Sensors and Actuators B 96 (2003) 372–378 Development of a novel FRET method for detection of Listeria or Salmonella Sungho Ko, Sheila A. Grant∗ Department of Biological Engineering, 250 Ag. Engineering Building, University of Missouri, Columbia, MO 65202, USA Received 28 March 2003; received in revised form 26 June 2003; accepted 1 July 2003 Abstract Listeria and Salmonella contamination in foods results in not only foodborne outbreaks, but also a large economical burden for the industry due to product recalls. As a result, a fluorescence resonance energy transfer (FRET)-based method was developed to detect Listeria and Salmonella. Sensors utilizing FRET switch their fluorescence wavelength between donor and acceptor fluorophores as the distance between the two fluorophores change. This change in distance is a result of the conformational change in the 3D structure of the antibody as it binds to the target antigen. Listeria or Salmonella antibodies were labeled with FRET donor fluorophores (Alexa Fluor 546 (AF546)) while protein G or A was labeled with the acceptor fluorophores (Alexa Fluor 594 (AF594)). The labeled antibody–protein G or A complex was formed via incubation of the labeled antibody with protein G/A which specifically attaches to the Fc fragment of antibodies. The labeled antibody–protein G or A complex was tested in solution and specific and nonspecific antigens were exposed to the in solution complex. Changes in fluorescence were monitored by a spectrofluorometer. For “in-solution” tests, the optimal acceptor/donor fluorophore (A/D) ratios were 0.2 and 1.0 for Listeria and Salmonella, respectively, and both of Listeria and Salmonella antigen detection limits were 2.0�g/ml. © 2003 Elsevier B.V. All rights reserved. Keywords: FRET; Listeria; Salmonella; Sensor; Antibody; Fluorophores 1. Introduction About 33 million cases of foodborne diseases, resulting in more than 5 billion dollars for treatment and about 9000 deaths occur in the US each year [1]. Listeria and Salmonella have been the main organisms causing the outbreaks of food- borne illnesses. Conventional detection methods, requiring 3–4 days for presumptive results and 5–7 days for confir- mation, are cumbersome and time-consuming. Although the Hazard Analysis Critical Control Point (HACCP) system is a popular approach to preventing foodborne disease, rapid, accurate, and real-time methods for detecting pathogens in food processing facilities are urgently needed. The poly- merase chain reaction (PCR) technology has proved to be a powerful method for the detection of pathogens in foods [2–5], but dot-blot hybridizations which prove the presence of specifically amplified products still require multistep proccessings and add considerable time and expense to the detection [6]. Although enzyme-linked immunosorbent ∗ Corresponding author. Tel.: +1-573-884-9666; fax: +1-573-882-1115. E-mail address:
[email protected] (S.A. Grant). assay (ELISAs) have been developed to detect Salmonella and Listeria in food products, most require time-consuming pre-enrichment steps [7,8]. Immunosensors that detect the binding between antibody and antigen have been developed for pathogens but still require long reaction times, multistep processing, and/or resulted in false positives [9,10]. While great strides have been made in biosensor tech- nology, many technological problems that hinder their widespread use for monitoring food safety remain, such as those listed above. In this study, the proposed biosensor takes advantage of the inherent conformation changes that occur when antibodies and antigens combine. Combining this with the basic principles of fluorescence resonance energy transfer (FRET), only viable analytes will induce a conformational change, and thus will be detected in real-time; drastically reduced false positives. Researchers have utilized FRET to detect biological agents. FRET probes were used to detect pagA gene in the causative agent of anthrax [11], and by analyzing FRET pairs a dual receptor detection method for enhanced biosen- sor monitoring was investigated [12]. Sensors utilizing FRET switch their fluorescence wavelength between donor and acceptor fluorophores as distance between the two 0925-4005/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0925-4005(03)00572-0 S. Ko, S.A. Grant / Sensors and Actuators B 96 (2003) 372–378 373 acceptor labeled Protein G donor labeled antibody specific antigens nonspecific antigens(BSA) donor fluorophore acceptor fluorophore λ : wavelength Fig. 1. Schematic of the FRET immunosensor showing the Y-shaped antibody labeled with the donor fluorophores and attached to acceptor labeled protein G. fluorophores change. This distance change is a result of the conformational changes in the 3D structure of an antibody as it binds to the target antigen [13]. The FRET process requires a donor fluorophore and a separate acceptor fluorophore. The acceptor fluorophore must have an excitation that spec- trally overlaps the donor’s emission spectrum but not the donor’s excitation band [14,15]. The basic principle is a transfer of energy from an excited donor fluorophore (λ0) to an acceptor fluorophore nonradiatively, which occurs when the two fluorophores are in close vicinity ( 374 S. Ko, S.A. Grant / Sensors and Actuators B 96 (2003) 372–378 2.3. Determination of the ratio of moles of fluorophore to moles of protein The degree of labeling was determined as the ratio of moles of fluorophore to moles of protein (F/P). The F/P ra- tio was determined for each fluorophore–protein conjuga- tion. The F/P ratio was calculated using known extinction coefficients for the fluorophores and Eq. (1): F P = Amax × dilution factor εdye × [M] (1) where Amax is the absorbance at the maximum exci- tation wavelength of the fluorophore used in conjuga- tion, and εdye the molar extinction coefficient at λmax in cm−1 M−1 (εAF546 = 104,000 cm−1 M−1 and εAF594 = 73,000 cm−1 M−1, according to the manufacture’s proce- dure sheet) and [M] the molarity of the protein. The F/P ratio obtained was used to determine an optimal acceptor to donor dye ratio. 2.4. Determination of the optimal acceptor to donor fluorophore ratio The optimal acceptor to donor fluorophore (A/D) ra- tio was derived from the number of labeled antibody molecules incubated with the number of labeled protein A or G molecules and the F/P ratio obtained. The A/D ra- tio can be defined as the number of acceptor fluorophores to donor fluorophores in an antibody–protein complex. To determine the optimal A/D ratio, each of complexes of LisAb-AF546/PA-AF594, LisAb-AF546/PG-AF594, or SalAb-AF546/PG-AF594 was prepared by co-incubation at various concentration ratios of labeled antibodies and labeled protein A or G overnight at 4 ◦C and then scanned using the spectrofluorometer, Jasco FP-750. Determining the optimal A/D ratio is critical for successful energy trans- fer between acceptor and donor dyes. Too little or too much of either fluorophore can produce nondetectable signals due to self-quenching and/or insufficient energy transfer. The optimal ratios were determined by examining the donor and acceptor emission peaks. As an example, the A/D ra- tio determination for the SalAb-AF546/PG-594 complex: SalAb-AF546 (0.045�g/�l) was varied from 0.2 to 0.25�g and was mixed with 0.8�g of PG-AF594 (0.053�g/�l) in a microcentrifuge tube. Each aliquot, examining a differ- ent A/D ratio, was incubated overnight at 4 ◦C in the dark and then placed in a 1 cm × 1 cm (4 ml volume) cuvette. The solution was not filter to remove incomplete complex formations since there was excess acceptor labeled protein G. An excess of the acceptor-protein G ensured that the donor labeled antibodies would be complexed within com- prising the signal. It was then diluted with 3 ml of PBS, stirred, and scanned with the spectrofluorometer set at the excitation wavelength of the donor (546 nm for AF546). The A/D ratio varied from 1 to 8. For all other com- plexes examined, the A/D ratio was determined similarly to SalAb-AF546/PG-AF594 complex. 2.5. In-solution detection of antigen with the labeled antibody–protein complex A/D ratios that showed acceptable energy transfer were utilized to determine if the “in solution” FRET technique can specifically distinguish the presence of specific and nonspecific antigens. Each of the mixed solutions (LisAb- AF546/PA-AF594, LisAb-AF546/PG-AF594, or SalAb- AF546/PG-AF594) with the labeled proteins and antibodies were prepared based on their optimal A/D ratios by in- cubation overnight at 4 ◦C in dark. For each complex, a dose of 50�l was carefully distributed into seven semi-microcuvettes. One cuvette was used for control (no antigen) and the other six cuvettes were divided into two groups of three; one group was utilized for the specific antigen (Listeria antigen (LisAg) or Salmonella antigen (SalAg)) and the other group was utilized for the nonspecific antigen (BSA). A 50�l dose of LisAg or SalAg was added into the three cuvettes using various concentrations of 0.04, 0.02, and 0.002�g/�l, respectively. For the nonspecific, BSA, 50�l of BSA was added at various concentrations of 0.04, 0.02, and 0.002�g/�l, into each of three cuvettes for the Listeria or Salmonella tests. The final volume in each semi-microcuvettes was 100�l. The solutions were gently shaken for several minutes to ensure that the solutions were well mixed and then incubated for 1 h at room temperature. After incubation, each sample was diluted and mixed for 1 min with PBS to 1 ml to fill a semi-microcuvette prior to scanning with the spectrofluorometer. The optical path length was 0.5 cm and the solution was not mixed during scanning. The excitation wavelength was set according to the donor fluorophore present in the experiment (546 nm for AF546/AF594). The change in fluorescence of the donor and acceptor fluorophores was recorded by spectrofluorometer. 2.6. Statistical analysis The experiments to examine the response of the values as an indication of FRET activity were repeated three times to determine the reproducibility of the results. The means and standard deviation of the resulting values from all samples were calculated by using the General Linear Models proce- dure of SAS (SAS Institute, 1982, Statistical Analysis Sys- tem, Cary, NC). Significance was expressed at the 5% level. 3. Results and discussion 3.1. Determining F/P ratio The determination of F/P ratio is important in order to ob- tain efficient energy transfer while preserving the biological activity. The resulting F/P ratio for each of the fluorophores S. Ko, S.A. Grant / Sensors and Actuators B 96 (2003) 372–378 375 Table 1 F/P ratios of conjugated proteins and antibodies Antibodies-fluorophores F/P ratio LisAb-AF546 2.23 SalAb-AF546 6.8 Protein G-AF594 5.8 conjugated to the proteins and antibodies is shown in Table 1. There was a F/P ratio of 2.23 for LisAb-AF546, 6.8 for SalAb-AF546, and 5.8 for protein G-AF594, re- spectively. A number of fluorophores might randomly bind near the Fc region of the antibodies as well as near the Fab region, which is the preferred site for the FRET technique. At a high F/P ratio, a loss in fluorescence is observed. This may be due to the binding of fluorophores near the Fc region of the antibody and the subsequent interaction between the fluorophores and fluorescently labeled protein A or G. Addi- tionally, resonance energy transfer between the fluorophores on the antibody could result in fluorescence quenching, and the nonfluorescent state caused by the fluorophore interac- tions can cause some additional reduction in fluorescence [17–19]. These interactions also contribute to background signals. On the other hand, a low F/P ratio may result in a corresponding low background signal but a nondetectable FRET signal may result. The F/P ratios obtained (Table 1) in this study were suitable to detect Lis and Sal antigens as shown in the next experimental step. It was essential for immunofluorescence to achieve intense fluorescence per labeled antibody while retaining specific binding of antigen. 3.2. Determination of an optimal acceptor to donor ratio Fig. 2 shows the emission scans of the complex of LisAb-AF546/PG-AF594 and SalAb-AF546/PG-AF594 for determination of an optimal acceptor to donor (A/D) ratio. The optimal A/D ratios were experimentally determined by comparing the FRET activity of several molar ratios of LisAb-Alexa546 and SalAb-Alexa594, respectively, in solution. The A/D ratios were then derived from the F/P ra- tios and the amount of proteins present in solution. An A/D ratio of 1.6 and 8.0 in Listeria and Salmonella complex, respectively, resulted in little donor fluorescence. Therefore, the electron transmission to the acceptor emission peak was too weak to be detected. The intensity of the donor fluorescence increased as the A/D ratio decreased from 1.6 to 0.2 in Listeria complex (Fig. 2a) and from 8.0 to 1.0 in Salmonella complex (Fig. 2b). Optimal A/D ratio of 0.2 for Listeria complex and 1.0 for Salmonella complex was determined in this study by the judicious choice of an A/D ratio displaying a donor peak that did not overwhelm the acceptor peak and had a higher intensity compared to the acceptor peak. Ideally, the intensity of emission peak of ac- ceptor fluorophore should not exit in the absence of specific antigen. However, in both the Listeria and Salmonella tests, some background FRET activity occurred since the labeled 0 50 100 150 200 570 590 610 630 650 nm In te n si ty A/D=0.2 A/D=0.4 A/D=0.8 A/D=1.6 A, only D, only PBS 0 5 10 15 20 25 30 35 40 45 50 570 590 610 630 650 nm In te n si ty A/D=1.0 A/D= 2.0 A/D= 4.0 A/D=8.0 A, only D, only (a) (b) Fig. 2. Determination of an optimal ratio of acceptor (A) to donor (D) fluorophore in LisAb-AF546/PG-AF594 complex (a) and SalAb-AF546/PG-AF594 complex (b). antibodies and protein G were freely floating in solutions resulting in random interaction of the fluorophores as they combined. Therefore, the spectra of optimal A/D ratios were used to provide a baseline (R1) for the results shown in Fig. 3. 3.3. In-solution detection of Listeria and Salmonella antigens with FRET activity In-solution tests were performed by adding a series of specific and nonspecific antigens to microcurvettes contain- ing the complex of LisAb-AF546/PG-AF594 and SalAb- AF546/PG-AF594 at the optimal A/D ratio. Results from the tests are presented in Fig. 3. Values displayed are the ratios to ratios, which were calculated as follows and monitored with a spectrofluorometer. R1 = I (λ = 570–575 nm) I (λ = 608–613 nm) , with no antigen present (baseline) 376 S. Ko, S.A. Grant / Sensors and Actuators B 96 (2003) 372–378 Fig. 3. Detection of Listeria antigen (A) and Salmonella antigen (B) in so- lutions containing LisAb-AF546/PG-AF594 and SalAb-AF546/PG-AF594 complex, respectively, in the form of R3 with BSA acting as the nonspe- cific. In (B), different letters (a and b) within the same group (concen- tration) indicate significantly different mean R3 values (n = 3, P < 0.05) as shown for the 2 and 4�g/ml concentrations. R2 = I (λ = 570–575 nm) I (λ = 608–613 nm) , with specific or nonspecific antigen present R3 = R2 R1 , the ratio of ratio used to determine change in FRET activity where I (λ = 570–575 nm) was the average fluorescence intensity of the donor fluorophore and I (λ = 608–613 nm) was the average fluorescence intensity of the acceptor. FRET activity was presented by measuring the change of the emission intensity of the donor and acceptor fluorophores in the form of a ratio, R3. The FRET activity is extremely sensitive to distance changes between donor and acceptor fluorophores. Energy transfer dramatically increases when the separation distance between fluorophores is close to the Försters distance (between 10 and 100 angstroms for most FRET pairs) which is defined as the distance where the transfer of energy from donor to acceptor fluorophores is 50% efficient [15]. As the specific Listeria and Salmonella antigens were introduced, the labeled antibody/PG complex reduced the distance between the dye pairs due to the con- formational change of antibody via binding of antigen. The resulting structure enhanced the FRET activity, increasing the emission intensity at I (λ = 608–613 nm) (the acceptor) and reducing the emission intensity at I (λ = 570–575 nm) (the donor), which resulted in decrease of a R2 value whereas R1 remained the same. As a result, the value of R3 decreased, indicating specific binding of the antigens with the antibodies when the specific antigen was introduced into the solution. Conversely, when the nonspecific anti- gen (BSA) was added into solution, there were minimum changes in fluorescence at 570 and 603 nm causing little change in R3, as expected. In Fig. 3A, at a low concentration (less than 2.0�g/ml) of specific Listeria antigens, a signif- icant reduction in R3 was not shown, but there was a sig- nificant difference in R3 values at the antigen concentration of 2.0�g/ml (P < 0.05), which resulted in a detection limit of Listeria antigen of 2.0�g/ml with this experimental set up and fluorophore pair (AF546 and AF594). The detection limit of Salmonella antigen is displayed in Fig. 3B. The spe- cific Salmonella antigen showed significantly larger reduc- tions in R3 values than that of nonspecific antigens (BSA) at the specific antigen concentrations of 2.0 and 4.0�g/ml (P < 0.05), whereas there was no significant reduction of R3 values below Salmonella concentrations of 2.0�g/ml (P < 0.05), indicating the detection limit of Salmonella antigen was also 2.0�g/ml. This result indicated that when the sample was excited at the maximum excitation wave- length of the donor molecule (546 nm), the donor molecule (AF546) donated energy to the acceptor molecule (AF594) as specific antigens were introduced. As a result, the emis- sion peak of donor is reduced, while the emission peak of acceptor increased. The feasibility of the FRET immunosen- sor technique to detect not only Listeria but also Salmonella antigens has been demonstrated. However, it is necessary to increase the sensitivity of this technique because the amount of foodborne pathogens on the food products are expected to nil. For simplicity, the tests performed were conducted in solution which inherently causes a reduced sensitiv- ity since random chance interaction between free-floating proteins and antibodies might occurred by Brownian mo- tion, micro thermal currents in the cuvette, and/or subtle difference in the path length of each cuvette. Therefore, the sensitivity is reduced. Immobilization onto optical waveguides would reduce random interactions and enhance sensitivity. 3.4. Comparison of the contribution of proteins A and G in FRET activity The contribution of proteins A and G to FRET activ- ity was also examined. When specific/nonspecific antigens S. Ko, S.A. Grant / Sensors and Actuators B 96 (2003) 372–378 377 Fig. 4. Comparison of the contribution of proteins A and G in FRET activity. were added to each solution containing the complex of LisAb-AF546/PA-AF594 or LisAb-AF546/PG-AF594, respectively, a larger reduction in R3 values was shown in solution containing protein G conjugates than in solutions containing protein A conjugates. This result indicated that the protein G contributed to more efficient detection of Listeria antigen than did protein A (Fig. 4). The binding affinity between the Fc region of goat IgG and protein G is stronger than that of between the Fc region of goat IgG and protein A, which may account for the lower sensitivity in the protein A complex [20]. Therefore, protein G, instead of protein A, was chosen for all experiments as an acceptor conjugate. 4. Conclusion The objective of this study was to demonstrate the feasi- bility of the FRET immunosensor technique for the detec- tion of foodborne pathogens, Listeria and Salmonella. The antibodies and proteins A and G were labeled with a FRET fluorophore pair, and then the presence of specific antigens was detected in solution utilizing spectrofluorometry. The results indicated the level of detection of 2.0�g/ml for Lis- teria and Salmonella antigens. It is critical to increase the sensitivity of this system for foodborne pathogens. There- fore, the immobilization of antibody onto a substrate will be applied in the future experiments to minimize the random interaction in solution, and the packing/grafting density of the antibodies onto the substrates will be optimized in order to achieve higher sensitivity. This biosensor may be im- plemented into an on-line instrumental pathogen detection system, which permits on-site analysis of samples, thereby reducing the large economical burden by food products recalls and medical treatments. Additionally, this method may be useful for detection other analytes, such as E. coli, Campylobacter, and HIV. Acknowledgements This research was funded by the National Pork Producers Council. References [1] Anonymous, From the Center for Disease Control and Prevention, Update: multi-state outbreak of Listeriosis, United States, 1998–1999, JAMA 281 (1999) 317–318. [2] C.M. Simon, D.I. Gray, N. Cook, DNA extraction and PCR methods for the detection of Listeria monocytogenes in cold-smoked salmon, Appl. Environ. Microbiol. 62 (1996) 822–824. [3] R. Karpiskova, M. Pejchalova, J. Mokrosova, J. Vytrasova, P. Smuharova, J. Ruprich, Application of a chromogenic medium and the PCR method for the rapid confirmation of Listeria monocyto- genes in foodstuffs, J. Microbiol. Methods 41 (2000) 267–271. [4] J.M.C. Luk, A PCR enzyme immunoassay for detection of Salmonella typhi, Biotechnology 17 (1994) 1038–1042. [5] S.D. Oliveira, L.R. Santos, D.M.T. Schuch, A.B. Silva, C.T.P. Salle, C.W. Canal, Detection and identification of salmonellas from poultry-related samples by PCR, Vet. Microbiol. 87 (2002) 25– 35. [6] R.D. Oberst, M.P. Hays, L.K. Bohra, R.K. Phebus, C.T. Yamashiro, C. Paszko-Kolva, S.J.A. Flood, J.M. Sargeant, J.R. Gillespie, PCR-based DNA amplification and presumptive detection of Escherichia coli O157:H7 with an internal fluorogenic probe and the 5-nuclease (TaqMan) assay, Appl. Environ. Microbiol. 64 (1998) 3389– 3396. [7] H.J. Beckers, P.D. Tips, P.S.S. Soentoro, E.H.M. Delfgou-Van Asch, R. Peters, The efficacy of enzyme immunoassays for the detection of salmonellas, Food Microbiol. 5 (1998) 147–156. [8] P.S. Hayes, L.M. Graves, G.W. Ajello, Comparison of cold enrich- ment and U.S. Department of Agriculture methods for isolating L. monocytogenes from naturally contaminated foods, Appl. Environ. Microbiol. 57 (1991) 2109–2113. [9] E. Prusak-Sochaczewski, J.H.T. Luong, Development of a piezo- electric immunosensor for the detection of Salmonella typhimurium, Enzyme Microb. Technol. 12 (1990) 173–177. [10] N.F. Starodub, P.J. Arenkov, A.N. Starodub, Fiber optic immunosen- sors based on enhanced chemiluminescence and their application to determine different antigens, Sens. Actuators, B 18–19 (1994) 161– 165. [11] J.R. Uhl, C.A. Bell, L.M. Sloan, M.J. Espy, T.F. Smith, J.E. Rosen- blantt, F.R. Cockerill, Application of rapid-cycle real-time poly- merase chain reaction for the detection of microbial pathogens: The Mayo–Roche Rapid Anthrax Test, Mayo Clin. Proc. 77 (2002) 673– 680. [12] S.A. Grant, J. Xu, E.J. Bergeron, J. Mroz, Development of dual receptor biosensors: an analysis of FRET pairs, Biosens. Bioelectr. 16 (2001) 231–237. [13] I.M. Roitt, P.J. Delves, Roitt’s Essential Immunology, 10th ed., Blackwell Scientific Publications, Oxford, 2001. [14] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, second ed., Kluwer Academic Publishers, New York City, New York, 1999. [15] P. Wu, L. Brand, Resonance energy transfer: methods and applica- tions, Anal. Biochem. 218 (1994) 1–13. 378 S. Ko, S.A. Grant / Sensors and Actuators B 96 (2003) 372–378 [16] R.P. Haugland, Coupling of monoclonal antibodies with fluorophores, in: W.C. Davis (Ed.), Methods in Molecular Biology, Monoclonal Antibodies Protocols, vol. 45, Humana Press, NJ, 1995, pp. 205–221. [17] N. Panchuk-Voloshina, R.P. Haugland, J. Bishop-Stewart, M.K. Bhal- gat, P.J. Milard, F. Mao, W.-Y. Leung, R.P. Haugland, Alexa dyes, a series of new fluorescent dyes that yield exceptionally bright, pho- tostable conjugates, J. Histochem. Cytochem. 47 (1999) 1179–1188. [18] I.Z. Steinberg, Long-range nonradiative transfer of electronic exci- tation energy in proteins and polypeptides, Annu. Rev. Biochem. 40 (1971) 83–114. [19] H.J. Gruber, C.D. Hahn, G. Kada, C. Riener, G.S. Harms, W. Ahrer, T.G. Dax, H.-G. Knaus, Anomalous fluorescence enhancement of Cy3 and Cy3.5 versus anomalous fluorescence loss of Cy5 and Cy7 upon covalent linking of proteins and noncovalent binding to avidin, Bioconjugate Chem. 11 (2000) 696–704. [20] http://www.probes.com/handbook/tables/0408.html, Binding profiles of protein A and protein G,
[email protected]. Biographies Sungho Ko received his Bachelor’s degree in food engineering from Kyungnam University, and his MS degree in food science from the University of Arkansas, Fayetteville, AR. Currently, he is a PhD stu- dent in biological engineering at the University of Missouri, Columbia, MO. His interests included the development of optical sensors for food safety. Sheila A. Grant received her Master’s in biomedical engineering and PhD in materials engineering from Iowa State University. She worked for 4 years at Lawrence Livermore National Laboratory in Livermore, CA, and 3 years at Michigan Technological University in Houghton, MI. Currently, she is an assistant professor in biological engineering at the University of Missouri, Columbia. Her interests involve diagnostic biosensors for medi- cal and homeland security applications as well as bioMEMS multianalyte sensors. Development of a novel FRET method for detection of Listeria or Salmonella Introduction Materials and methods Materials Covalent labeling of proteins with fluorophores Determination of the ratio of moles of fluorophore to moles of protein Determination of the optimal acceptor to donor fluorophore ratio In-solution detection of antigen with the labeled antibody-protein complex Statistical analysis Results and discussion Determining F/P ratio Determination of an optimal acceptor to donor ratio In-solution detection of Listeria and Salmonella antigens with FRET activity Comparison of the contribution of proteins A and G in FRET activity Conclusion Acknowledgements References