Porous Silicon-Based Biosensors: Towards Real-Time Optical Detection of Target Bacteria in the Food Industry

Abstract

Rapid detection of target bacteria is crucial to provide a safe food supply and to prevent foodborne diseases. Herein, we present an optical biosensor for identification and quantification of Escherichia coli (E. coli, used as a model indicator bacteria species) in complex food industry process water. The biosensor is based on a nanostructured, oxidized porous silicon (PSi) thin film which is functionalized with specific antibodies against E. coli. The biosensors were exposed to water samples collected directly from process lines of fresh-cut produce and their reflectivity spectra were collected in real time. Process water were characterized by complex natural micro-flora (microbial load of >10⁷ cell/mL), in addition to soil particles and plant cell debris. We show that process water spiked with culture-grown E. coli, induces robust and predictable changes in the thin-film optical interference spectrum of the biosensor. The latter is ascribed to highly specific capture of the target cells onto the biosensor surface, as confirmed by real-time polymerase chain reaction (PCR). The biosensors were capable of selectively identifying and quantifying the target cells, while the target cell concentration is orders of magnitude lower than that of other bacterial species, without any pre-enrichment or prior processing steps.

Figure 1

(a) Water samples from a Dutch fresh-cut produce company were evaluated for E. coli presence by three different methodologies: culturing techniques (upper left), PCR-based methods (upper right), and label-free optical biosensors (bottom). Note that the time indicated for each method refers to the total assay time. (b) Specific capture probes (antibodies) immobilized onto the PSiO₂ surface function as the active component of the biosensor. After exposure of the biosensor to process water spiked with the target bacteria, the bacteria cells were directly captured onto the antibody-modified PSiO₂ surface. (c) Light reflected from the porous nanostructure provides the monitored optical signal. Changes in the light intensity are correlated to specific immobilization of the bacteria onto the surface. Upper panel: reflectivity spectra of a typical Fabry-Pérot PSiO₂ nanostructure before (blue) and after (red) bacteria capture. Lower panel: applying a fast Fourier transform (FFT) of the raw reflectivity spectrum results in a single peak whose magnitude is monitored.

Figure 2

(a) IS-Pro bacterial profiles and (b) the corresponding E. coli-specific medium culture (EMB agar) of (1) pure E. coli K-12 culture; (2) water samples before spiking with E. coli K-12; (3) water samples after spiking with 10⁵ cells/mL E. coli K-12. Peak length, expressed in nucleotides, corresponds to IS-fragment length. Peak height, expressed as intensity, reflects quantity of fragments. The blue peaks represent Firmicutes, pink peaks represent Bacteroidetes, and yellow peaks represent Proteobacteria. The E. coli-specific peaks are indicated by arrows and amplicon length.

Figure 3

(a) A top-view high-resolution scanning electron microscope (HRSEM) image of a typical PSiO₂ film demonstrating the porous nanostructure morphology with typical pores in the range of 60–100 nm. (b) Schematic illustration of the synthesis steps for the biofunctionalizion of PSiO₂ with IgG. (I) PSiO₂ was reacted with APTES and catalyzed by an organic base to create an amine-terminated surface. (II) The amine-terminated PSiO₂ was reacted with one of the aldehyde groups of the cross-linker GluAld. (III) Grafting of SA onto the aldehyde-terminated surface. (IV) Biotinylated-IgG (E. coli) was conjugated via biotin-SA binding.

Figure 4. The biosensing concept.

(a) Specific capture probes (antibodies) were immobilized onto the porous surface to provide the active component of the biosensor. (b) Next, the biosensor was exposed to the target bacteria in order to directly capture the bacteria cells onto the antibody-modified PSiO₂ surface. (c) A drop in the intensity of the thin-film optical interference spectrum of the biosensor results from bacteria capture. (d) Microscopy tools (light microscope and HRSEM) and real-time PCR methods were used to confirm the presence of bacteria on the biosensor surface.

Figure 5. Representative biosensing experiments (left) and the corresponding HRSEM images (right) of the biosensors immediately after the experiments.

(a) Spiked process water (10⁴ cell/mL E. coli). (b) Control - original process water (no E. coli). (c) The corresponding HRSEM images (in two different magnifications) of the biosensor after a biosensing experiment with spiked process water, demonstrating bacteria capture. The inset presents enlargement of a captured bacterium on the biosensor surface. (d) A corresponding HRSEM image of the biosensor after the control experiment (original process water, no E. coli) showing a negligible amount of cells. The inset presents enlargement of the biosensor surface.

Figure 6. Averaged intensity changes of the biosensor upon introduction to process water spiked with different concentrations of E. coli bacteria (ranging from 10³ to 10⁵ cells/mL).

For the control experiments, the biosensors were incubated with the original process water (no E. coli). The incubation time was set to 15 min, after which the samples were washed with a buffer solution for 30 min (n ≥ 3 for each concentration), *Significantly different (t-test, p < 0.05).