| Resumo: | Fungal infections have greatly increased in risk populations, namely in immunocompromised patients, probabily because the diagnosis of fungal infections is delayed. Microbial metabolomics arises as a powerful feature screening the metabolites produced by microorganisms. It provides information regarding the state of biological organisms which can be used as a diagnostic tool for diseases through fungal metabolites pattern. Thus, this research aimed to in-depth study of the Aspergillus niger exometabolome, in order to establish a targeted metabolomic pattern that characterizes A. niger. A methodology based on headspace-solid phase microextraction combined with comprehensive two-dimensional gas chromatography coupled to mass spectrometry with a high resolution time of flight analyser (HS-SPME/GC×GC-ToFMS) was used. A. niger exometabolome was analysed in different growth conditions: temperature (25 and 37 °C), incubation time (3 and 5 days), and culture medium (solid and liquid medium). A. niger exometabolome included 430 metabolites, distributed over several chemical families, being the major ones alcohols, aldehydes, esters, hydrocarbons, ketones and terpenoids. Differences among volatile metabolites produced under different growth conditions were observed, being the major relative abundance determined for 5 days of growth, at 25 °C, using solid medium. These results indicated the high complexity of A. niger exometabolome. A subset of 44 metabolites, which were present in all previously tested growth conditions, was defined as the A. niger targeted metabolomic pattern. This pattern may be used in detection of fungal infections by this specie and be further exploited to fungal infections diagnosis. Furthermore, this subset of metabolites was compared with samples of Candida albicans (yeast) and Penicillium chrysogenum (filamentous fungi), and Partial Least Squares Discriminant Analysis (PLS-DA) was applied. The results clearly showed that this metabolites subset allowed the distinction between these microorganisms. In order to validate the PLS-DA model, permutation test was applied, and a statistically significant model for 44 metabolites was obtained with a predictive Q2 capability of 0.70 for A. niger. When the subset of compounds were reduced to 16 (obtained by Variables Importance in Projection (VIP) parameter), the obtained model had a predictive Q2 capability of 0.86 for A. niger, which was significantly higher, being more robust than the previous. The decrease of 44 to 16 metabolites, reduced the require analysis time and the conditions used were similar to the conditions used in clinical context, (solid medium, at 25 °C and ca. 1 week). However, in this study was possible to reduce the time for 3 days. In conclusion, these 44 volatile molecular biomarkers could be useful for diagnosis of fungal infections, and they can even be further exploited in clinical context.
|
|---|