Using Capacitance Measurements as the Detection Method in Antigen

Jan 31, 2007 - capable of molecular recognition of anti-pasteurellosis antibodies. Detection was carried out using a novel strategy entirely based upo...
0 downloads 0 Views 164KB Size
Download PDF
Anal. Chem. 2007, 79, 2163-2167

Using Capacitance Measurements as the Detection Method in Antigen-Containing Layer-by-Layer Films for Biosensing Valtencir Zucolotto,*,† Katia R. P. Daghastanli,‡ Caio O. Hayasaka,† Antonio Riul, Jr.,§ Pietro Ciancaglini,‡ and Osvaldo N. Oliveira, Jr.†

IFSC, Universidade de Sa˜o Paulo, CP 369, 13566-590, Sa˜o Carlos, SP, Brazil, DQ-FFCLRP, Universidade de Sa˜o Paulo, 14040-901, Ribeira˜o Preto, SP, Brazil, and Universidade Federal de Sa˜o Carlos, P.B. 3031, Sorocaba, SP, 18043-970, Brazil

The layer-by-layer technique is employed here to immobilize antigen-containing liposomes, so-called proteoliposomes, onto Au-interdigitated substrates, which are capable of molecular recognition of anti-pasteurellosis antibodies. Detection was carried out using a novel strategy entirely based upon capacitance measurements, and to enhance sensitivity, we combine the response of three different sensing units in a similar procedure used for taste sensors. With the three-electrode array immunoglobulin G (IgG) against pasteurellosis is detected at concentrations as low as nanograms per milliliter. Furthermore, because of the molecular recognition capability, a distinction can be made between specific and nonspecific IgG. The concepts behind the biosensors reported here may have a large impact for clinical tests, as the procedures to detect the antibody take only a few minutes and the biosensors are relatively low cost. The search for improved, fast, and low-cost clinical tests has motivated extensive research into biosensors that can be made with various strategies. One possibility is to employ nanostructured films incorporating biomolecules such as enzymes, as is the case with the layer-by-layer (LbL) films1-3 based on the physical adsorption of oppositely charged layers. LbL films are built at room temperature and from aqueous solutions whose pH may be controlled to suit the biomolecules. This method has been * To whom correspondence should be addressed. E-mail: [email protected]. † IFSC, Universidade de Sa˜o Paulo. ‡ DQ-FFCLRP, Universidade de Sa˜o Paulo. § Universidade Federal de Sa˜o Carlos. (1) Tang, Z.; Kotov, N. A.; Magonov, S.; Ozturk, B. Nat. Mater. 2003, 2, 413. (2) Decher, G. Science 1997, 277, 1232. (3) Oliveira, O. N., Jr.; He, J.-A.; Zucolotto, V.; Balasubramanian, S.; Li, L.; Nalwa, H. S.; Kumar, J.; Tripathy, S. K. Layer-by-layer polyelectrolyte films for electronic and photonic applications. In Handbook of Polyelectrolytes and Their Applications; Kumar, J., Nalwa, H. S., Eds.; American Scientific Publishers: Los Angeles, CA, 2002; Vol. 1, pp 1-37. (4) Lvov, Y.; Ariga, K.; Ichinose, I.; Kunitake, T. J. Am. Chem. Soc. 1995, 117, 6117. (5) Sukhorukov, G. B.; Montrel, M. M.; Petrov, A. I.; Shabarchina, L. I.; Sukhorukov, B. I. Biosens. Bioelectron. 1996, 11, 913. (6) Brynda, E.; Houska, M.; Skvor, J.; Ramsden, J. J. Biosens. Bioelectron. 1998, 13, 165. (7) Trau, D.; Yang, W.; Seydack, M.; Caruso, F.; Yu; N.-T., Renneberg, R. Anal. Chem. 2002, 74, 5480. 10.1021/ac0616153 CCC: $37.00 Published on Web 01/31/2007

© 2007 American Chemical Society

proven excellent for immobilization of biomolecules, whose activity is preserved for long periods of time.4-7 There are two main reasons for such success of the LbL method, namely, the possibility of choosing the adequate template or scaffolding material for the biomolecules and the mild conditions under which film fabrication occurs. An extensive body of work has indicated that suitable template materials for immobilization of biological species are mesoporous silicates,8 dendrimers,9,10 and liposomes.11-13 Due to their high entrapping or incorporation capacity, liposomes can be used as carrier agents for drug delivery, fluorescent markers, enzymes entrapped, or lipophilic protein incorporated during liposome preparation; some of these systems have been proven stable from hours to years.12,13 Pasteurella multocida is one of the most interesting bacterial pathogens, the etiological agent of pasteurellosis that causes various diseases, including hemorrhagic septicemia, fowl cholera, and secondary invasions in many animals. In spite of being a zoonosis, occasional pasteurellosis infections may occur in humans due to animal bites.14 The presence of pasteurellosis is difficult to determine though, which may occur by clinical analysis or necropsy. Since lipopolysaccharides and lipoproteins constitute major somatic antigens exposed on the cell surface of P. multocida, a biosensor can be devised with P. multocida antigens immobilized on a solid film and used to detect a specific antibody, improving the diagnostic at early stages of the disease. With the LbL technique, the biological activity of the antigens can be preserved, which also requires the choice of a scaffolding material, with dendrimers proving excellent in the present study. In a previous paper, we showed that, for enzyme-based biosensors, detection of specific analytes could be performed based on the analysis of electrical capacitance data.10 This new (8) Hudson, S.; Magner, E.; Cooney, J.; Hodnett, B. K. J. Phys. Chem. B 2005, 109, 19496. (9) Yoon, H. C.; Kim, H.-S. Anal. Chem. 2000, 72, 922. (10) Zucolotto, V.; Pinto, A. P. A.; Tumolo, T.; Moraes, M. L.; Baptista, M. S.; Riul, A., Jr.; Arau´jo, A. P. U.; Oliveira, O. N., Jr. Biosens. Bioelectron. 2006, 21, 1320. (11) Lee, H. Y.; Jung, H. S.; Fujikawa, K.; Park, J. W.; Kim, J. M.; Yukimasa, T.; Sugihara, H.; Kawaia, T. Biosens. Bioelectron. 2005, 21, 833. (12) Gregoriadis, G.; Florence, A. T. Drugs 1993, 45, 15. (13) Rongen, H. A. H.; Bult, A.; Van Bennekom, W. P. J. Immunol. Methods 1997, 204, 105. (14) Frost, A. J.; Adler, B. Vet. Microbiol. 2000, 72, 1.

Analytical Chemistry, Vol. 79, No. 5, March 1, 2007 2163

Figure 1. Idealized structure of the PAMAM/proteoliposome LbL films deposited onto interdigitated electrodes. The enlarged area illustrates the exposed portion of bilayers of PAMAM/proteoliposomes, in which interactions with IgG are expected to occur. Capacitance data were collected as a function of frequency of the ac signal with the electrodes immersed in different solutions containing buffer and the IgG to be detected. Three different electrodes were employed: (i) bare, (ii) electrode containing a five-bilayer PAMAM/PVS film, and (iii) electrode containing a five-bilayer PAMAM/proteoliposome film.

approach uses the concept of the taste sensors,15 bearing the advantage of molecular recognition and specific interaction, characteristic of biological systems. In this study, we show for the first time that the capacitance-based strategy can be successfully applied to the detection of specific antibodies, for a given antigen-antibody system. As a proof of concept, we show that LbL films of antigen-containing liposomes can be applied to detect anti-pasteurellosis antibodies using capacitance measurements. By using a three-electrode array architecture, immunoglobulin G (IgG) against pasteurellosis was detected at concentrations as low as nanograms per milliliter. Furthermore, the sensors were able to distinguish between specific and nonspecific IgG. EXPERIMENTAL SECTION Materials. All solutions were prepared using Millipore Milli-Q ultrapure water. Tris(hydroxymethyl)aminomethane (Tris), sodium dodecyl sulfate (SDS), dipalmitoylphosphatidylcholine (DPPC), and bovine serum albumin were purchased from Sigma (St. Louis, MO). Calbiosorb resin was acquired from Calbiochem (San Diego, CA). Polyamidoamine generation 4 dendrimer (PAMAM), poly(allylamine hydrochloride) (PAH) (used as polycations) and poly(vinylsulfonic acid (PVS) (polyanion) were purchased from Aldrich and used without further purification. Au-interdigitated electrodes (IDE) were lithographically fabricated at the National Syncroton Laboratory (LNLS) facilities (Campinas, Brazil). Estimation of Protein Concentration. Protein concentration was estimated with the procedure described by Hartree,16 in a solution with SDS 2% (w/v), using crystallized bovine serum albumin as standard. Antiserum Preparation. Antiserum against total P. multocida antigenic determinants (positive IgG) were obtained by applying four weekly intramuscular injections of thermally inactivated (15) Riul, A., Jr.; dos Santos, D. S., Jr.; Wohnrath, K.; Di Tommazo, R.; Carvalho, A. C. P. F. L.; Fonseca, F. J.; Oliveira, O. N., Jr.; Taylor, D. M., Mattoso, L. H. C. Langmuir 2002, 18, 239. (16) Hartree, E. F. Anal. Biochem. 1972, 48, 422. (17) Daghastanli, K. R. P.; Ferreira, R. B.; Thedei, G., Jr.; Ciancaglini, P. Biochem. Mol. Biol. Educ. 2000, 28, 256.

2164 Analytical Chemistry, Vol. 79, No. 5, March 1, 2007

Figure 2. (A) Electronic absorption of the PAMAM/proteoliposomes for films containing different numbers of bilayers. (B) Absorption at 260 nm vs number of PAMAM/proteoliposome bilayers. Contribution from the background due to light scattering was subtracted in all curves.

P. multocida cells (∼(6-9) × 1010 cfu) to healthy rabbits as described by Daghastanli et al.17 Mouse IgG anti-sheep erythrocytes (used here as negative IgG) were obtained by subcutaneous injection of sheep erythrocytes to mice as described in ref 18. The blood of these animals was collected without anticoagulant,

Figure 3. PCA plot built up using capacitance data from the three electrodes employed, at a fixed frequency of 100 Hz. The capacitance values were collected in triplicate with the electrodes immersed in buffer (4) and IgG anti-P. multocida (positive IgG) at concentrations of 10-2 (+), 10-3 (g), 10-4 (O), and 10-5 (0) mg/mL. Apart from the data from 10-4 mg/mL (O), all the concentrations can be clearly separated considering only the PC1. Also worth noting is the correlation among the clusters, going from the most concentrated 10-2 mg/mL (+) in the left side to the buffer (4) in the right side.

and the total serum was isolated from plasma by centrigugation at 3000 rpm. Then, it was precipitated by addition of ammonium sulfate up to 40% (w/v) final concentration under constant agitation and centrifuged at 110000g for 20 min; the pellet was suspended in 5 mM Tris-HCl pH 7.5 and dialyzed against 2 L of this buffer for 24 h at 4 °C, with two buffer changes. The dialyzed samples were loaded in an ion-exchange chromatograph using a DEAEcellulose column (40 × 2 cm) eluted with the same Tris buffer using a flow rate of 60 mL/h. Proteoliposomes Preparation. The isolation of P. multocida strain and the solubilization of bacterial membrane proteins were performed as described in ref 19. The antigenic lipophilic proteins from the outer membrane of the bacteria P. multocida were incorporated into liposomes (the so-called proteoliposomes) through the cosolubilization of lipids, proteins, and detergent with a relationship of 10:1:70 (w/w).19 Briefly, DPPC (10 mg) was dissolved in 1 mL of chloroform and dried under nitrogen flow; and the lipid film formed was maintained under vacuum for 1 h. Then, 2 mL of 5 mM Tris-HCl (pH 7.5) buffer, containing 25 mg/ mL SDS, was added to the film, incubated at 60 °C for 1 h, and vortexed at intervals of 10 min. In the following step, the solubilized lipids were sonicated, using a microtip, for 1 min at 240 W. Two milliliters of SDS-solubilized membrane proteins (0.5 mg/mL) were added to this mixture and incubated for 30 min at room temperature. The detergent was removed from the mixture by treatment with 200 mg/mL Calbiosorb resin with two changes at 2-h intervals. After the detergent had been removed, the suspension was centrifuged at 140000g for 1 h and the pellets, constituted of proteoliposomes, were resuspended in 1.0 mL of 5 mM Tris-HCl, pH 7.5. The final amount of SDS in the proteoliposomes is estimated at ∼0.5% (225 µM), below the cmc for this (18) Mantovani, B. J. Immunol. 1975, 11, 5.

surfactant (7-10 mM). Furthermore, previous investigations suggested that SDS molecules may coexist with the lipidic bilayer.19 The latter would result in the presence of negative charges on the vesicles. Proteoliposome Immobilization. PAMAM and proteoliposomes were used at 1 and 0.7 mg/mL, respectively, in a 5 mM Tris-HCl pH 7.5 buffer solution. Nanostructured LbL films containing up to 15 PAMAM/proteoliposome bilayers were assembled on quartz slides for UV-visible measurements and five-bilayer PAMAM/proteoliposome films were deposited onto the IDEs for capacitance measurements, as depicted in Figure 1. All substrates were previously cleaned in a NH4OH/H2O2/H2O (5:1:1 v/v) bath for 20 min. After cleaning, the substrates were covered with two bilayers of a PAH/PVS LbL film, where both PAH and PVS dipping solutions were used at concentrations of 1 mg/mL, pH 6.5. The latter procedure was employed to ensure a uniform distribution of charges onto the substrates prior to the PAMAM/proteoliposome film deposition. Deposition of the multilayers was carried out by immersing the quartz slides or IDEs alternately into the PAMAM (polycationic) and proteoliposomes (polyanionic) solutions for 5 and 15 min, respectively. After deposition of each layer, the substrate/film system was immersed for 1 min in the buffer solution. The deposition process was monitored at each deposited bilayer using UV-visible spectroscopy. Electrical Detection of Antiserum. Capacitance measurements were performed with a Solartron impedance/gain phase analyzer (model 1260A). Unlike the electrochemical experiments, the in-plane capacitance of the film deposited between the tracks of the IDE was collected in a frequency range from 10 Hz to 1 kHz, with no need of a reference electrode. The IDE employed contained 50 digit pairs, each having 10-µm width, 0.1-µm height, and 10 µm apart from each other. All measurements were taken using a set of three electrodes, viz.: (i) bare, (ii) electrode containing a five-bilayer PAMAM/PVS film (in the absence of proteoliposomes), and (iii) electrode containing a five-bilayer PAMAM/proteoliposome film. For LbL films, data acquisition was carried out with the electrodes immersed in (1) 5 mM Tris-HCl buffer, (2) positive IgG antibody solutions at concentrations of 10-2, 10-3, 10-4, 10-5, and 10-6 mg/mL, and (3) negative IgG solutions at 10-1, 10-2, and 10-3 mg/mL. Capacitance curves were taken three times for each sensor, after they were soaked for 20 min in the analytical solutions. After each measurement, the sensors were rinsed in buffer solution. RESULTS AND DISCUSSION The proteoliposomes were the active materials in biosensors via their immobilization onto Au-interdigitated electrodes using the PAMAM dendrimer as the support polyelectrolyte. The idealized PAMAM/proteoliposome film architecture is depicted in the expanded area of Figure 1. The successful preparation of PAMAM/proteoliposome LbL films is indicated in Figure 2A, which displays UV-visible spectra for films with increasing numbers of PAMAM/proteoliposome bilayers. Adsorption of PAMAM and proteoliposome layers is expected to be driven primarily by electrostatic interactions. For example, the membrane proteins exhibit hydrophilic moieties containing residues of charged amino acids, which may be exposed for interaction since they are located close to the lipid polar heads. In addition, even though DPPC has a neutral liquid charge, it is zwitterionic with Analytical Chemistry, Vol. 79, No. 5, March 1, 2007

2165

Figure 4. PCA plot obtained using the same electrodes from Figure 3. Again, the capacitance was collected at 100 Hz, with the electrodes immersed in Milli-Q water (O), buffer (b), IgG anti-P. multocida (positive IgG) (black symbols at concentrations of 10-2 (0), 10-3 (4), 10-4 (+), 10-5 (]), and 10-6 (f) mg/mL) and IgG anti-sheep erythrocytes (negative IgG) (red symbols at concentrations of 10-1 (3), 10-2 (0), and 10-3 (4) mg/mL). Although the data from negative IgG are not well separated, a good distinction is clearly visualized between the positive and negative IgG.

a negative charge from phosphate groups. Therefore, these negatively charged groups should interact electrostatically with the protonated amines from PAMAM. Nevertheless, other interactions such as H-bonding and hydrophobic interactions, which have been proven to play an important role in LbL film assembly,20-22 may also occur within the multilayers. The maximum in the absorption spectrum at 260 nm is due to the membrane proteins incorporated in the proteoliposomes. Figure 2B shows that the absorption increases almost linearly with the number of layers, and this is taken as evidence for the adsorption of equal amounts of proteoliposomes at each deposition step. A background signal due to light scattering was observed during spectra collection, which for the sake of clarity was subtracted in Figure 2A. Such a background signal may be indicative of the integrity of the proteoliposomes, most of which would be adsorbed in the form of vesicles. Three different electrodes, viz: (i) bare IDE, (ii) IDE covered with a five-bilayer PAMAM/PVS film, and (iii) IDE covered with a five-bilayer PAMAM/proteoliposome film, were employed in the electrical detection of the antibodies. The reason for using three electrodes is that the limit of detection of a sensor array is always lower than that of a single sensor.10 After immersing the electrodes for 20 min into various concentrations of the antibody, a stable reading was acquired and the capacitance was measured as a function of frequency. Capacitance curves were analyzed with an equivalent electric circuit representing the system under study (metal electrodes covered with thin films immersed in an

electrolyte), following the work of Taylor and MacDonald.23 Principal component analyses (PCA)24 was used to statistically correlate the samples, similar to what we have made in enzymebased biosensors.10 A detailed description of PCA correlating capacitance data in sensors can be found in ref 15. Briefly, it allows the extraction of the relevant information from a set of input data, by creating a set of principal components from the original data. Consequently, the correlation between the data can be graphically visualized in a plot such as those presented in Figures 3 and 4, where the first Principal Component (PC1) presents the highest variance, bearing most of the information to discriminate the data. In this study, PCA plots were built from capacitance values collected at 100 Hz because at this frequency the electrical response depends on the film/electrolyte interaction.23 The PCA plot in Figure 3 shows the detection of the specific IgG antibody against P. multocida down to 10 ng/mL. As expected, the data were shifted toward the buffer solution value as the solutions became more diluted (from the left to the right side). Such high sensitivity should be attributed not only to the large influence that trace amounts of a given analyte have on the electrical properties of thin films in contact with an electrolyte, as it has been exploited in taste sensors,25 but also due to the high antigen/antibody molecular recognition. We have confirmed the importance of this specific interactionsdue to molecular recognitionsin subsidiary experiments by checking different electrode combinations, such as bare electrode plus the electrode containing five bilayers of PAMAM/PVS (not shown), in the PCA

(19) Daghastanli, K. R. P.; Ferreira, R. B.; Thedei, G., Jr.; Maggio, B.; Ciancaglini, P. Colloids Surf,, B 2004, 36, 127. (20) Stockton, W. B.; Rubner, M. F. Macromolecules 1997, 30, 2717. (21) Clark, S. L.; Hammond, P. T. Langmuir 2000, 16, 10206. (22) Hammond, P. T. Curr. Opin. Colloid Interface Sci. 2000, 4, 430.

(23) Taylor, D. M.; MacDonald, A. G. J. Phys. D: Appl. Phys. 1987, 20, 1227. (24) Jackson, J. E. A. Users’ Guide to Principal Components; Wiley: New York, 1991. (25) Ferreira, M.; Riul, A., Jr.; Wohnrath, K.; Fonseca, F. J.; Oliveira, O.N., Jr.; Mattoso, L. H. C. Anal. Chem. 2003, 75, 953.

2166

Analytical Chemistry, Vol. 79, No. 5, March 1, 2007

analysis. The best distinguishing ability was achieved with the use of all three electrodes, reinforcing the importance of crosssensitivity in this sort of measurements. Interestingly, but not surprisingly, data from concentrations at 10-4 were located far from the other data, shifted in the Y direction. Such an effect is common in PCA analyses, but its origin is difficult to establish. It is important to note, however, that a good separation is still possible when the PC1 and PC2 components are considered. Although the high sensitivity provided by the combination of ac electrical measurements and ultrathin films has a positive impact in achieving good sensors,15 the presence of specific interactions between film-forming molecules and analytes improves considerably the distinguishing ability of the device. To be useful as a biosensor, the LbL films from immobilized antigens must be able to distinguish between specific antibodies for which a molecular recognition process occurs (positive IgG) and nonspecific antibodies (negative IgG), as the latter should not interact specifically with the antigen. Figure 4 shows that such discrimination is possible, with the data obtained for the IgG antisheep erythrocytes (negative IgG) significantly distant from the data for the IgG anti-P. multocida antibodies (positive IgG) in the PCA plot. Furthermore, we may note that the sensors are not efficient for separating the three different concentrations of negative IgG, due to the lack of specific interactions. In Figure 4, data for the positive IgG are the same as in Figure 3 and appear clustered because with the inclusion of data for the nonspecific antibody the data points were all lumped together. The mechanisms responsible for the changes in film capacitance upon interacting with IgG binding are difficult to identify because the electrical response depends on various inter-related parameters. While examining this issue, we noted that the presence of antibodies in a buffer solution may either increase or decrease the film capacitance (in the system studied here, a decrease was observed). Therefore, it is not possible at the

moment to unequivocally assign the antibody effect to a single type of interaction. In fact, we believe that the specific binding of the antibodies to the antigens may be affected by various forces, including van der Waals, hydrogen-bonding, hydrophobic, and electrostatic interactions. In addition to the fact that experiments are relatively fast (∼20 min), other features concerning the use of the LbL films as biosensors should be remarked. For example, similar results could be obtained for PAMAM/proteoliposome films stored at 5 °C for more than one week, which was expected since immobilization of biomolecules using the LbL technique may preserve their activity for long periods of time.4 CONCLUSIONS Proteoliposomes incorporating antigenic membrane proteins of P. multocida have been successfully immobilized in LbL films, and with the preserved activity of the antigen, it was possible to detect IgG anti-P. multocida at concentrations as low as nanograms per milliliter, using capacitance measurements. It was also shown that this high sensitivity can be accompanied by selectivity, since the electrical measurements allowed the distinction between specific and nonspecific antibodies. Due to its very high sensitivity, low-cost, and rapid performance, we believe that the approach presented here will be useful in clinical analyses, especially for those where antigen-antibody systems are employed. ACKNOWLEDGMENT This work was supported by FAPESP, CNPq, and IMMP/MCT (Brazil). The authors are also grateful to LNLS (project LMF 4385) regarding the fabrication of interdigitated electrodes. Received for review August 29, 2006. Accepted December 1, 2006. AC0616153

Analytical Chemistry, Vol. 79, No. 5, March 1, 2007

2167