Microarrays reveal that each of the ten dominant lineages of Staphylococcus aureus has a unique combination of surface-associated and regulatory genes

Abstract

Staphylococcus aureus is the most common cause of hospital-acquired infection. In healthy hosts outside of the health care setting, S. aureus is a frequent colonizer of the human nose but rarely causes severe invasive infection such as bacteremia, endocarditis, or osteomyelitis. To identify genes associated with community-acquired invasive isolates, regions of genomic variability, and the S. aureus population structure, we compared 61 community-acquired invasive isolates of S. aureus and 100 nasal carriage isolates from healthy donors using a microarray spotted with PCR products representing every gene from the seven S. aureus sequencing projects. The core genes common to all strains were identified, and 10 dominant lineages of S. aureus were clearly discriminated. Each lineage carried a unique combination of hundreds of "core variable" (CV) genes scattered throughout the chromosome, suggesting a common ancestor but early evolutionary divergence. Many CV genes are regulators of virulence genes or known or predicted to be expressed on the bacterial surface and to interact with the host during nasal colonization and infection. Within each lineage, isolates showed substantial variation in the carriage of mobile genetic elements and their associated virulence and resistance genes, indicating frequent horizontal transfer. However, we were unable to identify any association between lineage or gene and invasive isolates. We suggest that the S. aureus gene combinations necessary for invasive disease may also be necessary for nasal colonization and that community-acquired invasive disease is strongly dependent on host factors.

FIG. 1.

Conditional tree constructed using the Spearman correlation of 161 isolates and 728 CV genes. A selection of 30 CV PCR products is displayed representing insertions/deletions and gene variants of interest (see Fig. 2). Each vertical line represents an isolate, and its branch on the tree is colored according to MLST CC type, with the same colors repeated in the row of boxes at the bottom with corresponding CC numbers. Note that the six pale green clusters represent 13 “orphan” isolates that did not match an MLST CC and do not cluster with the most common lineages by microarray. *, two CC5 (ST6) isolates that do not cluster with the other CC5 (ST5) isolates, marked in dark green; †, two CC1 (ST188) isolates that do not cluster with the other CC1 isolates, marked in red. In the central region, each row of colored squares represents the hybridization signal to a microarray PCR product. The corresponding PCR products are listed in Fig. 2. The intensity of all the colors is an indicator of the total signal intensity, while the color is an indicator of test signal over reference signal ratio. Thus, PCR products colored yellow hybridized to both the test and reference (MRSA252) isolates, PCR products in blue hybridized to the reference strain only, and PCR products in red hybridized to the test strain only. Spots with fluorescent signals lower than two times the background are flagged and colored gray, indicating a gene absent in both the test and reference. Genes with white signals are very low intensity and regarded as negative for both isolates. The figure clearly shows CV genes that are present or absent in accordance with lineage, and the lineages correlate with MLST CC groups.

FIG. 2.

PCR products in Fig. 1. PCR products representing CV genes or gene variant regions are listed in the same order as Fig. 1 by putative gene product name and identifier or annotated gene number (R products are from MRSA252, N products are from N315, 8- products are from 8325-4, and the M product is from MW2). “v” denotes a PCR product designed to a specific variant region, and so multiple variant regions of some genes are included. A black box indicates that the gene is present in that CC. “u” indicates variation in gene distribution for that CC. For some genes that are found in all isolates, variant regions may not have been sequenced and therefore are not on the microarray, e.g., the coagulase gene variant region in isolates of CC15.

FIG. 3.

Representation of the MRSA252 genome with each gene colored according to whether it is a core gene or CV gene or whether it is found on an MGE. The figure was generated in GeneSpring using the same lists used for identifying CV genes. The outer circle represents genes on the forward coding strand, and the inner circle represents genes on the complementary strand.