Real-Time Monitoring of Industrial Biofilms using Confocal Raman Microscopy and Multivariate Analysis

Biofilm development in industrial settings can prove costly to manufacturing and consumer health. The presence of contaminants in raw materials, finished products and on process contact surfaces negatively impacts on product quality and safety. Rapid and accurate identification of spoilage and pathogenic microorganisms is crucial to implement effective biofilm control strategies that enhance product safety. The application of confocal Raman microscopy (CRM) for non-invasive and rapid characterisation of clinical and food isolates has been reported. The question remains whether the technique can be used as an online monitoring tool for real-time measurement of biofilm build-up in dynamic manufacturing conditions.

In this study we investigated if CRM could be used in the manufacturing environment as an alternative microbiological quality control method. We assessed if this technology is able to differentiate between bacterial species and their growth phenotype, as well as detect contaminants from process samples.

Laboratory and industrial isolates grown under different culture conditions (planktonic, agar plates, and CDC grown biofilms), and formulated products were analysed using a confocal Raman microscope (Thermofisher DXR2xi) under optimised settings. Reference and experimental Raman spectra were collected and analysed for all test conditions [1]. Spectral similarities were evaluated by developing a microbial multivariate predictive model using a two-way orthogonal partial least squares (O2PLS) regression for cluster analysis [2].

Optimal spectra for microbes were obtained in the fingerprint region at approximately 600 - 1800 cm^-1 where characteristic peaks could be assigned to different biological macromolecules (nucleic acids, proteins, lipids and carbohydrates). Cluster analysis showed good group separation with low variation within but high variation between bacterial strains, enabling bacterial differentiation. It also highlighted the variations observed in the spectral fingerprint for planktonic, agar and biofilm growth modes. Comparative studies suggest that the peak associated with nucleotide ring stretching (~ 700 cm^-1) could be used as a microbial marker for contamination in formulation.

Our findings indicate that confocal Raman microscopy can be used for at-line monitoring of contamination in product streams. Raman spectra provide biochemical data for microbial characterisation but variations in the spectra are often difficult to observe and interpret. Multivariate statistical methods permit rapid interrogation of spectral data, with the potential to improve microbial identification. In combination with multivariate analysis, CRM can be used as an analytical tool for rapid identification of industrial isolates and differentiation of their growth phenotype.

 

References

[1] Beier, B.D., R.G. Quivey, and A.J. Berger,Raman microspectroscopy for species identification and mapping within bacterial biofilms.AMB Express, 2012. 2(1): p. 35-35.

[2] Zou, X., et al., Automatic Spectroscopic Data Categorization by Clustering Analysis (ASCLAN): A Data-Driven Approach for Distinguishing Discriminatory Metabolites for Phenotypic Subclasses.Analytical Chemistry, 2016. 88(11): p. 5670-5679.