Staphylococcus aureus causes a wide range of hospital infections. Often, these infections involve epidemic methicillin-resistant S. aureus (MRSA) strains that are transferred by health care workers to patients. In order to detect outbreaks that are caused by epidemic strains, the clinical isolates have to be typed. Multilocus sequence typing (MLST) relies on the sequence analysis of housekeeping genes and is used to allocate the strains to sequence types (ST), which can be grouped into clonal complexes (CC). This method has provided a detailed insight into the population structure of MRSA and methicillin-susceptible S. aureus (MSSA) strains (1). Finer discrimination is achieved by pulsed-field gel electrophoresis (PFGE) and spa typing (2). All of these methods need additional experimentation.In matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS), different spectra or signatures of cell extracts (3, 4, 5, 6) or whole cells (7,8,9,10,11,12) could be identified for different strains or groups of strains. An increasing number of laboratories use MALDI-TOF MS for the identification of S. aureus. However, the differences in the signatures of the strains have not been evaluated so far, because the discriminatory threshold of the software used in clinical settings is set up to assign the isolate to a species. To this end, more subtle differences are ignored. Another reason is that, so far, MALDI-TOF MS of whole bacterial cells has been employed in a heuristic manner, and for most species, the identities of the compounds that are detected in the measurements are unknown. Thus, the spectra are not well understood, and the variations in the signatures cannot be interpreted.In principle, two spectra might differ by signal intensity, loss of a signal, or by the shift of a signal. Variations in signal intensity are probably caused by expression differences, which might be directly correlated with culture conditions and, therefore, do not give unambiguous information about a genotype of the strain. The loss of a signal is caused by the total failure to express a protein or peptide, which in turn might indicate a mutation causing a frameshift or stop codon, but might also depend on culture conditions, mutations of regulatory factors, or sample preparation. Thus, there is no clear-cut correlation between the loss of a signal and a genotype. In contrast, peak shifts (i.e., loss of a signal coupled to the appearance of a new signal, both of which are correlated to the same peptide) correspond to point mutations in the genes of the peptides detected in the analysis; the mutation leads to an amino acid exchange that alters the molecular weight of the corresponding gene product.In order to characterize the clonal lineages of S. aureus in the MALDI-TOF MS, this work was aimed at the identification of the peptides that are detected in the spectra and that show mass variations between the clonal complexes of S. aureus. To this end, we analyzed the spectra of 401 S. aureus strains, concentrating on...