Nanostructural Differentiation between Live and Dead Escherichia coli Cells Using FT-IR Spectroscopy and Comparison of Detection Limits to Quantitative PCR (qPCR)

Fourier-transform infrared spectroscopy (FT-IR) utilizes nanostructural differences between bacterial cells for the rapid identification, classification, and differentiation of many species of bacteria, but most studies have used only live cells. A rapid and reliable method for determining the total number of live and dead bacterial cells in a food could prevent the distribution of unsafe products to consumers. In this study, infrared spectra (4000-700 cm^-1) were collected using a Continuμm IR microscope attached to a ThermoNicolet Nexus 670 FT-IR spectrophotometer from varying concentrations of live and dead E. coli (ATTC^® 25922^™) cells. Spectra were analyzed using chemometric methods. Canonical variate analysis (CVA) and Mahalanobis distances (MD) quantified the spectral differences between 100% live and 100% dead bacteria, with linear discriminant analysis (LDA) using principal components able to successfully classify and differentiate between the two treatments (50 out of 50 correctly assigned). CVA and MD were also 100% successful in quantifying the spectral differences between 100% dead bacteria and samples containing as little as 0.1% live cells. The PLS method accurately quantified live cells in the presence of dead cells (R = 0.981), and Cooman plots showed clear separation between clusters of live and dead bacteria. To compare detection limits with qPCR, dead bacterial cells were treated with ethidium bromide monoazide (EMA) to prevent amplification of DNA from these cells and enable the detection of both live and dead cells using universal bacterial primer sets. The qPCR techniques using EMA-treated bacteria could also detect live cell concentrations as low as 0.1%. However, the total analysis time for detection using qPCR is ~6 hours in comparison to ~30 minutes using FT-IR. Therefore, FT-IR is a useful tool for the rapid differentiation between live and dead bacterial samples, even when the concen- - tration of live bacteria is as low as 0.1% of the target microbe.