Fourier transform infrared (FT-IR) spectroscopy is not only an efficient method for the quantitative and qualitative analysis of chemical substances but has also been introduced as a tool for classification and identification of microorganisms. In this thesis, the application of FT-IR spectroscopy to microbial analysis is illuminated under three aspects: (i) establishment and validation of a spectral database for the identification of coryneform bacteria and related taxa, (ii) assessment of the infraspecific diversity of three coryneform species, and (iii) the quantitative analysis of mixed microbial populations. (i) In order to allow the rapid, simple and cost-efficient identification of coryneform bacteria and related taxa from different habitats, an extensive FT-IR reference database was established, comprising bacteria from the two suborders Micrococcineae and Corynebacterineae (Actinomycetales, Actinobacteria) as well as other morphologically similar genera. The spectral library consists of averaged infrared spectra from 730 reference strains, covering 220 different species out of 46 genera. 192 species are represented by type strains. The identity of 352 reference strains was determined by comparative 16S rDNA sequence analysis and, if necessary, strains were reclassified accordingly. FT-IR frequency ranges, weights and reproducibility levels were optimized for this section of high G+C gram positive bacteria. In an internal validation, 98.1 % of the strains were correctly identified at the species level. An external validation which was carried out using 544 strains from 54 species out of 16 genera resulted in a correct identification of 87.3 % at the species level and 95.4 % at the genus level. The performance of this identification system is well within the range of those having been reported in the literature for the identification of coryneform bacteria by phenotypical methods. Apparently, coryneform bacteria and related taxa display a certain degree of overlap between different taxonomical markers, leading to a limited differentiation capacity of non-genotypical identification methods in general. However, easy handling, rapid identification within 25 h starting from a single colony, a satisfactory differentiation capacity and low cost render FT-IR technology clearly superior over other routine methods for the identification of coryneform bacteria and related taxa. (ii) The infraspecific diversity of 23 strains of Brevibacterium linens, 25 strains of Corynebacterium glutamicum and 26 strains of Rhodococcus erythropolis was determined by pairwise partial 16S rDNA sequence analysis and FT-IR spectroscopy in parallel. The infraspecific comparison of FT-IR spectral distances suggested that R. erythropolis and C. glutamicum strains are less diverse than B. linens, whereas at the level of pairwise 16S rDNA similarity, R. erythropolis strains had a very low diversity, B. linens displayed a high variability and C. glutamicum might have an intermediate position. Therefore, no correlation of FT-IR spectral similarity and 16S rDNA sequence similarity below the species level was observed, i.e. strains belonging to the same species which have only a low spectral distance between each other might not display an equivalent sequence similarity in their 16S rDNA and vice versa. Hence, diversification of 16S rDNA sequences and microevolutionary change of the cellular components detected by FT-IR spectroscopy appear to be de-coupled. In total, 27 strains derived from international culture collections carried invalidly described or wrong species denotations and had to be reclassified in accordance with polyphasic taxonomical analyses. It was found that FT-IR spectroscopy is a rapid and reliable method to screen for similar isolates. (iii) Furthermore, FT-IR spectroscopy was used for the first time to determine the ratios of different microorganisms in mixtures. Exemplarily, systems composed of two food-associated yeast species (Saccharomyces cerevisiae / Hanseniaspora uvarum) and two yoghurt lactic acid bacteria (Lactobacillus acidophilus / Streptococcus salivarius ssp. thermophilus) were investigated. Determination of the cell number ratio in the lactic acid bacteria system was possible with a minimal prediction accuracy of ± 16 ratio percentage points while the minimum accuracy of prediction in the yeast two-component system was ± 4 % (both at a 95 % confidence level). These results show that FT-IR spectroscopy is potentially a rapid method for the quantification of cell ratios in mixtures of two different microorganisms, provided that the cell ratio does not drop below a certain, system-specific threshold.
«
Fourier transform infrared (FT-IR) spectroscopy is not only an efficient method for the quantitative and qualitative analysis of chemical substances but has also been introduced as a tool for classification and identification of microorganisms. In this thesis, the application of FT-IR spectroscopy to microbial analysis is illuminated under three aspects: (i) establishment and validation of a spectral database for the identification of coryneform bacteria and related taxa, (ii) assessment of the...
»