M1 serotype (strain SF370): pharyngeal cell adherence vs. association

We developed spotted oligonucleotide arrays of the S. pyogenes SF370 (an M1 serotype) genome and compared the transcriptomes of streptococci that adhere to Detroit 562 human pharyngeal cells to non-adherent (“associated”) streptococci within the same experiment. We replicated experiments independently and used dye-swaps to incorporate biological and technical variation, biological replicate experiments. Following filtering and normalization, we analyzed data from four biological replicates with robust summary statistics, Bayesian statistics and permutation algorithms to identify genes differentially expressed with significance during pharyngeal cell adherence. This analysis identified 79 genes (4% of the genome) exhibiting statistically significant fold changes in expression (PF value < 0.05) during adherence from 1769 open reading frames represented on the array. The genes encoding several known virulence factors (e.g. cysteine protease SpeB and streptolysin O) were differentially expressed with significance; however, many were either never before associated with adherence (e.g. phage-encoded genes including speH), or of undefined function. Genes demonstrating up-regulation (n= 45) and down-regulation (n= 34) included virulence factors, prophage-encoded transcripts, metabolic genes and transcriptional regulators (Table 2). Undefined or hypothetical genes comprised 27% of differentially expressed genes (n= 21; 11 chromosomally-encoded, 10 phage-encoded). We conducted TaqMan (qRT-PCR) analysis of 11 differentially expressed genes to validate selected microarray hybridization results. We further analyzed the data with neighbor clustering, using a set of novel computational algorithms (termed “GenomeCrawler”) for evaluating bacterial microarray data, which couples the expression profiles of genes with their physical location on the chromosome. When applied to the microarray data from this study, neighbor clustering identified a greater number of differentially expressed genes, which facilitated the reconstruction of known multimeric proteins, complete metabolic pathways and intact virulence loci that would not have been possible without its application. These data sets and algorithms are freely available for download at www.rockefeller.edu/vaf. Keywords: Pharyngeal cell adherent vs. associated Series contains 4 biological replicates including dye-swap experiments