Ultrasensitive multiplex optical quantification of bacteria in large samples of biofluids

Abstract

Efficient treatments in bacterial infections require the fast and accurate recognition of pathogens, with concentrations as low as one per milliliter in the case of septicemia. Detecting and quantifying bacteria in such low concentrations is challenging and typically demands cultures of large samples of blood (~1 milliliter) extending over 24-72 hours. This delay seriously compromises the health of patients. Here we demonstrate a fast microorganism optical detection system for the exhaustive identification and quantification of pathogens in volumes of biofluids with clinical relevance (~1 milliliter) in minutes. We drive each type of bacteria to accumulate antibody functionalized SERS-labelled silver nanoparticles. Particle aggregation on the bacteria membranes renders dense arrays of inter-particle gaps in which the Raman signal is exponentially amplified by several orders of magnitude relative to the dispersed particles. This enables a multiplex identification of the microorganisms through the molecule-specific spectral fingerprints.

Figure 1. Microorganism optical detection system (MODS).

Conceptual view of MODS and its relevant components. Silver nanoparticles (NPs) are separately labelled with different Raman-active molecules and functionalized with bacteria-selective antibodies (1). A nanoparticle dispersion is mixed in a vessel (3 mL) with the sample fluid, possibly infected (2). Several types of bacteria are targeted using NPs prepared with different specific combinations of Raman molecules and antibodies. The presence of one of these microorganisms induces aggregation of antibody-matching NPs on its membrane, rapidly evolving towards full random coverage (3). The mixture is circulated through a millifluidic channel with a micropump (4) and passing through the focus of a 785 nm laser (5), which is in turn spectrally analysed to record the SERS signal generated by the Raman-active molecules (6). Targeted bacteria produce a large increase in SERS signal, whose spectral fingerprints allow us to identify the type of pathogen.

Figure 2. Encoded nanoparticles and their interaction with bacteria.

(A) SERS spectra of the different coded particles here used for each targeted pathogen, along with their corresponding labelling molecules. (B) Transmission electron microscope images of the targeted bacteria (E. coli, P. aeruginosa, S. aureus, and S. agalactiae) coated with their respective matching NPs.

Figure 3. MODS performance for contaminated samples.

(A) Correlation between a temporal series of spectra collected over 270 ms intervals and the SERS reference of the labelled NPs (Fig. 1B). The analysed serum samples contain either one pathogen (1–4, see labels) or no pathogen (5, blank). Series 6 shows the result for a blood sample spiked with a combination of three different bacteria and concentrations (S. aureus, E. coli, and S. agalactiae). Large correlation values reveal the passage of an individual bacteria or CFU. (B) Bacterial cultures (24–48 hours) for the microorganism inoculated in the blood samples (series 6). White spots correspond to CFUs. (C) Comparison of the bacteria concentrations (CFUs per mL) as determined by MODS (open squares) for the sample contaminated with three pathogens (series 6) versus traditional cultures (open circles). Averages over three runs of both MODS and culture experiments are shown by the corresponding solid symbols.

Figure 4. Kinetics of sensing enhancement through NP aggregation.

(A) Kinetics of NP aggregation as measured through the time-dependent SERS signal (symbols) after adding E. coli to the mixture of coded NPs. Solid curve: theory from (B). (B) Simulation of the temporal evolution of 60 nm Ag NP aggregation on the bacteria membrane produced by random NP-membrane encounters, resulting in a rapidly growing NP density (right scale) and SERS intensity (left scale, calculated per μm² of membrane area and normalized to the signal from an individual NP). The latter is given with (solid curve) and without (broken curve) inclusion of the effect of NP gap hotspots. The time is normalized to the average delay interval between consecutive NP arrivals. The theory curve in A is scaled to 17.3 arrivals per second, as estimated from kinetic theory (see ESI). (C) Near-electric-field intensity in a rod-like individual E. coli covered with Ag NPs (top, intensity plotted on the NP surfaces) and detail of the array (bottom, intensity on a surface passing by the NP centers), revealing the formation of optical hotspots. The intensity is averaged over light incidence directions and polarizations, the light wavelength is 785 nm, and the colour scale is saturated to improve visibility.