Silver and Gold Glyconanoparticles for Colorimetric Bioassays

Langmuir 2006, 22, 6707-6711

6707

Silver and Gold Glyconanoparticles for Colorimetric Bioassays Claire L. Schofield, Alan H. Haines, Robert A. Field, and David A. Russell* School of Chemical Sciences and Pharmacy, UniVersity of East Anglia, Norwich, Norfolk, NR4 7TJ United Kingdom ReceiVed January 30, 2006. In Final Form: May 9, 2006 The color changes associated with the aggregation of metal nanoparticles has led to the development of colorimetricbased assays for a variety of target species. We have examined both silver- and gold-based nanoparticles in order to establish whether either metal exhibits optimal characteristics for bioassay development. These silver and gold nanoparticles have been stabilized with a self-assembled monolayer of a mannose derivative (2-mercaptoethyl R-Dmannopyranoside) with the aim of inducing aggregation by exploiting the well-known interaction between mannose and the lectin Concanavalin A (Con A). Both metal glyconanoparticles were determined to be ca. 16 nm in diameter (using TEM measurements). Aggregation was observed on addition of Con A to both silver and gold nanoparticles resulting in a shift in the surface plasmon absorption band and a consequent color change of the solution, which was monitored using UV-visible spectrophotometry. Mannose-stabilized silver nanoparticles at a concentration of 3 nM provide an assay for Con A with the largest linear range (between 0.08 and 0.26 µM). Additionally, the kinetic rate of aggregation of the silver-nanoparticle-based bioassay was significantly greater than that of the gold-nanoparticle system. However, in terms of sensitivity, the mannose-stabilized gold-nanoparticle-based assay was optimum with a limit of detection of 0.04 µM Con A, as compared with a value of 0.1 µM obtained for the mannose-stabilized silver nanoparticles. Additionally, a lactose derivative (11-mercapto-3,6,9-trioxaundecyl β-D-lactoside) was used to stabilize gold nanoparticles to induce aggregation upon addition of the galactose specific lectin Ricinus communis agglutinin (RCA120). To examine the specificity of the bioassay, lactose-stabilized gold nanoparticles were mixed with a solution of mannose-stabilized silver nanoparticles to give an aggregation assay capable of detecting two different lectins. When either Con A or RCA120 was added to the mixed glyconanoparticles, selective recognition of the respective natural ligand was shown by aggregation of a single metal nanoparticle. Centrifugation and removal of the aggregated species enabled further bioassay measurements using the second glyconanoparticle system.

Introduction Numerous metals can be synthesized into nanoscale particles of varying shapes each with characteristic optical properties. When the dimensions of the metal are reduced to the nanoscale, the optical properties are dominated by a collective oscillation of conduction electrons in resonance with incident electromagnetic radiation.1 This phenomenon is termed surface plasmon resonance.2 For noble metals, this resonance frequency, the surface plasmon absorption band, lies in the visible part of the electromagnetic spectrum giving rise to intense colors from solutions of colloidal metal particles. 1 The surface plasmon absorption band, and consequently the color of a metal nanoparticle solution, is dependent on a number of parameters, viz., the size and shape of the particle, the type of metal, the dielectric properties of the medium, and the distance between particles.3,4 Gold nanoparticles, with an interparticle distance greater than the average particle diameter, appear red as a consequence of the surface plasmon absorption band centered at 520 nm. As the interparticle distance decreases to less than the diameter of the particles, coupling interactions result in a broadening and a shift to longer wavelengths of the surface plasmon absorption band that results in a solution of aggregated gold nanoparticles appearing blue. Monodisperse silver nanoparticles characteristically have a surface plasmon absorption band at ca. 400 nm. In solution dispersed silver nanoparticles are a yellow color, upon * To whom correspondence should be addressed. E-mail: d.russell@ uea.ac.uk. Tel/Fax: +44-1603-593012. (1) Liz-Marza´n, L. M. Langmuir 2006, 22, 32-41. (2) Schultz, D. A. Curr. Opin. Biotechnol. 2003, 14, 13-22. (3) Sun, Y.; Xia, Y. Analyst 2003, 128, 686-691. (4) Aslan, K.; Lakowicz, J. R.; Geddes, C. D. Anal. Biochem. 2004, 330, 145-155.

aggregation the nanoparticles appear orange. The color changes associated with nanoparticle aggregation have been exploited in the development of colorimetric assays. The first report of such colorimetric nanoparticle-based assays was for the detection of DNA using gold nanoparticles by Mirkin et al.5,6 Subsequently, other gold nanoparticle studies have focused on the detection of lectins,7,8 metal ions,9-13 antibodies,14 and other analytes.15-17 In each instance, the colorimetric assay was achieved by selfassembling a monolayer of a molecular recognition molecule on the nanoparticle surface. Controlled aggregation of the nanoparticles is achieved when specific recognition and binding of the target species occurs. Although not frequently reported, silver nanoparticles can also be used in such assays, as demonstrated by Sastry et al.18 who exploited the high affinity biotin-avidin (5) Mirkin, C. A.; Letsinger, R. L.; Mucic, R. C.; Storhoff, J. J. Nature 1996, 382, 607-609. (6) Storhoff, J. J.; Lazarides, A. A.; Mucic, R. C.; Mirkin, C. A.; Letsinger, R. L.; Schatz, G. C. J. Am. Chem. Soc. 2000, 122, 4640-4650. (7) Otsuka, H.; Akiyama, Y.; Nagasaki, Y.; Kataoka, K. J. Am. Chem. Soc. 2001, 123, 8226-8230. (8) Hone, D. C.; Haines, A. H.; Russell, D. A. Langmuir 2003, 19, 71417144. (9) Lin, S.; Liu, S.; Lin, C.; Chen, C. Anal. Chem. 2002, 74, 330-335. (10) Kim, Y.; Johnson, R. C.; Hupp, J. T. Nano Lett. 2001, 1, 165-167. (11) Obare, S. O.; Hollowell, R. E.; Murphy, C. J. Langmuir 2002, 18, 1040710410. (12) Reynolds, A. J.; Haines, A. H.; Russell, D. A. Langmuir 2006, 22, 11561163. (13) Lin, S.-Y.; Chen, C-h.; Lin, M.-C.; Hsu, H.-F. Anal. Chem. 2005, 77, 4821-4828. (14) Thanh, N. T. K.; Rosenzweig, Z. Anal. Chem. 2002, 74, 1624-1628. (15) Liu, J.; Lu, Y. Angew. Chem., Int. Ed. 2006, 45, 90-94. (16) Guarise, C.; Pasquato, L.; de Fillippis, V.; Scrimin, P. Proc. Natl. Acad. Sci. 2006, 103, 3978-3982. (17) Huang, C.-C.; Huang, Y.-F.; Cao, Z.; Tan, W.; Chang, H.-T. Anal. Chem. 2005, 77, 5735-5741. (18) Sastry, M.; Lala, N.; Patil, V.; Chavan, S. P.; Chittiboyina, A. G. Langmuir 1998, 14, 4138-4142.

10.1021/la060288r CCC: $33.50 © 2006 American Chemical Society Published on Web 06/15/2006

6708 Langmuir, Vol. 22, No. 15, 2006

interaction to develop an aggregation assay based on biotinylated silver nanoparticles. We are particularly interested in developing colorimetric bioassays for the detection of bacterial toxins, and other biological molecules, using carbohydrate-stabilized nanoparticles (glyconanoparticles). In this paper, glyconanoparticles formulated from both silver and gold have been assessed to establish the optimal metal upon which colorimetric aggregation bioassays should be based. To achieve this comparison, the well-known interaction between the lectin concanavalin A (Con A) and the monosaccharide mannose19 was studied by self-assembling a thiolated mannose derivative onto both silver and gold nanoparticles. The analytical bioassay performance of each metal nanoparticle system has been established. In a further element of this work, mannosestabilized silver and lactose-stabilized gold nanoparticles have been mixed in order to demonstrate the selectivity of each glyconanoparticle for its respective cognate ligand (Con A and Ricinus communis agglutinin (RCA120) respectively). Addition of either lectin resulted in the specific aggregation of one type of nanoparticle enabling both the selective detection of a lectin but also the simple separation of the two component sample. Experimental Section Reagents. Concanavalin A isolated from jack bean (CanaValia ensifromis), 20 hydrogen tetrachloroaurate trihydrate, and silver nitrate were purchased from Sigma-Aldrich (Gillingham, U.K.). Sodium borohydride was purchased from Lancaster Synthesis Ltd. (Morecombe, U.K.). Ricinus communis agglutinin was purchased from Vector Laboratories Ltd, (Peterborough, U.K.). All other chemicals were purchased from Fisher Scientific U.K. Ltd. (Loughborough, U.K.). All aqueous solutions were prepared using analytical reagent grade water. The syntheses of the thiolated mannose and lactose derivatives (2-mercaptoethyl R-D-mannopyranoside and 11-mercapto-3,6,9-trioxaundecyl β-D-lactoside respectively) have been reported previously.8,21,12 Preparation of Mannose-Stabilized Silver Nanoparticles. Silver nanoparticles were prepared by the reduction of a silver salt with sodium borohydride. A solution of sodium borohydride (2.2 mg, 30 mL) was placed in an ice-bath. Silver nitrate (1.7 mg, 10 mL) was added dropwise with vigorous stirring, producing a vivid yellow colored solution. The mannose derivative (30 mg) was added to the solution of silver nanoparticles (40 mL) and left to stir for 48 h to ensure self-assembly of the thiolated mannose onto the silver particle surface. To remove any excess mannose, purification was carried out by dialysis for 6 h using dialysis tubes (Genotech, Tube-Odialyzer tubes) with a molecular weight cutoff of 15 kDa. The resulting silver nanoparticle solution was of 1 nM concentration. To obtain a silver nanoparticle solution of 3 nM, silver nanoparticles were centrifuged, using Vivaspin concentrator centrifuge tubes (Camlab, ALC multispeed centrifuge) for 15 min at 328g, to remove the supernatant. The particles were then resuspended in tris buffer to the required concentration. More concentrated samples of silver nanoparticles were not produced as the magnitude of the surface plasmon absorption band would be too intense for quantitative UVvisible measurements. Preparation of Mannose- and Lactose-Stabilized Gold Nanoparticles. Gold nanoparticles were prepared, via the citrate reduction of hydrogen tetrachloroaurate as previously reported,8,12 to produce particles of 16 nm diameter, at a concentration of 3 nM. To obtain a 1 nM gold nanoparticle solution, the as-produced gold nanoparticle solution was diluted using tris buffer. A 5 nM gold nanoparticle solution was obtained by centrifugation (Beckman Avanti J-25 centrifuge) of the 3 nM gold nanoparticles for 25 min at 23 710g (19) Dam, T. K.; Brewer, C. F. Chem. ReV. 2002, 102, 387-429. (20) Goldstein. I. J.; Hollerma. C. E.; Smith, E. E. Biochemistry 1965, 4, 876-883. (21) Revell, D. J.; Knight, J. R.; Blyth, D. J.; Haines, A. H.; Russell, D. A. Langmuir 1998, 14, 4517-4524.

Schofield et al.

Figure 1. (a) Mannose-stabilized silver nanoparticles (3 nM, ca. 16 nm in diameter); (b) lactose-stabilized gold nanoparticles (3 nM, ca. 16 nm in diameter); and (c) a 50:50 mixture of mannose-silver and lactose-gold nanoparticles. and resuspending the particles in the required amount of tris buffer. To 40 mL of the gold nanoparticle solution was added 20 mg of either the mannose or lactose derivatives, and the solution was left to stir for 48 h to facilitate self-assembly. Nonbound carbohydrate was removed by centrifugation of the nanoparticle solution for 25 min at 23 710g. The carbohydrate-stabilized nanoparticles were resuspended in tris buffer solution (10 mM, pH 7.6). The centrifugation/resuspension process was repeated a total of three times to ensure removal of excess carbohydrate. Colorimetric Bioassays Based on Glyconanoparticle Aggregation. A range of concentrations of Con A were prepared (0.01-0.48 µM). Each Con A solution was added to mannose-stabilized gold or silver nanoparticles with stirring, and the reaction was monitored by UV-visible spectrophotometry. For the mixed metal nanoparticle bioassay, mannose-stabilized silver nanoparticles and lactosestabilized gold nanoparticles were mixed (both at ca. 3 nM). Con A (0.48 µM) was added to induce aggregation of the silver nanoparticles. Centrifugation of the solution at 2051g for 10 min separated the silver and gold glyconanoparticles. The supernatant, containing the lactose-gold nanoparticles, was removed and subsequently reacted with RCA120 lectin (0.80 µM) to aggregate these particles. A second mixture of mannose-silver and lactose-gold nanoparticles was produced and this time was initially reacted with RCA120 (0.80 µM) to induce aggregation of the gold nanoparticles. The solution was then centrifuged (2051g for 10 min) to separate the two metal glyconanoparticles. Con A (0.48 µM) was then added to the supernatant to aggregate the mannose-stabilized silver nanoparticles. Instrumentation. UV-visible absorption and kinetic measurements were performed using a Hitachi U3000 UV-visible spectrophotometer at room temperature (ca. 20 °C). Transmission electron microscopy (TEM) was used to provide images of the nanoparticles (JEOL 2000EX TEM using an accelerating voltage of 120 kV). Samples were prepared by drop casting 5 µL of the sample onto holey copper grids.

Results and Discussion Mannose-Stabilized Silver and Mannose/Lactose-Stabilized Gold Nanoparticles. The reduction of the silver nitrate with borohydride produced a solution of nanoparticles with a vivid yellow color with an average diameter of ca. 16 nm as determined by TEM. The silver nanoparticles exhibited an intense surface plasmon absorption band at 395 nm. Following addition of the 2-mercaptoethyl R-D-mannopyranoside to the silver particles, the surface plasmon absorption band broadened and the maximum shifted to 401 nm, the solution retaining the yellow coloration (Figure 1a). Similarly, gold nanoparticles were readily produced by reduction of the gold salt with sodium citrate. The solution of gold nanoparticles was bright red in color due to the surface plasmon absorption band centered at 520 nm which is typical of particles 16 nm in diameter.8 Following modification with the

Glyconanoparticles for Colorimetric Bioassays

Langmuir, Vol. 22, No. 15, 2006 6709

Figure 3. Changing UV-visible spectra following addition of increasing concentrations of Con A to (a) mannose-stabilized silver nanoparticles and (b) mannose-stabilized gold nanoparticles. Figure 2. TEM images of mannose-stabilized silver nanoparticles (a) as prepared, and (b) following addition of 0.48 µM Con A.

mannose or the lactose (11-mercapto-3,6,9-trioxaundecyl β-Dlactoside) derivatives, a shift in wavelength of the surface plasmon absorption band to 523 nm occurred. With either carbohydrate, the stabilized gold nanoparticles retained an intense red coloration (Figure 1b). Metal-Nanoparticle-Aggregation-Based Bioassays. At pH > 7, Con A is a tetrameric lectin (carbohydrate binding protein), with each of the four subunits containing an R-D-mannose binding site.22 These sites enable binding to the mannose ligands on a nanoparticle surface leading to aggregation of the nanoparticles in solution. Figure 2 shows the TEM images of dispersed mannose-stabilized silver nanoparticles (Figure 2a) and the formation of a large aggregate of silver nanoparticles following addition of 0.48 µM Con A (Figure 2b). With aggregation of the silver nanoparticles, there is a commensurate shift in the surface plasmon absorption band. Figure 3a shows the changing absorption spectrum of the mannose-stabilized silver nanoparticles with increasing concentrations of Con A. As the concentration of Con A increases, the surface plasmon absorption band broadens with the band maximum at 401 nm decreasing in intensity. The spectra show a clear isosbestic point at 440 nm upon addition of the lectin demonstrating that the aggregation of mannosestabilized nanoparticles is directly related to the concentration of the Con A. The red-shift of the surface plasmon absorption band changes the color of the silver-nanoparticle solution from yellow to orange. We have previously reported that similar large networks of nanoparticle aggregates are formed when Con A is added to mannose-stabilized gold nanoparticles.8 Upon aggregation of the gold nanoparticles, a red-shift in wavelength and a broadening (22) Sanders, J. N.; Chenoweth, S. A.; Schwarz, F. P. J. Inorg. Biochem. 1998, 70, 71-82.

of the surface plasmon absorption band occurs which is accompanied by a solution color change from red to purple. Figure 3b shows the typical UV-visible spectra of the mannosestabilized gold nanoparticles on addition of Con A. From the data shown in Figure 3, it is clear that there is a significant change in absorbance intensity, at 480 and 620 nm for silver and gold nanoparticles, respectively, that provides a means for the quantitative measurement of the Con A lectin. To optimize the bioassay, varying concentrations of both silver and gold mannose-stabilized nanoparticles were assessed. Figure 4 shows the quantitative relationship between changing intensity of absorbance and the concentration of Con A for both silverand gold-based nanoparticles. It is clear that both metals can be used in the development of assays based on colorimetric changes. It is also apparent from Figure 4 that the metal and the concentration of the metal nanoparticles both have a profound effect on the analytical sensitivity, limit of detection, and linear dynamic range of the colorimetric bioassay of Con A. The largest change in absorbance (ca. 0.7) from dispersed to fully aggregated nanoparticles was observed for the 3 nM silver nanoparticles (Figure 4a). This concentration of mannose-stabilized silver nanoparticles provides an assay for Con A with a good sensitivity and one with the largest linear range, between 0.08 and 0.26 µM Con A. For the gold particles (Figure 4b), the greatest sensitivity was again observed for the 3 nM nanoparticles although the overall change in absorbance (ca. 0.3), and the linear range of 0.04-0.10 µM Con A was significantly less than that observed for the silver nanoparticles. By comparing the mannose-stabilized silver and gold nanoparticles at 3 nM concentrations for the colorimetric assay of Con A (Figure 4c), it is clear that while the silver nanoparticles exhibited a significantly longer linear dynamic range, the sensitivity of the gold nanoparticle based assay is optimal with a limit of detection of 0.04 µM (cf. ca. 0.1 µM for the silver nanoparticles). In a further assessment of the bioassay characteristics of the two metal-nanoparticle systems, the rate of change of the absorbance signal versus concentration8

6710 Langmuir, Vol. 22, No. 15, 2006

Schofield et al.

Figure 5. Rate of change in A480 for mannose-stabilized silver (blue square) and A620 for gold (red square) nanoparticles versus Con A concentration. Absorbance changes were measured 30 s after addition of the Con A.

Figure 4. (a) Change in A480 of mannose-silver nanoparticles (3 nM (red square); 1 nM (blue triangle)) with varying concentrations of Con A. (b) Change in A620 of mannose-gold nanoparticles (5 nM (red square); 3 nM (blue triangle); and 1 nM (green diamond)) with varying concentrations of Con A. (c) Comparison of change in A480 (for 3 nM mannose-silver nanoparticles (blue square)) and A620 (for 3 nM mannose-gold nanoparticles (red triangle)) with varying concentrations of Con A. Each point is an average of three separate measurements.

was determined (Figure 5). Here the mannose-stabilized silver nanoparticles proved to respond faster than their gold counterparts. It is known that self-assembled monolayers on silver and gold surfaces have different surface orientations23 and this may account for the significant difference in the rate of aggregation between the mannose-stabilized silver and gold nanoparticles. Mixed Metal Nanoparticle/Ligand Bioassays. To demonstrate the selectivity of the bioassay, the mannose-stabilized silver and the lactose-stabilized gold nanoparticles were mixed at equal (23) Harder, P.; Grunze, M.; Dahint, R.; Whitesides, G. M.; Laibinis, P. E. J. Phys. Chem. B 1998, 102, 426-436.

Figure 6. UV-visible absorption spectra of (a) Mannose-stabilized silver nanoparticles (blue line), lactose-stabilized gold nanoparticles (red line), mixed silver and gold nanoparticles (green line), and mixed silver and gold nanoparticles following addition of Con A (black line). (b) Orange pellet consisting of Con A-mannose silver nanoparticle aggregates (blue line), solution supernatant consisting of lactose-stabilized gold nanoparticles (red line), and aggregated lactose-stabilized gold nanoparticles following addition of RCA120 to the supernatant (black line).

concentrations. As anticipated, the mixed solution of nanoparticles was orange in color (Figure 1c). The RCA120 lectin is known to bind to galactose terminated molecules,24,25 implying that it should bind to the lactose (a disaccharide which consists of a glucose unit and a terminal galactose unit) on the gold nanoparticles. RCA120 has two galactose binding sites per molecule;26,27 therefore, on addition of this lectin to lactose-stabilized gold nanoparticles, aggregation should occur. When Con A was added (24) Sweeney, E. C.; Tonevitsky, A. G.; Temiakov, D. E.; Agapov, II.; Saward, S.; Palmer, R. A. Proteins-Struct. Funct. Genet. 1997, 28, 586-589. (25) Endo, Y.; Mitsui, K.; Motizuki, M.; Tsurugi, K. J. Biol. Chem. 1987, 262, 5908-5912. (26) Sharma, S.; Bharadwaj, S.; Surolia, A.; Podder, S. K. Biochem. J. 1998, 333, 539-542. (27) Hegde, R.; Podder, S. K. Eur. J. Biochem. 1998, 254, 596-601.

Glyconanoparticles for Colorimetric Bioassays

to the mixed nanoparticle solution, the mannose-stabilized silver nanoparticles aggregated. Figure 6a shows the UV-visible spectra for separate silver- and gold-stabilized nanoparticles (with surface plasmon absorption maxima at 401 and 523 nm respectively), the mixed nanoparticle solution, and the mixed nanoparticle solution following the aggregation of the mannose-stabilized silver nanoparticles with the Con A lectin. As expected, the spectrum of the mixed nanoparticle solution is simply a summation of the two individual spectra for the silver and gold nanoparticles. Upon addition of the Con A, the combined nanoparticle spectrum broadens and red-shifts, predominantly that part of the spectrum that originates from the mannose-stabilized silver nanoparticles, indicating that selective recognition of the Con A lectin has been achieved. To separate the aggregated species following aggregation of the mannose-stabilized silver nanoparticles with Con A, the solution was centrifuged. An orange pellet of the Con A-silver nanoparticle complex was obtained leaving the lactosestabilized gold nanoparticles in the supernatant (Figure 6b). RCA120 was then added to the supernatant inducing aggregation of the lactose-stabilized gold nanoparticles that remained in the supernatant (Figure 6b). The separation and subsequent interaction of the lactose-gold nanoparticles indicates that only limited nonspecific interaction occurs with the Con A, thereby enabling subsequent specific interaction with the RCA120 lectin. Similarly, when the mixed mannose-silver and lactose-gold nanoparticles were reacted first with RCA120, aggregation of the lactose-stabilized gold nanoparticles was induced. The aggregate could be readily separated from the mannose-stabilized silver nanoparticles remaining in the supernatant. When Con A was added to the supernatant, aggregation of the mannose-stabilized silver nanoparticles occurred. These combined results show that

Langmuir, Vol. 22, No. 15, 2006 6711

mixed metal/ligand nanoparticles can be used to specifically bind to individual cognate substrates and that the centrifugation process demonstrates the potential for separation of the two nanoparticle systems thereby enabling further bioassay measurements.

Conclusions Our results have shown that thiolated derivatives of carbohydrates can be readily assembled onto both silver and gold nanoparticles and that both metal glyconanoparticles can be used to develop aggregation based colorimetric bioassays. In each instance, the concentration of the metal-nanoparticle system used is critical in terms of analytical performance of the developed assay. Although it has been established that mannose-stabilized silver nanoparticles exhibit a longer linear dynamic range and faster reaction kinetics for the target lectin, the gold nanoparticle counterparts proved to provide the bioassay that was most sensitive. Mixing lactose-stabilized gold nanoparticles and mannose-stabilized silver nanoparticles has shown that the selectivity of each glyconanoparticle is specific for the respective cognate ligand. Finally, it is clear that future aggregation-based colorimetric bioassays should be devised using either silver or gold nanoparticles dependent upon the analytical parameter that is of paramount importance for the measurement of the target analyte. Acknowledgment. Financial support from the RSC/EPSRC Analytical Ph.D. Studentship scheme for C.L.S. is gratefully acknowledged. LA060288R