INTRODUCTION
Bioconjugation of nanoparticles (NPs)[1-14] is the attachment of specific biological molecules or components to nanoparticles. The resulting structures represent the convolution of biotechnology and nanotechnology, and yield hybrid materials, processes, and devices that can utilize both the unique optical and magnetic properties of NPs and highly selective binding of biological interactions. The combination of these features can potentially make a prominent impact in current biomedical technologies, and possibly in nanoelectronics, microphotonics and related fields. Antigen and antibody interaction is a naturally occurring immunology interaction of the biological defense system, and has been widely used as molecular recognition method for medical diagnostics and detection of biological threat agents.[15] The bioconjugation of NPs with antigens and antibodies are interesting and important as a method for their organization in more complex structures, and as a pathway to new sensing and imaging technologies. In the viewpoint of biologically programmed assembly of nanostructures,[12'16'17] the similarity of sizes of NP and proteins limits the overall number of affinity ligands around one semiconductor core to a few proteins leading to, for instance, 1-, 2-, 3-, and 4-valent quantum dots. The practical aspect of NP-antigen/antibody conjugation is related to the further development of immuno-luminescence as a technique that affords highly sensitive and specific detection of various biological and nonbio-logical analytes of military and civilian importance.
Colloid gold nanoparticles have been used in labeling antigens and antibodies as contrasting agents of various immunoassays;1-18-20-1 however, their sensitivity is limited by the nature of colorimetric detection. Luminescence spectroscopy is more sensitive than absorption spectros-copy because of its substantially better signal-to-noise ratio. We observed the excited-state dipole-dipole coupling between NPs of different sizes in the constructed immunocomplex, which results in the excitation energy transfer from NPs of smaller diameters to those with bigger diameters, i.e., Forster resonance energy transfer (FRET). The detection limits of analytical processes based on FRET can be as low as 10 ppt, with a linear dynamic range of 0.1-1000 ppb,[21,22] while the utilization of antibodies enables the selective detection of substrates, which may differ only by a few atoms. Highly luminescent semiconductor NP, or quantum dots, are a new class of luminescence materials providing a number of advantages over organic dyes. They can be made in a single-step synthesis, with precisely controlled size distribution and tunable emission spectrum. Compared to organic dyes such as Rhodamine, this new class of luminescent material are 20 times brighter and 100 times more stable against photobleaching, and one-third as wide in spectra linewidth.[3] Therefore luminescence nanopar-ticles will have tremendous potential in biomedical imaging and diagnostic applications.
Typically, NPs are synthesized by arrested precipitation in the presence of organic molecules strongly coordinating to metal ions, such as thiols or phosphines.[23,24] The simplified description of the product is a semiconductor core coated with a monolayer of organic molecules attached to surface metal sites. The core of the NPs is highly crystalline, which is a prerequisite for strong luminescence. Many different thiol derivatives may be used for the stabilization of II-VI nanoparticles. The terminal end of the thiol can be conveniently used for further func-tionalization of nanoparticles and for conjugation to biological molecules by either nonspecific electrostatic absorption or specific conjugation reactions. Most antigens and antibodies are proteins, have a number of amino acids that provide functional groups which can be chemically conjugated, such as -NH2 (lysine, and N-terminal), -COOH (aspartic acid, glutamic acid, and C-terminal), and -SH (cysteine).
The direct NP-protein conjugation method frequently used is the 1-ethyl-3(3-dimethylaminopropyl) carbodii-mide hydrochloride (EDC)/N-hydroxysulfo-succinimide (sulfo-NHS) reaction (Scheme 1).[20,25-27] NHS-conjugat-ed proteins have the highest bioactivity among other conjugates, as established by several comparative stud-ies.[28-30] The carboxylic acid group of thioglycolic acid-stabilized NPs will form amide bond with the primary amine groups of the protein, or, the amine groups of a cysteine-stabilized NPs can link to the carboxylic acid group of protein.
Scheme 1 Schematics of the sulfo-NHS/EDC conjugation reaction.
To explore and demonstrate the potential of NP Bioconjugates, we used EDC/sulfo-NHS reaction to conjugate different-sized CdTe NPs to two antigen/ antibody systems: 1) bovine serum albumin (BSA) and anti-BSA IgG (immunoglobulin G)[8] and 2) Brucella suis and its antibody Bru-38 (Anti-GBa-O side chain). Antigens were conjugated to red-emitting CdTe NPs, while green-emitting NPs were attached to the corresponding antibodies. FRET was observed upon the formation of immunocomplex between the complementary antigens/ antibodies. the competitive inhibition of FRET by unla-beled antigens was investigated.
METHODOLOGY
The optimized EDC/sulfo-NHS conjugation protocol is as follows: a reaction mixture containing 0.05 mM CdTe NP, 1.5-2.5 mg/mL antigen or antibody, 0.05 M NHS, and 0.05 M EDC in pH 7.0 PBS buffer was prepared and kept in room temperature for 2-4 hr, then stored overnight at 4°C. This allows the unreacted EDC to hydrolyze and lose its activity. After that, a small amount of precipitate is formed, likely consisting of unconjugated NPs, which are known to agglomerate and become nonemissive at fairly low pH. The precipitate (if any) is removed by centrifu-gation. The stock, ready-to-use solution of the product was stored at 4°C. Optionally, the conjugates can be dialyzed with Spectra/Por® 4 membrane with molecular weight cut-off 12,000-14,000 (Spectrum Laboratories, Inc.) in pH 7.4 phosphate-buffered saline (PBS) to remove the small molecules.
The NP bioconjugates were characterized by gel electrophoresis, gel permeation high-performance liquid chromatography (HPLC), circular dichroism (CD), fluorescence spectroscopy, and enzyme-linked immunosorbent assay (ELISA) test. Optical absorption spectra were obtained on a Hewlett-Packard 8453 diode array spectrophotometer using 1-cm quartz cuvettes. Fluorolog 3 and Fluoromax 2 from JY SPEX were used to register the luminescence spectra. The right-angle registration mode with no intermediate filters was utilized in all measurements. All FRET experiments were carried out after centrifugation of the solutions to remove all particles aggregated as a result of oxidation and coagulation. The quantum yield of NP luminescence was determined by using Rhodamine B in ethylene glycol (1max=580 nm, quantum yield 1) as a standard as described elsewhere.[7,8] The luminescence intensity does not decrease even after the samples were stored for over a month at 4°C. Gel permeation HPLC experiment was carried out on an HP-1090-II instrument equipped with Jordi GPC column 500A with physical dimensions of 300 x 7.8 mm. Deionized 18 MO water was used as an eluent. The optical adsorption signal was monitored at 254 nm by a diode array detector.
RESULTS AND DISCUSSION
Investigate the Formation of NP-BSA and NP-Anti-BSA Bioconjugates by Electrophoresis
The native electrophoresis results (Fig. 1, left panel) show that both NP-conjugated BSA (well 2) and anti-BSA IgG (well 4) bands become more mobile in the electric field than the unlabeled biospecific ligands (wells 1 and 3). The BSA monomer band shifts from the relative marker of 65 to 47 kDa, which demonstrates that the high negative charge of the NP and their compactness overcomes the increase of their mass as a result of labeling. Note that the commercial BSA shows two other bands at 100 and 150 kDa, corresponding to BSA-BSA dimer and globulins, respectively. This observation agrees with the specifications of Sigma; these compounds have minor influence on the biospecific reactions discussed below. Both of these bands shift synchronously with BSA monomer to smaller masses after the NP conjugation. By mixing NP with BSA and IgG without adding coupling reagents (EDC and sulfo-NHS), no evidence of nonspecific binding was observed in native gel electrophoresis.
Fig. 1 Native and SDS-PAGE electrophoresis of CdTe bioconjugates on 4-20% gradient Tris-HCl precast gels (Bio-Rad). Left panel-native, stained by Coomassie Blue. Center panel—SDS-PAGE, stained by Coomassie Blue. Right panel—SDS-PAGE, luminescence image (excitation 360 nm). Wells: 1) BSA; 2) green-emitting NP-BSA; 3) anti-BSA IgG; 4) red-emitting NP-IgG; N) Free CdTe NPs; L) standard protein ladder, molecular weight are marked on the side in kiloDaltons (kDa). Note that in native electrophoresis, the position of the band is not linearly proportional to the molecular weight, because of the different charge status of each sample.
The mobility of the proteins in sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) (Fig. 1, center panel) is determined by the mass/charge ratio of denatured protein chains carrying SDS, which imparts negative charge to them. Interestingly, the band positions of NP-labeled and unlabeled proteins virtually coincide at the 65-kDa marker (Fig. 1, center panel). Similar to the native gel results, the increase of mass resulting from the addition of NPs to the protein is compensated by the increase of the overall charge density of the conjugate. Unlike NP-BSA, the SDS-PAGE band of NP-IgG conjugate remains at 150 kDa, while the free IgG are mostly broken apart into small molecular weight fragments as a result of SDS and heating treatment (Fig. 1, center panel, wells 3-4). Considering the SDS-PAGE data, we emphasize three factors. 1) Estimates of molecular masses based on gel electrophoresis results can give erroneous outcomes for bioconjugates from highly charged NPs. Different experimental techniques must be used for this purpose. 2) A brief heat treatment (150 sec) at 96°C was used for the preparation of SDS-PAGE samples. Because high-temperature denaturation destroys the tertiary structure of the protein, little nonspecific binding of NP and BSA is to be expected after that. Therefore the observation of NP luminescence in the SDS-PAGE bands demonstrates covalent linkage between protein and the semiconductor units. 3) Bioconjugation to NP may increase the stability of antibodies. The relative intensity of the SDS-PAGE bands indicates that the conjugated antibodies are more resilient to this temperature than the unlabeled ones. The unlabeled antibodies (well 3) are mostly broken apart by SDS and heating into small fragments, showing up as band at 40-50 kDa; whereas for NP-IgG conjugates, this band has significantly lower intensity—the antibody remains mostly intact. The latter effect was reproduced in several control experiments.
The direct evidence of the successful conjugation of CdTe to BSA and IgG can be found from the luminescence analysis of the gel plates. The NP conjugate bands show strong luminescence, while the free proteins do not show any detectable signal in the luminescence image (Fig. 1, right panel). As before,[7] the gel pieces cut out of the gel plates in the area of the conjugate bands reveal the luminescence spectra with identical peaks to the original NPs.
Confirm the Formation of NP-BSA and NP-Anti-BSA Bioconjugates by HPLC
The formation of NP bioconjugates was also confirmed by HPLC size exclusion chromatography (Fig. 2), which is complementary to electrophoresis. The original BSA revealed three HPLC peaks congruent to the three bands seen in gel electrophoresis (Fig. 1). Because the species with higher molecular weight are eluted at shorter retention times, the observed HPLC peaks at retention times 7.8, 6.8, and 5.4 min should be attributed to BSA monomer, BSA dimer, and globilins, respectively. After conjugation to CdTe, the same three peaks can be seen at 6.5, 5.3, and 4.2 min, all being shifted to higher molecular weight, as expected for the attachment of NPs to proteins. Considering that the approximate molecular weight of 5.0-nm red-emitting CdTe attached to BSA is 240 kDa (specific density of CdTe is 6.2 g/cm3), the HPLC peak shift for monomeric BSA shows that not more than one NP is attached to this protein. Similar behavior can be observed for the bioconjugates of IgG. An HPLC peak attributed to free IgG was observed at 6.7 min, while shifting to 6.1 min after conjugation to green-emitting NP of ca. 2.5 nm in diameter. Their molecular mass is estimated to be 30 kDa; therefore, the HPLC molecular weight shift corresponds to the attachment of about one quantum dot per IgG as well. We want to emphasize that the estimates of NP-protein composition of the conjugates should be taken with caution and should be treated as preliminary data. The retention of NPs and polymers/ proteins in gel permeation media may follow different scaling schemes, and therefore introduce an unforeseen error in them. No free BSA or large oligomeric NP-BSA agglomerates with gradually increasing molecular masses can be detected in elution curves (Fig. 2). Besides such oligomers, the luminescent bands seen at the starting line in gel electrophoresis can possibly originate from aggregates produced by accumulation on dust particles, rather than multiple chemical cross-linking because of BSA’s tendency to adsorb on many surfaces.[31]
Fig. 2 HPLC gel permeation elution curves for: 1) BSA and 2) NP-BSA.
Study Structure of NP-BSA and NP-Anti-BSA Bioconjugates by CD Spectroscopy
The integration of antigen and antibody structure after NP conjugation was confirmed by CD spectroscopy, which shows little change between both BSA and NP-BSA conjugate, as well as the anti-BSA and NP-anti-BSA pair. These data show that the tertiary structure of both proteins remains mostly intact after conjugation.1-8-1
Examine Biological Activity of NP-BSA and NP-Anti-BSA Bioconjugates by ELISA
The antigen-antibody binding affinity was retained for both conjugates as evaluated by standard ELISA experiments.1-8-1 NP-IgG retained 25-50% of binding affinity to free BSA, as compared to free IgG, while NP-BSA retained almost 100% of binding capability to unmodified antibody. The decrease in NP-IgG binding affinity is believed to be a result of partial blocking of binding sites 1) by NP positioned close to theN terminal of the IgG and 2) by other bioconjugates in forming dynamic aggregates. The luminescence properties of NPs remained unchanged in the conjugates as measured by fluorescence spectroscopy.
FRET Study of NP-BSA and NP-Anti-BSA Bioconjugates
When NP-IgG (anti-BSA) with green luminescence are combined with NP-labeled BSA with red luminescence, the NP-IgG/BSA-NP immunocomplex should form. As expected, a significant enhancement of the NP-BSA’s red emission at 611 nm and the corresponding quenching of the green emission of NP-IgG at 555 nm are observed after the self-assembly of the labeled biospecific ligands in the immunocomplex (Fig. 3). The mutual affinity of the antigen and antibody brought the NPs close enough together to allow the resonance dipole-dipole coupling required for FRET to occur. Thus the energy of the ex-citonic state in the green-emitting NP was transferred to the similar state of the red-emitting NP with lower exciton energy. FRET efficiency is particularly high for green/red NP pairs because of the strong overlap of their emission and absorption spectra. Importantly, when unlabeled BSA was added to the immunocomplex, it competitively bound to NP-IgG, and replaced NP-BSA in the immunocomplex, thereby inhibiting the FRET process. Consequently, the green emission peak of NP-IgG at 555 nm can be observed again, while the red emission peak at 610 nm shows decreased intensity. Following this approach, we were able to clearly detect as low as 1.5 x 10" 8 M BSA (Fig. 4). Unlike ELISA, the described detection process does not require the multiple binding and washing steps. High- concentration egg albumin, which does not bind to anti-BSA IgG, shows no influence on FRET between NP-IgG and NP-BSA. It demonstrates the high biospecificity of the prepared conjugates and virtually absent interference from the noncomplimentary proteins.
Fig. 3 FRET-based detection of BSA in solution. 1) NP-BSA; 2) NP-BSA+NP-IgG; 3) BSA + NP-BSA+NP-IgG. Concentrations: NP-BSA 5 x 10" 7 M, NP-IgG 5 x 10" 7 M, BSA 2 x 10"6 M.
Study of NP-Conjugated B. suis and Its Antibody
After successfully demonstrating FRET-based detection in the BSA/anti-BSA model system, we examined it with a more practical antigen-antibody system. We obtained whole-cell killed B. suis and the corresponding antibody Bru 38 (anti-GBa-O side chain) and prepared the green NP-conjugated B. suis and red NP-conjugated Bru 38 according to the NP-anti-BSA protocol. We used CdTe/ CdS NPs with green emission (561 nm) and red emission (601 nm) for the conjugation experiment. The concentration of Bru38 used in the synthesis was 3 mg/mL and the final concentration in the conjugation was 1.2 mg/mL. The concentration of B. suis (estimated from optical density, OD 600 nm absorption value) used in the synthesis was 2.7 x 10"12 M (1.6 x 109 cells/mL) and the final concentration in the conjugation was 7 x 10"13 M (4 x 108 cells/mL). The conjugated B. suis was purified by centrifuge and washed with 0.1 M PBS buffer three times. Native and SDS-PAGE gel electrophoresis results of the conjugated Bru 38 are very similar to the NP-anti-BSA conjugate. Although the surface of the bacteria is complex, it should contain numerous primary amine sites for the NP binding, and many of these binding sites should be close enough to the antibody-binding site to allow FRET to occur. Considering the much larger size of bacteria, multiple attachments of NPs on a single bacterium is expected; however, we did not verify it at this stage.
Fig. 4 Result of FRET-based detection limit—concentrations: 1.5 x 10"8 M BSA, NP-BSA 3.8 x 10" 9 M NP-BSA and 1.6 x 10" 9 M NP-IgG.
Similar to BSA results, we observed (Fig. 5) FRET as expected when we mixed NP-anti-B. suis with the NP-B. suis solution. It results as a fluorescence enhancement of the red-emission NP-anti-B. suis at 601 nm. The present unlabeled B. suis bound to NP-Bru 38 and competitively inhibited the FRET process, and dramatically reduced the fluorescence enhancement at 601 nm.
Binding kinetics, as well as specificity of the FRET, has been studied, and the results are shown in Fig. 6.
Fig. 5 FRET-based detection of Brucella suis in solution. 1) NP-anti-B. suis; 2) NP-B. suis; 3) NP-B. suis+NP-anti-B. suis; 4) B. suis + NP-B. suis + NP = anti-B. suis. Concentration: NP-anti-B. suis (Bru38) 2 x 10"7 M, NP-B. suis 3.3 x 10"15 M (2 x 106 cells/mL), B. suis 1.3 x 10"14 M (8 x 106 cells/mL).
Fig. 6 FRET transfer kinetics between NP-B. suis and NP-anti-B. suis showing the influence of the presence of B. suis. The FRET is qualified by the intensity of the red-emission peak (601 nm). Concentrations: NP-anti-B. suis (Bru38) 8 x 10"8 M; NP-B. suis 1.6 x 10"15 M (106 cells/mL); B. suis 6.5 x 10"15 M (4 x 106 cells/mL).
The presence of unlabeled B. suis inhibited the FRET process. The full reaction takes about an hour to finish, because of the antigen/antibody binding nature. However, FRET and competitive inhibition effect can be clearly observed a few minutes after mixing the components together.
CONCLUSION
Luminescence NPs represent a new class of biological labeling reagents with advantages over traditional dyes. To demonstrate the potential of this class of materials, two complementary antigens and antibodies systems were labeled with thiol-stabilized green- and red-emitting CdTe NPs. They retain substantial bioactivity and can form the corresponding immunocomplex, which pairs NPs with different emission properties in a supramolecular assembly. Quenching of the green-emitting NPs and the enhancement of the luminescence of their red-emitting counterparts signify the presence of FRET in the immu-nocomplex. The introduction of unlabeled antigen into the complex solution induced the competitive inhibition of FRET. The exceptional specificity of immunocomplex reactions affords unique possibilities for assembling organized assemblies of NPs and other nanocolloids. After synthetic optimization such as enhancement of the luminescence quantum yield of the bioconjugates, FRET effects can be exploited in biosensing as well as in hybrid photoelectronic devices.
