Raman spectra identify vancomycin-resistant phenotypes and their transcriptomic features in Staphylococcus aureus

Staphylococcus aureus is a pathogenic bacterium that has been linked to multiple epidemics of novel antibiotic resistance. Vancomycin is the first-line antibiotic treatment for methicillin-resistant S. aureus (MRSA) infections; however, various vancomycin-non-susceptible MRSA strains have emerged globally. Rapid assessment of the resistance levels and the underlying molecular profiles is crucial for preventing their spread and for counteracting resistance. Nevertheless, a broad resistance spectrum and diverse genetic alterations have hindered these efforts. Here, we demonstrated that vancomycin resistance levels of various MRSA strains can be determined using dimension-reduced Raman spectra obtained from single cells. Furthermore, transcriptome profiles of different strains can be predicted from these spectra using simple linear regression. This Raman-transcriptome correspondence enables mapping of transcriptomic components onto a dimension-reduced Raman space, allowing for characterization of gene groups associated with different phenotypes. Additionally, single-cell Raman spectroscopy has successfully identified a strain with higher phenotypic heterogeneity, which was subsequently validated by single-cell growth analysis. Overall, our results demonstrate the efficacy of Raman spectroscopy for identifying resistant phenotypes, their associated features for gene expression, and intrapopulation phenotypic heterogeneity.

The spread of antibiotic-resistant bacteria threatens human health worldwide. Rapid identification of antibiotic resistance and molecular-level features of an infected pathogen is crucial for treating patients and counteracting the emergence and spread of new resistant bacteria. Here, we demonstrate that Raman spectra, measurable by exposing individual cells to a focused laser, can be used to infer the resistance phenotypes and global gene expression profiles of different Staphylococcus aureus strains. Furthermore, we present a method to visualize transcriptome components using the major axes of the dimension-reduced Raman space linked to important phenotypic features, such as resistance and virulence. This method identifies several gene sets related to crucial phenotypic features of pathogenic bacteria. Our results demonstrate a novel application of spectroscopy in microbiology by enabling the simultaneous identification of resistant and virulent phenotypes, as well as the underlying gene expression features.