Subspecies typing of Streptococcus agalactiae based on ribosomal subunit protein mass variation by MADLI-TOF MS

J. Rothen, J. F. Pothier, F. Foucault, J. Blom, D. Nanayakkara, C. Li, M. Ip, M. Tanner, G. Vogel, V. Pflüger, C. Daubenberger
2019
A ribosomal subunit protein (rsp)-based matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) method was developed for fast subspecies-level typing of Streptococcus agalactiae (Group B Streptococcus, GBS), a major cause of neonatal sepsis and meningitis. Methods: A total of 796 GBS whole genome sequences, covering the genetic diversity of the global GBS population, were used to in silico predict molecular mass variability of 28 rsp and to identify unique
more » ... mass combinations, termed "rsp-profiles". The in silico established GBS typing scheme was validated by MALDI-TOF MS analysis of GBS isolates at two independent research sites in Europe and South East Asia. Results: We identified in silico 62 rsp-profiles, with the majority (>80%) of the 796 GBS isolates displaying one of the six rsp-profiles 1-6. These dominant rsp-profiles classify GBS strains in high concordance with the core-genome based phylogenetic clustering. Validation of our approach by in-house MALDI-TOF MS analysis of 248 GBS isolates and external analysis of 8 GBS isolates showed that across different laboratories and MALDI-TOF MS platforms, the 28 rsp were detected reliably in the mass spectra, allowing assignment of clinical isolates to rsp-profiles at high sensitivity (99%) and specificity (97%). Our approach distinguishes the major phylogenetic GBS genotypes, identifies hyper-virulent strains, predicts the probable capsular serotype and surface protein variants and distinguishes between GBS genotypes of human and animal origin. Conclusion: We combine the information depth of whole genome sequences with the highly cost efficient, rapid and robust MALDI-TOF MS approach facilitating highthroughput, inter-laboratory, large-scale GBS epidemiological and clinical studies based on pre-defined rsp-profiles.
doi:10.5451/unibas-ep69789 fatcat:cyl6yc7fkragfcfgjekx4gsmju