Microarray-based detection of Salmonella enterica serovar Typhimurium transposon mutants that cannot survive in macrophages and mice

Abstract

DNA microarrays provide an opportunity to combine the principles of signature-tagged mutagenesis (STM) with microarray technology to identify potentially important bacterial virulence genes. The scope of DNA microarrays allows for less laborious screening on a much larger scale than possible by STM alone. We have adapted a microarray-based transposon tracking strategy for use with a Salmonella enterica serovar Typhimurium cDNA microarray in order to identify genes important for survival and replication in RAW 264.7 mouse macrophage-like cells or in the spleens of BALB/cJ mice. A 50,000-CFU transposon library of S. enterica serovar Typhimurium strain SL1344 was serially passaged in cultured macrophages or intraperitoneally inoculated into BALB/cJ mice. The bacterial genomic DNA was isolated and processed for analysis on the microarray. The novel application of this approach to identify mutants unable to survive in cultured cells resulted in the identification of components of Salmonella pathogenicity island 2 (SPI2), which is known to be critical for intracellular survival and replication. In addition, array results indicated that a number of SPI1-associated genes, currently not associated with intracellular survival, are negatively selected. However, of the SPI1-associated mutants individually tested for intracellular survival, only a sirA mutant exhibited reduced numbers relative to those of wild-type bacteria. Of the mutants unable to survive in mice, significant proportions are either components of the SPI2 pathogenicity island or involved in lipopolysaccharide synthesis. This observation is in agreement with results obtained in the original S. enterica serovar Typhimurium STM screen, illustrating the utility of this approach for the high-throughput identification of virulence factors important for survival in the host.

FIG.1.

Data corresponding to (A) the SPI2 island and (B) the SPI1 island were extracted from the full macrophage microarray data set and organized by genome order. Features corresponding to sirA and barA, neither of which are in SPI1 but have been shown to regulate the island, are included at the bottom of the SPI1 image. All seven arrays, including the average for each feature, are represented. Arrays that are independent biological replicates are grouped at the top by color bars and designated by the letters A, B, and C, corresponding to experiments at an MOI of 1, an MOI of 10, and an MOI of 10, respectively. If multiple features are present for a given gene, they are grouped by the presence of a vertical bar next to the gene name. Asterisks indicate features that are present in the SAM macrophage SGL. The color scheme for change (n-fold) is indicated at the bottom. Generally, yellow indicates genes whose corresponding mutants are absent from the macrophage-selected library, while blue/black indicates their presence.

FIG. 2.

RAW 264.7 macrophage survival and replication assays with individual SPI2 mutants. (A) Mutants were placed in a background where the luxCDABE operon, under the control of an rpsM promoter, was constitutively expressed. The level of intracellular, gentamicin-protected bacteria was tracked by light production. Individual mutants are indicated on the x axis. The ssrA and ssrB genes were the only genes tested as a double mutant, as they are part of the same SPI2 regulatory unit. Each strain was tested in quadruplicate. The average log₁₀ relative light units (RLU) are reported with normalization to wild-type SL1344 levels and to the levels at 3 h postinfection as described in Materials and Methods. These results are representative of multiple assays. (B) The same strains were assayed by traditional colony counts. Each strain was tested in triplicate, and the average log₁₀ CFU/milliliter are reported after normalization to wild-type SL1344 levels and to the levels at 3 h postinfection. Open circles indicate genes that appear in the macrophage SGL as negatively selected, closed circles indicate genes that are represented in the macrophage microarray data set but did not appear in the SGL, and strains without any symbols are not represented in the microarray data set. Hours postinfection are indicated, mutants are ordered as they appear in the genome, and none of the strains displayed a growth defect in LB. Standard errors are reported, and asterisks indicate a P value of <0.01 as calculated using an unpaired t test comparing each mutant to the wild-type strain for each respective time point.

FIG. 3.

Individual SPI1-associated mutants that constitutively produce light were assayed for replication defects in RAW 264.7 macrophage cells as described in the legend of Fig. 2. Each strain was tested in quadruplicate, and the average log₁₀ relative light units (RLU) are reported after normalization to wild-type SL1344 levels and levels at 3 h postinfection. The ssrAB strain was included as a control. These results are representative of multiple assays. “p” indicates polar mutants, while “np” indicates nonpolar mutants. Standard errors are reported, and asterisks indicate significance (P < 0.01) as calculated using an unpaired t test comparing each mutant to the wild-type strain for each respective time point. Hours postinfection are indicated, and none of the strains displayed a growth defect in LB.

FIG. 4.

The distribution of S. enterica serovar Typhimurium LT2 genes across the Enterobacteriaceae and how this distribution in (A) the macrophage SGL compares to (B) the complete macrophage array data set of 5,044 features. The features in the SGL and the macrophage data set were associated with various categories as defined by McClelland et al. according to whether or not a homolog for each LT2 gene was present in various Enterobacteriaceae genomes including Klebsiella pneumoniae, E. coli K-12, E. coli O257:H7, S. bongori, and multiple S. enterica serovars (42). The “unknown” category comprised a negligible portion of the genes and was removed. In addition, the original “some or all Salmonella only,” “some Salmonella only,” and “all Salmonella only” categories were combined into the single “some or all Salmonella only” category for simplification. This category represents serovar Typhimurium LT2 genes that are present in most if not all the Salmonella serovars represented. The “all nine genomes” category represents genes that are conserved across all the Enterobacteriaceae sampled. The “other distributions” category encompasses genes that may have a varied distribution, being present in some of the non-Salmonella Enterobacteriaceae but not others. The remaining categories are self-explanatory. The key indicates color scheme and category association, and the percent values indicate the proportion of either the SGL or array data set that is comprised by each category.