Biosensors and Bioelectronics 19 (2004) 1331–1335

Short communication

A generic approach for the detection of wholeListeria monocytogenescells in contaminated samples using surface plasmon resonance

Paul Leonarda,b, Stephen Heartya,b, John Quinna,1, Richard O’Kennedya,b,∗a School of Biotechnology, Dublin City University, Glasnevin, Dublin 9, Ireland

b National Centre for Sensor Research (NCSR), Dublin City University, Dublin 9, Ireland

Received 26 May 2003; received in revised form 26 September 2003; accepted 18 November 2003

Abstract

The opportunistic food pathogenListeria monocytogenes is of great concern to the food industry and its rapid detection is of majorimportance. This paper describes the detection ofL. monocytogenes with a polyclonal antibody by means of a new subtractive inhibitionassay using a BIAcore 3000 biosensor. IncubatingL. monocytogenes cells and antibody for a short period of time, followed by subsequentseparation of free unbound antibody with a stepwise centrifugation process, allowed the detection of 1× 105 L. monocytogenes cells/ml inless than 30 min. Free antibody was passed over an anti-Fab ligand-coated sensor chip surface with the generated response being inverselyproportional to the inhibiting cell concentration. The method was simple, rapid and needed minimum sample preparation. This assay formathas the potential for the quick and sensitive detection of pathogens with limited sample handling and preparation.© 2003 Elsevier B.V. All rights reserved.

Keywords: Surface plasmon resonance (SPR); Subtractive inhibition assay;Listeria monocytogenes

1. Introduction

Since its discovery in 1926 (Murray et al., 1926), Lis-teria monocytogenes has been regarded as an importantfood borne pathogen and was the cause of many recentwell-publicised food poisoning outbreaks (Schlech, 2000;Donnelly, 2001). L. monocytogenes is a gram-positive fac-ultatively anaerobic rod-shaped bacterium that grows be-tween 1 and 45◦C. Optimal growth occurs between 30 and37◦C within the pH range of 5.0–9.0 in a NaCl concen-tration of 10% (w/v) (Jones and Seeliger, 1992). Pregnantwomen and newborns, elderly people and people with weak-ened or suppressed immune systems caused by chemother-apy, AIDS, diabetes and kidney disease are especially at riskto listeriosis. Indeed, immunocompromised patients can beup to 500–1000 times more susceptible to listeriosis thanthe general population (Jensen et al., 1994; Hof, 2003). Inthe majority of cases, infection byL. monocytogenes re-sults in symptoms such as mild fever, diarrhoea, nausea orvomiting accompanied by a slight headache. However, in

∗ Corresponding author. Tel.:+353-1-7005319; fax:+353-1-7005412.E-mail address: richard.okennedy@dcu.ie (R. O’Kennedy).1 Texas Instruments Inc., Austin, Texas, USA.

more severe casesListeria infection can cause septicemia,meningitis and meningo-encephalitis. Listeriosis can alsocause stillbirths and spontaneous abortions (Doganay, 2003).Given the severity ofListeria infections and its ability togrow at refrigeration temperatures, the US FDA has adopteda ‘zero-tolerance’ policy for the presence of the bacteriumin food and with the use of HACCP strategies aims to ef-fectively eliminateL. monocytogenes from food process-ing environments and, therefore, processed food products(Donnelly, 2001). The critical issue facing the implemen-tation of any “zero-tolerance” policy relates to the lack ofreliable procedures for the detection of low numbers ofLis-teria in foods. Standard methods such as the ISO method11290 for the detection and enumeration ofL. monocyto-genes in foods, can take up to 1 week for confirmation ofresults (Scotter et al., 2001a,b).

Over the last decade, a great deal of research has cen-tred on the development of sensors for the detection of mi-croorganisms, allowing rapid and “real-time” identificationof pathogens (Leonard et al., 2003). Biosensors are analyti-cal instruments that use a combination of biological receptorcompounds (antibody, enzyme, nucleic acid, etc.) and thephysical or physicochemical transducer directing, in manycases, “real-time” observation of a specific biological event

0956-5663/$ – see front matter © 2003 Elsevier B.V. All rights reserved.doi:10.1016/j.bios.2003.11.009

1332 P. Leonard et al. / Biosensors and Bioelectronics 19 (2004) 1331–1335

(Invitski et al., 1999; Fitzpatrick et al., 2000). The BIA-core biosensor is a commercial biosensor based on the phe-nomenon of surface plasmon resonance (SPR) (Quinn andO’Kennedy, 1999), allowing the “real-time” detection ofbiomolecular interaction analysis (Jönsson et al., 1991). InSPR, energy carried by photons of light can be “coupled” ortransferred to electrons in a metal (usually gold or silver).When a transfer of energy occurs, all the light at most wave-lengths is reflected except at the resonant wavelength, wherea distinct minimum reflectivity is reported. Changes in massat the surface (e.g. antibody–antigen interaction) can causea change in resonant wavelength or angle, which can be de-tected in “real-time”. Biosensors based on SPR have provenuseful for diagnostic, environmental and food safety analysisand have being used to detect a range of molecules such aswhole cells (Fratamico et al., 1998; Quinn and O’Kennedy,2001), toxins (Daly et al., 2000; Rasooly, 2001; Naimushinet al., 2002) and small analytes such as illicit drugs (Dillonet al., 2003) and steroids (Fitzpatrick et al., 2003).

The objective of this study was to develop a new“real-time” subtractive inhibition assay to detect low num-bers ofL. monocytogenes cells in solution using a BIAcore3000 biosensor. By removing cells and bound antibody fromsolution with a stepwise centrifugation step, free unboundantibody could be detected by an immobilised anti-Fab anti-body on the sensor chip surface. A decrease in free antibodyconcentration was observed with increasingL. monocyto-genes cell concentrations. Removing the cells and boundantibody by gradually increasing the centrifugation speedwas the key to the success of the approach. Centrifugationat 1800× g without the gradual increase in speed (resultsnot shown) seemed to shear the antibodies from the cells,possibly due to the high centrifugal forces, resulting in noinhibition of antibody being observed. Removing cells andantibody by centrifugation as described in this paper, is rel-atively rapid, is easy to perform and will allow the detectionof cells in more complex and viscous matrices which maycause problems for other reported methods of cell removalsuch as filtration (Haines and Patel, 1995). In this paper apolyclonal rabbit anti-wholeL. monocytogenes cell anti-body was mixed with various concentrations of heat-killedL. monocytogenes cells and free antibody detected using apolyclonal goat anti-rabbit Fab antibody immobilised onthe sensor chip surface.

2. Materials and methods

TheL. monocytogenes cells (L. monocytogenes 4b (NCTC4885)) used in this work were donated by Dr. Gary Wyatt,Institute of Food Research, Norwich, UK. Cells were dilutedin phosphate-buffered saline solution (PBS-pH 7.3, 0.15 MNaCl) and mixed with antibody.

Antibody was purified from serum taken from an adultNew Zealand white female rabbit immunised with wholeheat-killed (76◦C for 20 min) L. monocytogenes cells.

Briefly, the immunoglobulin G (IgG) fraction of the serumwas purified by precipitation with 45% (w/v) saturatedammonium sulphate, followed by protein G-affinity chro-matography. Antibody purity was assayed by SDS-PAGEcharacterisation and antibody titre (1/1,000,000) determinedby ELISA.

Analysis was carried out on a BIAcore 3000TM instrumentusing a CM 5 sensor chip. HBS (10 mM HEPES, 150 mMNaCl, 3.8 mM EDTA, 0.05% (v/v) Tween) was used as arunning buffer, which was filtered (0.02�M Watman filters)and degassed (Milipore sintered glass funnel) prior to use.

The CM dextran surface was activated by injecting 50�lof 100 mM NHS mixed with 50�l of 400 mM EDC at10�l/min for 8 min. A 50�g/ml solution of commercialgoat anti-rabbit (Fab portion) polyclonal antibody (Fitzger-ald Industries International Inc., Concord, MA 01742-3049,USA) in 10 mM acetate buffer, pH 4.5, was injected over thesurface at 2�l/min for 15 min. Unreacted sites were subse-quently deactivated by injecting 1 M ethanolamine, pH 8.5,at 10�l/min for 7 min.

To assess the stability of immobilised anti-Fab antibodysurface, a known concentration of anti-L. monocytogenesantibody was passed over the chip surface, and the surfaceregenerated by passing over 30 s pulses of 20 mM HCl and20 mM NaOH. This cycle of regeneration was completed forgreater than 32 cycles, and the binding signal measured toassess the stability and suitability of the immobilised surfacefor use in an assay to detectL. monocytogenes.

Approximately 10,000 RUs of anti-Fab antibody was im-mobilised on a CM5 sensor chip surface. 300�l of a 1/250dilution of protein-G purified anti-L. monocytogenes wholecell polyclonal antibody solution diluted in ultra pure water(Millipore) was added to 300�l of decreasingL. monocyto-genes cells diluted in PBS. Each mixture was incubated for20 min at 37◦C, inverting the mixture occasionally to allowgood mixing of cells and antibodies. After incubation, thecells were centrifuged in a stepwise fashion separating cellsand bound antibody from unbound antibody in the super-natant. Cells were centrifuged for 1 min intervals at 50, 200,450, 800, 1200, 1800 and 3200× g. Gradually increasingthe centrifugation speed seemed less abrasive on the anti-bodies bound to cells, forming pelleted layers in which theantibodies would not dissociate from the cells. Centrifuga-tion at higher speeds seemed to shear the antibodies from thecells resulting in no inhibition being observed. The super-natants from each sample were then randomly passed overthe sensor surface in triplicate and an average value taken.

3. Results and discussion

Regeneration conditions affect the performance and life-time of the sensor chip surface and therefore optimumregeneration is essential for the analysis of large numbersof samples, to evaluate the reproducibility of each mea-surement and to reduce the cost of the assay. To determine

P. Leonard et al. / Biosensors and Bioelectronics 19 (2004) 1331–1335 1333

27600

27700

27800

27900

28000

28100

28200

28300

28400

650 700 750 80 850Time (Secs)

Res

pons

e U

nits

.

30900

30950

31000

31050

31100

31150

31200

31250

31300

825 850 875 925

Time (Secs)R

espo

nse

Uni

ts.

27000

27500

28000

28500

29000

29500

30000

30500

31000

31500

600 700 800 900 1000

Time (Secs)

Res

pons

e U

nits

.s

I II

900

Fig. 1. Overlay plot demonstrating the decrease in binding of affinity purified anti-Listeria monocytogenes polyclonal antibody to the immobilisedanti-Fab antibody surface with increasingL. monocytogenes cell concentrations. Plot (I) shows the binding of increasing free antibody with decreasingL. monocytogenes cell concentrations, while plot (II) shows the base line shift when buffer is passed over the surface after each sample injection. Curvesmarked A–F represent antibody mixed with 0, 1× 105, 1× 106, 1× 107, 1× 108 and 1× 109 L. monocytogenes cells/ml, respectively.

that the immobilised anti-Fab antibody sensor chip surfacecould be regenerated without loss of binding response, a1/500 dilution anti-L. monocytogenes antibody in PBS waspassed over the surface for 3 min at 10�l/min and thesurface regenerated with 30 s pulses of 20 mM HCl and20 mM NaOH. A binding response of approximately 360RU was observed for each binding cycle with less than 2%reduction in binding capacity observed over the course of32 binding-regeneration cycles studied. The immobilisedanti-Fab surface went through a further 120 binding cyclesover a 2-week period before a significant (<20% originalbinding capacity) decrease in binding capacity was observed.

In the present study, the applicability of using a subtractiveinhibition assay to detect low numbers ofL. monocytogenescells in “real-time” was assessed. An anti-L. monocytogenespolyclonal antibody was mixed with decreasing concen-trations of L. monocytogenes cells and unbound antibodyseparated from bound antibody by centrifugation. Super-natants from each sample containing unbound antibody wasassayed in triplicate using the BIAcore 3000 biosensor andthe binding responses and baseline shifts observed (Fig. 1).The average response for each sample was divided by theaverage of the maximum response (sample with no cells)and the normalised values plotted againstL. monocytogenescell concentrations (Fig. 2). A significant difference inbinding response was observed for each cell concentration(Fig. 1 (II)) and a limit of detection (the concentration atwhich no further inhibition was apparent) of approximately1×105 L. monocytogenes cells/ml obtained (Fig. 2) which iscomparable with ELISA for the same antibody (results notshown). Mixing 10-fold decreasing dilutions ofL. monocy-

togenes cells in PBS with anti-L. monocytogenes antibodyresulted in increased binding of free antibody to the anti-Fabantibody immobilised surface (195, 297, 346, 376 and 439RU observed from supernatants from samples incubatedwith 109, 108, 107, 106 and 105 L. monocytogenes cells,respectively).

Several factors have to be taken into account when devel-oping a SPR-based assay to detect whole pathogens in so-lution. The detection system of a SPR biosensor essentiallyconsists of a monochromatic and p-polarised light source, aglass prism, a thin metal film in contact with the base of theprism, and a photodiode array. When the surface plasmonsbecome excited at the metal–liquid interface, an evanescentelectromagnetic field is formed. This field exponentially de-cays from the metal film surface into the interfacing medium(Stenberg et al., 1990). The effective penetration depth ofthe evanescent field which arises under conditions of totalinternal reflection is approximately 300 nm. Therefore, onlyrefractive index changes occurring within this distance fromthe surface will cause a change in the generated SPR signal.Bacteria such asL. monocytogenes probably do not pene-trate the dextran layer that coats the gold surface and there-fore only a small portion of the cell which is in close contactwith the sensor surface will produce a measurable signal. Inaddition, the BIAcore biosensor measures an average SPRangle over an area of approximately 0.25 mm2 of the sensorsurface, and therefore since bacterial cells are large andmay not evenly cover an area measured, the signal responseis decreased (Fratamico et al., 1998). Enhancement of thegenerated SPR signal by passing free antibody over the cellcaptured surface in a sandwich format or by immobilising

1334 P. Leonard et al. / Biosensors and Bioelectronics 19 (2004) 1331–1335

0.4

0.5

0.6

0.7

0.8

0.9

1

1.1

1.E+04 1.E+05 1.E+06 1.E+07 1.E+08 1.E+09 1.E+10

Cell Concentration (Cells/ml)

R/R

0

Fig. 2. The SPR response observed by binding between immobilised anti-Fab polyclonal antibody and free anti-Listeria monocytogenes polyclonalantibody. The results shown are the average of triplicate results and the range of detection ofL. monocytogenes cells was found to be between 1× 105

and 1× 109 cells/ml. The binding response at each cell concentration (R) was divided by the antibody binding response determined in the presence ofzero L. monocytogenes cells (R0) to give a normalised binding response (R/R0).

layers of capture antibody on the surface have proven toincrease the detection limits of two SPR biosensors for thedetection ofE. coli O157:H7 ((5–7) × 107 cells/ml) andL.monocytogenes (1 × 106 cells/ml), respectively (Fratamicoet al., 1998; Koubavá et al., 2001). However, cell bindingrequires bulk transport of the cells to the interaction surfaceand the generation of sufficient binding avidity to overcomehydrodynamic resistive forces. The assay format for thedetection ofL. monocytogenes presented in this paper canbe used to detect any pathogen, simplifies sample handlingin comparison to direct detection methods and may alsobe enhanced with the use of a second injection of suitableantibody (e.g. mouse anti-rabbit antibody) and further opti-misation. The assay took less than 30 min from incubationof cells to the generation of a binding response. The wholeprocedure is simple, automated (BIAcore control softwareand automated sample handling) and relatively sensitive(1 × 105 cells/ml, which was comparable with ELISA).Hence, the assay format could be used for the routine analy-sis of pathogens with minimal sample preparation required.

4. Conclusion

SPR biosensors are potentially very useful for food safetyanalysis as they are sensitive, real-time interactive devicesthat can allow the timely response to food borne pathogen ortoxin contamination, as well as permitting the rapid detectionof trace drug and hormone residues in food. The assay formatpresented in this paper offers a new simple method for thedetection of pathogens using SPR. Various concentrations

of the well known food pathogen,L. monocytogenes, weremixed with an anti-L. monocytogenes polyclonal antibodyand unbound antibody detected by immobilised anti-Fab lig-and using a BIAcore 3000 biosensor. Using a stepwise cen-trifugation process, cells and bound antibody were separatedfrom unbound antibody allowing the rapid detection of 1×105 L. monocytogenes cells/ml. The assay format describedin this paper can be used for the detection of any pathogenwith appropriate antibody and has the potential to allow thesimple detection of pathogens in complex food matrices.

Acknowledgements

The authors would like to thank INCO COPERNICUSproject No. PL979012 and Enterprise Ireland for fundingthis project. We also wish to thank Gary Wyatt (Institute ofFood Research, Norwich), Professor Pavel Raunch, Lud-mila Karasová (ICT, Prague, Czech Republic), ProfessorKatarına Horáková, Vladimır Mastihuba, Mária Greifová,Andrea Šovèiková (Slovak Technical University), PetrRoubal (MILCOM, Dairy Research Institute, Czech Repub-lic) and Martin Tomáška (Dairy Research Institute, Žilina,Slovakia) for their various contributions in this research.

References

Daly, S.J., Keating, G.J., Dillon, P.P., Manning, B.M., O’Kennedy, R.,Lee, H.A., Morgan, M.R., 2000. Development of a surface plasmon

P. Leonard et al. / Biosensors and Bioelectronics 19 (2004) 1331–1335 1335

resonance-based immunoassay for Aflatoxin B(1). J. Agric. FoodChem. 48, 5097–5104.

Dillon, P.P., Daly, S.J., Manning, B.M., O’Kennedy, R., 2003. Im-munoassay for the determination of morphine-3-glucuronide usinga surface plasmon resonance-based biosensor. Biosens. Bioelectron.18, 217–227.

Doganay, M., 2003. Listeriosis: clinical presentation. FEMS Immunol.Med. Microbiol. 1485, 1–3.

Donnelly, C.W., 2001.Listeria monocytogenes: A continuing challenge.Nutr. Rev. 59, 183–194.

Fitzpatrick, J., Fanning, L., Hearty, S., Leonard, P., Manning, B.M., Quinn,J.G., O’Kennedy, R., 2000. Applications and recent developments inthe use of antibodies for analysis. Anal. Lett. 33 (13), 2563–2609.

Fitzpatrick, J., Manning, B.M., O’Kennedy, R., 2003. Development ofELISA and sensor-based assays for the detection of ethynyl estradiolin bile. Food Agric. Immunol. 15, 55–64.

Fratamico, P.M., Strobaugh, T.P., Medina, M.B., Gehring, A.G., 1998.Detection of Escherichia coli O157:H7 using a surface plasmonresonance biosensor. Biotechnol. Tech. 12, 571–576.

Haines, J., Patel, P.D. 1995. Detection of food borne pathogens usingBIA. Biacore AB Biotechnology note, p. 31.

Hof, H., 2003. History and epidemiology of listeriosis. FEMS Immunol.Med. Microbiol. 1489, 1–4.

Invitski, D., Abdel-Hamid, I., Atanasov, P., Wilkins, E., 1999. Biosensorsfor the detection of pathogenic bacteria. Biosens. Bioelectron. 14,599–624.

Jensen, A., Frederiksen, W., Gerner-Smidt, P., 1994. Risk factors for Lis-teriosis in Denmark, 1989–1990. Scand. J. Infect. Dis. 26, 171–178.

Jones, D., Seeliger, H., 1992. The genus Listeria. In: C.A. Balows, H.G.Truper, M. Dworkin, W. Harder, K-H. Schleur, (Eds.), The Prokary-otes, 2nd Edition. Springer Verlag, Heidelberg, pp. 1595-1616.

Jönsson, U., Fägerstam, L., Ivarsson, B., Lundh, K., Löfås, S., Persson,B., Roos, H., Rönnberg, I., Sjölander, S., Stenberg, E., Stånhlberg,R., Urbaniczky, C., Östlin, H., Malmqvist, M., 1991. Real-timebiospecific interaction analysis using surface plasmon resonance anda sensor chip technology. Biotechniques 11, 620–627.

Koubavá, V., Brynda, E., Karasová, L., Škvor, J., Homola, J., Dostálek, J.,Tobiška, P., Rošický, J., 2001. Detection of foodborne pathogens usingsurface plasmon resonance biosensors. Sens. Actuators B 74, 100–105.

Leonard, P., Hearty, S., Brennan, J., Dunne, L., Quinn, J., Chakraborty,T., O’Kennedy, R., 2003. Advances in biosensors for the detectionof pathogens in food and water. Enzyme Microb. Technol. 32, 3–13.

Murray, E.G.D., Webb, R.A., Swann, H.B.R., 1926. A disease of rab-bits characterized by a large mononuclear leucocytosis caused by ahitherto undescribed bacillusBacterium monocytogenes (n. sp.). J.Pathol. Bacteriol. 29, 407–439.

Naimushin, A.N., Soelberg, S.D., Nguyen, D.K., Dunlap, L., Bartholomew,D., Elkind, J., Melendez, J., Furlong, C.E., 2002. Detection ofStaphylococcus aureus entertoxin B at femtomolar levels with aminiature integrated two-channel surface plasmon resonance (SPR)sensor. Biosens. Bioelectron. 17, 573–584.

Quinn, J.G., O’Kennedy, R., 1999. Transduction platforms and bioint-erfacial design of biosensors for “real-time” biomolecular interactionanalysis. Anal. Lett. 32 (7), 253–258.

Quinn, J.G., O’Kennedy, R., 2001. Detection of whole cell:antibodyinteractions using Biacore’s SPR technology. BIA J. 1, 22–24.

Rasooly, A., 2001. Surface plasmon resonance analysis of staphylococcalenterotoxin B in food. J. Food Prot. 64, 37–431.

Schlech III, W.F., 2000. Foodborne listeriosis. Clin. Infect. Dis. 31,770–775.

Scotter, S.L., Langton, S., Lombard, B., Schulten, S., Nagelkerke, N.,In’t Veld, P.H., Rollier, P., Lahellec, C., 2001a. Validation of ISOmethod 11290 Part 1-Detection ofListeria monocytogenes in foods.Int. J. Food Microbiol. 64, 295–306.

Scotter, S.L., Langton, S., Lombard, B., Lahellec, C., Schulten, S.,Nagelkerke, N., In’t Veld, P.H., Rollier, P., 2001b. Validation ofISO method 11290 Part 2-Enumeration ofListeria monocytogenes infoods. Int. J. Food Microbiol. 70, 121–129.

Stenberg, E., Persson, B., Roos, H., Urbaniczky, C., 1990. Quantitativedetermination of surface concentration of protein with surface plas-mon resonance using radiolabeled proteins. Colloid Interf. Sci. 143,513–526.