A bacteriophage detection tool for viability assessment of Salmonella cells.

TL;DR: This work presents and validates a novel bacteriophage (phage)-based microbial detection tool to detect and assess Salmonella viability and shows the phage selectivity in cell recognition minimizes false-negative and false-positive results often associated with most detection methods.

Abstract: Salmonellosis, one of the most common food and water-borne diseases, has a major global health and economic impact. Salmonella cells present high infection rates, persistence over inauspicious conditions and the potential to preserve virulence in dormant states when cells are viable but non-culturable (VBNC). These facts are challenging for current detection methods. Culture methods lack the capacity to detect VBNC cells, while biomolecular methods (e.g. DNA- or protein-based) hardly distinguish between dead innocuous cells and their viable lethal counterparts. This work presents and validates a novel bacteriophage (phage)-based microbial detection tool to detect and assess Salmonella viability. Salmonella Enteritidis cells in a VBNC physiological state were evaluated by cell culture, flow-cytometry and epifluorescence microscopy, and further assayed with a biosensor platform. Free PVP-SE1 phages in solution showed the ability to recognize VBNC cells, with no lysis induction, in contrast to the minor recognition of heat-killed cells. This ability was confirmed for immobilized phages on gold surfaces, where the phage detection signal follows the same trend of the concentration of viable plus VBNC cells in the sample. The phage probe was then tested in a magnetoresistive biosensor platform allowing the quantitative detection and discrimination of viable and VBNC cells from dead cells, with high sensitivity. Signals arising from 3 to 4 cells per sensor were recorded. In comparison to a polyclonal antibody that does not distinguish viable from dead cells, the phage selectivity in cell recognition minimizes false-negative and false-positive results often associated with most detection methods.

Summary

1. Introduction

• The ingestion of food, its derivatives and water contaminated with microbial pathogens (e.g. Escherichia coli, Campylobacter sp. or Salmonella sp.) is responsible for about 2.2 million deaths annually.
• “Dormant” bacteria have therefore been called viable but non-culturable (VBNC) cells.
• Significant progress has been reported in the phage-based detection of foodborne and waterborne pathogens (Hagens and Loessner, 2007; Singh et al., 2012; Smartt et al., 2012).

2.2. Bacteriophages and bacterial strains

• PVP-SE1 was isolated from a Regensburg wastewater plant in the context of a European Project (Phagevet-P).
• Salmonella Enteritidis strain S1400 was used as host (Sillankorva et al., 2010).
• Campylobacter coli phage vB_CcoM-IBB_35, isolated from poultry intestines, was used as negative control (Carvalho et al., 2010a).

2.3. Phage propagation and buffer exchange

• The phages were produced using the double layer agar technique as described by Sambrook and Russell (2001) and resuspended in SM buffer.
• Exchange of SM buffer by MOPS buffer was needed to avoid the presence of amine groups from SM buffer, which may interfere with the surface chemistry adopted for phage immobilization on solid substrates.
• Buffer exchange was made using a Vivaspin 500 centrifugal concentrator (MW 100 kDa).
• Following the buffer exchange the concentration of phage was verified using the double layer agar technique.

2.4. Induction of Salmonella into viable but non-culturable (VBNC) state

• Bacteria were induced to enter the VBNC state by using sodium hypochlorite (commercial bleach—stock concentration 5%) at different concentrations.
• The serial dilutions of bleach were done with milli-Q water.
• The samples were mixed at 200 rpm for 1 min at room temperature.
• Following chlorination, the suspensions were centrifuged at 3420xg for 10 min at 4 1C and washed twice with cold PB.

2.5. Determination of cell viability

• Cell viability was assessed after submitting bacteria to different bleach concentrations using the LIVE/DEADs BacLight™ Bacterial Viability and Counting Kit (Molecular Probes).
• SYTO9 and PI dyes were used, accordingly to manufacturer's instructions.
• Upon staining, cells were analyzed either by epifluorescence microscopy (OLYMPUS BX51 EXTREMO microscope) or by flow cytometry (BD LSRII flow cytometer using FACS DIVA software for acquisition; BD Biosciences).
• For absolute cell quantification, 6 μm diameter microspheres were used at a known concentration in the flow cytometry acquisition.
• Flow cytometry data was analyzed using the FlowJo software (Tree Star, Ashland, OR).

2.6. Phage lysis time and adsorption studies

• 1 mL of each Salmonella sample was infected with PVP-SE1 phage at a multiplicity of infection (MOI) of 0.001, which refers to the number of phages that were added per cell.
• Samples were taken immediately after infection (time 0) and after 20 min and 40 min of phage inoculation, followed by 10-fold dilution in MOPS and centrifugation at 10,000g for 10 min.
• The supernatant was 10-fold serially-diluted in MOPS and plated to assess the concentration of PFU (plaque forming unit).
• The phage adsorption fraction was calculated by dividing the PFU concentration at each time point by the initial phage concentration.
• To assess the phage lysis time viable exponential phase grown Salmonella cells were used.

2.7. Phage immobilization on Au surfaces

• Cr 5 nm/Au 40 nm thin film layers were sputtered (Kenosistec sputtering tool) over a silicon wafer.
• The wafer was then spincoated with a photoresist (PR) polymer (AZ1505 AZ Electronic Materials) for surface protection and diced in 7 7 mm2 dies using an automatic dicing saw (Disco, DAD3350).
• Substrates were then rinsed with isopropanol (IPA) and milli-Q water and dried under a nitrogen stream.
• The gold surface was then functionalized with a heterobifunctional linker, the sulfo-LC-SPDP (sulfosuccinimidyl 6-[3′-(2-pyridyldithio)-propionamido] hexanoate).
• Spot pictures were taken with an optical stereomicroscope (Nikon SMZ 1500) equipped with a CCD camera and analyzed using the image processing software ImageJ.

2.8. MR-biochip measurement

• The MR-biochip was produced at INESC MN through a dedicated microfabrication process (Martins et al., 2009) and wirebonded to a PCB chip-carrier.
• The probe sites on the MR biochip terminate with exposed Cr/Au pads, underneath which lie the magnetoresistive sensors that will detect the magnetic nanoparticle labels.
• Briefly, the MR chip architecture comprises two distinct sensing areas arranged in two columns.
• A 1 mL droplet of Salmonella-specific phage was spotted over the left column of sensors (12 sensors) and a non-specific phage (Campylobacter phage) on the right column of sensors (12 sensors).
• The difference between the signal acquired after washing and the baseline signal is proportional to the number of cells bound to the sensor surface.

2.9. Antibody-conjugated MNPs preparation

• Commercial 250 nm Protein A modified MNPs (Nanomag, Micromod) were used.
• The unbound antibody was removed by the same magnetic separation procedure.
• The functionalized MNPs were finally resuspended in 5 mL of PB Tw20 and injected over the chip.

2.10. Statistical analysis

• All data are represented as mean7SD (standard deviation).
• For Figs. 2 and 3, means were compared using two-way ANOVA followed by the Bonferroni post hoc test.

3.1. Induction of VBNC physiological state in Salmonella

• Since the goal of this work was to prove the phage ability to detect the VBNC state of bacterial cells, a process was first developed capable of affecting cell viability in a controlled manner that would not lead to killing or lysing the entire cell population.
• For this purpose different bactericidal and bacteriostatic compounds, known to induce the VBNC state in Salmonella cells, were tested (data not shown).
• When exposed to fresh liquid medium under adequate growth conditions all tested concentrations of bleach, even above the break-point, showed cell growth (Supplementary data, Fig. S2.1).
• In order to quantitatively determine the relative and absolute proportion of the different cell populations (classified as live, dead or compromised), flow cytometry analyses were conducted for the different cell samples (Fig. 1A bars and 1C).
• Results confirmed that, despite being present in sub-optimal host infection conditions, the phage adsorption capability was conserved, maintaining its potential to be used as a detection tool.

3.3. Phage performance as a biorecognition element

• After optimization of the surface chemistry (Supplementary data, Fig. S3.1 and S3.2), the phage was immobilized on an Au surface at discrete areas by manual spotting.
• Also according to phage adsorption rates in solution, the immobilized phages were able to discriminate between viable and dead cells.
• This resulted in reduced cell densities for samples with increasing number of dead cells (Fig. 3A) but proportional to the relative concentration of viable plus VBNC cells (compromised population) obtained by flow cytometry analysis (Fig. 1A—bars plot).
• Identical biorecognition elements may hinder each other's proper attachment.
• This is a common scenario in standard immunoassays where a labeling antibody may block the epitopes to the capture antibody or vice versa.

3.4. Phage-based magnetoresistive biochip for cell viability assessment

• The feasibility of developing a “sandwich” phage-based biosensing system and its potential as a cell viability determination tool was assessed making use of an existent magnetoresistive (MR) biochip (Freitas et al., 2012; Martins et al., 2009, 2010) and respective electronic reader (Germano et al., 2009).
• The biomolecular recognition strategy used on the biochip combines the phage and a magnetically-labeled antibody as recognition and labeling elements, respectively.
• After the functionalization of the biochip with PVP-SE1 bacteriophage, each cell solution was loaded over the chip surface and incubated.
• After washing, the magnetic fringe field created by the labels was detected as a variation on the sensor resistance.
• Fig. 4A (dashed line and black dots) shows the biosensor normalized output for decreasing concentrations of viableþ VBNC cells.

4. Conclusions

• The lytic phage PVP-SE1 was explored as an alternative biorecognition element for bacterial detection and viability assessment.
• Taking into account the problematic occurrence of false positives associated with DNA-chips and the high production costs, poor stability and cross-reactivity related to immuno-chips, the development of phage-based biochips emerges as a valuable tool.
• The feasibility to immobilize phages on sensing surfaces and conjugate this biomolecular tool with electronic analytical devices without losing functionality was proven.
• The combined use of the magnetoresistive sensor with the phage probes allowed a clear detection of viable from dead Salmonella cells.

Figures

Fig. 3. Phage vs antibody performance as a biorecognition element. Density of cells specifically captured by phage and antibody spots, covalently immobilized on gold solid surface. (A) The surface coverage was analyzed by ImageJ software. (B) Stereoscope pictures for representative spots of Salmonella cells specifically recognized by phage PVP-SE1 (top) and anti-Salmonella antibody (bottom). Statistical analysis for phage vs antibody results: *po0.05, ***po0.0001, two-way ANOVA.

Fig. 2. Profile of phage adsorption to cells at different physiological states. Studies of bacteriophage PVP-SE1 adsorption to Salmonella Enteritidis in phosphate buffer at room temperature. (A) Adsorption profile of phage to viable, VBNC and dead cells (upper curves) and phage growth curve in viable cells (bottom curve). For phage growth, average values result from triplicates of one single assay and standard deviations are all below 20%. (B) Picture of plaque forming units (PFU) on a culture plate for phage adsorption counts. (C) Zoom-in over PFU on the bacterial lawn. Statistical analysis: ***po0.0001 for viable vs dead, þ þ þpo0.0001 for VBNC vs dead; two-way ANOVA.

Fig. 1. Induction of VBNC physiological state in Salmonella. Assessment of the physiological state of Salmonella Enteritidis cells after treatment with different concentrations of bleach by three different evaluation methods: (A) CFU counts on standard solid culture media (black line) compared to flow cytometry results for relative percentage of cells in the viable (green bars), compromised (yellow bars) and dead (red bars) states. The culturability test was collected from several trials over time and performed in triplicate; thus all CFU counts were normalized to the same initial cell concentration of 2.39 108 cells/mL. Results presented for the flow cytometry analysis are the mean of independent triplicates from three independent experiments (n¼9). (B) Epifluorescence microscopy images of Salmonella Enteritidis cells treated with different concentrations of bleach (0%, left panel; 0.006% middle panel; 0.01% right panel) where the different physiological states may be distinguished based on each cell's coloration (green, viable; yellowish, compromised; red, dead). (C) Representative flow cytometry plots showing the percentage of live, dead and compromised cells present after each treatment. For both assays, depicted in (B) and (C), cells were stained with SYTO 9 (green) and propidium iodide (PI; red); for flow cytometry analysis, the gating strategy adopted was the one described in the LIVE/DEAD BacLight kit's inset (Berney et al., 2007). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 4. Phage-based magnetoresistive biochip for cell viability assessment. Phage-based magnetoresistive biochip measurements performed on an electronic reader. (A) Normalized differential voltage signal for Salmonella samples subjected to different treatments (dashed line and black dots). The acquired sensor signals are differential values (ΔVbinding), calculated from the difference between the sensor baseline (Vsensor) and the signal from the specifically bound MNPs over the sensor (Vparticles) as previously described (Martins et al., 2009). The shaded area in blue on the bottom refers to the average signal from the control sensors using an unspecific phage. Each point and shaded area represents the average value of at least 10 independent sensors with respective associated standard deviation (error bars). Bars represent the concentration of cells and respective cell state (viable, compromised (comp.) and dead) for each cell sample measured on the MR-chip. Cell concentrations were obtained by flow cytometry using the LIVE/DEAD counting Baclight kit, in which each data point results from the analysis of triplicate samples from one representative experiment out of three. (B) Schematic representation of the “sandwich” type biomolecular recognition strategy adopted in the MR-biochip measurements. (C) Picture of the MR-biochip PCB used for the detection of Salmonella cells, shown on a standard cell culture plate.

Frequently asked questions

Q1. What have the authors contributed in "A bacteriophage detection tool for viability assessment of salmonella cells" ?

This work presents and validates a novel bacteriophage ( phage ) -based microbial detection tool to detect and assess Salmonella viability. This ability was confirmed for immobilized phages on gold surfaces, where the phage detection signal follows the same trend of the concentration of viable plus VBNC cells in the sample. Salmonella Enteritidis cells in a VBNC physiological state were evaluated by cell culture, flow-cytometry and epifluorescence microscopy, and further assayed with a biosensor platform.