Surface-Enhanced Raman Spectroscopy of Bacterial Metabolites for Bacterial Growth Monitoring and Diagnosis of Viral Infection

Abstract

Bacterial metabolites are intermediate products of bacterial metabolism and their production reflects metabolic activity. Herein, we report the use of surface-enhanced Raman spectroscopy (SERS) for detection of both volatile and nonvolatile metabolites and the application of this approach for bacterial growth quantification and diagnosis of viral infection. The timedependent SERS signal of the volatile metabolite dimethyl disulfide in the headspace above bacteria growing on an agar plate was detected and quantified. In addition, SERS signals arising from the plate reflected nutrient consumption and production of nonvolatile metabolites. The measurement of metabolite accumulation can be used for bacterial quantification. In the presence of bacteriophage virus, bacterial metabolism is suppressed, and the relative decrease in SERS intensity reflects the initial virus concentration. Using multivariate analysis, we detect viral infection with a prediction accuracy of 93%. Our SERS-based approach for metabolite production monitoring provides new insights toward viral infection diagnosis.

Investigation of the Influence of Stress on Label-Free Bacterial Surface-Enhanced Raman Spectra

“…These results indicate that the SERS intensity of 733 cm –1 (adenine), 1030 cm –1 (phenylalanine), and 1330 cm –1 (typical phospholipids) decreases with time, thus reflecting cellular consumption. For example, adenine containing molecules can be consumed as nutrients during bacterial cell growth and lead to the decrease in the SERS intensity at 733 cm –1 . In contrast, positive signs for the remaining cross peaks suggest an increase in the SERS signal due biomolecule release.…”

“…The raw spectral data were pre-processed by the instrument-embedded software (Project Five) for cosmic ray removal, Savitzky–Golay smoothing, and baseline subtraction (asymmetric least-squares smoothing method). Subsequently, the SERS data were normalized using the localized intrinsic internal standard, the pseudo plasmon-enhanced electronic Raman scattering (ERS) peak at ∼77 cm –1 ( I 77 ), to minimize signal variability arising from spatial hot spot heterogeneity. − The ERS peak is enhanced within the hot spots along with the analyte and has been applied as an internal standard for spatial and temporal signal normalization. ,, …”

“…A major challenge in label-free SERS bacterial detection is the lack of standardized sample handing procedures, which results in the poor reproducibility of the technique. ,, Label-free SERS spectra reflect signals that arise from both bacterial cells and metabolites − and as such the spectra are affected by the bacterial growth state and metabolite production. Common bacterial growth parameters (e.g., nutrient components and availability, temperature, oxygen, pH, and bacterial growth phase) − and external stress stimuli (e.g., antibiotic exposure, phage infection, UV irradiation, and the presence of environmental contaminants) − can greatly affect the label-free SERS/Raman features of bacterial cells or metabolites. To minimize the effect of such parameters and to acquire stable SERS signals, bacterial growth conditions should ideally be maintained as consistently as possible.…”

“…Additionally, bacterial samples are sometimes held at room temperature or an optimal growth temperature for some time during label-free SERS/Raman probing to evaluate the effects of external environmental stimuli (antibiotic treatment, viral infection, nutrient availability, etc.) on bacteria cells ,,, and the SERS spectral changes following stimuli are then compared. However, the active metabolism of bacterial cells stored at room/optimal growth temperature can potentially result in cellular changes, even in the absence of external stimuli, and may result in label-free SERS signal alterations.…”

Biomimetic Transparent Nanoplasmonic Meshes by Reverse‐Nanoimprinting for Bio‐Interfaced Spatiotemporal Multimodal SERS Bioanalysis

“…This normalization method had previously been applied for various biological applications. [19,28,52] Data truncation was performed using MATLAB. Lastly, R software was used for performing PCA and LDA.…”