Destruction-free procedure for the isolation of bacteria from sputum samples for Raman spectroscopic analysis

作者：Kloss Sandra; Lorenz Bjoern; Dees Stefan; Labugger Ines; Roesch Petra; Popp Juergen^*

来源：Analytical and Bioanalytical Chemistry, 2015, 407(27): 8333-8341.

DOI：10.1007/s00216-015-8743-x

• Raman spectroscopy
• Single-cell pathogen identification
• Sputum
• Isolation

Lower respiratory tract infections are the fourth leading cause of death worldwide. Here, a timely identification of the causing pathogens is crucial to the success of the treatment. Raman spectroscopy allows for quick identification of bacterial cells without the need for time-consuming cultivation steps, which is the current gold standard to detect pathogens. However, before Raman spectroscopy can be used to identify pathogens, they have to be isolated from the sample matrix, i.e., sputum in case of lower respiratory tract infections. In this study, we report an isolation protocol for single bacterial cells from sputum samples for Raman spectroscopic identification. Prior to the isolation, a liquefaction step using the proteolytic enzyme mixture Pronase E is required in order to deal with the high viscosity of sputum. The extraction of the bacteria was subsequently performed via different filtration and centrifugation steps, whereby isolation ratios between 46 and 57 % were achieved for sputa spiked with 6 center dot 10(7) to 6 center dot 10(4) CFU/mL of Staphylococcus aureus. The compatibility of such a liquefaction and isolation procedure towards a Raman spectroscopic classification was shown for five different model species, namely S. aureus, Staphylococcus epidermidis, Streptococcus pneumoniae, Klebsiella pneumoniae, and Pseudomonas aeruginosa. A classification of single-cell Raman spectra of these five species with an accuracy of 98.5 % could be achieved on the basis of a principal component analysis (PCA) followed by a linear discriminant analysis (LDA). These classification results could be validated with an independent test dataset, where 97.4 % of all spectra were identified correctly.

• 出版日期2015-11