Appl Microbiol Biotechnol (2014) 98:1795–1805 DOI 10.1007/s00253-013-5434-4

METHODS AND PROTOCOLS

Phagomagnetic immunoassay for the rapid detection of Salmonella Tamara Laube & Pilar Cortés & Montserrat Llagostera & Salvador Alegret & María Isabel Pividori

Received: 9 October 2013 / Revised: 21 November 2013 / Accepted: 23 November 2013 / Published online: 21 December 2013 # Springer-Verlag Berlin Heidelberg 2013

Abstract This work explores the use of the phage P22 in a phagomagnetic immunoassay for the rapid detection of Salmonella . The covalent attachment of wild-type phages was performed on two different magnetic carriers: carboxylactivated magnetic nanoparticles (300 nm) and tosyl-activated magnetic microparticles (2.8 μm). The bacteria were captured and preconcentrated by the phage-modified magnetic particles, followed by the detection using specific anti-Salmonella antibodies conjugated to horseradish peroxidase as an optical reporter. Outstanding selectivity and sensitivity was obtained with this approach, achieving detection limits of 19 CFU mL−1 in 2.5 h without any pre-enrichment, in milk samples. Moreover, if the samples were pre-enriched for 6 h, the method was able to detect as low as 1.4 CFU in 25 mL of milk. Therefore, the proposed strategy based on the combined use of phagomagnetic separation with immunological labeling is promising as a rapid and simple method for food safety.

Keywords Magnetic particles . P22 . Bacteriophage . Salmonella . Food safety

Electronic supplementary material The online version of this article (doi:10.1007/s00253-013-5434-4) contains supplementary material, which is available to authorized users. T. Laube : S. Alegret : M. I. Pividori (*) Grup de Sensors i Biosensors, Departament de Química, Universitat Autònoma de Barcelona, 08193 Bellaterra, Catalonia, Spain e-mail: [email protected] P. Cortés : M. Llagostera Grup de Microbiologia Molecular, Departament de Genètica i Microbiologia, Universitat Autònoma de Barcelona, 08193 Bellaterra, Catalonia, Spain

Introduction Detection of pathogenic bacteria is an area of prime interest since infectious diseases spreading every day through food have become a life-threatening problem for millions of people around the world (Dwivedi and Jaykus 2011). There are many programs like good agricultural and manufacturing practices, the hazard analysis and critical control points (HACCP) and the food code indicating approaches, which can significantly reduce the pathogenic microorganisms in food. However, the technology for pathogen detection is still the key for the prevention of the outbreaks (Velusamy et al. 2009). In recent years, considerable effort has been directed towards the development of rapid and simple detection methods such as polymerase chain reaction (PCR) and enzyme-linked immunosorbent assay (ELISA) in order to screen-out samples, as well as to reduce the time-consuming and laborious conventional culture-dependent techniques (Feng 2007). There are many reports exhaustively reviewing the rapid methods for bacteria detection (Arora et al. 2011; Dwivedi and Jaykus 2011; Liébana et al. 2013; Theron et al. 2010). Immunological detection of bacteria has become more sensitive, specific, reproducible, and reliable with many commercial immunoassays available for the detection of a wide variety of bacteria (Upmann and Bonaparte 2000). Quantitative real-time PCR has made DNA amplification more attractive and some commercial kits were also developed. However, problems associated with Taq enzyme inhibition and DNA extraction make direct detection of low numbers of bacteria in foods by PCR difficult to achieve (Gervais et al. 2007). Moreover, food samples are a complex and heterogeneous matrix consisting of various components including particulate matter, biochemical, and inorganic food components, fats and non-target (harmless) background microflora. Many of these components produce severe interferences in biological reactions and biosensing, for instance, fat

1796

and particulates can interfere with antibody binding, and complex carbohydrates can inhibit nucleic acid amplification (Dwivedi and Jaykus 2011). Additionally, some degree of selective enrichment remains essential in almost all cases to ensure that low detection limits (1 CFU/25–375 g) are met (Dwivedi et al. 2013). Regarding the sample processing, further improvement involves immunomagnetic separation approaches (IMS), that is, the use of biofunctionalized magnetic particles to capture target bacteria throughout an immunological reaction. The IMS allows both the separation from contaminating microflora and interfering components, as well as the preconcentration into smaller volume of buffer (Cudjoe 1999). In addition, due to the improved washing and separation steps, the matrix effect is minimized in complex samples (Zacco et al. 2006). Furthermore, the magnetic carriers can be easily integrated in microfluidic devices and cartridges (Bruls et al. 2009). We have previously reported methodological improvements due to the integration of magnetic particles in bioanalytical assays for the rapid detection of pathogenic bacteria in food, including specific antibodies and DNA as biorecognition elements (Lermo et al. 2009; Liébana et al. 2009). Recently, the use of bacteriophages for the improved detection of Salmonella has been reported, by coupling a “phagomagnetic separation" (PMS) step with a doubletagging PCR amplification of the DNA of the captured bacteria followed by electrochemical magneto-genosensing (Liébana et al. 2013). Phage-based techniques are attracting much interest for bacteria detection due to their high specificity as well as their high stability, cost-efficient and animal-free production, which makes them suitable for in situ monitoring of food and environmental contaminants. They are extremely resistant in a range of harsh conditions including pH and temperature, and can even be used in the presence of nucleases or proteolytic enzymes, without degradation. Beside this, they provide a large amount of viral coat proteins with a large surface for further chemical modification (Van Dorst et al. 2010; Singh et al. 2013; Smartt and Ripp 2011). Different strategies using phages for bacteria specific biorecognition were reported; for instance, the measurement of the activity of reporter genes carried by the phages and expressed only after infection (Chen and Griffiths 1996; Kuhn et al. 2002), the phage amplification approaches with the progeny being detected by different methods (Favrin et al. 2003; De Siqueira et al. 2003), and the amperometric detection of enzymatic activity after phage infection and bacteria lysis (Neufeld et al. 2003). Phages can be also immobilized on solid supports for bacteria detection by different strategies, such as physical adsorption (Li et al. 2010; Nanduri et al. 2007), covalent reaction (Shabani et al. 2008; Arya et al. 2011), and affinity interaction involving peptides expressed

Appl Microbiol Biotechnol (2014) 98:1795–1805

in the capsid through phage display techniques (Tolba et al. 2010). Recently, the use of native bacteriophages as a replacement of the primary capturing antibody in a sandwich format standard ELISA test for the detection of bacteria was also reported (Galikowska et al. 2011). In this work, we report the use of bacteriophagemodified magnetic particles for the preconcentration of Salmonella followed by the rapid detection of the whole bacterial cells by a specific antibody, without the need of any amplification step and thus simplifying the analytical procedure. The bacteriophage P22 was used as a model towards Salmonella enterica serovar Thyphimurium, for the biorecognition and preconcentration of the pathogenic bacteria. The P22 bacteriophages were covalently immobilized throughout the amine moieties present in their capsid proteins to tosylactivated magnetic microparticles or carboxylic nanoparticles. After the PMS of the target bacteria, the detection was based on an immunoassay using specific anti-Salmonella antibodies conjugated to HRP as optical reporter. Both magnetic supports were compared in terms of their analytical performance. Other features of this approach are discussed and compared with classical culture methods, immunological approaches and PCR-based assays.

Materials and methods Instrumentation The P22 bacteriophages lysates were concentrated using 25×89-mm ultracentrifuge tubes (Ultra-ClearTM Tubes, Beckman, California, USA) in an ultracentrifuge (OptimaTM L-80, Beckman) using the SW28 Ti rotor (Beckman). The immobilization to the magnetic particles was performed under shaking using an Eppendorf Thermomixer compact. Polypropylene and polystyrene microtiter plates were purchased from Corning (Catalogue no. 153364) and Nunc (Catalogue no. 269787, Roskilde, DK), respectively. The magnetic separation of the particles during the washing steps were carried out using a magnetic separator for Eppendorf tubes Dynal MPC-S (product no. 120.20D, Dynal Biotech ASA, Norway) or a 96-well plate magnet (product no. 21358, Thermo Fisher Scientific, Waltham, USA). The incubation and washing steps with the microtiter plates were performed under shaking conditions using a Minishaker MS1 (IKA, Germany). Optical measurements were performed on a TECAN Sunrise microplate reader with Magellan v4.0 software. The scanning electron microscopy (SEM) images were taken with the scanning electron microscope MERLIN FE from Carl Zeiss Microscopy GmbH.

Appl Microbiol Biotechnol (2014) 98:1795–1805

Chemicals and immunochemicals Tosylactivated magnetic particles (MP) Dynabeads M-280 (product no. 142.03, diameter 2.8 μm, concentration 2×109 particles mL−1, 30 mg mL−1) were purchased from Life Technologies, Invitrogen Dynal AS (Oslo, Norway). The carboxyl-Adembeads (nMP) (product no. 0213; diameter 300 nm, concentration 1.1 × 10 1 2 particles mL − 1 , 30 mg mL−1) were obtained from Ademtech SA (Pessac France). Anti-Salmonella antibodies conjugated to HRP (product no. ab20771) were purchased from Abcam (Cambridge, UK). Bovine serum albumin, EDC, and sulfoNHS were obtained from Sigma-Aldrich. The peroxide and TMB (3,3′,5,5′- tetrametylbenzidine) solutions used for the optical measurements (TMB Substrate Kit, Product no. 34021) were purchased from Pierce. All other reagents were of the highest available grade, supplied from Sigma or Merck and all buffer solutions were prepared with Milli-Q water (Millipore Inc., Ω =18 MΩ cm) and are detailed in the “Electronic supplementary material” (ESM). Bacterial strains and phage lysate Salmonella enterica serovar Typhimurium LT2 (ATCC® 700720™ courtesy of J.L. Ingraham, Bacteriology Department, University of California, Davis, USA) and Escherichia coli K12 (CGSC 5073, The Coli Genetic Stock Center) strains were routinely grown in Luria Bertani (LB) broth or on LB agar plates for 18 h at 37 °C and bacterial viable counts were determined by plating on LB plates followed by incubation at 37 °C for 24 h. The P22 bacteriophage lysate (ATCC 19585-B1) was obtained by infecting exponential cultures of Salmonella Typhimurium LT2 (108 CFU mL−1) and by further purification with cesium chloride gradient (Sambrook and Russell 1989), as detailed in the ESM. The phage stock solutions were maintained in MgSO4 10 mM solution at 4 °C retaining a constant titer for several months. Covalent immobilization of bacteriophages on magnetic micro and nanocarriers P22 bacteriophages were covalently coupled to 2.8-μm tosylactivated magnetic particles (P22-MP) as well as to 300-nm carboxylic magnetic nanoparticles (P22-nMP) through the amine moieties of their capsid proteins (gp5). No pre-activation step was needed for the tosylactivated MP, while when using the nMP, a previous activation step of the carboxylic groups of the magnetic nanoparticles which EDC and sulfo-NHS was performed. The immobilization protocols are detailed in the ESM. The phage/particle ratio was also optimized analyzing the effect of increasing the magnetic particles concentration or the phage concentration up to

1797

approximately 1012 plaque-forming units (PFU) mL−1. After the immobilization, the efficiency of the coupling strategies was evaluated by the double agar layered method. In this approach, tenfold serial dilutions of the supernatants after the covalent attachment were plated onto lawns of the bacterial strain, as detailed in the ESM. Evaluation of phage infectivity and biorecognition towards Salmonella of the immobilized P22 bacteriophages The availability and integrity of the tailspike proteins (TSP) gp9 after the immobilization is an important issue to be considered since the conjugation to the magnetic carriers could hinder the bacterial recognition. For the evaluation of phage orientation and infectivity, several tenfold dilutions of the phage-modified magnetic particles were cultured by double agar-layered method. Afterwards, the capture abilities of the modified magnetic carriers were evaluated by SEM microscopy. A bacterial solution of approximately 106 CFU mL−1 in LB broth was added to P22-MP or P22nMP and the PMS of Salmonella Typhimurium LT2 in these samples was performed. The samples (500 μL) were mixed with 10 μg of P22-MP or P22-nMP and incubated 20 min at room temperature with agitation followed by 20 min at 37 °C without agitation. After that, the magnetic particles with the attached bacteria were separated with a magnet, and then two washes were performed with PBST for 5 min at room temperature. Finally, the modified magnetic particles were resuspended in 100 μL PBS and after adding it to 5 mL of Milli-Q water the solutions were filtered through a Nucleopore membrane (25 mm Ø, 0.2-μm pore size). The filters were then fixed with glutaraldehyde overnight, post-fixed with osmium tetraoxide for 2 h, dehydrated in a graded ethanol series and dried by CO2 critical point. Samples were then mounted on metallic stubs with adhesive carbon films and observed with the scanning electron microscope operating at 15 kV (Brambilla et al. 2012). Phagomagnetic immunoassay for the detection of Salmonella in milk The phagomagnetic immunoassay was performed in 96-well microtiter plates as schematically outlined in Fig. 1 and comprised the following steps (all the referred quantities are “the amounts added per well”): (1) PMS of 100 μL Salmonella for attaching the target bacteria on the magnetic carriers as previously described; (2) labeling with 100 μL of anti-Salmonella antibody (HRP) incubating 30 min at room temperature and 700 rpm; and finally, the last step, (3) optical detection with 100 μL of substrate solution (0.004 %v/v H2O2 and 0.01 %w/v TMB—3,3′,5,5′-tetramethylbenzidine—in citrate buffer) incubated for 30 min at room temperature (RT) in darkness. The enzymatic reaction was stopped by adding 100 μL of H2SO4

1798

Appl Microbiol Biotechnol (2014) 98:1795–1805

Fig. 1 Schematic representation of the phagomagnetic immunoassay, comprising the phagomagnetic separation (PMS) of the bacteria, followed by the labeling step using anti-Salmonella antibody (HRP) and the optical detection

(2 mol L−1) and the absorbance measurement of the supernatants was performed at 450 nm. For the evaluation of the results, the exact concentration of the initial inoculum coming from an overnight culture in LB broth was found by dilution and plating in LB agar. A negative control of LB broth was always also processed. Different parameters (for instance the concentration of the anti-Salmonella antibody and the magnetic carriers) were optimized in order to find the better conditions for achieving high positive signals and low background values. In an attempt to shorten as well as to simplify the analytical procedure, four different protocols were also evaluated, by varying the number of washing steps, as detailed explained in Table 1. Other experimental parameters such as temperature and agitation during the PMS, surfactant concentration, ionic strength, and pH were used as optimized in previous works. The evaluation of the matrix effect on artificially inoculated Salmonella samples (from 101 to 108 CFU mL−1) in LB broth and in milk diluted 1/10 in LB were compared for both kinds

of P22 bacteriophage modified magnetic carriers (P22-MP and P22-nMP). Specificity study The specificity of the system was also evaluated by comparing the response to Salmonella with the signal obtained in the presence of an equivalent amount of another gram-negative bacterium as E. coli artificially inoculated in the samples, as well as the negative control. Phagomagnetic immunoassay for the detection of Salmonella in pre-enriched milk As cells are injured when exposed to adverse conditions during food processing, a pre-enrichment step is usually included in classical methods to achieve the proliferation of stressed Salmonella spp. cells, since otherwise bacteria that have not fully repaired may be missed (Amaguaña and

Table 1 Different procedures performed with the P22-MP as a magnetic carrier for the optimization of the phagomagnetic immunoassay Stepwise protocol

No washing between incubation Antibody addition without One-step protocol steps supernatant discard

1. Phagomagnetic separation (PMS) 1. PMS (20 min., RT, 700 rpm+ 1. PMS (20 min., RT, (20 min., RT, 700 rpm+20 min., 20 min., 37 °C, no shaking) 700 rpm+20 min., 37 °C, no shaking) 37 °C, no shaking) 2. Washing steps (2×) 2. Supernatant discard+ 2. Addition of Ab-HRP+ enzymatic labeling with Abenzymatic labeling HRP (30 min, RT, 700 rpm) (30 min, RT, 700 rpm) 3. Enzymatic labeling with Ab-HRP 3. Washing steps (3×) 3. Washing steps (3×) (30 min, RT, 700 rpm) 4. Washing steps (3×)

1. PMS and enzymatic labeling with Ab-HRP (20 min., RT, 700 rpm+20 min., 37 °C, no shaking+30 min, RT, 700 rpm) 2. Washing steps (3×)

Appl Microbiol Biotechnol (2014) 98:1795–1805

1799

Andrews 2004). The pre-enrichment step is performed with a nonselective broth medium, like LB broth, leading to a 1/10 dilution of the food matrix. In order to fulfill the legislation requirements for milk, which establish the absence of Salmonella in 25 g, sampled in five portions of 5 g each in different points (Real Decreto 1679/1994, BOE 24-09-94), the detection system should be able to detect 0.04 CFU mL−1, which is only possible if a pre-enrichment step is included. As a result the milk samples were pre-enriched in LB broth at 37 °C, and assayed at 4, 6, 8 and 16 h. Safety considerations All the procedures involving the manipulation of potentially infectious materials or cultures were performed following the safe handling and containment of infectious microorganism's guidelines (Chosewood and Wilson 2007). According to these guidelines, the experiments involving Salmonella Typhimurium and E. coli were performed in a Biosafety Level 2 Laboratory. The ultimate disposal was performed according to local regulations.

Results Covalent immobilization of bacteriophages on magnetic micro and nanocarriers The chemical reaction for the covalent immobilization does not specifically control the orientation since amine groups exist both on the head and TSP domains of P22. However, the shape could also play a role, since the large and smooth capsid head seems to be favored over the pointy TSP in immobilization procedures (Handa et al. 2008). The immobilization efficiency of the P22 phage on the magnetic micro and nanoparticles was studied and compared by immobilizing phage amounts of approximately 1010 PFU to 1 mg magnetic carriers and evaluating the phage concentration before and after the immobilization step. The amount of phages in the supernatants was compared with the initial amount before immobilization, obtaining coupling

a

1.5

Evaluation of phage infectivity and biorecognition towards Salmonella of the immobilized P22 bacteriophages The bacteriophages bind to specific receptors on the bacterial surface in order to inject the genetic material inside the bacteria. In the case of P22, six homotrimeric tailspike molecules

P22-nMP

b

P22-MP

0.9 0.6 0.3

2.5

41 phages/ MP 190 phages/ MP

2.0

Abs (450nm)

1.2

Abs (450nm)

Fig. 2 Comparative response of both magnetic carriers (P22-MP and P22-nMP) at 0, 105, and 107 CFU mL−1 of Salmonella (a) and evaluation of different phage per magnetic particle ratios (b)

efficiencies of 92.4 and 83.8 % for MP and nMP, respectively, as detailed shown in Table S1, ESM. When comparing the analytical performance of both magnetic micro and nanoparticles at a Salmonella concentration of 105 and 107 CFU mL−1 (Fig. 2a), although a higher signal was obtained with the P22-nMP at 107 CFU mL−1, a better signal to non-specific adsorption ratio was obtained with the P22MP for the lower bacterial concentration (at 105 CFU mL−1), suggesting improved features of the magnetic microparticles in terms of non-specific adsorption and LODs. Further studies about covalent immobilization were performed on P22-MP, by immobilizing 1.44×1010 PFU on increasing amounts of MP, in detail 7.0×107 (1 mg) and 3.5×108 (5 mg). Coupling efficiencies of 92.4 and 98.9 % and phage/ MP ratios of 190 and 41 were respectively obtained, as expected. On the other hand, when the amount of phage was increased by immobilizing 2.01×1011 and 6.42×1011 PFU on 7.0×107 MPs, a decrease in the coupling efficiency was observed. Nevertheless, the number of phages per MP increased considerably obtaining a plateau in the immobilization efficiency at a phage/MP ratio of approximately 1,650 (by the immobilization of up to 2.0×1011 phages on 7.0×107 MPs), in agreement with previous studies (Liébana et al. 2013). The detailed comparative results are presented in Table 2. Furthermore, the signals obtained by increasing the phage/MP ratios from 41 to 1,665 and assaying at 105 and 107 CFU mL−1 Salmonella, showed better analytical performance –in terms of higher signal to background ratio—for fully covered magnetic particles (approximately 1,650 phage per MP), as shown in Fig. 2b and Table S2, ESM. All these results suggest that a better response was obtained by immobilizing higher phage titer (up to 1011 PFU) on lower MP amount (7×107 MPs), achieving full coverage of the phages on the magnetic particles, and thus increased PMS efficiency.

1665 phages/ MP

1.5 1.0 0.5

0.0

0.0 107 -1 Salmonella (CFU mL )

(-)

105

(-)

105

107

Salmonella (CFU mL-1)

1800

Appl Microbiol Biotechnol (2014) 98:1795–1805

Table 2 Coupling efficiency and final number of phages per magnetic particle for different phages and magnetic particles amounts Initial phage amount (PFU)

Magnetic particles (amount/number)

Coupling efficiency (%)

Phage/ MP ratio

1.44×1010 1.44×1010 2.01×1011 6.42×1011

5 mg/3.5×108 1 mg/7.0×107 1 mg/7.0×107 1 mg/7.0×107

98.9 92.4 58.0 18.0

41 190 1,665 1,651

(gp9) are part of the viral adhesion protein which specifically recognizes the O-antigenic repeating units of the cell surface lipopolysaccharide of Salmonella (Steinbacher et al. 1997) In order to evaluate the phage infectivity after the covalent attachment to the magnetic carriers, both the P22-MPs and the P22-nMPs were cultured by the double agar-layered method and enumeration of plaques. By this method, it is not possible to establish the number of bacteriophages per magnetic particle since each modified magnetic particle is able to

produce a unique plaque, regardless of how many bacteriophages are correctly oriented on its surface. Nevertheless, by plating the modified magnetic carriers it is possible to evaluate their global lytic activity. The P22-MPs showed lytic activity of approximately 100 % at a phage/MP ratio of 1651 (fullcoverage), but it dropped appreciably to 43 % for a phage/MP ratio of 41. On the other hand, the P22-nMPs showed lytic activity in as low as 0.001 % of the nanoparticles. This result is not concordant with the immobilization efficiency above 80 %, and could be explained by the high particle agglomeration during culturing which was also observed in the SEM images (Fig. 3). The captured bacteria on the magnetic carriers after the PMS are shown in Fig. 3. While the magnetic microparticles were able to attach more than one bacteria (Fig. 3b, d), the pattern observed with the magnetic nanoparticles was the opposite, showing single bacteria tagged with more than one P22-nMP (Fig. 3a, c). Due to multivalency in both magnetic carrier and bacteria, aggregates were clearly observed in both

Fig. 3 Evaluation of the PMS by SEM at a Salmonella concentration of 3.2× 106 CFU mL−1 using carboxyl magnetic nanoparticles (a and c) and tosylactivated magnetic particles (b and d). In all cases, identical acceleration voltage (15 kV) was used. The arrows in c show the phages immobilized on the magnetic nanoparticles

a

b

c

d

Appl Microbiol Biotechnol (2014) 98:1795–1805

1801

types of particles. However, in the case of the magnetic nanoparticles, the degree of aggregation is higher, in accordance with the poor lytic activity obtained by culturing. Figure 3c also shows the P22 phages immobilized on the surface of the nMP. Phagomagnetic immunoassay for the detection of Salmonella in milk First of all, the concentration of the anti-Salmonella antibody (HRP) as optical reporter for the phagomagnetic immunoassay was optimized by testing three different dilutions (1/1,000, 1/2,000 and 1/4,000), obtaining better results at a 1/1,000 dilution, as detailed in the ESM (Fig. S1 and Table S3). The concentrations of the magnetic particles P22-MP and P22nMP were also optimized by analyzing the signal to background ratio at two bacteria concentrations (10 5 and 107 CFU mL−1) using three different particle concentrations: 0.05, 0.1 and 0.2 mg mL−1 (corresponding to 3.5×106, 7×106 and 1.4×107 P22-MP mL−1 or 1.9×109, 3.8×109 and 7.7× 109 P22-nMP mL−1, respectively), as shown in detail in Fig. S2, ESM. The highest signal to background values were obtained in both kinds of magnetic particles at a concentration of 0.05 mg mL−1 (3.5×106 P22-MP mL−1 and 1.9×109 P22nMP mL−1). To establish the optimal protocol for the phagomagnetic immunoassay, different procedures were evaluated by comparing the response at two bacteria concentrations (105 and 107 CFU mL−1) using P22-MP as a magnetic carrier. Four different protocols were assayed by varying the incubation and washing steps, as shown in Fig. 4. Although in all cases the bacteria were clearly detected, improved results were obtained with the strategy A, by performing the two incubation steps separately with washings in between, showing lower background values and higher positive signals and, as a result, better signal-to-background ratios (as shown in Table S4, ESM). LB broth

1.5

10 5 CFU mL-1 Abs (450nm)

10 7 CFU mL-1 1.0

Finally, the phagomagnetic immunoassay was evaluated for artificially inoculated Salmonella (ranged from 0 to 108 CFU mL−1) in LB broth and in milk diluted 1/10 in LB for both kinds of magnetic carriers, P22-MP and P22-nMP, as shown in Fig. 5a, b. A slight matrix effect became visible at high bacteria concentrations showing a decrease in the signals in the presence of the milk as a matrix, being the effect more evident in the case of the P22-nMP. The cut-off and LOD values were evaluated by processing 10 negative samples (0 CFU mL−1) obtaining in the case of the P22-MP a mean value of 0.204 (s =0.007) and 0.184 (s = 0.007) absorbance units (a.u.) for the assay performed in LB broth and in milk diluted 1/10 in LB, respectively. The cut-off values were then extracted with a one-tailed t test at a 99 % confidence level, obtaining 0.224 (solid line) and 0.203 a.u. (dotted line) in LB and in milk, respectively. From these data, the LOD values were finally calculated by interpolation of the cut-off signals in the sigmoidal dose response curve obtained after plotting the absorbance vs. the logarithm of Salmonella concentration (R 2 =0.9950) (Fig. S3B, ESM), obtaining LOD values of 10 CFU mL−1 in LB broth and 19 CFU mL−1 in milk. To confirm this, a zoom shot of the signals in the lower bacteria concentrations near the LOD, with the cut-off values in LB (solid line) and milk (dotted line) using both kinds of magnetic carriers, is also shown in Fig. S3A of the ESM. As shown in the Figure S3A, the system is able to give a positive signal for the first point of the calibration curve in both matrixes, which corresponds to a concentration of 101 CFU mL−1 Salmonella. On the other hand, the mean values of the negative samples (0 CFU mL−1) for the P22-nMP obtained in LB broth and milk diluted 1/10 were 0.173 and 0.144 a.u. with standard deviations of 0.014 and 0.011, respectively. When comparing these results with the values obtained with the P22-MP, higher coefficients of variation were obtained with the nanoparticles (8.3 % in LB and 7.4 % in milk) than with the P22-MP (3.5 % in LB and 3.8 % in milk). Regarding the cut-off values for the P22-nMP, the results were 0.215 and 0.176 a.u. in LB and milk, respectively (Figs. 5b and S3C of the ESM), indicating that the system is able to detect 105 CFU mL−1. When interpolating the cut-off signals in the sigmoidal dose response curve (Fig. S3D, ESM), the following LODs were obtained: 1.09×105 CFU mL−1 in LB and 3.35×105 CFU mL−1 in milk.

0.5

Specificity study 0.0 A

B

C

D

Strategy

Fig. 4 Phagomagnetic immunoassay performed with the P22-MP following four different procedures. The bars represent the signals obtained at a Salmonella concentration of 0, 105 and 107 CFU mL−1. In all cases n =3, P22-MP 0.05 mg mL−1 and anti-Salmonella antibody (HRP) 1/1,000

Figure 5c, d shows the results of the phagomagnetic immunoassay performed in milk diluted 1/10 in LB and artificially inoculated with 106 and 107 CFU mL−1 of both E. coli and Salmonella, as well as a negative control. In the case of the P22-MP (Fig. 5c), the signal obtained for the E. coli samples are almost identical to the blank, regardless the concentration, indicating a high specificity of the assay as

1802

a

b 2.5

LB broth

Milk 1/10 in LB broth

Milk 1/10 in LB broth

2.0

Abs (450nm)

1.5 1.0 0.5

1.5 1.0 0.5

0.0 (-)

101

102

103

104

105

106

107

0.0

108

(-)

101

102

Salmonella (CFU mL-1)

103

104

105

106

107

108

Salmonella (CFU mL -1 )

c

d 1.25

1.2

1.00

Phagomagnetic immunoassay for the detection of Salmonella in pre-enriched milk The pre-enrichment step was studied with a nonselective broth medium, in this case LB broth using the P22-MP, since better results and a higher specificity were obtained with this system. In order to contaminate a milk sample with a proportion of around 1 CFU of Salmonella in 25 mL, to test the capability of the system to fulfill the established regulations, 250 mL of milk were spiked with around 10 CFUs. From there, five portions of 5 mL were taken and pre-enriched in 45 mL LB broth at 37 °C (samples S1 to S5). A positive control containing 10× Salmonella concentration was also evaluated, as well as a negative control (0 CFU mL−1). Finally, aliquots were taken at 4, 6, 8, and 16 h to test each sample by the developed method, as well as to perform plating in LB agar in order to control the amount of bacteria in each sample and preenrichment time by the classical culture method. As shown in Fig. 6, the system was not able to detect the bacteria after just 4 h pre-enrichment. This result was also in agreement with the colony counting, in which not even the

7

10 la

li

el on lm

el on

co Sa

li lm Sa

10

6

10

(-) co

el on lm

well as a low non-specific binding. However, in the case of the P22-nMP (Fig. 5d), a small increase in the signal was observed when increasing the E. coli concentration in comparison with the background adsorption, suggesting a slight nonspecific binding.

E.

10 la

li co E. Sa

lm Sa

10

10 on

el

co

li

la

10

(-) E.

7

0.00 7

0.0 6

0.25

6

0.3

7

0.50

10

0.6

0.75

E.

0.9

6

Abs (450nm)

1.5

la

Abs (450nm)

2.5

LB broth

2.0

Abs (450nm)

Fig. 5 Top: Phagomagnetic immunoassay performed with the P22-MPs (a) and P22-nMP (b) by increasing the amount of Salmonella from 0 to 108 CFU mL−1, artificially inoculated in LB broth and in milk diluted 1/10 in LB. The cut-off values for LB broth and milk diluted 1/10 are represented with a solid and dotted line, respectively. In all cases, n =3, except for the 0 CFU mL−1 negative control in which n =10. Bottom: Specificity study for the phagomagnetic immunoassay for LB artificially inoculated, respectively, with 0 CFU mL−1, 1.9× 106 CFU mL−1 E. coli, 2.5× 106 CFU mL−1 Salmonella, 2.3× 107 CFU mL−1 E. coli, and 3.2× 107 CFU mL−1 of Salmonella for the P22-MP (c) and P22-nMP (d)

Appl Microbiol Biotechnol (2014) 98:1795–1805

positive control reached a bacteria concentration above 10 CFU mL−1, being thus below the detection limit in milk of the method (19 CFU mL−1). However, after 6 h of preenrichment, one of the five milk samples was clearly positive and another was around the cut-off value, as shown in Fig. 6(a). The evolution over the time of the negative control, the sample S1 and the positive control are detailed in Fig. 6(b). The exact initial inoculum was found by classical culture method obtaining that in fact the milk sample was spiked with around 14 CFUs in 250 mL. As can be seen in the results presented in Fig. 6, the bacteria in both the S1-positive sample (containing 1.4 CFU in 25 mL) and the positive control (with 14 CFU in 25 mL) were able to be detected after a preenrichment of as low as 6 h, showing exponential growth over the time, while the negative control shows no growing at all, as could be expected. Although the method was able to detect around 0.06 CFU mL−1 in milk (1.4 CFUs in 25 mL) according to the legislation after 6 h pre-enrichment in LB, remarkable improvement of the signal was achieved between 8 and 16 h of pre-enrichment, as shown in Fig. 6(b).

Discussion Bacteriophages are promising candidates to be used as a biorecognition element for the detection of pathogenic microorganisms. They provide many advantageous features such as

Appl Microbiol Biotechnol (2014) 98:1795–1805

1803 Preenrichment

b

a 0.3

2.5

6h

4h

4h

6h

8h

16 h

Abs (450nm)

Abs (450nm)

2.0

0.2

0.1

1.5 1.0 0.5

0.0

0.0

(-)

S1

S2

S3

S4

S5

(+)

Sample

(-)

S1

(+)

Sample

Fig. 6 a Phagomagnetic immunoassay after a pre-enrichment step of 4 and 6 h of artificially inoculated milk containing 1.4 CFU in 25 mL (0.06 CFU mL−1). Five portions (S1 to S5) of the sample, a negative control (0 CFU mL−1), and a 10× positive control are also shown. The dotted line represents the cut-off calculated by processing 16 negative

controls. In all cases, n =3. b Phagomagnetic immunoassay after a preenrichment step of 4, 6, 8, and 16 h for milk artificially inoculated with Salmonella. The S1-positive sample of milk (0.06 CFU mL−1), a negative control (0 CFU mL−1) and a positive control (0.6 CFU mL−1) are shown

outstanding selectivity, high sensitivity, and stability, which are three ideal attributes for any biorecognition probe that makes them suitable for in situ monitoring of food and environmental contaminants (Van Dorst et al. 2010; Mao et al. 2009). Compared to antibodies, phages have distinct advantages as recognition receptors. On one hand, they are less fragile and less sensitive to environmental stress such as pH and temperature fluctuation reducing the environmental limitations, and on the other, their production besides being animal-free can be less complicated and less expensive than antibody production (Handa et al. 2008; Zourob and Ripp 2010). In this work, the detection of Salmonella was demonstrated using both magnetic micro and nanoparticles modified with the bacteriophage P22 by a phagomagnetic immunoassay with optical detection. Although the covalent immobilization of P22 bacteriophages was successfully performed on both magnetic carriers achieving excellent coupling efficiencies, magnetic microparticles showed improved performance in terms of sensitivity and specificity, as well as lower matrix effect. These results could be related with the tagging pattern of the nanoparticles, showing a distribution all along the surface of the bacteria that could hinder the biorecognition event with the specific antibodies. Another probable explanation for the poorer performance of the nanoparticles in this case could be the fact that the higher surface area per volume ratio given by their smaller size could also increase the nonspecific adsorption, raising thus the influence of the matrix components during the assay. Outstanding sensitivity was achieved using the P22-MP, being able to detect as low as 19 CFU mL−1 in milk with no pre-enrichment step. Compared with other reported phage-based methodologies for detecting pathogenic bacteria in food, excellent detection

limits were reached with this system, achieving values that to the best of our knowledge were only obtained using phageintegrated reporter genes (Chen and Griffiths 1996; Wolber 1993), labeling of phage DNA with fluorescent dyes (Goodridge et al. 1999), labeling of biotin-tagged phages with streptavidin-modified quantum dots (Edgar et al. 2006), phage amplification techniques (Favrin et al. 2003), or PCR-based approaches coupled to DNA biosensors (Liébana et al. 2013). Some examples of immunomagnetic (instead of phagomagnetic) separation for Salmonella detection are summarized in Table S5 of the ESM. In some of this approaches (Croci et al. 2004; Wang et al. 2011; Starodub and Ogorodnijchuk 2012; Cho and Irudayaraj 2013), the LODs reported were similar, ranging from 1 to 102 CFU mL−1. Nevertheless, all the aforementioned methods require longer assay times and/or more expensive and labor-intensive DNA extraction and amplification steps or phage engineering techniques, and in some cases also expensive instrumentation, having thus the present method the advantage of being more rapid, simple, and cost-efficient. Moreover in the presented strategy, if the milk sample was pre-enriched for 6 h in LB, as low as 1.4 CFUs Salmonella could be detected in 25-mL milk sample, fulfilling the requirements of the legislation. Finally, since a native phage is used in this strategy instead of an antibody, additional advantages are the high stability, as well as the cost-efficient and animal-free production of the biorecognition element, which makes thus this approach suitable for in situ monitoring of food and environmental contaminants. It should be highlighted that this approach was able to clearly distinguish between pathogenic bacteria such as Salmonella and E. coli. The specificity can be ascribed to the double biorecognition, coming from the PMS with the specific interaction between the membrane receptor of the

1804

bacteria and the TSP of the P22 bacteriophage immobilized on the magnetic carrier, and the immunological labeling step, involving the anti-Salmonella antibody (HRP) towards the H or flagellar antigen and the O or somatic antigen (part of the LPS moiety). Therefore, phagomagnetic separation followed by the immunological detection with a specific antibody as optical reporter, can be considered as good alternative candidates to the gold standard tandem, differential plating and the biochemical/serological assays, reducing considerably the time of the assay from 3 to 5 days to 2.5 h, in order to screen-out food samples. A positive result will involve further investigation with the gold standard conventional culturing. Future work will be focused on the analytical validation of this promising methodology with a high number of dairy samples. Moreover, further studies are focused on the application of the developed strategy to a phagomagnetic immunoassay with electrochemical detection using disposable, lowcost screen-printed electrodes, which could be more suitable for the rapid and on-site screening-out of Salmonella in HACCP programs. Acknowledgments Financial support from Ministry of Science and Innovation (MEC), Madrid (Project BIO2010-17566) and from Generalitat de Catalunya (Projects SGR323 and SGR1106), are acknowledged. T.L. also acknowledges the PhD grant of the Comissionat per a Universitats i Recerca del DIUE de la Generalitat de Catalunya i del Fons Social Europeu. Finally, we want to thank Joan Colom and Susana Escribano (Microbiology group) for her excellent technical assistance.

References Amaguaña RM, Andrews WH (2004) In: Robinson RK, Batt CA, Patel P (eds) Encyclopedia of food microbiology. Academic Press, Oxford, p 1948 Arora P, Sindhu A, Dilbaghi N, Chaudhury A (2011) Biosensors as innovative tools for the detection of food borne pathogens. Biosens Bioelectron 28:1–12 Arya SK, Singh A, Naidoo R, Wu P, McDermott MT, Evoy S (2011) Chemically immobilized T4-bacteriophage for specific Escherichia coli detection using surface plasmon resonance. Analyst 136:486– 492 Brambilla C, Sánchez-Chardi A, Pérez-Trujillo M, Julián E, Luquin M (2012) Cyclopropanation of α-mycolic acids is not required for cording in Mycobacterium brumae and Mycobacterium fallax . Microbiology 158:1615–1621 Bruls DM, Evers TH, Kahlman JAH, Van Lankvelt PJW, Ovsyanko M, Pelssers EGM, Schleipen JJHB, De Theije FK, Verschuren CA, Van der Wijk T, Van der Zon JBAD, Dittmer WU, Immink AHJ, Nieuwenhuis JH, Prins MWJ (2009) Rapid integrated biosensor for multiplexed immunoassays based on actuated magnetic nanoparticles. Lab Chip 9:3504–3510 Chen J, Griffiths M (1996) Salmonella detection in eggs using Lux(+) bacteriophages. J Food Prot 59:908–914

Appl Microbiol Biotechnol (2014) 98:1795–1805 Cho IH, Irudayaraj J (2013) In-situ immuno-gold nanoparticle network ELISA biosensors for pathogen detection. Int J Food Microbiol 164: 70–75 Chosewood LC, Wilson DE, U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Institutes of Health (2007) Biosafety in microbiological and biomedical laboratories. U.S. Government Printing Office, Washington DC, pp 44–49 Croci L, Delibato E, Volpe G, De Medici D, Palleschi G (2004) Comparison of PCR, electrochemical enzyme-linked immunosorbent assays, and the standard culture method for detecting Salmonella in meat products. Appl Environ Microbiol 70(3): 1393–1396 Cudjoe KS (1999) Detection by immunomagnetic particle-based assay. In: Robinson RK, Batt CA, Patel P (eds) Encyclopedia of food microbiology. Academic Press, Oxford, p. 1968 De Siqueira RS, Dodd CER, Rees CED (2003) Phage amplification assay as rapid method for Salmonella detection. Braz J Microbiol 34:118–120 Dwivedi HP, Jaykus L-A (2011) Detection of pathogens in foods: the current state-of-the-art and future directions. Crit Rev Microbiol 37: 40–63 Dwivedi HP, Smiley RD, Jaykus L-A (2013) Selection of DNA aptamers for capture and detection of Salmonella Typhimurium using a whole-cell SELEX approach in conjunction with cell sorting. Appl Microbiol Biotechnol 97:3677–3686 Edgar R, McKinstry M, Hwang J, Oppenheim AB, Fekete RA, Giulian G, Merril C, Nagashima K, Adhya S (2006) High-sensitivity bacterial detection using biotin-tagged phage and quantum-dot nanocomplexes. Proc Natl Acad Sci U S A 103:4841–4845 Favrin SJ, Jassim S a, Griffiths MW (2003) Application of a novel immunomagnetic separation–bacteriophage assay for the detection of Salmonella enteritidis and Escherichia coli O157:H7 in food. Int J Food Microbiol 85:63–71 Feng P (2007) Rapid methods for the detection of foodborne pathogens, current and next generation technologies. In: Doyle M, Beuchat L (eds) Food microbiology, fundamentals and frontiers, 3rd edn. ASM Press, Washington, pp 911–934 Galikowska E, Kunikowska D, Tokarska-Pietrzak E, Dziadziuszko H, Loś JM, Golec P, Węgrzyn G, Loś M (2011) Specific detection of Salmonella enterica and Escherichia coli strains by using ELISA with bacteriophages as recognition agents. Eur J Clin Microbiol Infect Dis 30:1067–1073 Gervais L, Gel M, Allain B, Tolba M, Brovko L, Zourob M, Mandeville R, Griffiths M, Evoy S (2007) Immobilization of biotinylated bacteriophages on biosensor surfaces. Sens Actuators B 125:615–621 Goodridge L, Chen J, Griffiths M (1999) Development and characterization of a fluorescent-bacteriophage assay for detection of Escherichia coli O157:H7. Appl Environ Microbiol 65:1397–1404 Handa H, Gurczynski S, Jackson MP, Auner G, Mao G (2008) Recognition of Salmonella Typhimurium by immobilized phage P22 monolayers. Surf Sci 602:1392–1400 Kuhn J, Suissa M, Chiswell D, Azriel A, Berman B, Shahar D, Reznick S, Sharf R, Wyse J, Bar-On T, Cohen I, Giles R, Weiser I, LubinskyMink S, Ulitzur S (2002) A bacteriophage reagent for Salmonella: molecular studies on Felix 01. Int J Food Microbiol 74:217–227 Lermo A, Fabiano S, Hernández S, Galve R, Marco M-P, Alegret S, Pividori MI (2009) Immunoassay for folic acid detection in vitaminfortified milk based on electrochemical magneto sensors. Biosens Bioelectron 24:2057–2063 Li S, Li Y, Chen H, Horikawa S, Shen W, Simonian A, Chin B a (2010) Direct detection of Salmonella Typhimurium on fresh produce using phage-based magnetoelastic biosensors. Biosens Bioelectron 26: 1313–1319 Liébana S, Lermo A, Campoy S, Cortés MP, Alegret S, Pividori MI (2009) Rapid detection of Salmonella in milk by electrochemical magneto-immunosensing. Biosens Bioelectron 25:510–513

Appl Microbiol Biotechnol (2014) 98:1795–1805 Liébana S, Spricigo DA, Cortés MP, Barbe J, Alegret S, Pividori MI (2013) Phagomagnetic separation and electrochemical magnetogenosensing of pathogenic bacteria. Anal Chem 85:3079–3086 Mao C, Liu A, Cao B (2009) Virus-based chemical and biological sensing. Angew Chem Int Ed Engl 48:6790–6810 Nanduri V, Sorokulova I, Samoylov A, Simonian A, Petrenko V, Vodyanoy V (2007) Phage as a molecular recognition element in biosensors immobilized by physical adsorption. Biosens Bioelectron 22:986–992 Neufeld T, Schwartz-Mittelmann A, Biran D, Ron EZ, Rishpon J (2003) Combined phage typing and amperometric detection of released enzymatic activity for the specific identification and quantification of bacteria. Anal Chem 75:580–585 Sambrook S, Russell DW (1989) Molecular cloning: a laboratory manual, 2nd edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, pp 2.73–2.76 Shabani A, Zourob M, Allain B, Marquette C, Lawrence MF, Mandeville R (2008) Bacteriophage-modified microarrays for the direct impedimetric detection of bacteria. Anal Chem 80:9475–9482 Singh A, Poshtiban S, Evoy S (2013) Recent advances in bacteriophage based biosensors for food-borne pathogen detection. Sensors 13: 1763–1786 Smartt AE, Ripp S (2011) Bacteriophage reporter technology for sensing and detecting microbial targets. Anal Bioanal Chem 400:991–1007 Starodub NF, Ogorodnijchuk JO (2012) Immune biosensor based on the ISFETs for express determination of Salmonella Typhimurium. Electroanalysis 24(3):600–606 Steinbacher S, Miller S, Baxa U, Budisa N, Weintraub A, Seckler R, Huber R (1997) Phage P22 tailspike protein: crystal structure of the head-binding domain at 2.3 Å, fully refined structure of the

1805 endorhamnosidase at 1.56 Å resolution, and the molecular basis of o-antigen recognition and cleavage. J Mol Biol 267:865–880 Theron J, Eugene Cloete T, De Kwaadsteniet M (2010) Current molecular and emerging nanobiotechnology approaches for the detection of microbial pathogens. Crit Rev Microbiol 36:318–339 Tolba M, Minikh O, Brovko LY, Evoy S, Griffiths MW (2010) Oriented immobilization of bacteriophages for biosensor applications. Appl Environ Microbiol 76:528–535 Upmann M, Bonaparte C (2000) Rapid methods for food hygiene inspection. In: Robinson RK, Batt CA, Patel PD (eds) Encyclopedia of food microbiology. Academic Press, Oxford, pp 1887–1895 Van Dorst B, Mehta J, Bekaert K, Rouah-Martin E, De Coen W, Dubruel P, Blust R, Robbens J (2010) Recent advances in recognition elements of food and environmental biosensors: a review. Biosens Bioelectron 26:1178–1194 Velusamy V, Arshak K, Korostynska O, Oliwa K, Adley C (2009) An overview of foodborne pathogen detection: in the perspective of biosensors. Biotechnol Adv 28:232–254 Wang H, Li Y, Wang A, Slavik M (2011) Rapid, sensitive, and simultaneous detection of three foodborne pathogens using magnetic nanobead-based immunoseparation and quantum dot-based multiplex immunoassay. J Food Prot 74(12):2039–2047 Wolber PK (1993) Bacterial ice nucleation. Adv Microb Physiol 34:203– 237 Zacco E, Pividori MI, Alegret S, Galve R, Marco M-P (2006) Electrochemical magnetoimmunosensing strategy for the detection of pesticides residues. Anal Chem 78:1780–1788 Zourob M, Ripp S (2010) Bacteriophage-based biosensors. In: Zourob M (ed) Recognition receptors in biosensors. Springer, New York, pp 415–448

Phagomagnetic immunoassay for the rapid detection of Salmonella.

This work explores the use of the phage P22 in a phagomagnetic immunoassay for the rapid detection of Salmonella. The covalent attachment of wild-type...
625KB Sizes 0 Downloads 0 Views