Quantification and viability assessment of waterborne bacteria using Raman spectroscopy

The two objectives of this work were (1) to explore the potential of Raman spectroscopy as a tool for quantification of bacteria in water and (2) to evaluate its suitability to detect cell viability.; The intensity of the Raman response correlated with number of cells present in drops of sample water on aluminum-coated slides. However, concentrations of pathogens in environmental waters are low, requiring a concentration step (i.e. filtering). Subsequent evaluation of filtering approaches for Raman detection showed significant background signal from the membranes; though there was a linear correlation between Raman Spectroscopy and other quantification methods at high bacterial concentrations.; Secondly, detection of naturally occurring and induced metabolic changes related with loss of viability in bacteria was considered. Natural metabolic variability expected during growth phases in batch cultures of Escherichia coli and Staphylococcus epidermidis were studied by Raman spectroscopy. Spectra were analyzed by Principal Component Analysis and Discriminant Analysis to identify similarities and differences and to estimate a classification performance based on growth phases. The average percentage correct classification based on growth phase was 73 % for E.coli (four growth cycles) and 68% for S.epidermidis (two growth cycles). The ability of the model to discriminate between species (E.coli versus S.epidermidis) when all the growth phases were considered was 78%.; Finally, Raman spectroscopy was applied to study Escherichia coli and Staphylococcus epidermidis cells that were inactivated by different chemicals and stress conditions. E.coli cells exposed to starvation conditions over several days lost viability at the same rate spectral bands assigned to DNA and RNA bases decreased in intensity. Band intensity correlated with Standard Plate Counts (R2=0.99 and R2=0.97). Significant changes were observed in the spectra of treated cells in comparison with their respective controls (samples without treatment). Discriminant Analysis was used along with PCA to estimate a classification rate based on viability status of the cells. The classification rate obtained considering all the treatments (non-viable cells) and controls (viable cells) for each of the species studied was 86%. The classification rate based on species differentiation when all the spectra (viable and non-viable) were used was 87%.