Nanoencapsulated Microcrystalline Particles for Superamplified

Sep 25, 2002 - Aldrich Chemical Co. (Gillingham, U.K.). Affinity-purified poly- clonal goat anti-mouse IgG (Gt R-M IgG) (Fc specific), Gt R-M. IgG-FIT...
0 downloads 0 Views 212KB Size
Download PDF
Anal. Chem. 2002, 74, 5480-5486

Nanoencapsulated Microcrystalline Particles for Superamplified Biochemical Assays Dieter Trau,*,†,‡ Wenjun Yang,† Matthias Seydack,§ Frank Caruso,| Nai-Teng Yu,† and Reinhard Renneberg†

Department of Chemistry and Bioengineering Graduate Program, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, SAR Hong Kong, 8sens.biognosticAG, Robert Roessle-Strasse 10, D-13125 Berlin, Germany, and Max Planck Institute of Colloids and Interfaces, D-14424 Potsdam, Germany

We report on the preparation and utilization of a novel class of particulate labels based on nanoencapsulated organic microcrystals with the potential to create highly amplified biochemical assays. Labels were constructed by encapsulating microcrystalline fluorescein diacetate (FDA; average size of 500 nm) within ultrathin polyelectrolyte layers of poly(allylamine hydrochloride) and poly(sodium 4-styrenesulfonate) via the layer-by-layer technique. Subsequently, the polyelectrolyte coating was used as an “interface” for the attachment of anti-mouse antibodies through adsorption. A high molar ratio of fluorescent molecules present in the microcrystal core to biomolecules on the particle surface was achieved. The applicability of the microcrystal-based label system was demonstrated in a model sandwich immunoassay for mouse immunoglobulin G detection. Following the immunoreaction, the FDA core was dissolved by exposure to organic solvent, leading to the release of the FDA molecules into the surrounding medium. Amplification rates of 70-2000-fold (expressed as an increase in assay sensitivity) of the microcrystal label-based assay compared with the corresponding immunoassay performed with direct fluorescently labeled antibodies are reported. Our approach provides a general and facile means to prepare a novel class of biochemical assay labeling systems. The technology has the potential to compete with enzyme-based labels as it does not require long incubation times, thus speeding up bioaffinity tests. Biochemical assays, such as enzyme-linked immunosorbent assays (ELISA), radioimmunoassays (RIA), immunoagglutination assays (IAA), fluorescent immunoassays (FIA), and those based on DNA hybridization or receptor ligand interactions play an important role in medical diagnostics, food, and environmental analyses.1,2 A common feature of most assay techniques is that they require a label to detect the interaction of a biocompound * To whom correspondence should be addressed. E-mail: [email protected]. Fax: 852-3106-4857. † Department of Chemistry, The Hong Kong University of Science and Technology. ‡ Bioengineering Graduate Program, The Hong Kong University of Science and Technology. § 8sens.biognosticAG. | Max Planck Institute of Colloids and Interfaces.

5480 Analytical Chemistry, Vol. 74, No. 21, November 1, 2002

with an analyte. The basic requirements for a biolabel system are high sensitivity, low limit of detection, and preserved biomolecule functionality. In the present trend of applying nonradioactive labeling strategies, fluorescent molecules and enzymes are widely employed as labels. Both strategies offer good sensitivity and limits of detection. These parameters can be improved by increasing the ratio of fluorescent dyes to biomolecules (i.e., the F/P ratio) for fluorescent-based label techniques and by increasing the substrate incubation time for enzyme-based label systems. The F/P ratio is a key parameter that has an impact on sensitivity and is widely used for evaluation of a fluorescent label in biochemical assay technologiessa high number of labeling molecules per biomolecule is desired. For example, the F/P value is typically 4-8 for a conventional covalently coupled fluorescent immunolabel, e.g., an immunoglobulin G (IgG)-FITC conjugate. Since conjugation of the fluorescent molecules to the biomolecules is typically achieved by covalent binding, the fluorescent probes used in FIAs should possess a reactive group for bioconjugation, high water solubility, low nonspecific adsorption, and good photostability.3 Additionally, they should have a high molar extinction coefficient and fluorescence quantum yield. However, covalent binding of a large number of label molecules usually leads to a decrease of specific binding activity of the biomolecule and the quantum yield due to self-quenching effects. Antibodies labeled with more than four to six fluorophores per protein may also exhibit reduced specificity and binding affinity.4 To prevent such complications, the optimal dye/protein ratio is normally kept below ∼4, which highly limits the potential sensitivity improvement of FIAs.5 These aforementioned demands placed on the performance of fluorescent probes vastly limit the number of suitable candidates. The sensitivity of fluorescence assays is determined by the number of light quanta emitted per analyte molecule. Increasing the dye/biomolecule ratio while minimizing dye self-quenching to obtain signal amplification (and hence sensitivity) is highly (1) Tijssen, P. Practice and theory of enzyme immunoassays, 8th ed.; Elsevier: New York, 1985. (2) Kessler, C. Nonradioactive Labeling and Detection of Biomolecules; SpringerVerlag: New York, 1992. (3) Hemmila, I. A. Applications of fluorescence in immunoassays; J. Wiley & Sons: New York, 1991. (4) Haugland, R. P. Handbook of Fluorescent Probes and Research Chemicals, 6th ed.; Molecular Probes, Inc.: Eugene OR, 1996. (5) Johnson, G. D. Antibodies. A Practical Approach; IRL: Oxford, U.K., 1989. 10.1021/ac0200522 CCC: $22.00

© 2002 American Chemical Society Published on Web 09/25/2002

Figure 1. Schematic illustration of the preparation of biolabeled, polyelectrolyte-encapsulated organic microcrystals. FDA was ball milled into micrometer-sized crystals in an aqueous surfactant (SDS) medium (introducing charges onto the originally uncharged microcrystal surface), followed by encapsulation with polyelectrolyte multilayers of nanometer thickness (a), and the attachment of a specific immunoreagent (b).

desirable. There exist several methods based on fluorescence detection that permit signal amplification: (1) avidin-biotin or antibody-hapten secondary detection; (2) enzyme-labeled secondary detection reagents in conjunction with fluorogenic substrates; or (3) probes that contain multiple fluorophores.4 In the first method, the target-specific molecule contains multiple binding sites that can be selectively recognized by secondary detection reagents, which are already labeled with fluorophores, thus providing a simple system for amplifying the signal. In these systems, however, the problem of covalent conjugation of fluorescent probes with biomolecules still exists, and the amplification factor is also very low. The second method exploits the enzyme conjugated with the target-specific molecule to catalyze the transformation of the fluorogenic substrate into a highly fluorescent product. The amplification is due to the extremely high turnover rate of the enzyme. However, special fluorogenic reagents must be added, and hence, additional time for incubation, usually between 15 min to 2 h is required. The last method exhibits high versatility. Fluorophore-loaded latex beads,6 liposome-encapsulated fluorophore,7-9 chemiluminescent microemulsions,10 fluorescent-conjugated dendrimers,11 and other related systems12-14 all display a high dye/protein ratio and are suitable for use in immunoassays. Immunospecies-conjugated fluorescent latex particles were employed to increase the dye/protein ratio.4 Commercial fluorescent beads are manufactured using polystyrene (PS) microspheres loaded with specific dyes. Hall et al.15 reported the use of such fluorescent beads as probes in sandwich immunoassays, resulting in an 10-fold signal amplification. The application of fluorescent microspheres together with the technique of time-resolved fluorescence spectroscopy was reported to detect 60 zmol (10-21) of analyte.16 However, this technique requires very sophisticated and expensive instrumentation. Another approach of the multiple fluorophore system is to use liposome-encapsulated fluorescent molecules. Carboxyfluorescein and rhodamine B were encapsulated in either the lipid phase or aqueous phase of a liposome (6) Bangs, L. B. Pure Appl. Chem. 1996, 68, 1873-1879. (7) Truneh, A.; Machy, P.; Horan, P. K. J. Immunol. Methods 1987, 100, 5971. (8) Schott, H.; Von Cunow, D.; Langhals, H. Biochim. Biophys. Acta 1992, 1110, 151-157. (9) Imai, K., Namura, Y. U.S. Patent 4916080, 1986. (10) Kamyshny, A., Magdassi, S. Colloids Surf. B 1998, 11, 249-254. (11) Goeran, K. DE Patent 19703718, 1997. (12) Fribnau, T. C. J.; Roeles, F.; Leuvering, J. H. W. U.S. Patent 4,373,932, 1981. (13) Mandle, R. M.; Wong, Y. N. U.S. Patent 4,372,745, 1983. (14) Kamyshny, A.; Magdassi, S. Colloids Surf. B 2000, 18, 13-17. (15) Hall, M.; Kazakova, I.; Yao, Y. M. Anal. Biochem. 1999, 272, 165-170. (16) Harma, H.; Soukka, T.; Lovgren, T. Clin. Chem. 2001, 47, 561-568.

and labeled with an antigen or antibody as immunoprobe.17,18 When activated by an immune complex resulting from the specific antigen-antibody reactions, the liposomes are lysed and the fluorescent component is expelled from the liposomes into the surrounding liquid, emitting fluorescence proportional to the analyte concentration. Although this approach is attractive, the preparation of microcapsules with liposomes is rather elaborate and the stability of the systems is critical. Another approach employs perylene microparticles prepared by precipitation from solution in the presence of human IgG for the quantitative detection of anti-human IgG. The analyte concentration is determined by measuring either the chemiluminescence intensity of the microemulsions10 or the fluorescence intensity.14 The F/P ratio was calculated to be as high as ∼1400 in these systems. The microparticle preparation method and assay principle are quite straightforward, but the precipitation method to prepare perylene microparticles requires a large amount of antigen or antibody, which is not economical for most biological analyses. Given the limitations of existing fluorescence-based biochemical assays, the development of new strategies and systems to perform FIAs is desirable. In this article, we report on the stepwise preparation of a novel class of stable, particulate biolabels based on polyelectrolyteencapsulated microcrystals of a fluorogenic precursor material, fluorescein diacetate (FDA) (Figure 1). The encapsulation was achieved by using the layer-by-layer (LbL) technology.19 Microcrystals enveloped in polyelectrolyte multilayers of nanometer thickness exhibit a suitable interface for the subsequent attachment of biorecognition molecules. As the entire microcrystal core is composed of precursor fluorescent molecules (for fluorescein) and the biomolecule forms a layer on the encapsulated crystal surface, the potentially reachable F/P ratio is exceptionally high. The general concept of the application of the biolableled, nanoencapsulated FDA microcrystalline particles as fluorescent labels in FIAs is depicted in Figure 2. Following immunoreaction, the microcrystals are dissolved by a release reagent and the solubilized molecules are released into the surrounding medium (Figure 2c). This leads to a dilution of the fluorescence molecules in a desired volume of release reagent to overcome the problems of fluorescence quenching associated with other FIAs. The dye molecules are solubilized in a large volume of the release reagent (17) Rongen, H. A. H.; Bult, A.; van Bennekom, W. P. J. Immunol. Methods 1997, 204, 105-133. (18) Singh, A. K.; Schoeninger, J. S.; Carbonell, R. G. Biosensors and their applications: Liposomes as signal-enhancement agents in immunodiagnostic applications; Kluwer Academic: Dordrecht, The Netherlands, 1999. (19) Decher, G.; Hong, J. D. Ber. Bunsen-Ges. Phys. Chem. 1991, 95, 1430.

Analytical Chemistry, Vol. 74, No. 21, November 1, 2002

5481

Figure 2. Principle of a sandwich immunoassay using FDA particulate labels. The analyte is first immobilized by the capture antibody preadsorbed on the solid phase (a) and then exposed to antibody-labeled microparticle detectors (b). Every microparticle contains ∼108 FDA molecules. High signal amplification is achieved after solubilization, release, and conversion of the precursor FDA into fluorescein molecules by the addition of DMSO and NaOH (c). Fluorescence intensity is proportional to the analyte concentration. The surfactant is not shown in the last step; some surfactant may be adsorbed to the polyelectrolyte layers and some may be released with the fluorescein.

instead of being confined within the neighborhood of the immunoreagent. We chose FDA as a model precursor fluorescer since it can be easily converted to fluorescein, which has high absorptivity, excellent fluorescence quantum yield, and good water solubility. The low solubility of FDA in aqueous media minimizes leakage once encapsulated, providing highly stable particulate labels. EXPERIMENTAL SECTION Materials. Fluorescein diacetate and dimethyl sulfoxide (DMSO) were purchased from Sigma Chemical Co. (St. Louis, MO). The polyelectrolytes poly(allylamine hydrochloride) (PAHpolyelectrolyte; Mw 15 000) and poly(sodium 4-styrenesulfonate) (PSS-polyelectrolyte; Mw 70 000) as well as sodium dodecyl sulfate (SDS) and β-morpholinoethansulfonic acid (MES) were from Aldrich Chemical Co. (Gillingham, U.K.). Affinity-purified polyclonal goat anti-mouse IgG (Gt R-M IgG) (Fc specific), Gt R-M IgG-FITC conjugate (Catalog No. F 0257, protein concentration 1.1 mg/mL, F/P ratio 4.2), and bovine serum albumin (BSA; fraction V) were obtained from Sigma-Aldrich (Steinheim, Germany). Affinity-purified polyclonal mouse IgG (M IgG), goat IgG (Gt IgG), and Gt R-M IgG (whole molecule) were supplied by Arista Biologicals Inc. All other chemicals were obtained from Sigma and were of analytical grade. The water used in all experiments was prepared in a Millipore Milli-Q Plus 185 purification system with a resistivity higher than 18.2 MΩ cm. Preparation of FDA Particulate Labels. FDA microcrystal suspensions were prepared in a Retsch S1000 (Haan, Germany) ball mill, using a 50-mL agate milling beaker (Retsch Mahlbecher “S”) and a mixture of 10-mm agate and 2-mm titanium dioxide balls as the milling medium. A suspension of 4 g of FDA in 20 mL of 1% (w/v) SDS aqueous solution was milled for 6 h. The temperature was kept below 30 °C to prevent the material from hydrolyzing. The morphology of milled FDA microparticles was examined with a JEOL 6300F ultrahigh-resolution scanning electron microscope, operating at 10 kV. Particle size distributions were measured based on laser scattering using the Mie theory and a Helios H9241 (Sympatec GmbH) particle analyzer. The sample suspensions were sonicated for 30 s prior to particle size analysis in the instrument. 5482 Analytical Chemistry, Vol. 74, No. 21, November 1, 2002

Polyelectrolyte multilayers were assembled onto the microcrystals by the sequential deposition of PAH-polyelectrolyte and PSS-polyelectrolyte. Typically, 0.2 mL of the FDA particle suspension (1.0%, w/v) was added to 1 mL of PAH-polyelectrolyte solution (5 mg/mL, containing 0.5 M NaCl). The suspension was mixed at constant intervals for 15 min. The excess polyelectrolyte was removed by three repeated centrifugation/buffer washing (0.5 M NaCl) and redispersion cycles. For the subsequent assembly of negatively charged PSS-polyelectrolyte, 1 mL of PSS-polyelectrolyte solution (5 mg/mL, containing 0.5 M NaCl) was added. The centrifugation/washing and polyelectrolyte incubation steps were repeated, as above, until the desired number of layers (typically four) was assembled. The polyelectrolyte-coated FDA microparticles with an outermost layer of PSS-polyelectrolyte were conjugated to antibodies as follows: Particle suspensions (0.13%, w/v) were incubated with four different concentrations (37.5, 75, 150, and 300 µg/mL) of Gt R-M IgG (whole molecule) in 0.05 M MES buffer (pH 6.0) at 20 °C for 1 h. After centrifugation at 6000g for 3 min, the supernatant was removed and its UV absorption was measured at 280 nm (3000 array spectrophotometer, Milton Roy). The antibody surface coverage on the microparticles was determined by the difference of the absorption at 280 nm between the supernatant and the original protein solution. Finally, the IgGcoated particles were separated from soluble IgG by three centrifugation/washing cycles and diluted to 0.05% (w/v). Microelectrophoresis. The surface charge of the FDA microcrystals was examined by electrophoresis using a Zetasizer 4 (Malvern Instruments) by taking the average of five measurements at the stationary level. The mobility u was converted to the zeta potential (ζ-potential) using the Helmholtz-Smoluchowski relation ζ ) 4πuη/, where η is the viscosity of the solution and  is the permittivity.20 Solid-Phase Sandwich Fluorescence Immunoassay. Nunc Maxisorp 96-well microplates (Nunc International, Rochester, NY) were coated with 2 µg/mL Gt R-M IgG (Fc specific), 100 µL/well (20) Hunter, R. J. Zeta potential in colloid science: principles and applications; Academic Press: London, 1981.

in carbonate/bicarbonate buffer (pH 9.6). The plate was incubated at 20 °C for 2 h on a plate shaker and then at 4 °C overnight. After rinsing three times with washing buffer (0.1 M PBS, 0.1% (w/v) BSA, 0.5% (w/v) Tween-20), the wells were blocked with 200 µL/well of 1.0% (w/v) BSA solution for 1 h at 20 °C. The plate was then washed three times and incubated with 100 µL/well of dilutions of M IgG as the analyte, washing buffer as the blank control, or Gt IgG (500 ng/mL) as a negative control. The plate was incubated with the samples at 20 °C for 2 h and subsequently washed three times. Anti-mouse-coated polyelectrolyte-encapsulated microcrystal suspensions (1:500-1:2000 dilutions of 0.05% (w/v) stock) were dispensed into the wells (100 µL/well), and the microplate was incubated again at 20 °C for 2 h. The soluble fluorescent label Gt R-M IgG-FITC at a dilution of 1:800 and 1:1500 (measurements seen in Figure 5 and Figure 4, respectively) was used for comparison. After incubation, excess detector antibody conjugates were washed off by three washing cycles with buffer. An aliquot of 50 µL of release reagent per well (DMSO and 1 M NaOH in a 1:1 ratio) was added to wells incubated with the antibody-coated nanoencapsulated microcrystals. Complete release of the fluorescent molecules from the microcrystals was achieved by shaking the plate for 5 min. The fluorescence intensity was measured using a fluorescence microplate reader (Fluoromark, Bio-Rad) with an excitation/emission wavelength of 485/ 535 nm. A fluorescence reader gain setting of 10 was used for measurements with FDA-labeled antibodies, a gain setting of 50 for direct FITC-labeled antibodies. From our calibration (data not shown), a gain setting of 50 leads to a 53-fold increase of instrument sensitivity compared with a gain 10 setting. Readings for the direct FITC-labeled antibodies were corrected to gain 10, from prior comparison of the assay sensitivities (the gain has no influence on the S/N ratio). The detection limit is defined as the concentration of mouse IgG corresponding to the mean fluorescence of the zero dose response plus two times the standard deviation (SD) of this measurement. RESULTS AND DISCUSSION Preparation of Biolabeled, Nanoencapsulated FDA Microcrystals. The particle size distribution of the FDA microcrystals, determined by light-scattering measurements is shown in the inset of Figure 3. Approximately 50% of the particles were found to be smaller than 0.6 µm, 90% smaller than 1.0 µm, and 99% smaller than 1.47 µm. This result is in agreement with scanning electron microscopy (SEM) analysis of the particles, as shown in Figure 3. SEM shows that the particles have a range of shapes. During the ball milling process in the presence of SDS surfactant solution, newly generated FDA/water interfaces were immediately coated with a layer of SDS molecules. The hydrocarbon chains are most likely associated with the hydrophobic microcrystal surface and the ionic groups exposed to the solution phase.21 The hydrophobic interactions between the surface of the uncharged FDA microcrystals and the SDS chains provide the driving force for adsorption.22 The adsorbed, negatively charged SDS layer introduces surface charges to the FDA crystals, making them dispersible in water and preventing their aggregation, hence conferring colloidal stability. This was verified by microelectro(21) Caruso, F.; Yang, W.; Trau, D.; Renneberg, R. Langmuir 2000, 16. (22) Myers, K. R.; Nemirovsky, A. M.; Freed, K. F. J. Chem. Phys. 1992, 97.

Figure 3. SEM micrograph of ball-milled SDS-stabilized FDA microcrystals. The image shows the shapes of the particles with the size distribution function seen in the inset. The biggest particle in the image shown has a major axis length of 0.95 µm, while the smallest particles are ∼0.15 µm in diameter.

phoresis measurements (expressed as a ζ-potential), which is an effective method to characterize the surface charge of colloid particles.23 The sign of the ζ-potential indicates whether the particle surface is positively or negatively charged, and the magnitude indicates the stability of the colloidal suspension. Generally, the greater the absolute magnitude of the ζ-potential, the more stable is the suspension. Following the ball milling process, the FDA microcrystals exhibited a ζ-potential of -50 mV, indicating a negatively charged, highly stable particle suspension. The subsequent alternate adsorption of PAH-polyelectrolyte and PSS-polyelectrolyte layers onto the SDS-coated FDA microcrystals yielded ζ-potentials of about +45 and -45 mV, respectively. As noted for other colloidal systems, this “surface recharging” indicates polyelectrolyte deposition with each adsorption step.21,23-25 The thickness of the polyelectrolyte coating, four layers total of PAH/PSS-polyelectrolyte, assembled onto amphiphilemodified FDA microparticles is only about 6-8 nm.23,25 This thickness is negligible when compared to the total size of the FDA microcrystals. The polyelectrolyte coatings did not cause any noticeable change in the morphology and size distribution of the FDA microcrystals (as observed by light scattering and SEM). Immunolabeling. The polyelectrolyte coating on the FDA microcrystals has to fulfill two functions. The first is to impart sufficient colloidal stability, and the second is to provide a suitable interface for the attachment of biomolecules (e.g., antibodies) to the microcrystals. As the polyelectrolytes may contain a variety of functional groups, such as carboxyl or amino moieties, antibodies can be assembled onto the microcrystal surface via various covalent binding protocols.26,27 Another possible approach is the use of biotinylated polyelectrolytes as an outer layer to enable the binding of various avidin-antibody conjugates.28 However, in this work, the antibody (IgG) was immobilized onto the polyelec(23) Caruso, F.; Lichtenfeld, H.; Donath. E.; Mo¨hwald, H. Macromolecules 1999, 32, 2317-2328. (24) Joanny, J. F. Eur. Phys. J. B 1999, 9, 117-122. (25) Caruso, F.; Mo ¨hwald, H. J. Am. Chem. Soc 1999, 121, 6039-6046. (26) Schwendener, R. A.; Trub, T.; Schott, H.; Langhals, H.; Barth, R. F.; Groscurth, P.; Hengartner, H. Biochim. Biophys. Acta 1990, 1026, 69-79. (27) Hermanson, G. T. Bioconjungate Techniques; Academic Press: New York, 1996.

Analytical Chemistry, Vol. 74, No. 21, November 1, 2002

5483

Table 1. IgG Surface Coverage on Polyelectrolyte-Encapsulated FDA Microcrystals (0.13% (w/v)) and Corresponding Calculated F/P Values of the Novel Immunolabel sample incubation concn of Gt R-M IgG (µg/mL) protein adsorbed (µg/mL) protein adsorbed (%) protein surface coverage Γ (mg/m2) assay concn of Gt R-M IgG after 500-fold dilution of 0.05% (w/v) stock (µg/mL) calculated F/P value

(a) FDA1.3

(c) FDA3.0

(d) FDA4.9

37.5 26 68 1.3 0.019

75 32 42 1.6 0.023

150 60 40 3.0 0.045

300 100 33 4.9 0. 077

1.9 × 105

1.6 × 105

8.3 × 104

5.1 × 104

trolyte-coated FDA microcrystal surface by adsorption. We chose this approach because adsorptive immobilization of antibody molecules onto polyelectrolyte-coated colloid particles has been extensively studied and has shown to be a successful means for the stable attachment of biomolecules by retaining their specific immunorecognition ability.25,29-31 For instance, IgG adsorption onto polyelectrolyte multilayers assembled onto polystyrene microparticles has been confirmed by monitoring the change in ζ-potentials as a function of particle suspension pH30 and quantified by singleparticle light-scattering experiments.25 IgG-coated latex microspheres have also shown to be colloidally stable.25,30,32,33 In the current study, the amount of IgG adsorbed onto the encapsulated FDA microcrystals was determined spectrophotometrically (see Experimental Section). The protein surface coverage is calculated by assuming an average size of 500 nm for the FDA microcrystals (Figure 3). Four different IgG concentrations were chosen for incubation with the encapsulated FDA microcrystals. The experimental data are summarized in Table 1 and indicate that the protein surface coverage on the FDA microcrystals increases with increasing protein incubation concentration, while the protein adsorption efficiency (percentage of protein adsorbed) decreases. For example, following incubation of the encapsulated FDA microcrystals with an IgG solution of 37.5 µg/mL in 0.05 M MES buffer for 1 h at 20 °C, 68% of the protein added was adsorbed on the surface of the FDA particles, giving a protein surface coverage of 1.3 mg/m2. In comparison, incubation of the encapsulated FDA microcrystals at the highest IgG concentration of 300 µg/mL resulted in 33% of the total protein immobilized onto the particle surface, yielding a surface coverage of 4.9 mg/m2. The theoretically calculated surface coverage value for a close-packed IgG monolayer is in the range of 2.0-5.5 mg/m2, depending on the different orientations of the adsorbed IgG molecules.34,35 Under the present experimental conditions, it is possible that an IgG monolayer may be formed for the samples with 3.0 and 4.9 mg/ m2 surface coverage; however, it is also likely that protein aggregates may have formed on the FDA particle surface. To (28) Savage, M. D. Avidin-Biotin Chemistry: A Handbook; Pierce Chemical Co.: Rockford, IL, 1992. (29) Lvov, Y.; Moehwald, H. Protein architecture: Electrostatic layer-by-layer assembly of proteins and polyions; Marcel Dekker: New York, 2000. (30) Yang, W. J.; Trau, D.; Renneberg, R.; Yu, N.-T.; Caruso, F. J. Colloid Interface Sci. 2001, 234, 356-362. (31) Lvov, Y.; Caruso, F. Anal. Chem. 2001, 73, 4212. (32) Peula, J. M.; de las Nieves, F. J. Colloid Surf. A 1994, 90, 55-62. (33) Davalos-Pantoja, L.; Ortega-Vinuesa, J. L.; Bastos-Gonzalez, D.; HidalgoAlvarez, R. Colloids Surf. B 2001, 20, 165-175. (34) Caruso, F.; Furlong, D. N.; Ariga, K.; Ichinose, I.; Kunitake, T. Langmuir 1998, 14, 4559-4565. (35) Buijs, J.; Lichtenbelt, J. W. T.; Norde, W.; Lyklema, J. Colloids Surf. B 1995, 5, 11-23.

5484

(b) FDA1.6

Analytical Chemistry, Vol. 74, No. 21, November 1, 2002

calculate the number of fluorescence molecules released by a single encapsulated FDA microcrystal label, cubic crystal morphology with dimensions of 500 × 500 × 500 nm3 (the average value obtained from the light-scattering measurements) was assumed. As each FDA molecule can be converted into one fluorescein molecule, the number of fluorescein molecules per microcrystal is calculated to be ∼1.5 × 108. This value is ∼100 times larger than the fluorescein equivalent (FE)36 value of commercial fluorescent latex particles of the same size (yellowgreen FluoroSphere, Molecular Probes, Inc.) and ∼10 000 times larger than those of liposomes37 that carry up to 104 dye molecules in a solubilized state. This difference arises from the fact that in our approach the FDA microparticle is totally composed of precursor fluorescent molecules in a crystalline state, providing the highest possible packing density. From the calculated number of fluorescein molecules per FDA microcrystal and the measured values of the protein surface coverage, F/P values can be calculated. Protein surface coverage and calculated F/P values are summarized in Table 1. Calculated F/P values for microcrystal labels range from 5 × 104 to 2 × 105 and are ∼104 times higher than those of directly covalent labeled antibodies (which carry normally 4-8 fluorescence molecules) and about 100-104 times higher than those obtained with microspheres and liposomes (by assuming a similar size and protein surface coverage). The F/P value of an immune detection system reflects its potential amplification rate. In practice, the achieved performance (limit of detection and sensitivity) is not only a function of the calculated F/P values. Other factors, for example, unspecific binding, changes in sensitivity and specificity of the labeled biorecognition molecule, and assay concentration of the biorecognition molecule strongly influence the assay performance. Utilization of the FDA Particulate Labels in Immunoassays. Sandwich assays of M IgG as an analyte and Gt R-M IgGFDA particles as fluorescent labels were performed in a 96-well microplate. After performing the immunoreaction, solubilization, and release, conversion of the FDA molecules into soluble fluorescein was achieved by treatment with a 1:1 mixture of DMSO and 1 M NaOH. Fluorescein fluorescence intensity (in DMSO, 1 M NaOH) was linear from a concentration of (0-5) × 10-6 mol/L (data not shown). The final fluorescein concentration after release must be kept in this range. Figure 4 shows calibration curves of Gt R-M IgG-FDA microcrystal labels having different surface (36) The FE value defines the number of fluorescein molecules necessary to produce the same light intensity as fluorescein-labeled particles under ideal nonquenching conditions. Self-quenching effects often result in the actual number of fluorescein molecules in the particle to be much higher. (37) Singh, A. K., Harrison, S. H.; Schoeniger, J. S. Anal. Chem. 2000, 72, 60196024.

Figure 4. Sandwich fluorescence immunoassay of M IgG using (a) FDA1.3, (b) FDA1.6, (c) FDA3.0, (d) FDA4.9 (all at 1:500 dilution of 0.05% (w/v) stock measured at gain 10) microcrystals, and (e) Gt R-M IgGFITC (1:1500 dilution, measured at gain 50) as labels. The FIU reported were obtained after subtraction of the blank background. Error bars correspond to standard deviations (( SD, n ) 2).

Figure 5. Sandwich fluorescence immunoassay of M IgG using three different dilutions of FDA microcrystal-labeled detector antibody. Measurements of assays using FDA microcrystal labels (a-c) were carried out with a fluorescence reader sensitivity setting of gain 10 and measurements using direct FITC labels (d) with gain 50. From our instrument calibration (data not shown), gain 50 settings result in 53-fold sensitivity increase compared with gain 10. The mean blank fluorescence background intensities for FDA-labeled antibodies and for direct FITC-labeled antibodies were 214 and 26 FIU, respectively and were subtracted from the reported values. Error bars correspond to standard deviations (( SD, n ) 2).

coverages of detector antibody (Gt R-M IgG) in comparison to a direct FITC-labeled detector antibody. The fluorescence signals increase with increasing M IgG concentration direct proportional in the range from ∼4 to 100 ng/mL and show a further increase until a concentration of 500 ng/mL is reached. FDA microcrystals having a high IgG surface coverage (Figure 4c,d) performed better than those with lower surface coverage (Figure 4a,b). The system was further studied in terms of its signal-to-noise ratio and amplification by using FDA microcrystal labels having high surface coverage but different dilution factors (Figure 5). In this study, the assay sensitivity in relative fluorescence intensity units (FIU) per analyte concentration change (in ng/ mL; in the proportional region of the assay between 4 and 100 ng/mL analyte) was compared between the FDA microcrystal labeled systems and a direct labeled antibody system. Direct comparison of assay performance of different bioassay systems (microcrystal labeled versus direct labeled) is a difficult task. A suitable “measure” that indicates similar assay conditions has to be defined. In this work, we define the total detector antibody concentration in the assay as this measure. To ensure that the microcrystal-based label system is not favored over the direct label system, an ∼20-fold higher R-M IgG concentration (1.38 µg/mL) was chosen for the direct label system. For the chosen sandwich assay (with sequential incubation of the analyte and detector antibody), we can assume that a higher concentration of detector antibody will result in higher signals and a higher FIU value until saturation is achieved. Figure 5 shows calibration curves for Gt R-M IgG-FDA microcrystal labels in comparison to a direct FITC-labeled detector antibody. The signal-to-noise ratio of assays using FDA-labeled antibodies was higher than for those using the direct labeled antibody at analyte concentrations below 30 ng/mL. For example,

a limit of detection of 4 ng/mL was achieved for the FDA labeled and 16 ng/mL for the direct labeled antibody (Table 2). The assay sensitivity observed using FDA-labeled antibodies was divided by the sensitivity observed using direct labeled antibodies as summarized in Table 2 (after normalization as explained in the Experimental Section). Our comparison resulted in a 70-2000-fold higher sensitivity of the assays using FDA microcrystal labels, depending on analyte concentration and dilution. For example, more than 1200-fold increase of sensitivity was observed between 1 and 30 ng/mL analyte (linear range) and more than 380-fold increase for higher analyte concentrations (Table 2, 1:500 dilution). The current 1200-fold amplification is still 1 order of magnitude lower than the theoretical value of ∼104, calculated from the F/P values of the two systems. However, the presented approach is still in an early phase of its development, and considerable improvement in sensitivity and limit of detection can be expected with optimization in the future. Another important quality of any immunoassay is the degree of nonspecific binding. Usually, the nonspecific binding defines the limit of detection that can be reached. To optimize nonspecific binding, typically low concentrations of the fluorescent-labeled detector antibody are favorable. Nonspecific binding is defined as the signal that is observed in absence of the analyte. The mean nonspecific binding intensity of the measurements presented in Figure 5 are 214 FIU and 26 FIU for FDA-labeled antibodies and for direct FITC labeled antibodies, respectively. No increase of nonspecific binding was observed by using Gt IgG as a negative control (instead of the analyte M IgG), confirming that the detector antibody adsorbed onto the FDA microparticles is immunoactive and recognizes its specific antigen. Analytical Chemistry, Vol. 74, No. 21, November 1, 2002

5485

Table 2. Signal-to-Noise Ratios (S/N) and Sensitivity-to-Sensitivity Ratios (S/S) of FDA-Labeled to FITC-Labeled Antibodiesa S/N of Gt R-M IgG detector antibody having different labels

S/S of Gt R-M-FDA/ Gt R-M-FITC

analyte (ng/mL)

(a) FDA 1:500 dil

(b) FDA 1:1000 dil

(c) FDA 1:2000 dil

(d) FITC

(a) FDA 1:500 dil

(b) FDA 1:1000 dil

(c) FDA 1:2000 dil

500.00 250.00 125.00 62.50 31.25 15.63 7.81 3.91 1.95 0.98

9.43 10.02 8.82 7.42 5.93 4.29 3.35 2.20 1.98 1.46

6.83 7.00 6.53 5.51 4.59 3.27 2.63 1.71 1.67 1.47

5.20 6.43 5.83 5.08 4.22 2.87 2.42 1.65 1.68 1.39

14.38 12.50 8.82 5.50 3.52 2.36 1.75 1.49 1.28 1.20

384 478 609 869 1191 1471 1906 1502 2162 1391

202 242 329 465 661 775 1008 678 1129 1090

71 106 139 204 287 309 425 301 556 434

a

Raw data were extracted from the measurements presented in Figure 5 and were normalized to the same fluorescence reader gain setting of

10.

Our concept of using a polyelectrolyte coating as an interface for bioconjugation onto microcrystalline particles provides the following advantages: (1) The fluorescent material (or its precursor) making up the particles does not need to be water-soluble or possess groups for bioconjugation, as this functionality is provided by the polyelectrolytes. The decoupling of the chromogenic and bioconjugation requirements of fluorescent molecules provides a way to optimize each requirement separately and permits the selection of a large number of organic fluorescent dyes. (2) As the fluorescent dye (or precursor) occupies the majority of the volume of the particle and is densely packed, the dye equivalent of one particulate label and the dye/protein ratio is inherently high. (3) Since the fluorescent dye can be released from the encapsulated label with organic solvent, the quenching problem normally arising from high dye/protein ratio labels can be prevented. The fluorescence measurement is taken after the dye molecules are solubilized in a large volume of the release reagent, instead of being confined within the neighborhood of the immunoreagent. (4) The preparation of the microparticles and the subsequent biolabeling are straightforward to perform and controllable. (5) High signal amplification rates are realistic without the need of long incubation times, as is required for enzyme labels.

were confirmed to be immunologically active. The protein surface coverage for incubation with four different initial IgG concentrations was determined to be 1.3, 1.6, 3.0, and 4.9 mg/m2 leading to a calculated F/P value of the order of 105. The sensitivity of the prepared FDA labels was compared with a conventional FITClabeled IgG in a sandwich immunoassay for M IgG and was found to be 70-2000-fold higher. From a theoretical viewpoint, the presented approach has the potential to reach even higher amplification. The reported method is restricted to test formats carried out in cavities to prevent mixing of fluorescent molecules released from different samples. More than one assay can be carried out in one cavity by using different fluorophores for encoding. However, the use of more than three to five different fluorophores is unrealistic. The format of the here demonstrated label provides some inherent advantages that solve the problem of quenching and to achieve a high F/P ratio. Theoretically, F/P ratios up to 108 are possible and the method is expected to outperform existing technologies after optimization. Using visible dyes, the high sensitivity of these systems may also lead to bioanalytical tests that can be detected by the naked eye. The speed and robustness makes the reported technique potentially suitable for field tests where immediate results are required.

CONCLUSION We have demonstrated the general applicability of a new class of particulate labels for biochemical assays. The technology presented is based on microcrystalline particles encapsulated within a nanoscale polyelectrolyte coating that provides an interface for the conjugation with biorecognition molecules. FDA microcrystals, a nearly water-insoluble precursor of fluorescein, with an average size of 500 nm were chosen as labels for utilization in a sandwich immunoassay for mouse IgG. IgG molecules were adsorbed onto polyelectrolyte-encapsulated FDA microcrystals and

ACKNOWLEDGMENT This work was supported by grants of the Research Grant Committee (RGC) of Hong Kong SAR, the German Academic Exchange Program (DAAD), and the BMBF (BioFuture Program).

5486

Analytical Chemistry, Vol. 74, No. 21, November 1, 2002

Received for review January 28, 2002. Accepted July 19, 2002. AC0200522