Comparative Global Gene Expression Profiles of Wild-Type Yersinia pestis CO92 and Its Braun Lipoprotein Mutant at Flea and Human Body Temperatures

Abstract

Braun/murein lipoprotein (Lpp) is involved in inflammatory responses and septic shock. We previously characterized a Deltalpp mutant of Yersinia pestis CO92 and found that this mutant was defective in surviving in macrophages and was attenuated in a mouse inhalation model of plague when compared to the highly virulent wild-type (WT) bacterium. We performed global transcriptional profiling of WT Y. pestis and its Deltalpp mutant using microarrays. The organisms were cultured at 26 and 37 degrees Celsius to simulate the flea vector and mammalian host environments, respectively. Our data revealed vastly different effects of lpp mutation on the transcriptomes of Y. pestis grown at 37 versus 26 degrees C. While the absence of Lpp resulted mainly in the downregulation of metabolic genes at 26 degrees C, the Y. pestis Deltalpp mutant cultured at 37 degrees C exhibited profound alterations in stress response and virulence genes, compared to WT bacteria. We investigated one of the stress-related genes (htrA) downregulated in the Deltalpp mutant relative to WT Y. pestis. Indeed, complementation of the Deltalpp mutant with the htrA gene restored intracellular survival of the Y. pestis Deltalpp mutant. Our results support a role for Lpp in Y. pestis adaptation to the host environment, possibly via transcriptional activation of htrA.

Figure 1

Hierarchical clustering of genes determined to be significantly altered in the Δlpp mutant of Y. pestis CO92 cultured at 26°C relative to the WT bacteria. (a) Heat map showing clustering of genes differentially expressed between the Y. pestis Δlpp mutant, compared to WT bacteria, is presented. Clustering was performed on normalized and log-transformed hybridization signals using CLUSFAVOR 6.0 (Baylor College of Medicine, Houston, TX). The three replicate samples representing the two experimental conditions (Y. pestis CO92 or its Δlpp mutant) are labeled as WT and Lpp, respectively. Note that the two experimental conditions clustered apart from one another, and altered genes collectively exhibited a pronounced difference in signal intensity. The vertical dendrograms indicate relative similarity between samples (columns), while the horizontal dendrograms indicate clusters of genes (rows). Bright red indicates the highest normalized intensity value, bright green the lowest, and black median values. (b) Graphical representation of the cluster shown in panel (a). Normalized signal intensity values are shown on the ordinate, and experimental conditions are listed on the abscissa. The blue bars represent the range of normalized, log-transformed signal intensities for the entire group of genes while the red line indicates the median signal and thus the trend of gene expression differences. As shown, the average and median signal intensities for this group of genes is lower in the Y. pestis Δlpp mutant, compared to the WT bacteria.

Figure 2

Venn diagram showing the overlap of major functions of genes identified as significantly altered in a Δlpp mutant of Y. pestis CO92 and the WT strain. Functions were obtained from the CMR online database (http://cmr.jcvi.org) as well as from the literature. Numbers of genes significantly altered (at least 1.5-fold, Benjamini and Hochberg-corrected P value ≤.05) exclusively and commonly in the Δlpp mutant of Y. pestis CO92 cultured at 26°C and at 37°C, compared to its respective WT control, are also shown.

Figure 3

Complementation of the lpp mutant of Y. pestis CO92 with the htrA gene restores intracellular survival in macrophages. Intracellular survival of WT with pBR322, Δlpp mutant with pBR322, and the Δlpp with pBR322htrA was determined by infecting RAW 264.7 murine macrophages with an MOI of 1 for 45 minutes, followed by a 60-minute gentamicin wash, and plating the surviving intracellular bacteria at 4, 8, and 12 hours. The Δlpp mutant with pBR322htrA has a significant increase in percent survival compared to the Δlpp mutant with pBR322 alone, as determined by ANOVA and Holm-Sidak method.