Anal. Chem. 2006, 78, 3758-3764
Aptamer-Based Detection and Quantitative Analysis of Ricin Using Affinity Probe Capillary Electrophoresis Amanda J. Haes, Braden C. Giordano, and Greg E. Collins*
Naval Research Laboratory, 4555 Overlook Avenue, SW, Chemistry Division, Code 6112, Washington, D.C. 20375-5342
The ability to detect sub-nanomolar concentrations of ricin using fluorescently tagged RNA aptamers is demonstrated. Aptamers rival the specificity of antibodies and have the power to simplify immunoassays using capillary electrophoresis. Under nonequilibrium conditions, a dissociation constant, Kd, of 134 nM has been monitored between the RNA aptamer and ricin A-chain. With use of this free-solution assay, the detection of 500 pM (∼14 ng/mL) or 7.1 amol of ricin is demonstrated. The presence of interfering proteins such as bovine serum albumin and casein do not inhibit this interaction at sub-nanomolar concentrations. When spiked with RNAse A, ricin can still be detected down to 1 nM concentrations despite severe aptamer degradation. This approach offers a promising method for the rapid, selective, and sensitive detection of biowarfare agents. Ricin is a plant lectin from the castor bean plant Ricin communis that is toxic (LD50 5 µg/kg and 90 h to time of death by intravenous injection and 3-5 µg/kg and 60 h to time of death by inhalation).1 It is identified as a class II ribosome-inactivating protein, meaning it consists of two chains, an A chain and B chain (linked by a single disulfide bond).2,3 The A chain is toxic to cells while the B chain is necessary for attachment to the cell surface. The relative ease of its production and wide availability make ricin a potential threat as a terrorist weapon. The traditional methods for ricin detection are using antibodybased immunoassays4-6 and enzyme-linked immunosorbant assays (ELISA).7-10 While each technique is superior in either sensitivity, * To whom correspondence should be addressed. E-mail: greg.collins@ nrl.navy.mil. Phone: (202) 404-3337. Fax: (202) 404-8119. (1) Eitzen, E. Medical Management of Biological Casualties Handbook, 4th ed.; U.S. Army Medical Research Institute of Infectious Diseases: Frederick, MD, 2001. (2) Endo, Y.; Tsurugi, K. J. Biol. Chem. 1988, 263, 8735-8739. (3) Endo, Y.; Tsurugi, K. J. Biol. Chem. 1987, 262, 8128-8130. (4) Ligler, F. S.; Taitt, C. R.; Shriver-Lake, L. C.; Sapsford, K. E.; Shubin, Y.; Golden, J. P. Anal. Bioanal. Chem. 2003, 377, 469-477. (5) Rubina, A. Y.; Dyukova, V. I.; Dementieva, E. I.; Stomakhin, A. A.; Nesmeyanov, V. A.; Grishin, E. V.; Zasedatelev, A. S. Anal. Biochem. 2005, 340, 317-329. (6) Stine, R.; Pishko, M. V.; Schengrund, C.-L. Anal. Chem. 2005, 77, 28822888. (7) Koja, N.; Shibata, T.; Mochida, K. Toxicon 1980, 18, 611-618. (8) Griffiths, G. D.; Newman, H.; Gee, D. J. J. Forens. Sci. Soc. 1986, 26, 349358. (9) Leith, A. G.; Griffiths, G. D.; Green, M. A. J. Forens. Sci. Soc. 1988, 28, 227-236.
3758 Analytical Chemistry, Vol. 78, No. 11, June 1, 2006
ease of use, or required assay time, improvement of these methods is required. For instance, Koja reported a detection limit of 4 ng/ mL ricin using an ELISA method; however, the assay took several hours and required highly trained personnel to perform the assay.7 A sandwich immunoassay when combined with gold colloids has demonstrated ricin detection within 10 min; however, the sensitivity of the assay suffers (10 ng/mL ricin).11 Ligler has demonstrated 8 ng/mL ricin detection using an array-based immunosensor with analysis times of 15-30 min as a result of significant user intervention.4 Capillary electrophoresis (CE) has been shown to be a viable alternative to traditional immunoassays when coupled with laserinduced fluorescence (LIF) detection. Capillary-based immunoassays12 take two primary forms: competitive13,14 and noncompetitive.15 In the noncompetitive assay, an antibody for the protein of interest is fluorescently labeled. Typically, a mobility shift is noted between the labeled antibody and antigen/antibody complex. This allows for direct detection of the antigen, due to the fact that the antigen concentration is proportional to the signal associated with the antigen-antibody complex band. The competitive assay, on the other hand, incorporates a known amount of labeled antigen into the background electrolyte. As with the noncompetitive assay, two species are observed, a band associated with free labeled antigen and a band associated with the labeled antigen-antibody complex. When unlabeled antigen is added to the sample, the signal associated with the free antigen band increases, while the signal associated with the complex decreases due to the unlabeled antigen competing for binding sites on the antibody. As with the noncompetitive assay, both qualitative and quantitative results are attainable. Conceptually, capillary-based immunoassays or affinity probebased CE (APCE) is simplistic; however, in practice, that is not always the case. Perhaps the most significant problem is inconsistencies with respect to the labeling process.16 Antibodies and (10) Poli, M. A.; Rivera, V. R.; Hewetson, J. F.; Merrill, G. A. Toxicon 1994, 32, 1371-1377. (11) Shyu, R.-H.; Shyu, H.-F.; Liu, H.-W.; Tang, S.-S. Toxicon 2001, 40, 255258. (12) Heegaard, N. H. H. Electrophoresis 2003, 24, 3879-3891. (13) Wan, Q.-H.; Le, X. C. J. Chromatogr. B 1999, 734, 31-38. (14) Lam, M. T.; Wan, Q. H.; Boulet, C. A.; Le, X. C. J. Chromatogr. A 1999, 853, 545-553. (15) Hafner, F. T.; Kautz, R. A.; Iverson, B. L.; Tim, R. C.; Karger, B. L. Anal. Chem. 2000, 72, 5779-5786. (16) Harvey, M. D.; Bandilla, D.; Banks, P. R. Electrophoresis 1998, 19, 21692174. 10.1021/ac060021x CCC: $33.50
© 2006 American Chemical Society Published on Web 05/06/2006
antigens typically have multiple reactive sites to common fluorescent dyes or electrochemical tags. As a result, this exhibits microheterogeneity with respect to labeling efficiency, which presents itself as artificially broad or even multiple bands associated with either labeled antigen or antibody.17 In addition, unless the specific location for labeling is carefully chosen and achieved, active binding sites (whether on labeled antibody or labeled antigen) can be adversely affected. Efficient, homogeneous labeling is possible; however, this effort is time-consuming. Beyond potential labeling difficulties, the noncompetitive assay can be especially difficult due to only slight differences in the mobility of the antibody versus an antigen-antibody complex.18 An increasingly popular alternative to antibody-based APCE is the application of aptamers instead of antibodies.19-28 Aptamers are single-stranded DNA/RNA molecules that have been selected from synthetic nucleic acid libraries.29-31 Aptamers have several advantages over antibodies including the following: (1) a straightforward and reproducible synthesis without the use of animals, (2) labeling simplicity that ensures the absence of heterogeneity, and (3) stability over a wide range of pH and temperatures. Furthermore, antibodies have a narrow stability in terms of pH, ionic strength, and temperature. Conversely, aptamers exhibit reversible denaturation in harsh conditions and their only functional limitation is degradation due to the presence of nucleases. Finally, aptamers have been developed to have binding constants for target species that rival and/or exceed the specificity of antibodies. Several groups have demonstrated that aptamers can be selected and used to detect specific target proteins including IgE,26,32 thrombin,32 and HIV-1 reverse transcriptase25 for APCEbased separations. Kennedy and co-workers have published a series of papers on the use of aptamers in CE for detecting IgE and thrombin.26,32 In the case of IgE detection, fluorescein isothiocyanate (FITC) was linked to the 5′ end of a 37 base pair sequence via an ethylene glycol linker.18 In the absence of IgE a single band is observed and was attributed to the fluorescently tagged aptamer. In the presence of IgE, a distinct, well-resolved band attributed to IgE bound to aptamer is observed. In subsequent publications, the authors evaluated buffer effects, assay compatibility with electroosmotic flow suppression, and electric field strength.32 In general, the authors found that limiting diffusion (17) Schultz, N. M.; Kennedy, R. T. Anal. Chem. 1993, 65, 3161-3165. (18) German, I.; Buchanan, D. D.; Kennedy, R. T. Anal. Chem. 1998, 70, 45404545. (19) Drabovich, A.; Krylov, S. N. J. Chromatogr. A 2004, 1051, 171-175. (20) Berezovski, M.; Krylov, S. N. J. Am. Chem. Soc. 2003, 125, 13451-13454. (21) Berezovski, M.; Nutiu, R.; Li, Y.; Krylov, S. N. Anal. Chem. 2003, 75, 13821386. (22) Rehder, M. A.; McGown, L. B. Electrophoresis 2001, 22, 3759-3764. (23) Kotia, R. B.; Li, L.; McGown, L. B. Anal. Chem. 2000, 72, 827-831. (24) Baldrich, E.; Restrepo, A.; O’Sullivan, C. K. Anal. Chem. 2004, 76, 70537063. (25) Pavski, V.; Le, X. C. Anal. Chem. 2001, 73, 6070-6076. (26) Stadtherr, K.; Wolf, H.; Lindner, P. Anal. Chem. 2005, 77, 3437-3443. (27) Huang, C.-C.; Cao, Z.; Chang, H.-T.; Tan, W. Anal. Chem. 2004, 76, 69736981. (28) Wang, H.; Lu, M.; Le, X. C. Anal. Chem. 2005, 77, 4985-4990. (29) Bock, L. C.; Griffin, L. C.; Latham, J. A.; Vermaas, E. H.; Toole, J. J. Nature 1992, 355, 564-566. (30) Famulok, M. Curr. Opin. Struct. Biol. 1999, 9, 324-329. (31) Jayasena, S. D. Clin. Chem. 1999, 45, 1628-1650. (32) Buchanan, D. D.; Jameson, E. E.; Perlette, J.; Malik, A.; Kennedy, R. T. Electrophoresis 2003, 24, 1375-1382.
via short separation times and minimized capillary lengths are necessary for sensitive complex detection. This presumably limits the dissociation of aptamer from the protein target. An RNA aptamer specific to the A chain of ricin was developed by Ellington and co-workers.33,34 A 31-nucleotide long RNA sequence was developed using in vitro selection. The authors demonstrated that it was possible to use this aptamer in sensor arrays for ricin. The aptamer was immobilized onto streptavidin agarose beads via a biotinylated linker. The beads were loaded into a silicon microchip and utilized in both a direct capture configuration and sandwich assay. The authors noted that the limit of detection (ricin) was 320 ng/mL (∼11 nM) for the sandwich assay, results comparable to those observed in traditional antibodytype assays.34 In this paper, we demonstrate for the first time that capillary electrophoresis can be used to detect ricin by monitoring its interaction with a fluorescently tagged aptamer under nonequilibrium conditions. The assay will be demonstrated to provide sensitive detection of ricin in a contaminated matrix containing bovine serum albumin, casein, and most importantly, RNA nucleases. The quantitative response reveals detection limits down to 500 pM. EXPERIMENTAL SECTION Reagents and Chemicals. Bovine serum albumin (BSA), casein, hydrochloric acid, glycerol, RNAse A, sodium tetraborate, and sodium hydroxide were purchased from Sigma-Aldrich (St. Louis, MO). All materials used for capillary coatings were purchased form Sigma-Aldrich unless otherwise noted. Water was purified to a resistivity of 18.2 MΩ cm-1 using Millipore cartridges (Marlborough, MA). The RNA aptamer (5′-GGC GAA UUC AGG GGA CGU AGC AAU GAC UGC C-3′)33,34 specific for the ricin-A chain was synthesized by Integrated DNA Technologies (Coralville, IA). The 5′ end was labeled with a fluorophore dye (FAM). Fused silica capillary was purchased from Polymicro (Phoenix, AZ) with an internal diameter of 50 µm. All capillaries had an outer diameter of 360 µm and had an exterior coating of polyimide to impart mechanical stability. The total capillary length was 30 cm with a 20 cm effective length. Sample Preparation. Ricin A chain (ricin) from Ricinus communis was purchased from Sigma. Stock dilutions of ricin were made in 10 mM sodium phosphate buffer (pH 6.0) that contained 0.15 mM sodium chloride, 10 mM galactose, 0.5 mM dithiothiothreitol, and 40% glycerol (all from Sigma-Aldrich). All ricin dilutions were performed in a Class II Biosafety hood. In the experiments for ricin detection in the presence of BSA, casein, and RNAse A, ricin solutions were incubated with the interfering species overnight prior to their analysis. The aptamer was reconstituted in 10 mM Tris and 0.1 mM EDTA buffer (pH 8.0), aliquoted, and stored at -20 °C until use. Stock solutions of aptamer were then diluted in 33 mM tetraborate buffer (pH 8.5) and stored at 4 °C until use. The aptamer samples remained stable under these conditions and were used up to 3 months. (33) Hesselberth, J. R.; Miller, D.; Robertus, J.; Ellington, A. D. J. Biol. Chem. 2000, 275, 4937-4942. (34) Kirby, R.; Cho, E. J.; Gehrke, B.; Bayer, T.; Park, Y. S.; Neikirk, D. P.; McDevitt, J. T.; Ellington, A. D. Anal. Chem. 2004, 76, 4066-4075.
Analytical Chemistry, Vol. 78, No. 11, June 1, 2006
3759
Buffer Preparation. Tetraborate buffers were prepared at 33, 66, and 125 mM concentrations (pH 8.5). The pH of the buffers was adjusted to 8.5 using concentrated HCl. Sample matrixes were spiked with 300 pM sulforhodamine G. The bands associated with this internal standard were used to monitor injection variability and coating stability. All solutions were filtered with 0.2 µm filters prior to use. Capillary Coatings. Capillaries were coated with epoxy poly(dimethylacrylamide) (EPDMA) or poly(dimethylacrylamide) (PDMA). The EPDMA was prepared according to Chiari.35 The capillary was first rinsed with 1 M NaOH (10 min, 30 psi) and water (10 min, 30 psi). The capillary was then filled with a 0.2% (w/v) EPDMA solution (in water, 10 min, 30 psi) and stored at room temperature overnight. The capillary was rinsed with water (20 min, 30 psi) prior to use. The EPDMA capillary coating remained stable for ∼2.5 h of use. Before disposal, a 10% bleach solution was flushed through the capillary to inactivate any residual ricin. The PDMA capillary coating was prepared via a modified Hjerten method.36 Briefly, the capillary was rinsed for 15 min (20 psi) thoroughly with both 1 M NaOH and water. Next, a 1% (v/v) solution of 3-(trimethoxysilyl)propyl methacrylate in 0.2% (v/v) glacial acetic acid was flushed through the capillary for 15 min (20 psi). After the capillary was thoroughly rinsed with water for 20 min (20 psi), it was filled with a solution containing 4% (v/v) dimethylacrylamide in Tris-Borate-EDTA (TBE) buffer (44.5 mM Tris and borate, 1 mM EDTA, pH 8.3), 0.4% N,N,N′,N′-tetramethylethylenediamine, and 0.06% ammonium persulfate. This solution was adjusted to pH 7.0 using 1 M HCl. The capillary was then stored at 4 °C for a minimum of 24 h. Prior to use, the capillary was rinsed with water for 30 min at 20 psi. The PDMA capillary coating exhibited excellent stability for ∼5 days of continuous use. Before disposal, a 10% bleach solution was flushed through the capillary to inactivate any residual ricin. LIF-CE Equipment. All separations were performed on a Beckman Coulter PACE MDQ capillary electrophoresis (CE) instrument equipped with a laser-induced fluorescence (LIF) detector (Fullerton, CA). The fluorescent samples were excited with an Ar+ laser (488 nM) and emission was monitored at 520 nm. Capillary temperature was maintained at 25 °C and the instrument was utilized at all times per manufacturer recommendations. Premixed aptamer/ricin samples were allowed to incubate for a minimum of 30 min and were then electrokinetically injected into the capillary by applying an injection voltage of 8 kV (reverse polarity) for 15 s. In the case of the on-capillary mixing experiments, the injection scheme was as follows: 15 s ricin (8 kV, reverse polarity), 1 s sample matrix plug (no aptamer or ricin) (8 kV, reverse polarity), and 15 s 1 nM aptamer (8 kV, reverse polarity). Separations were performed by applying a voltage of 8 kV (reverse polarity). Following each separation, the capillary was rinsed with water for 1 min and separation buffer for 3 min (at 30 psi). All capillary electrophoresis data were analyzed using Grams/ AI (Version 7.02, Thermo-Galactic).
RESULTS AND DISCUSSION Optimization of Aptamer Ricin Complex Formation. For the successful complexation of ricin to the previously selected aptamer, conditions that facilitate their free-solution interaction at 25 °C must be determined. Additionally, the free aptamer and aptamer/ricin species must have different mobilities for a successful separation using capillary electrophoresis. Based on experimental observations and previously published work, four experimental parameters were targeted for optimization of the ricin/aptamer complex formation. The implementation of capillary coatings, Mg2+ concentration (sample matrix), glycerol concentration (sample matrix), and ionic strength of the separation buffer proved to be crucial parameters for the success of ricin detection by CE. In initial experiments, aptamer and ricin were pressure-injected into a bare capillary and separated using normal polarity. The migration time of the free aptamer was ∼45 min. When 245 nM ricin was premixed with aptamer, a complex band (height ∼0.4 RFU) was detected at ∼35 min. In an effort to increase the efficiency and performance of the assay, separations were performed with absorbed and covalent coatings under reverse polarity. The coatings have been demonstrated to reduce nonspecific interactions between proteins and capillary walls as well as decrease electroosmotic flow. EPDMA, an adsorbed coating, and PDMA, a covalent coating, performed equally well in terms of separation efficiency (less than 10 min per separation) and suppression of electroosmotic flow. The EPDMA coating resulted in consistent separation conditions for ∼10 assays. After this, the coating began to degrade. While the EPDMA coating has been shown to be effective for up to 50 runs of a DNA ladder,35 we hypothesize that the buffer conditions used in our work caused coating degradation. The PDMA coating was found to give consistent separation conditions for ∼250 runs before degradation was observed. For this reason, the PDMA coating was used for the majority of the experiments. The presence of Mg2+ has been shown to be necessary for the interaction between aptamers and proteins.37 In the absence of the Mg2+ salt, no complex species is formed, while the addition of 0.5 mM MgCl2 permits interaction between the species. To optimize the complex band area, the MgCl2 concentration was varied from 0.5 to 60 mM (data not shown). While the complex band height increased steadily with increasing salt concentration, complex band area maximized and produced the most symmetric band shape at 5 mM MgCl2 concentrations. It was observed that aptamer injection into the capillary was inefficient in the absence of ricin. Glycerol, when added to the sample matrix, facilitates the electrokinetic injection of aptamer into the capillary. It is hypothesized that glycerol stabilizes the protein and aptamer. Additionally, the presence of glycerol increases the viscosity of the sample matrix, thereby increasing its injection efficiency presumably due to further electroosmotic flow suppression. The aptamer injection was maximized by adding 5% glycerol to the sample matrix. At concentrations higher than 5%, the interaction between the aptamer and ricin decreased as determined from the reduction in complex band area. For this
(35) Chiari, M.; Cretich, M.; Horvath, J. Electrophoresis 2000, 21, 1521-1526. (36) Hjerten, S. J. Chromatogr. 1985, 347, 191-198.
(37) Andre, C.; Xicluna, A.; Guillaume, Y.-C. Electrophoresis 2005, 26, 32473255.
3760
Analytical Chemistry, Vol. 78, No. 11, June 1, 2006
Figure 1. (A) Electropherograms that demonstrate the optimization of the aptamer/ricin assay by varying the ionic strength of the separation buffer: (1) 33 mM tetraborate, pH 8.5; (2) 66 mM tetraborate, pH 8.5; and (3) 125 mM tetraborate, pH 8.5. The ricin and aptamer mixture was allowed to incubate for 30-90 min prior to injection. Sample matrix: 5 nM aptamer, 245 nM ricin, 33 mM tetraborate (pH 8.5), 5 mM MgCl2, 0.4% glycerol. Injection conditions: 15 s, 8 kV, reverse polarity. Separation conditions: 8 kV, reverse polarity, EPDMA coating. (B) Electropherograms for the optimized separation of aptamer (1 nM) and different concentrations of ricin: (1) 0, (2) 100, and (3) 500 nM ricin. The left band corresponds to free aptamer and the right band corresponds to the aptamer ricin complex. Sample matrix: 33 mM tetraborate (pH 8.5), 5 mM MgCl2, 5% glycerol. Injection conditions: 15 s, 8 kV, reverse polarity. Separation buffer: 125 mM tetraborate, pH 8.5. Separation conditions: 8 kV, reverse polarity, EPDMA or PDMA coating. The complex band is starred in each electropherogram.
reason, 5% glycerol was added to the sample matrix for all optimized experimental conditions. Finally, the ionic strength of the separation buffer relative to that of the sample matrix is critical for optimal complex band area and resolution. The separation of free aptamer from the aptamer/ ricin complex in separation buffers with increasing ionic strengths is displayed in Figure 1. The sample matrix composition is fixed at 33 mM tetraborate (pH 8.5), 5 mM MgCl2, 0.4% glycerol, 5 nM aptamer, and 245 nM ricin. When the sample matrix and separation buffers are matched in borate concentration and pH, the aptamer and complex bands are resolved (Figure 1A-1); however, the complex band is rectangular in shape. When the ionic strength of the run buffer is doubled by increasing the tetraborate concentration to 66 mM, the complex band shape
becomes more Gaussian (Figure 1A-2). Increasing the ionic strength of the run buffer relative to the sample matrix allows for field-amplified stacking, thereby changing the band shape. When the tetraborate concentration is quadrupled relative to the original separation buffer conditions (125 mM), the complex band area doubles while still remaining Gaussian (Figure 1A-3). Further increases in separation buffer ionic strength resulted in Joule heating which was detrimental to this ricin assay. The increase in band area can be attributed to a combination of increased suppression of electroosmotic flow that permits more complex to be injected, and the formation of a more stable complex in the higher ionic strength buffer. Evaluation of the Dynamic Range of the Sensor’s Response. Successful separations for a 15 s injection of aptamer and ricin are shown in Figure 1B. The aptamer and aptamer/ ricin complex bands are well-resolved (resolution ∼7.3). These electropherograms have been obtained from the CE-LIF analysis of a mixture of 1 nM aptamer and 0 nM (Figure 1B-1), 100 nM (Figure 1B-2), and 500 nM (Figure 1B-3) ricin. The samples were allowed to incubate for at least 30 min prior to analysis to ensure full complexation. The free aptamer sample reveals one primary band and two small shoulders/secondary bands. The simplicity of the electropherogram indicates a relatively homogeneous aptamer sample. The sidebands can be attributed to either conformation differences in the aptamer structure or slight heterogeneity of aptamer length. As the concentration of ricin increases, a complex band grows in while the aptamer band decreases in both height and area. The full dynamic range of complex band formation is summarized in Figure 2. The interaction between the aptamer and ricin species was monitored by evaluating the complex band area as a function of varying ricin concentration versus a fixed aptamer concentration. Because the separation is occurring under fieldamplified stacking conditions, evaluation of the complex band area produced systematic results whereas the complex band height data did not. For this reason, all data presented will be compared using complex band area. This will be discussed in further detail in the next section. As shown in Figure 2A, a dissociation constant, Kd, of 134 nM is revealed for the complex. Applying a 95% confidence level, the value ranges between 131.5 and 136.1 nM. This value is ∼20 times larger than the magnitude of the originally reported Kd value (7.3 nM)33 for the ricin/aptamer complex. This discrepancy can be attributed to a variety of sources. First, aptamer/protein interactions, in general, are greatly influenced by the sample matrix/buffer conditions. The buffers used in the work herein and those in the original ricin/aptamer literature differ in composition, pH, and ionic strength. Second, the fluorescent label used in these experiments might influence the complex formation because of steric interactions. Third, given the length of the capillary and the time it takes the complex to reach the detection window, the aptamer/ricin complex could be dissociating prior to its detection. Finally, the separation in the included work is performed in an electric field. This field might influence/ encourage the dissociation between the aptamer and ricin complex. The original Kd value was attained after the ricin and aptamer had formed an equilibrium complex. Because of the nature of this CE-based assay, the binding interaction is not at equilibrium and may contribute to the difference in values. Analytical Chemistry, Vol. 78, No. 11, June 1, 2006
3761
Figure 2. Effects of varying ricin concentration on the area of the aptamer ricin complex band. (A) Complex band area variation as a function of ricin concentration. A dissociation constant of 133.8 ( 8.2 nM was measured. Each data point represents an average response from 3 to 7 assays. (B) Complex band area variation at low concentrations (500 pM to 75 nM) of ricin. The data follow a linear response, y ) 0.0547*x + 0.0561, R2 ) 0.9888. (C) Electropherogram demonstrating the detection of 500 pM ricin. Same separation and injection conditions as in Figure 1B.
At low ricin concentrations, the complex band area increases linearly as the ricin concentration is increased from 0 to 80 nM (Figure 2B). As demonstrated in Figure 2C, 500 pM (∼14 ng/ mL) or 7.1 amol of ricin can easily be detected using the free solution assay with the fluorescently labeled aptamer and the capillary electrophoresis separation. It should be noted that the assay limit of detection16 (LOD) rivals the sensitivity of traditional antibody-based immunoassays for ricin. While the instrumental LOD of the LIF system would provide a lower detection limit, given the dissociation constant of the complex, the assay LOD more accurately describes the potential detection limitations of the system. On-Capillary Mixing for the Detection of Ricin. In all previously discussed experiments, the ricin and aptamer samples were mixed and allowed to incubate for at least 30 min prior to examination. If this incubation time could be eliminated, the potential usefulness of this assay could be improved by decreasing the time scale required for analysis. Furthermore, the aptamer/ protein interaction with potential degradation sources (i.e., nucleases) in a sample would be limited. For this reason, methods to sequentially inject the aptamer and ricin samples were evaluated. The individual mobilities of aptamer and ricin were determined using LIF and UV detection, respectively. Because ricin 3762
Analytical Chemistry, Vol. 78, No. 11, June 1, 2006
has a slower migration time in comparison to the aptamer, ricin must be injected before the aptamer to ensure species interaction. If a sample matrix plug (without ricin) is injected prior to the aptamer sample, one primary band with two small shoulders is observed in the electropherogram (Figure 3A). In comparison to an aptamer injection without the preliminary sample matrix plug, the aptamer band is slightly broader and nonsymmetric. Additionally, when ricin is included, the complex and free aptamer bands migrate closer together in comparison to the premixed samples. When a 10 nM ricin plug is injected before the aptamer sample, a complex band is observed ∼40 s after the free aptamer band (Figure 3B). For equivalent comparisons to the premixed aptamer/ ricin samples, the individual components were both injected for 15 s. This injection timing was demonstrated to be optimal for the on-capillary mixing experiments. The complex band areas for all concentrations less than 50 nM ricin were approximately doubled in comparison to the premixed samples. Band area increased linearly with a slope of 0.128 and R2 ) 0.99942 (data not shown). This indicates that the aptamer and ricin species have less time to dissociate after their on-capillary mixing in comparison to the premixed samples. Detection of Ricin in Protein-Contaminated Sample Matricies. A potential problem with using any aptamer-based assay
Table 1. Limits of Detection of Ricin in the Presence of Interfering Species Using On-Capillary Mixing; Same Sample Matrix and Separation Conditions as in Figure 3
a
interfering species
ricin concentration
complex band area
complex band height
signal to noise ratio
none none 100 µg/mL BSA 150 µg/mL casein 100 ng/mL RNAse A 100 ng/mL RNAse A
250 pM 500 pM 500 pM 500 pM 500 pM 1 nM
undetectable 0.1116 0.0492 0.0427 undetectable 0.0616
undetectable 0.5514 RFU 0.1601 RFU 0.4436 RFU undetectable 0.3922 RFU
NAa 4.81b 5.13a 6.50a NAa 14.21b
Gain ) 10. b Gain ) 100.
Figure 3. Demonstration of on-capillary mixing for the detection of 10 nM ricin. Sequential injections: 15 s sample matrix (A) no ricin and (B) 10 nM ricin at 8 kV followed by 15 s of 1 nM aptamer at 8 kV. Same sample matrix and separation conditions as in Figure 1B. The complex band is starred in each electropherogram.
is the decreased sensitivity of the aptamer to specific protein binding in the presence of interfering proteins. The on-capillary mixing experimental setup presents an ideal situation in which protein interferences can be assessed. Interferants such as bovine serum albumin (BSA) and casein can easily be mixed with the ricin samples to mimic natural samples. To observe the disruption between the aptamer and ricin, varying concentrations of BSA were premixed with 50 nM ricin (Figure 4A). This contaminated ricin sample is injected into the capillary prior to the aptamer plug (as in Figure 3). The detection of a “clean” ricin sample (Figure 4A-1) has been included for a direct comparison. BSA-contaminated ricin samples do, in fact, form a complex with the aptamer (Figure 4A). The major effect observed in the electropherograms for clean and contaminated ricin assays is a decrease in band area with increasing BSA concentration. Specifically, in the presence of 50 µg/mL BSA, the complex band area decreases by 44% in comparison to the clean ricin samples. A 53% decrease in band area is observed when the BSA concentration is increased to 100 µg/mL. Clearly, increasing BSA concentrations does decrease the intermolecular interactions between the ricin and aptamer species. It should be noted that while a decrease in band area is observed, ricin is still easily detected in the presence of BSA. In fact, in the presence of 100 µg/mL BSA, 500 pM ricin can still be detected without a significant loss in signal (Table 1). While the mass quantity of the aptamer/ ricin complex decreases, the signal-to-noise ratio associated with the complex increases (Table 1).
Figure 4. Detection of ricin in protein mixtures using the sequential injection and on-capillary mixing procedure. (A) Electropherograms demonstrating detection of (1) 50 nM ricin, (2) 50 nM ricin, and 50 µg/mL BSA, (3) 50 nM ricin and 100 µg/mL BSA. (B) Electropherograms demonstrating detection of (1) 50 nM ricin, (2) 50 nM ricin, and 50 µg/mL casein, (3) 50 nM ricin and 100 µg/mL casein, and (4) 50 nM ricin and 150 µg/mL casein. Same sample matrix and separation conditions as in Figure 3. The complex band is starred in each electropherogram.
Similar experiments were performed by premixing 50 nM ricin solutions with varying concentrations of casein, a protein found in milk (Figure 4B). Just as with BSA, the ricin/aptamer complex forms in the presence of casein. When ricin samples are premixed with 50 µg/mL casein, no change in complex band area is observed; however, the bandwidth is doubled. Complexes formed in the presence of high concentrations of casein exhibit decreasing band areas with increasing casein concentrations. As the casein concentrations increases from 100 µg/mL to 150 µg/ mL, the Analytical Chemistry, Vol. 78, No. 11, June 1, 2006
3763
and to 1 µg/mL, the complex band decreases in area by 30%, 38%, and 73%, respectively. While ricin can still easily be detected at 1 nM concentration in the presence of 100 ng/mL RNAse (Table 1), the loss of viable aptamer concentration prevents the detection of ricin at sub-nanomolar concentrations.
Figure 5. On-capillary mixing and detection of 1 nM aptamer and 50 nM ricin that has been mixed with (A) 0 ng/mL, (B) 10 ng/mL, (C) 100 ng/mL, and (D) 1 µg/mL RNAse A. Same sample matrix and separation conditions as in Figure 3.
complex band area decreases by 7% and 40%, respectively. Just as in the case of complex formation in the presence of BSA, the 500 pM ricin detection limit can still be observed in the presence of 100 µg/mL casein (Table 1). Detection of Ricin in Nuclease-Contaminated Sample Matrixes. The success of the ricin/aptamer interaction in the presence of interfering proteins encouraged the analysis of ricin samples in the presence of RNAse. RNAse, an RNA nuclease, is found in nature and has the potential to be present in any real ricin sample analyzed. As a result, RNA nucleases present the largest challenge in developing an RNA aptamer assay for protein detection because they digest RNA. As the aptamer is digested, all specificity and interactions with ricin are eliminated. To address this concern, varying concentrations of RNAse A, a specific RNA nuclease, were premixed with 50 nM ricin solutions. As observed in Figure 5, RNAse A does degrade the aptamer sample. The primary band associated with undigested aptamer (Figure 5, Band 2) and a minor band (Figure 5, Band 4) steadily decreases in area with increasing RNAse concentration. Concurrent to the degradation of the original aptamer bands, however, two new bands form. A band that migrates faster than the original aptamer band (Figure 5, Band 1) steadily grows in while the second new band (Figure 5, Band 3) increases in area and then decreases. While the aptamer is severely damaged in the presence of RNAse, a complex between the aptamer and ricin still forms. This is the first demonstration of successful protein detection using RNA aptamer in the presence of nucleases. In comparison to the uncontaminated ricin sample (Figure 5A), the complex band remains Gaussian but does decrease in height and area. As the RNAse concentration increases from 10 ng/mL, to 100 ng/mL,
3764
Analytical Chemistry, Vol. 78, No. 11, June 1, 2006
CONCLUSIONS The integration of LIF-based capillary electrophoresis with fluorescently labeled aptamers provides a novel approach for the detection of ricin. This free solution assay offers an alternative technique for protein detection in comparison to standard immunoassay and ELISA methods. Aptamers provide a more stable alternative to antibodies in terms of room-temperature self-lives and more consistent syntheses. The lowest concentration of ricin detected is 500 pM (∼14 ng/mL) using this separation technique, limits that rival that of standard antibody-based detection schemes. This reproducible assay can be performed in less than 10 min, consumes sub-nanoliter quantities of material, and generates less than a milliliter of waste over 24 h of continuous use (∼70 assays). The highly specific interaction between the aptamer and ricin is influenced by the presence of interfering proteins such as BSA and casein; however, the 500 pM limit of detection of ricin is retained. Aptamer degradation by nucleases is observed when the ricin sample is premixed with RNAse A; however, even in the presence of 100 ng/mL RNAse, ricin can easily be detected at 1 nM concentrations. Future miniaturization of this assay onto a microchip will further this approach by decreasing the time scale required for analysis and offers options that would allow for the analysis of multiple analytes. The straightforward nature of the integration of the capillary electrophoresis-based separation with highly specific aptamers offers new potential avenues for the rapid analysis and detection of proteins in environmental samples. ACKNOWLEDGMENT This publication was made possible by Grant Number AI056047 from the National Institute of Allergy and Infectious Diseases (NIAID). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health (NIH). This research was performed while A.J.H. held a National Research Council Research Associateship Award at the Naval Research Laboratory. B.C.G. wishes to acknowledge the American Society for Engineering Education Postdoctoral Fellowship at the Naval Research Laboratory. The authors wish to thank Jerome Ferrance of the University of Virginia, Chemistry Department, for providing us with EPDMA.
Received for review January 4, 2006. Accepted April 7, 2006. AC060021X
