Identifying Bacterial Pathogens in Wastewater Samples with Liquid Surface Enhanced Raman Spectroscopy

The ability to accurately identify and monitor bacterial pathogens in environmental samples such as rivers, streams, and wastewater is critical to public health. Our group has recently shown that Raman spectroscopy combined with machine learning (ML) can be a powerful tool for identifying bacterial species, strain, and antibiotic susceptibility [1, 2, 3]. Gold nanorods can enhance the Raman scattering signature, owing to their ability to concentrate the incident light and selectively bind to bacteria via electrostatic interactions. In wastewater, however, anions mask the surface charges of the nanorods, disrupting interactions with bacteria and weakening the Raman enhancement.<br/><br/>Here, we achieve bacterial surface enhanced Raman spectroscopy (SERS) in wastewater by conjugating 12-residue gold-binding peptides (GBPs) to the bacterial surface. These peptides interact with gold through van der Waals interactions, overcoming the need for positive surface charges on the nanorods, and form bonds with gold 10 times more stable than electrostatic interactions [4]. As a proof of concept, we design a plasmid encoding a GBP conjugated to a circularly permuted outer membrane protein X (CPX) scaffold and use this plasmid to express GBPs on the <i>E. coli</i> cell surface. We describe the effects of GBP surface expression on adsorption affinity of bacteria to nanorods. Next, we use filtered wastewater samples and spike in GBP-expressing <i>E. coli</i> at a concentration of 10⁷ cells/ml, a realistic total bacterial concentration for wastewater. With our surface expression of GBPs, SERS can detect <i>E. coli</i> at this concentration using only 10 second integration times in a liquid wastewater sample. Finally, to demonstrate broad application to any bacterial species, we conjugate synthetic GBPs to dibenzocyclooctyne (DBCO). These conjugates can be selectively cross-linked to azides via copper-free click chemistry, enabling GBP cross-linking to any bacteria that have been metabolically labeled with azide-conjugated cell wall and outer membrane components. Hence, by combining advances in nanophotonics, ML, and biological engineering, we demonstrate SERS can be used to detect a wide variety of bacterial species and strains in complex environmental samples, allowing rapid and label-free monitoring of a community’s bacterial infections through wastewater.<br/><br/>[1] Ho et al., Nat. Comm. 10, 4927 (2019).<br/>[2] Tadesse et al., Nano Lett. 20, 7655-7661 (2020).<br/>[3] Safir et al., ArXiv (2022).<br/>[4] Hnilova et al., Langmuir 24, 12440-12445 (2008).