Genome Plasticity of agr-Defective Staphylococcus aureus during Clinical Infection

Abstract

Therapy for bacteremia caused by Staphylococcus aureus is often ineffective, even when treatment conditions are optimal according to experimental protocols. Adapted subclones, such as those bearing mutations that attenuate agr-mediated virulence activation, are associated with persistent infection and patient mortality. To identify additional alterations in agr-defective mutants, we sequenced and assembled the complete genomes of clone pairs from colonizing and infected sites of several patients in whom S. aureus demonstrated a within-host loss of agr function. We report that events associated with agr inactivation result in agr-defective blood and nares strain pairs that are enriched in mutations compared to pairs from wild-type controls. The random distribution of mutations between colonizing and infecting strains from the same patient, and between strains from different patients, suggests that much of the genetic complexity of agr-defective strains results from prolonged infection or therapy-induced stress. However, in one of the agr-defective infecting strains, multiple genetic changes resulted in increased virulence in a murine model of bloodstream infection, bypassing the mutation of agr and raising the possibility that some changes were selected. Expression profiling correlated the elevated virulence of this agr-defective mutant to restored expression of the agr-regulated ESAT6-like type VII secretion system, a known virulence factor. Thus, additional mutations outside the agr locus can contribute to diversification and adaptation during infection by S. aureus agr mutants associated with poor patient outcomes.

Keywords: Staphylococcus aureus; gene regulation; genome analysis.

FIG 1

Genomic comparison of study strains with block alignments. Shown is a maximum likelihood phylogenetic tree based on core genome SNVs of all patient clones (left), with a graphical representation of the complete genome alignment (right). Branches in the phylogenetic tree are colored according to the patient from whom each strain originated. Bars indicating the number of substitutions per site in the phylogenetic tree or the alignment block length are shown at the bottom. Dotted lines are included in the tree as guides and do not reflect genetic distance. Core colinear blocks present in all isolate genomes are shown as solid rectangles in the multiple alignment and are colored according to the block location in each genome to highlight inversions (key at the bottom). Noncore regions present in only a subset of genomes are each represented with a unique striped fill pattern.

FIG 2

Map of SNVs and indels found in infecting strains. Shown is a mutation matrix of genes (rows) affected by variants of <5 nt that are unique to infecting strains (columns) or for which an ancestral state could not be determined. Genes mutated in multiple patients are grouped at the top, and the color key for different mutation types is shown at the bottom. RAST subsystem classifications associated with each gene are shown in the center (gray). PROVEAN scores (90) were calculated for each nonsynonymous mutation to assess the impact of a variant on the biological function of the encoded protein and are displayed in the rightmost column, with the key shown below. In cases where multiple variants were found in a gene, the PROVEAN score with the highest absolute value for each gene is shown. PG's prosthetic groups; TE's, transposable elements; PTS, phosphotransferase system.

FIG 3

Phenotypic characterization of clinical and genetically manipulated strains from patient 53. (A) Exoprotein profiles of strains grown in TSB for 5 h. Extracts were prepared from culture supernatants and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Coomassie blue staining. M, protein ladder. (B) Intoxication of primary human neutrophils (hPMNs) with culture filtrates from the indicated S. aureus strains and controls (USA300 LAC wild type and agr mutant) as a percentage (vol/vol). Results represent the standard errors of the means (SEM) of data from 5 donors and 2 independent colonies. (C) Survival among mice infected with the indicated strains via intravenous inoculation (1 × 10⁸ CFU). Mouse survival results are for 15 mice per group. P values for differences in survival were determined by Bonferroni-corrected log rank (Mantel-Cox) tests (P = 1.03 × 10⁻¹ for agr-defective LAC versus agr⁺ LAC, P = 1.20 × 10⁻³ for the agr-defective nares strain versus the agr⁺ nares strain, P = 1.92 × 10⁻² for the agr-defective blood strain versus the agr⁺ blood strain, P = 1 for the agr-defective nares strain versus agr-defective LAC, P = 1.80 × 10⁻³ for the agr-defective blood strain versus agr-defective LAC, and P = 2.40 × 10⁻³ for the agr-defective blood strain versus the agr-defective nares strain).

FIG 4

Identification of patient 53 strain-specific changes in gene expression. (A) Bar plot showing expression levels in counts per million (CPM) of RNAIII and agrBDCA during late-exponential-phase growth in natural and laboratory-derived agr⁺ or agr-defective strains of isolates from patient 53. Experiments were performed in duplicate, and levels are plotted for each replicate individually. (B) Overview of expression changes between natural and laboratory-derived agr⁺ and agr-defective strains of blood and nares origins for 29 toxins, proteases, surface proteins, transporters, and regulatory genes implicated in S. aureus virulence and pathogenesis (42). The left column shows variants found in or near each gene. Center columns indicate the average changes in expression under the experimental conditions labeled at the top. Rightmost columns indicate false discovery rate (FDR)-corrected P values for the expression changes shown in the center columns. Identifications and descriptions are shown on the sides, and color keys are shown at the bottom. The position of each gene in the nares strain reference genome is indicated on the far left, and alternating colors of position markers and descriptions are used to denote directly adjacent genes on the same strand (i.e., putative operons). Known agr-regulated genes are highlighted by asterisks. Results are derived from the same experiment as the one for panel A. (C) Summary of expression changes for 103 genes with significant differences in expression between agr-defective blood (wild-type [wt]) and nares (Δagr) isolates from patient 53. The figure layout is the same as for panel B, and strain comparisons are indicated at the top. Selected genomic regions are annotated on the far right, and virulence genes are boxed on the left. Column numbers and names are shown at the top. HTH, helix-turn-helix; agr-def., agr-defective.

FIG 5

Rearrangement of the ess locus in patient 53 blood and nares strains. Matching ess regions in the nares (top) and blood (bottom) strains of patient 35 are indicated by shaded areas (gray) and connecting lines. The 4-kb inserted element and candidate regions for homologous recombination are highlighted by a horizontal line and yellow shading, respectively. Gene colors correspond to the type of encoded proteins, according to the key at the bottom.