Identification de micro-organismes par spectrometrie de masse pyrolytique.

摘要

Traditional techniques currently being used to identify microorganisms require more than a day to produce results, a delay that we are forced to accept because of the lack of faster alternative methods. However, recent developments in the fields of mass spectrometry and computer science now allow us to consider new approaches to this problem. Analytical fingerprinting techniques using mass spectrometry have a great potential to reduce the time required for analysis to just minutes. The analytical fingerprints used in this work reflect the global chemical composition of the sample. They were generated by analytical pyrolysis, a method that volatizes the sample to allow ionization. Recorded by a mass spectrometer, these fingerprints can then be compared to those of reference organisms in order to achieve identification.;The Py-MAB-TOF is a mass spectrometer designed to optimize each step involved in this identification process. The instrument features a pyrolysis unit, a MAB ion source and a time-of-flight mass analyzer, three technologies that, in our opinion, are key to the rapid analysis of microorganisms. Various parameters were optimized during the experimental procedure, which involved three different instruments. The preliminary experiments were conducted on a magnetic sector mass spectrometer, to which a MAB ion source and a house-built pyrolysis module was fitted. That initial setup eventually evolved to a prototype instrument that used a TOF analyzer and a commercial pyrolysis probe, before the final instrument was built.;The potential of the technique was first assessed and the main factors susceptible to interfere with the identification process were verified. A large number of bacteria and yeasts of various genera, species and strains were correctly identified using the Py-MAB-TOF. The age of the colonies, their conservation time and the nature of the media they were grown on are the main factors that may induce some variability in the analytical fingerprints. In general, the magnitude of those effects does not greatly impede the identification process. In the cases where the induced variation is problematic, corrective measures can be implemented.;Following the optimization of the instrumental parameters and methods, the applicability of the final instrument was evaluated. It was found that it is possible to detect bacteria at a concentration of 105 cfu/ml in aqueous suspensions and at a concentration of 106 cfu/ml in urine samples. The Py-MAB-TOF has also been successfully used to identify simulants of biological warfare agents. Those samples, closely related to different types of bacteria, toxin and viruses, susceptible to be used as biological warfare agents produced characteristic fingerprints that are easily distinguishable by visual inspection. Finally, a library search software was specially designed for microorganism identification. By comparing the analytic fingerprints to those in a mass spectral database, the software suggests a list of possible identities for the unknowns in descending order of likelihood, allowing minimally-trained personnel to use the system.