes in relation to invasion, phosphatidyl-inositol phospholipase C (PI-PLC) activity, plaque formation and in vivo virulence. Molecular analyses revealed the same mutations in the plcA, inlA and inlB genes in all five strains. The Thr262Ala substitution in the PI-PLC protein was responsible for the absence of PI-PLC activity. This residue, conserved in certain L. monocytogenes species, is located at the outer rim of the active site pocket and could impair the cleavage activity of the enzyme. The low invasion rate of these strains was due to a nonsense codon leading to a lack of InlA protein synthesis, and to an Ala117Thr substitution in the leucine-rich repeat of InlB, which altered the interaction with the Met receptor. Single trans complementation with the inlA(EGDe), inlB(EGDe) or plcA(EGDe) genes restored the capacity of low-virulence strains either to enter epithelial and fibroblastic cells or to express PI-PLC activity. Complementation by allelic exchange of the plcA(EGDe) gene on the chromosome and trans complementation with either the inlA(EGDe) or the inlB(EGDe) gene restored the ability to form plaques, but only partly restored the in vivo virulence, suggesting that there were other gene mutation(s) with consequences that could mainly be observed in vivo. These results indicate that the low virulence of L. monocytogenes strains can be explained by point mutations in a number of virulence genes; these could therefore be important for detecting low-virulence strains. Moreover, the fact that all the strains had the same substitutions suggests that they have a common evolutionary pathway.

A naturally occurring mutation K220T in the pleiotropic activator PrfA of Listeria monocytogenes results in a loss of virulence due to decreasing DNA-binding affinity.

The sequencing of prfA, encoding the transcriptional regulator of virulence genes, in 26 low-virulence field Listeria monocytogenes strains showed that eight strains exhibited the same single amino-acid substitution: PrfAK220T. These strains exhibited no expression of PrfA-regulated proteins and thus no virulence. This substitution inactivated PrfA, since expression of the PrfAK220T mutant gene in an EGDDeltaprfA strain did not restore the haemolytic and phosphatidylcholine phospholipase C activities, in contrast to the wild-type prfA gene. The substitution of the lysine at position 220 occurred in the helix alphaH. However, the data showed that the PrfAK220T protein is dimerized just as well as its wild-type counterpart, but does not bind to PrfA-boxes. PrfAK220T did not form a PrfA-DNA complex in electrophoretic mobility shift assays, but low concentrations of CI complexes (PrfAK220T-RNA polymerase-DNA complex) were formed by adding RNA polymerase, suggesting that PrfA interacted with RNA polymerase in solution in the absence of DNA. Formation of some transcriptionally active complexes was confirmed by in vitro runoff transcription assays and quantitative RT-PCR. Crystallographic analyses described the structure of native PrfA and highlighted the key role of allosteric changes in the activity of PrfA and especially the role of the Lys220 in the conformation of the helix-turn-helix (HTH) motif.

Modification of a virulence-associated phenotype after growth of Listeria monocytogenes on food.

AIM: To assess the effect of different foods, which have been implicated or not in cases of listeriosis, on the in vitro virulence-associated phenotype level of different Listeria monocytogenes strains. METHODS AND RESULTS: The virulence-associated phenotype level of L. monocytogenes was studied with the in vitro cell test based on a plaque-forming assay with a human adenocarcinoma cell line (HT-29) monolayer. Three strains of L. monocytogenes were grown in preparations (homogenate, 1-mum filtrate or 0.2-mum filtrate) of different food extracts ['rillettes' (potted minced pork), milk, raw salmon and cold-smoked salmon] or in a control medium, brain heart infusion (BHI). The bacterial suspensions grown in food extracts or in BHI at 37 degrees C were diluted with their growth medium (food extract or BHI) or with minimum essential medium before seeding on confluent HT-29 cell monolayers. Filtration of food extracts had no significant effect on the plaque numbers formed by the bacteria. A significant decrease in the plaque numbers was noted for the three strains when they grew in the rillettes extracts, compared with the other food extracts and BHI. The levels of in vitro virulence-associated phenotype of the strains after growth in the rillettes extract were similar to or lower than that of the hypovirulent internal reference strain L. monocytogenes 442. After growth in milk and cold-smoked salmon, the impact on virulence-associated phenotype depended on the strain. In contrast, plaque-forming assay indicated increased virulence-associated phenotype when the strains were switched from a nutrient-rich medium (food extract or BHI) to a minimum essential medium. CONCLUSIONS: In vitro virulence-associated phenotype level of the studied strains grown in BHI or cold-smoked salmon was the same as the control virulent strain EGD. In contrast, the nutrients present in rillettes may therefore substantially reduce the number of plaques but not the growth of L. monocytogenes. The utilization of minimum essential medium as diluent attenuates changes the effect of the food extract on virulence-associated phenotype in vitro. SIGNIFICANCE AND IMPACT OF THE STUDY: In the experimental design of this study, we showed that the nature of the food could affect the in vitro virulence-associated phenotype level of L. monocytogenes.

Vaccination and early protection against non-host-specific Salmonella serotypes in poultry: exploitation of innate immunity and microbial activity.

A recent European Union Directive required member states to put monitoring and control programmes in place, of which vaccination is a central component. Live Salmonella vaccines generally confer better protection than killed vaccines, because the former stimulate both cell-mediated and humoral immunity. Administering Salmonella bacteria orally to newly hatched chickens results in extensive gut colonization and a strong adaptive immune stimulus but broiler chickens are immunologically immature. However, colonization exerts a variety of rapid (within 24 h) protective effects. These include specific colonization-inhibition (competitive exclusion) in which the protective bacteria exert a profound resistance to establishment and colonization by other related bacteria. This is thought to be primarily a metabolic attribute of the vaccinating bacteria but may also involve competition for attachment sites. The presence of large numbers of bacteria originating from a live Salmonella vaccine in the intestine can also induce infiltration of polymorphonuclear cells into the intestinal wall, which confers resistance to invasion and systemic spread by virulent Salmonella strains. This opens new perspectives for vaccine usage in broilers, layers and breeding poultry but also in other animals which show increased susceptibility to infection because of their young age or for other reasons, such as oral chemoprophylaxis or chemotherapy, where the lack of established normal gut flora is an issue. We recommend that all live vaccines considered for oral administration should be tested for their ability to induce the two protective effects described above. Further developments in live Salmonella vaccines are, however, currently hindered by fears associated with the use and release of live vaccines which may be genetically modified.

Investigation of specific substitutions in virulence genes characterizing phenotypic groups of low-virulence field strains of Listeria monocytogenes.

Several models have shown that virulence varies from one strain of Listeria monocytogenes to another, but little is known about the cause of low virulence. Twenty-six field L. monocytogenes strains were shown to be of low virulence in a plaque-forming assay and in a subcutaneous inoculation test in mice. Using the results of cell infection assays and phospholipase activities, the low-virulence strains were assigned to one of four groups by cluster analysis and then virulence-related genes were sequenced. Group I included 11 strains that did not enter cells and had no phospholipase activity. These strains exhibited a mutated PrfA; eight strains had a single amino acid substitution, PrfAK220T, and the other three had a truncated PrfA, PrfADelta174-237. These genetic modifications could explain the low virulence of group I strains, since mutated PrfA proteins were inactive. Group II and III strains entered cells but did not form plaques. Group II strains had low phosphatidylcholine phospholipase C activity, whereas group III strains had low phosphatidylinositol phospholipase C activity. Several substitutions were observed for five out of six group III strains in the plcA gene and for one out of three group II strains in the plcB gene. Group IV strains poorly colonized spleens of mice and were practically indistinguishable from fully virulent strains on the basis of the above-mentioned in vitro criteria. These results demonstrate a relationship between the phenotypic classification and the genotypic modifications for at least group I and III strains and suggest a common evolution of these strains within a group.

Bacteria-host interactions of Salmonella Paratyphi B dT+ in poultry.

In recent years, a dramatic increase in incidence of the dextro-rotatory tartrate-positive variant (dT+) of Salmonella enterica subspecies enterica serovar Paratyphi B has been observed in poultry and poultry products. In the present study the interactions of this bacterium with the host were studied in vivo and in vitro in an attempt to explain the preferential association of this serotype with poultry. The ability of this organism to invade and multiply in chicken intestinal epithelial cells and the intracellular behaviour in chicken macrophages was studied in vitro using chicken cell lines. In vivo challenge experiments in specific pathogen-free chickens were carried out to determine the level of colonization of caeca and internal organs early after experimental infection. An in vivo trial with commercial broiler chickens, using a seeder model, was performed to determine whether S. Paratyphi B dT+ could persist and spread in broilers until slaughter. S. Paratyphi B dT+ invaded and multiplied in the chicken epithelial cell line and survived in a chicken macrophage cell line. The strain used colonized caeca and internal organs of chickens to a high extent 1 week after infection with a low-dose inoculum. Moreover, the strain was efficiently transmitted within a group of broilers and persisted until slaughter. It was concluded that S. Paratyphi B dT+ was well adapted to poultry and therefore it is suggested that specific control measures against this serotype should be considered.

Invasion of Salmonella enteritidis in avian intestinal epithelial cells in vitro is influenced by short-chain fatty acids.

Fermentation reactions in the caeca of chickens, the predominant place for Salmonella colonization, result in high concentrations of short-chain fatty acids (SCFA). Thus Salmonella bacteria are in close contact with SCFA during their life cycle. A study was carried out to analyse the effects of SCFA on invasion of Salmonella enteritidis in an avian intestinal epithelial cell line. Preincubation of S. enteritidis for 4 h in growth media supplemented with various concentrations of propionate or butyrate resulted in decreased invasion compared to bacteria, preincubated in nonsupplemented media, and to bacteria, preincubated in media supplemented with formate or acetate. Incubation of the S. enteritidis bacteria in media supplemented with mixtures of SCFA mimicking the in vivo caecal concentrations resulted in increased invasion compared with butyrate-exposed bacteria, but equal invasion compared with nonexposed bacteria. Increasing the butyrate concentration in these mixtures did not modify invasion compared with the original mixtures.

Experimental validation of low virulence in field strains of Listeria monocytogenes.

Several reports have described Listeria monocytogenes strains which were nonpathogenic or weakly pathogenic, but little is known about these low-virulence strains. We found that 9 field L. monocytogenes strains were hypovirulent and 17 were avirulent, based on the number of mice contaminated and the colonization of their spleens after subcutaneous inoculation. All these strains possessed the known virulence genes. We have now assessed the low virulence of these strains in other assays before determining how they differ from virulent strains. We have shown that the low-virulence strains exhibited a phenotypic stability and were not a mixture of virulent and avirulent bacteria. They did not recover virulence after many passages in mice and colonized the spleens of mice more poorly than virulent strains after i.v. inoculation. Their lethal capacities, determined by 50% lethal dose (LD(50)), were lower than those of virulent strains. Like Listeria innocua, 14 of 17 avirulent strains had no LD(50) and were eliminated by the lymph nodes after subcutaneous inoculation. The virulent, hypovirulent, and avirulent strains were always significantly different, whatever the tests of virulence used, confirming the importance of these low-virulence field strains in identifying the proteins involved in virulence.

Interactions of Salmonella enterica subsp. enterica serovar Muenchen with intestinal explants of the turtle Trachemys scripta scripta.

Salmonella infections in reptiles, in contrast to those in birds and mammals, are limited to the intestinal tract. In this study, interactions of a strain of Salmonella enterica subsp. enterica serovar Muenchen (SEEM) with intestinal explants of the turtle Trachemys scripta scripta were examined by scanning electron microscopy (SEM). Adhesion and invasion in the chelonian intestinal explants at 30 degrees C and 37 degrees C were evaluated quantitatively. For purposes of comparison, the invasive capacity of SEEM in the continuous avian epithelial cell line DIV-1 at 30 degrees C and 37 degrees C was determined. Small numbers of M-like cells were found in the ileum of the turtles. The bacteria adhered mainly to the mucus of the intestinal explants. Only small numbers of salmonellae were associated with epithelial cells. Higher numbers of bacteria adhered at 30 degrees C than at 37 degrees C. Epithelial damage, embedding of bacteria in the epithelial surface and a ruffling-like process were noted only at 37 degrees C. Minimal numbers of salmonellae invaded the explants at 30 degrees C and 37 degrees C. Invasion of DIV-1 cells was greater at 37 degrees C than at 30 degrees C. The study suggested that the intestinal mucous layer provides an important site of colonization for salmonellae in the chelonian host and protects the underlying epithelial cells.