Anal. Chem. 2007, 79, 9150-9159
Cellular Uptake of Gold Nanoparticles Passivated with BSA-SV40 Large T Antigen Conjugates Joseph A. Ryan,† K. Wesley Overton,† Molly E. Speight,† Christine N. Oldenburg,† LiNa Loo,† Wayne Robarge,‡ Stefan Franzen,† and Daniel L. Feldheim*,§
Departments of Chemistry and Soil Science, North Carolina State University, Raleigh, North Carolina 27695, and Department of Chemistry and Biochemistry, University of Colorado, Boulder, Colorado 80309
Internalization and subcellular localization in HeLa cells of gold nanoparticles modified with the SV40 large T antigen were quantified using inductively coupled plasma optical emission spectroscopy (ICP-OES). Internalization was monitored as a function of incubation time, temperature, nanoparticle diameter, and large T surface coverage. Increasing the amount of large T peptides per gold nanoparticle complex, by either increasing the coverage at constant nanoparticle diameter or by increasing the nanoparticle diameter at constant large T coverage, resulted in more cellular internalization. In addition, nuclear fractionation was performed to quantify nuclear localization of these complexes as a function of large T coverage. In contrast to our prior qualitative investigations of nuclear localization by video-enhanced color differential interference contrast microscopy (VEC-DIC), ICP-OES was able to detect nanoparticles inside fractionated cell nuclei. Although increasing the large T coverage was found to afford higher cell internalization and nuclear targeting, quantitative evaluation of cytotoxicity revealed that higher large T coverages also resulted in greater cytotoxicity. The ICP-OES and nuclear fractionation techniques reported here are valuable tools that can add important quantitative information to optical and electron imaging methods such as VEC-DIC and transmission electron microscopy regarding the fate of nanoparticles in cells. Polyvalent interactions are ubiquitous in biology. The valency of a particle (protein, virus, cell, etc.) is the number of connections it can make with another particle. It has been proposed that biological systems exploit polyvalent interactions because they allow an organism to take advantage of an existing set of monovalent (and perhaps weak) ligands rather than evolving completely new, higher affinity ligands for a given function.1-3 Indeed, polyvalent interactions can be very favorable; binding of a trivalent oligosaccharide ligand to its asialoglycoprotein cell * To whom correspondence should be addressed. E-mail: Daniel.Feldheim@ Colorado.edu. † Department of Chemistry, North Carolina State University. ‡ Department of Soil Science, North Carolina State University. § University of Colorado. (1) Mammen, M.; Choi, S. K.; Whitesides, G. M. Angew. Chem., Int. Ed. 1998, 37, 2754-2794. (2) Huskens, J. Curr. Opin. Chem. Biol. 2006, 10, 537-543. (3) Lundquist, J. J.; Toone, E. J. Chem. Rev. 2002, 102, 555-578.
9150 Analytical Chemistry, Vol. 79, No. 23, December 1, 2007
surface receptor occurs with a binding constant of 108 M-1 even though the binding constant of the corresponding monovalent interaction is only 103 M-1.4 Many synthetic systems are being explored for their ability to serve as polyvalent binders. These include dendrimers,5 liposomes,6 and gold surfaces (both planar and nanoparticles).7,8 Additionally, it has been observed that many viruses use multiple binding sites in order to facilitate membrane fusion and enhance endocytosis.9 Gold is advantageous as a multifunctional and polyvalent synthetic architecture as it can be synthesized in colloidal form with excellent control over shape and size and its surfaces are readily modified with small molecules, peptides, oligonucleotides, and polymers.10-22 The visible light extinction properties of gold nanoparticles also make them traceable inside (4) Sigal, G. B.; Mammen, M.; Dahmann, G.; Whitesides, G. M. J. Am. Chem. Soc. 1996, 118 (16), 3789-3800. (5) Landers, J. J.; Cao, Z.; Lee, I.; Piehler, L. T.; Myc, P. P.; Myc, A.; Hamouda, T.; Galecki, A. T.; Baker, J. R. J. Infect. Dis. 2002, 186, 1222-1230. (6) Jubeh, T. T.; Barenholtz, Y.; Rubinstein, A. Pharm. Res. 2004, 21 (3), 447453. (7) de Jong, M. R.; Huskens, J.; Reinhoudt, D. N. Chem.sEur. J. 2001, 7, 41644170. (8) Boal, A. K.; Rotello, V. M. J. Am. Chem. Soc. 2000, 122, 734-735. (9) Raja, K. S.; Wang, Q.; Finn, M. G. ChemBioChem 2003, 4, 1348-1351. (10) Tkachenko, A. G.; Xie, H.; Liu, Y.; Coleman, D.; Ryan, J.; Glomm, W. R.; Shipton, M. K.; Franzen, S.; Feldheim, D. L. Bioconjugate Chem. 2004, 15, 482-490. (11) Tkachenko, A. G.; Xie, H.; Coleman, D.; Glomm, W.; Ryan, J.; Anderson, M. F.; Franzen, S.; Feldheim, D. L. J. Am. Chem. Soc. 2003, 125 (16), 47004701. (12) Feldherr, M. C.; Akin, D. J. Cell Biol. 1990, 111, 1-8. (13) Xie, H.; Tkachenko, A. G.; Glomm, W. R.; Ryan, J. A.; Brennaman, M. K.; Papanikolas, J. M.; Franzen, S.; Feldheim, D. L. Anal. Chem. 2003, 75, 5797-5805. (14) Marinakos, S. M.; Novak, J. P.; Brousseau, L. C., III; House, A. B.; Edeki, E. M.; Feldhaus, J. C.; Feldheim, D. L. J. Am. Chem. Soc. 1999, 121, 85188522. (15) Liu, Y.; Shipton, M. K.; Ryan, J.; Kaufman, E. D.; Franzen, S.; Feldheim, D. L. Anal. Chem. 2007, 79, 2221-2229. (16) Rosi, N. L.; Giljohann, D. A.; Thaxton, C. S.; Lytton-Jean, A. K. R; Han, M. S.; Mirkin, C. A. Science 2006, 312 (5776), 1027-1030. (17) Wuelfing, W. P.; Gross, S. M.; Miles, D. T.; Murray, R. W. J. Am. Chem. Soc. 1998, 120, 12696-12697. (18) Lin, S. Y.; Tsai, Y. T.; Chen, C. C.; Lin, C. M.; Chen, C. H. J. Phys. Chem. B 2004, 108, 2134-2319. (19) Chithrani, B. D.; Chan, W. C. W. Nano Lett. 2007, 7 (6), 1542-1550. (20) Joshi, H. M.; Bhumkar, D. R.; Joshi, K.; Pokharkar, V.; Sastry, M. Langmuir 2006, 22, 300-305. (21) Xu, X. H. N.; Brownlow, W. J.; Kyriacou, S. V.; Wan, Q.; Viola, J. J. Biochemistry 2004, 43, 10400-10413. (22) Huff, T. B.; Hansen, M. N.; Zhao, Y.; Cheng, J. X.; Wei, A. Langmuir 2007, 23, 1596-1599. 10.1021/ac0715524 CCC: $37.00
© 2007 American Chemical Society Published on Web 10/31/2007
of cells using video-enhanced color differential interference contrast microscopy (VEC-DIC). For example, we have used VECDIC previously to monitor the subcellular localization of gold nanoparticles modified with cell and nuclear targeting peptides. VEC-DIC has revealed that particle internalization is dependent upon peptide sequence and cell type10 and that a combination of receptor-mediated endocytic (RME) and nuclear localization (NLS) peptides displayed on a single nanoparticle qualitatively afford more efficient nuclear targeting than particles modified with either peptide alone or even the analogous full-length (RME-NLS) sequence.11 Our prior work illustrated qualitatively the potential benefits of a multipeptide approach to cell targeting. It was then of interest to develop protocols for quantifying the effects of peptide coverage and gold nanoparticle size on cell uptake and nuclear localization. The peptide chosen for these studies was the large T antigen of the SV40 virus. It was known previously that large T is an effective NLS peptide if injected directly into the cytoplasm.12 During the course of our investigations we have found that it also behaves as an RME signal, affording rapid nanoparticle internalization into several cell lines studied (e.g., HeLa, 3T3, HepG2). Here we have characterized nanoparticle internalization into HeLa cells using inductively coupled plasma optical emission spectroscopy (ICPOES). While ICP analytical techniques have been previously used to study nanoparticle complexes of a different nature,19 in this research ICP-OES has revealed an increase in the number of nanoparticles internalized as the number of large T peptides per particle and nanoparticle size was increased. In addition, by fractionating cell nuclei following nanoparticle delivery it was possible to determine the ability of large T-modified nanoparticles to target the nucleus from outside of the cell. In contrast to prior work using VEC-DIC, which suggested that large T/gold nanoparticle conjugates were internalized into HeLa cells but were unable to escape the endosomal/lysosomal pathway and translocate the nuclear membrane,11 ICP-OES was able to detect nanoparticles inside the nucleus. ICP-OES and the nuclear fractionation techniques reported here are thus valuable tools that can add important quantitative information to optical imaging methods such as VEC-DIC. EXPERIMENTAL SECTION Materials. HeLa (human cervical cancer cells) cell line was purchased from the American Type Culture Collection (Rockville, MD). Minimal essential medium Eagle’s (EMEM), fetal bovine serum (FBS), Dulbecco’s phosphate-buffered saline (DPBS), T-25 and T-75 cell culture flasks, 12-well cell culture plates, and trypsin were purchased from Bio-Whittaker, Inc. (Walkersville, MD). All gold nanoparticles were purchased from Ted Pella, Inc. (Redding, CA). The modified large T peptide sequence (rhodamine-CysGly-Gly-Gly-Pro-Lys-Lys-Lys-Arg-Lys-Val-Gly-Gly-OH) was synthesized at the University of North Carolina Microprotein Sequencing and Peptide Synthesis Facility (Chapel Hill, NC). Bovine serum albumin (BSA) and subcellular fractionation kit were purchased from Pierce Co. (Rockford IL). [3-(4,5-Dimethylthiazol2-yl)-2,5-diphenyltetrazolium bromide] (MTT) assay was purchased from Promega Corp. Dimethylformamide (DMF), heparin sulfate, 4-(N-maleimidomethyl)-cyclohexane-1-carboxylic acid N-hydroxysuccinimide ester (SMCC), sodium chloride, mono basic sodium phosphate (NaH2PO3), dibasic sodium phosphate
(Na2HPO3), Optima-grade HCl, Optima-grade HNO3, Centricons (MWCO: 30 000), and glass coverslips (18 mm) were all purchased from Fisher Scientific. FluorSave cell mounting media was purchased from Calbiotech (San Diego, CA). All UV-vis absorption measurements were acquired using a Hewlett-Packard 8453 Chemstation photodiode array spectrophotometer with attached Chemstation software. Fluorescence spectra were acquired using a BioTek FL-600 plate reader with 550 ( 40 nm (excitation) and 590 ( 40 nm (emission) filters. All ICP experiments were performed with a Perkin-Elmer Optima 2100DV optical emission spectrometer equipped with a Meinhard type C glass nebulizer, unbaffled cyclonic chamber, and an alumina injection tube (2 mm opening). Other instrument settings include the following: 18 L/min plasma flow, 0.2 L/min auxiliary flow, 0.62 L/min nebulizer flow, 1.00 L/min pump rate, and 1500 W rf power. All ICP standards were made from 100 mg/mL SpexCertiPrep stock solution (lot no. CL3-19AU), and all ICP samples were analyzed at 242.795 nm with a read delay of 80 s and an integration time from 2 to 5 s. Methods. Conjugation of Peptide to BSA. Conjugation was performed in accordance with a previously published protocol.10 Briefly, 15.2 µL aliquots of a 20.0 mg/mL SMCC solution (in DMF) were added to six individual 1.0 mL BSA samples (1 mg/mL, in sodium phosphate buffer, pH ) 7.8) to achieve an SMCC/BSA molar ratio of 60:1 and allowed to mix at room temperature for 60 min. Excess unreacted SMCC was then removed from each of the samples via centrifugation (Centricon, MWCO: 30 000), and each sample was diluted to 1 mL with sodium phosphate buffer (pH ) 7.0). Varying volumes (49.5, 99.1, 148.6, 198.1, 247.6, and 297.2 µL) of rhodamine-labeled modified large T peptide (rhodamine-CGGGPKKKRKVGG; 3 mg/mL in sodium phosphate buffer; pH ) 7.0) were subsequently added to each BSA-SMCC sample to achieve six large T/BSA-SMCC molar ratios (5:1, 10:1, 15:1, 20:1, 25:1, and 30:1, respectively), and all samples were allowed to mix at room temperature for 24 h. Excess unreacted peptide was removed via centrifugation (Centricon, MWCO: 30 000). All samples were then diluted to 1 mL with sodium phosphate buffer (pH ) 7.0). Following dilution, the degree of peptide conjugation was then determined via examination of rhodamine fluorescence at 595 nm for each sample. Synthesis of Large T-BSA/Gold Nanoparticle Complexes. Each large T-BSA conjugate was added to separate 1 mL aqueous aliquots of 15 nm diameter gold nanoparticles (1.15 nM final concentration) in a large T-BSA conjugate/nanoparticle molar ratio of 250:1 and allowed to mix at room temperature for 30 min. The stability of large T-BSA/gold nanoparticle complexes was assessed by measuring the critical flocculation concentration. This assay was accomplished by adding 10 µL of a 10% NaCl solution to each sample of large T-BSA/gold nanoparticle complex. The samples were allowed to equilibrate for 30 min at room temperature and were then mixed thoroughly with a vortex mixer. For each sample, the visible spectrum of each sample was acquired and compared to the spectrum acquired in pure water. A critical flocculation concentration (CFC) value for each sample was then generated by repeated additions of 10% NaCl until either a significant red shift in the λmax was observed (indicating flocculation of the nanoparticles) or stability was observed at 1 M concentration of NaCl. Analytical Chemistry, Vol. 79, No. 23, December 1, 2007
9151
Cell Culture. HeLa cells were maintained in T-75 cell culture flasks using EMEM growth media containing 10% FBS at 37 °C and at 5% CO2. Prior to experimentation, cells were either plated on sterile glass coverslips in 12-well plates or seeded into T-25 cell culture flasks. In either case, cells were grown to approximately 75% confluency before beginning experiments. It should be noted that 6 h was chosen as a maximum incubation time for all experiments in this research to avoid potential effects on any cellular internalization due to overconfluency of the HeLa cells. Cell counting was performed using a fluorescent-activated cell sorting (FACS) instrument. ICP-OES Experiments. Four sets of control experiments were performed to probe the utility of ICP-OES as an analytical technique for use in subsequent determinations of nanoparticles internalized by HeLa cells. The first set of experiments was done to ensure that ICP-OES could accurately determine the concentration of gold nanoparticles in solution and that no matrix effects were introduced by the use of aqua regia in sample preparation. This was accomplished by placing 100 µL of 1 nM citrate-stabilized 20 nm diameter gold nanoparticles in each well of a 12-well cell culture plate and allowing the solutions to evaporate to dryness at room temperature. Six of the wells were then further treated with 0.5 mL of aqua regia while the remaining six wells received 0.5 mL of ultrapure water. The well plate was allowed to remain at room temperature for 2 h, after which 0.4 mL of each sample was placed in a 15 mL centrifuge tube, diluted to 3.9 mL with ultrapure water, and analyzed via ICP-OES. A second set of experiments was conducted to examine the relative degree to which nanoparticle complexes might associate with the glass coverslips in the absence of HeLa cells. Sterile glass coverslips were individually placed in each well of two 12-well cell culture plates. Six of these wells then received 150 µL of ultrapure water and were diluted to 1.5 mL with EMEM growth media, six wells received 150 µL of citrate-stabilized 15 nm diameter gold nanoparticles and were then diluted to 1.5 mL with growth media (0.23 nM final concentration), six wells received 1.5 mL of growth media containing 10% 15 nm diameter nanoparticle complexes stabilized with native BSA (0.23 nM final complex concentration), and the remaining six wells received 1.5 mL of EMEM growth media containing 10% 15 nm diameter nanoparticle complexes stabilized with a large T-BSA conjugate (15:1 large T/BSA experimentally determined molar ratio; 0.23 nM final complex concentration). Both well plates were then incubated at 37 °C and 5% CO2 for 6 h thus simulating the conditions used in later incubations with HeLa cells. After incubation, each well was rinsed with 1 mL of DPBS three times, and the coverslips were removed from their wells and allowed to air-dry in a sterile cell culture hood. These dried coverslips were then placed in a new well plate, treated with 0.5 mL of aqua regia for 2 h, after which 0.4 mL of each sample was placed in a 15 mL centrifuge tube, diluted to 3.9 mL with ultrapure water, and analyzed via ICP-OES. The third set of ICP-OES control experiments was performed to determine both if the HeLa cells themselves would interfere with the resulting observed ICP-OES signal and if the ICP-OES signal detected from combining HeLa cell samples prior to treatment with aqua regia was linear. The 15 nm gold nanoparticles were passivated with large T-BSA conjugate (15:1 large T/BSA experimentally determined molar ratio) using a large 9152
Analytical Chemistry, Vol. 79, No. 23, December 1, 2007
T-BSA/nanoparticle molar ratio of 250:1, and the final concentration of gold nanoparticles in the solution was approximately 2.3 nM. HeLa cells which had been previously cultured on sterilized coverslips placed in two 12-well plates were then allowed to incubate with either 1.5 mL of pure growth media (6 wells) or 1.5 mL of growth media containing 10% large T-BSA/gold nanoparticle complexes (18 wells; 0.23 nM nanoparticle concentration) per well for 6 h. Each well was then rinsed with 1 mL of DPBS three times, and the coverslips were removed from their wells and allowed to air-dry in a sterile cell culture hood. These dried coverslips were then placed in a new well plate; each of the 6 the coverslips containing HeLa cells which had not been incubated with the nanoparticle complexes was placed in its own well, 6 of the remaining coverslips were similarly placed in their own well, while the remaining 12 coverslips were placed in wells in pairs. Each well was then treated with 0.5 mL of aqua regia for 2 h, after which 0.4 mL of each sample was placed in a 15 mL centrifuge tube, diluted to 3.9 mL with ultrapure water, and analyzed via ICP-OES. A final set of experiments was performed to evaluate binding of nanoparticles to the exterior of the outer cell membrane. Gold nanoparticles, 15 nm in diameter, were passivated with large T-BSA conjugate (15:1 large T/BSA experimentally determined molar ratio) using a large T-BSA/nanoparticle molar ratio of 250:1 and a final concentration of gold nanoparticles of 2.3 nM. HeLa cells, which had been previously cultured on sterilized coverslips and placed in one 12-well plate, were allowed to incubate with 1.5 mL of growth media containing 10% large T-BSA/gold nanoparticle complexes (0.2 nM nanoparticle concentration) per well for 6 h. After incubation, each well was rinsed three times with 1 mL of DPBS, once with 1 mL of DPBS containing heparin sulfate (5 U/mL), and one final time with 1 mL of DPBS. Heparin sulfate is commonly used to desorb structures (e.g., molecules, biomolecules, particles) adhered to cell outer membranes. The samples were then treated with aqua regia and analyzed by ICP-OES as described above. Cellular Internalization of Large T-BSA/Gold Nanoparticle Complexes as a Function of Incubation Time. Gold nanoparticles 15 nm in diameter were passivated with large T-BSA conjugate (15:1 large T/BSA experimentally determined molar ratio) using a large T-BSA/nanoparticle molar ratio of 250:1, and the final concentration of gold nanoparticles in the solution was approximately 2.3 nM. HeLa cells which had been previously cultured on sterilized coverslips placed in 12-well plates were then allowed to incubate with 1.5 mL of growth media containing 10% large T-BSA/gold nanoparticle complexes (0.23 nM nanoparticle concentration) per well for varying amounts of time from 0 to 6 h. Six wells of HeLa cells were used for every time period under scrutiny. After the desired incubation time had elapsed, each well was rinsed with 1 mL of DPBS three times, and the coverslips were removed from their wells and allowed to air-dry in a sterile cell culture hood. These dried coverslips were then placed in a new well plate, combining two coverslips of identical samples per well. Each well was then treated with 0.5 mL of aqua regia for 2 h, and the resulting solution was prepared for ICP-OES analysis. Cellular Internalization as a Function of Large T/Gold Nanoparticle Ratio. Separate 15 nm diameter gold nanoparticle samples were passivated with one of six large T-BSA conjugates, each
one containing a different molar ratio of large T per BSA. The molar ratio of large T-BSA conjugates/gold nanoparticles was 250:1, and the final concentration of gold nanoparticles in the solution was 2.3 nM. HeLa cells which had been previously cultured on sterilized coverslips and placed in 12-well plates were then allowed to incubate with 1.5 mL of growth media containing 10% large T-BSA/gold nanoparticle complexes (0.23 nM nanoparticle concentration) per well for 1, 3, or 6 h. Additional well plates of HeLa cells which had been previously cultured on sterilized coverslips and placed in 12-well plates were removed from the incubator and cultured at 4 °C for 3 h prior to experimentation (along with the corresponding aliquots of DPBS and growth media containing nanoparticle complexes) in order to allow one replicate set of incubations to take place for 6 h at this temperature. Six wells of HeLa cells were used for every time period under scrutiny. After the desired incubation time had elapsed, each well was rinsed with 1 mL of DPBS three times, and then the coverslips were removed from their wells and allowed to air-dry in a sterile cell culture hood. These dried coverslips were then placed in a new well plate, combining two coverslips of identical samples per well. Each well was then treated with 0.5 mL of aqua regia for 2 h, and the resulting solution was prepared for ICP-OES analysis. Additional experiments under identical conditions were performed on separate samples of cells for the purposes of obtaining cell count and toxicity data. Cellular Internalization as a Function of Gold Nanoparticle Size. Experiments designed to probe the effect of nanoparticle size on cellular internalization were performed under two different experimental conditions. In both sets of conditions, separate samples of gold nanoparticles of varying diameter (10, 15, and 20 nm) but identical concentration (1.2 nM) were passivated with a large T-BSA conjugate (15:1 peptide/BSA experimentally determined molar ratio) prepared in the same fashion as described earlier. However, in one set of experiments, the large T-BSA conjugate was added to each nanoparticle sample in a large T-BSA/ nanoparticle molar ratio of 500:1 irrespective of nanoparticle size, while in the other set of experiments, conjugate was added to the samples of nanoparticles in a large T-BSA/nanoparticle molar ratio commensurate with nanoparticle surface area; that is, 20 nm diameter colloids were exposed to large T-BSA conjugates in a large T-BSA/nanoparticle molar ratio of 500:1, while the molar ratio used for 15 and 10 nm diameter colloids was 250:1 and 125:1, respectively. An additional sample of 20 nm diameter gold nanoparticles was prepared using native BSA in a BSA/nanoparticle molar ratio of 500:1. HeLa cells, which had been previously cultured on sterilized coverslips and placed in 12-well plates, were allowed to incubate with 1.5 mL of growth media containing 10% large T-BSA/gold nanoparticle complexes (0.12 nM nanoparticle concentration) per well for 6 h. Six wells of HeLa cells were used for every nanoparticle size under scrutiny. After the desired incubation time had elapsed, each well was rinsed with 1 mL of DPBS three times, and then the coverslips were removed from their wells and allowed to air-dry in a sterile cell culture hood. These dried coverslips were then placed in a new well plate, combining two coverslips of identical samples per well. Each well was then treated with 0.5 mL of aqua regia for 2 h, and the resulting solution was prepared for ICP-OES analysis.
Cellular Internalization as a Function of Excess Passivating Conjugate. In this experiment, separate samples of 15 nm gold nanoparticles were passivated with a peptide-BSA conjugate (15:1 peptide/BSA experimentally determined molar ratio) in varying large T-BSA/nanoparticle molar ratios (250:1 to 3000:1), and the final concentration of gold nanoparticles in the solution was 2.3 nM. Additional nanoparticle complexes were prepared using native BSA instead of large T-BSA conjugates. HeLa cells, which had been previously cultured on sterilized coverslips and placed in 12-well plates, were allowed to incubate with 1.5 mL of growth media containing 10% large T-BSA/gold nanoparticle complexes (0.23 nM nanoparticle concentration) per well for 6 h. Six wells of HeLa cells were used for every time period under scrutiny. After the desired incubation time had elapsed, each well was rinsed with 1 mL of DPBS three times, and then the coverslips were removed from their wells and allowed to air-dry in a sterile cell culture hood. These dried coverslips were then placed in a new well plate, combining two coverslips of identical samples per well. Each well was then treated with 0.5 mL of aqua regia for 2 h, and the resulting solution was prepared for ICP-OES analysis. Pulse-Chase Experiments. Gold nanoparticles 15 nm in diameter were passivated with a large T-BSA conjugate (15:1 peptide/ BSA experimentally determined molar ratio) in a large T-BSA/ nanoparticle molar ratio of 250:1, and the final concentration of gold nanoparticles in the solution was 2.3 nM. HeLa cells, which had been previously cultured on sterilized coverslips and placed in 12-well plates, were allowed to incubate with 1.5 mL of growth media containing 10% large T-BSA/gold nanoparticle complexes (0.23 nM nanoparticle concentration) per well for 3 h. After this “pulse,” each well was rinsed with 1 mL of DPBS three times, 1.5 mL of fresh growth media without nanoparticles was applied to each well, and the cells were allowed to further incubate for varying amounts of time (“chase” times) over a period of 12 h. Six wells of HeLa cells were used for every time period under scrutiny. After the desired chase time had elapsed, each well was again rinsed with 1 mL of DPBS three times, and then the coverslips were removed from their wells and allowed to air-dry in a sterile cell culture hood. These dried coverslips were then placed in a new well plate, combining two coverslips of identical samples per well. Each well was then treated with 0.5 mL of aqua regia for 2 h, and the resulting solution was prepared for ICPOES analysis. Subcellular Fractionation and Nuclear Targeting Efficiency. Separate 5 nm diameter gold nanoparticle samples were passivated with one of the six previously prepared large T-BSA conjugates containing varying amounts of large T per BSA in a peptideBSA/nanoparticle molar ratio of 10:1, and the final concentration of gold nanoparticles in the solution was 16.6 nM. HeLa cells, which had been previously cultured in T-25 cell culture flasks, were allowed to incubate with 5 mL of growth media containing 10% large T-BSA/gold nanoparticle complexes (1.6 nM nanoparticle concentration) per well for 6 h. Nanoparticles 5 nm in diameter were chosen for these experiments to minimize the potential accumulation of nanoparticles in nuclear fractions due solely to high-speed centrifugation (used in the cell fractionation protocol). T-25 flasks were chosen for use in these experiments in the interest of increasing the number of cells per experiment and thereby increasing the resulting gold emission signal detectAnalytical Chemistry, Vol. 79, No. 23, December 1, 2007
9153
Table 1. Number of Large T Peptides Per BSA and Per Gold Nanoparticle Complex large T/BSA mole ratio in reaction mixture
large T/BSA determined experimentallya
est no. of large T/15 nm diameter particlesb
est no. of large T/5 nm diameter particlesb
5.0 10.0 15.0 20.0 25.0 30.0
3(1 7(2 8(1 11 ( 2 13 ( 2 15 ( 2
300 ( 100 600 ( 100 700 ( 100 1000 ( 200 1200 ( 200 1300 ( 200
30 ( 10 70 ( 20 80 ( 10 110 ( 20 130 ( 20 150 ( 20
a Uncertainty represents the result of three measurements from three independent samples. b Estimate was based on the number of BSA molecules per 20 nm diameter gold nanoparticle determined using both fluorescence labeling and time-correlated single-photon counting (TCSPC) (ref 13) and adjusting for surface area. The standard deviation provided was calculated from the % error in the corresponding entry in column 2 and is not a propagated error.
able via ICP-OES. Consequently, six T-25 flasks of cells were used for each large T-BSA/nanoparticle complex under scrutiny. After the desired incubation time had elapsed, the cells in each flask were rinsed with 1 mL of DPBS three times and then trypsinized. Small aliquots (100 µL) of cells were removed from each flask for cell counting and toxicity assays using FACS, and the remaining cells were concentrated via centrifugation. Three of the six resulting samples per construct were then air-dried in a sterile cell culture hood, treated with 1.0 mL of aqua regia for 2 h, and each subsequent solution was prepared for ICP-OES analysis. The remaining three samples were subjected to fractionation using the Pierce Co. subcellular fractionation kit to obtain nuclear and cytosolic fractions of each sample. Each individual fraction was then air-dried in a sterile cell culture hood, treated with 1.0 mL of aqua regia for 2 h, and each resulting solution was prepared for ICP-OES analysis. Last, 12 additional T-25 flasks of HeLa cells, which had been incubated for 6 h with standard growth media (not containing colloidal constructs), were trypsinized and concentrated via centrifugation. These 12 flasks were divided into four groups containing three flasks each. All three of the flasks in a particular group had one of four different 5 nm nanoparticle complexes (passivation layers: native BSA, 7:1, 11:1, or 15:1 large T/BSA experimentally determined molar ratios) introduced to the fractionated cells prior to the final separation (via centrifugation) of nuclei from nonnuclear material. The samples created in such fashion are referred to as “spiked” samples. All of these fractions were then allowed to air-dry in a sterile cell culture hood, treated with 1.0 mL of aqua regia for 2 h, and each resulting solution was prepared for ICP-OES analysis. RESULTS Characterization of Large T-BSA/Gold Nanoparticle Complexes. In order to determine trends in the number of nanoparticles internalized into HeLa cells versus large T peptide coverage, it was first necessary to establish methods for modulating and quantifying the number of large T peptides per nanoparticle. This was accomplished by measuring the molar ratio of large T/BSA using rhodamine-labeled large T and fluorescence spectroscopy (Table 1). The number of large T-BSA conjugates per nanoparticle was subsequently calculated using a previously determined value for 9154 Analytical Chemistry, Vol. 79, No. 23, December 1, 2007
Table 2. ICP-OES Intensities and Calculated Number of Citrate-Stabilized 20 nm Diameter Nanoparticles for Standards Prepared with and without Acid Digestiona
analyte
intensity (× 105)
number of nanoparticles (× 1010)
Au nanoparticles Au nanoparticles (aqua regia)
1.537 ( 0.005 1.550 ( 0.009
6.220 ( 0.002 6.264 ( 0.004
a The standards contained 7.0 × 1010 gold nanoparticles. The intensity values from the two sets of samples were compared using Student’s t test and were observed to be statistically different from each other at the 95% confidence limit (tcalc ) 2.99). The uncertainty listed is the standard deviation of six replicate samples.
Table 3. ICP-OES Intensities of 15 nm Diameter Nanoparticle Complexes Constructed with Varying Passivating Layers and Incubated on Glass Coverslips for 6 h at 37 °C and in 5% CO2 in the Absence of HeLa Cellsa nanoparticle passivating layer
intensity (× 103)
18 MΩ H2O (no Au) citrate native BSA large T-BSA conjugate (14.6:1)
0.05 ( 0.03 0.18 ( 0.08 1.2 ( 0.1 4.0 ( 0.1
a The uncertainty listed is the standard deviation of six replicate samples.
the number of BSA molecules per 20 nm diameter gold nanoparticle.13 In that work, it was determined that 160 ( 8 [Ru(2,2′bipyridine)3]2+-labeled BSA complexes were associated with each 20 nm diameter gold nanoparticle using time-correlated singlephoton counting (TCSPC). For the experiments described here, it was presumed that large T-BSA conjugates associate with gold nanoparticles with a constant surface density irrespective of particle diameter and should thus scale in proportion to the total nanoparticle surface area. The first set of measurements was designed to test the reproducibility of gold nanoparticle analysis by ICP-OES, assess matrix effects potentially introduced by aqua regia sample digestion, and examine the statistical correlation between the number of nanoparticles detected by ICP-OES and the known value of nanoparticles contained in a standard. Table 2 shows the results of these analyses. One area of concern when using glass coverslips with proteincoated gold nanoparticles is nonspecific binding of particles to the glass substrate. To quantify this interaction, standard samples were incubated with bare coverslips (i.e., no cells) and then analyzed by ICP-OES. The data in Table 3 show that nonspecific binding of gold nanoparticles to the glass coverslips does occur to a small degree and increases with the addition of large T-BSA conjugates to the particles. The next set of experiments was designed to optimize ICPOES signals and determine the extent of nanoparticle nonspecific binding to the exterior cell membrane. In order to optimize ICPOES signals, the effects of combining two identically treated coverslips into one ICP-OES sample were examined. It was expected that coverslips containing identical cell coverage and incubated with an identical concentration of nanoparticle com-
Figure 1. Average number of 15 nm diameter gold nanoparticle complexes (determined via ICP-OES analysis) per HeLa cell (determined via cell counting) vs incubation time. The nanoparticles were modified with large T-BSA conjugates of a single, experimentally determined molar ratio, 15:1 large T/BSA. All incubations were performed at 37 °C and in 5% CO2. (Error bars not immediately visible are smaller than the data point.)
plexes could be combined to approximately double the resulting ICP-OES signal without introducing nonlinearity in the measured signal. ICP-OES analysis of HeLa cells not exposed to gold nanoparticles gave only a slightly larger signal than pure water (Table 4). HeLa cells incubated with large T-BSA/gold nanoparticle complexes yielded signals that were proportional to the number of coverslips combined prior to sample digestion. Thus, by combining coverslips the ICP-OES signal was increased without introducing artifacts from matrix effects or variability in cell coverage. A Student’s t test examining the variance in analyzing one coverslip versus two yielded no statistical difference at the 95% confidence limit (tcalc ) 2.15), and an analysis of the intensity of two coverslips with and without a heparin wash demonstrated a small but statistically significant difference at the 95% confidence limit (tcalc ) 2.36). Considering the average ICP-OES intensity values of the samples analyzed with and without a heparin wash, nonspecific binding of nanoparticles to the outer cell membrane contributes a maximum of 7% to the internalization of these nanoparticle complexes. Cellular Internalization of Peptide-BSA/Gold Nanoparticle Constructs as a Function of Incubation Time. Having established that stable large T-BSA/gold nanoparticle complexes can be synthesized, the uptake of these complexes by HeLa cells was investigated in time using nanoparticles that had constant coverage by large T. Results from this set of experiments are summarized in Figure 1. The average number of nanoparticle complexes per cell was observed to increase in time for ∼150 min, at which time a plateau was reached. The plateau continued for ∼90 min, after which time a second increase was observed. Cellular Uptake as a Function of Large T/Nanoparticle Ratio. The effects of large T coverage on gold nanoparticles on internalization were studied for coverages ranging from 0 to 2160 large T antigens per nanoparticle complex. Figure 2 shows that the number of nanoparticles internalized into HeLa cells increased as the number of large T peptides per nanoparticle was increased. The number of nanoparticles internalized was also dependent upon temperature. At 4 °C, internalization was dramatically reduced at
Figure 2. Plot of the average number of 15 nm diameter nanoparticle complexes per well as determined via ICP-OES analysis vs number of large T peptides per gold nanoparticle complex. All incubations were performed in 5% CO2 and under one of the following additional conditions: 1 h, 37 °C (diamond), 3 h, 37 °C (square), 6 h, 37 °C (triangle), and 6 h, 4 °C (×). (Error bars not immediately visible are smaller than the data point.)
Figure 3. Results of ICP-OES analyses of 6 h HeLa incubations with large T-BSA/nanoparticle complexes of varying size. The unshaded data set above corresponds to 6 h HeLa incubations with large T-BSA/nanoparticle complexes of varying size but constant large T-BSA conjugate concentration. The shaded data set corresponds to 6 h HeLa incubations with large T-BSA/nanoparticle complexes of varying size and varying large T-BSA conjugate concentration in an attempt to maintain constant surface coverage of the nanoparticles in accordance with size. All incubations were performed at 37 °C and in 5% CO2.
all large T/nanoparticle ratios compared to identical experiments performed at 37 °C. Cellular Uptake as a Function of Gold Nanoparticle Size. Results from experiments involving complexes of varying nanoparticle size are summarized in Figure 3. Although both sets of experiments involve incubations of HeLa cells with a single concentration of gold nanoparticles, the unshaded bars correspond to experiments in which a single concentration of large T-BSA conjugate was used, thus leading to a potentially dramatic excess of large T-BSA conjugate with the large T-BSA/gold nanoparticle complexes of smaller diameter. The shaded bars correspond to experiments performed with a varying concentration of large T-BSA conjugate such that the concentration of this conjugate is scaled according to available surface area on the gold nanoAnalytical Chemistry, Vol. 79, No. 23, December 1, 2007
9155
Figure 4. Plot of the average number of 15 nm diameter nanoparticle complexes per well as determined via ICP-OES analysis vs number of BSA (unshaded squares) or large T-BSA (shaded squares) per 15 nm diameter gold nanoparticle complex. All incubations were performed for 6 h at 37 °C and in 5% CO2. (Error bars not immediately visible are smaller than the data point.)
particles of a given diameter; that is, the surface areas of the 10 and 15 nm diameter nanoparticles are smaller than that of the 20 nm diameter nanoparticles by a factor of 4 and 1.78, respectively, so a proportionally decreased amount of large T-BSA conjugate was used in synthesizing these particular large T-BSA/gold nanoparticle complexes. It can been seen in the data that larger-diameter nanoparticle complexes were internalized more readily by HeLa cells than smaller-diameter nanoparticle complexes and the excess large T-BSA conjugate present in certain 10 and 15 nm diameter gold nanoparticle complex samples (unshaded bars) appears to have marginally increased cellular internalization. Note also that BSA/ gold nanoparticle complexes lacking the large T peptide exhibited a decrease in cellular internalization of over one order of magnitude. Cellular Uptake as a Function of Excess Passivating Conjugate. Since internalization of 15 nm diameter gold nanoparticle complexes in the previous experiment appeared to increase in the presence of excess large T-BSA conjugate, a further examination of this phenomenon was warranted. It can be concluded from the data in Figure 4 that having excess native BSA or large T-BSA present during cellular incubation with nanoparticle complexes did not inhibit cellular uptake of these complexes, even when that excess was increased over 10-fold. Additionally, nanoparticle complexes using large T-BSA once again exhibited greatly enhanced cellular uptake compared to the nanoparticles passivated with native BSA. Cellular Uptake: Pulse-Chase. As seen in Figure 5, once the nanoparticle constructs were taken up by the HeLa cells they remained associated with the cells in similar concentration for up to 6 h. Subcellular Fractionation, Nuclear Targeting, and Toxicity. The data presented in Table 5 show a direct relationship between the number of large T peptides per nanoparticle and cellular internalization. The next objective of this study was to determine if large T-BSA/gold nanoparticle complexes were localized in the nucleus. As shown in Table 5, nanoparticles were detected in the nuclear fractions, and the general trend of increased nuclear targeting with increased large T/BSA coverage 9156 Analytical Chemistry, Vol. 79, No. 23, December 1, 2007
Figure 5. Plot of the average number of 15 nm diameter gold nanoparticle complexes per well as determined via ICP-OES analysis vs chase time. The pulse was a 3 h incubation with large T-BSA/ gold nanoparticle complexes, and all incubations were performed at 37 °C and in 5% CO2. (Error bars not immediately visible are smaller than the data point.) Table 4. ICP-OES Intensities of HeLa Cell Samples Incubated for 6 h at 37 °C and in 5% CO2 with and without 15 nm Diameter Gold Nanoparticle Complexes Constructed Using a 15:1 Large T/BSA (Experimentally Determined Ratio) Conjugatea analyte
intensity (× 104)
HeLa cells (no Au) 1 coverslip 2 coverslips 2 coverslips (heparin wash)
0.010 ( 0.005 1.051 ( 0.064 2.247 ( 0.105 2.099 ( 0.112
a The uncertainty listed is the standard deviation of six replicate samples.
Table 5. Number of 5 nm Diameter Gold Nanoparticle Complexes Per HeLa Cell (as Determined via ICP-OES) as a Function of Large T Coveragea large T per nanoparticle 0 (native BSA) 30 70 80 110 130 150
nanoparticles nanoparticles nanoparticles per cell; per cytosolic per nuclear % unfractionated fraction fraction nuclear 3.48 × 104 3.52 × 104 1.21 × 105 4.97 × 105 1.16 × 106 1.80 × 106 5.44 × 106
1.33 × 104 5.22 × 103 1.54 × 104 3.21 × 104 4.74 × 104 8.97 × 104 1.20 × 105
8.18 × 103 9.29 × 103 7.04 × 103 1.73 × 105 9.61 × 105 1.47 × 106 3.11 × 106
37.5 62.5 82.4 84.4 95.3 94.2 96.3
a All incubations were performed for 6 h at 37 °C and in 5% CO . 2 The standard deviations of all listed values are all within (10%.
was observed. Note from Table 5 that the number of nanoparticles detected per unfractionated cell does not equal the sum of nanoparticles detected in the nuclear and cytosolic fractions. This most likely indicates that the fractionation process leads to a loss of nanoparticles, a loss that cannot be assigned definitively to the cytosolic or nuclear fraction. However, it has been considered that the values reported in Table 5 could be artificially inflated by particles that did not reach the nucleus during delivery but were instead simply pelleted with the nuclei during fractionation and centrifugation. To quantify this potential source of error, replicate
Table 6. Percentages of 5 nm Diameter Gold Nanoparticle Complexes Observed in Cytosolic and Nuclear Fractions (as Determined via ICP-OES) Following Injection of These Complexes Postincubationa large T per nanoparticle
% cytosolic
% nuclear
% nuclear (corrected)
0 (native BSA) 70 110 150
80.86 77.57 61.15 59.09
19.06 22.38 38.81 40.86
18.4 60.0 56.5 55.4
a HeLa cells were incubated for 6 h at 37 °C and in 5% CO , lysed, 2 spiked with nanoparticle complexes, and then centrifuged to recover nuclear and cytosolic fractions. Standard deviations of all listed values are all within (10%.
Table 7. Cell Counts Per Flask as a Function of Number of Large T Peptides Per Gold Nanoparticle as Determined by FACSa
b
large T per 5 nm nanoparticle
cell count per flask (× 104)
% viableb
control (no gold) 0 (native BSA) 30 70 80 110 130 150
98.6 ( 4.9 96.7 ( 4.4 96.3 ( 5.1 94.7 ( 5.5 87.0 ( 9.8 80.4 ( 7.0 68.5 ( 7.3 65.0 ( 10.1
100 98.1 97.7 96.0 88.2 81.5 69.5 65.9
a All incubations were performed for 6 h at 37 °C and in 5% CO . 2 % viability was referenced to the cell count of the control sample.
samples of cells not incubated with nanoparticle complexes were fractionated and then spiked with nanoparticle complexes prior to centrifugation. The number of nanoparticles in both the cytosolic and nuclear fractions was analyzed by ICP (Table 6). Note that in these experiments the sums of the nanoparticles determined in each of the two fractions were virtually identical to the total number of nanoparticles originally injected into the fractionated sample. Furthermore, a different concentration of nanoparticle complexes was used in spiking these samples than was used in the previous internalization samples, so the results in Table 6 are listed in percentages both to avoid confusion and to allow direct comparison to those results listed in Table 5. The largest recoveries of nanoparticles in this control experiment were for the highest large T coverage (38.81% and 40.86%, after correction; Table 6). As the percentages of nanoparticle complexes detected in the nuclear fractions in the standard deliveries for these same complexes were substantially higher (95.3% and 96.3%; Table 5), it is unlikely that the results of nuclear targeting reported in Table 5 are mere artifacts of centrifugation alone. These controls thus mark a lower limit on the number of large T-BSA/gold nanoparticle complexes that are determined to have reached the nucleus from the extracellular environment. Finally, toxicity levels of the nanoparticle complexes were estimated as a function of cell count per T-25 flask (Table 7). All “% viable” values were comparisons of cell counts for cells incubated with large T-BSA/gold nanoparticle complexes with counts for cells grown in the absence of nanoparticle complexes.
These data indicate that larger numbers of large T peptides per nanoparticle complex increased cytotoxicity significantly. DISCUSSION ICP-OES has demonstrated its efficacy as a valuable analytical tool for quantifying cellular internalization of metal nanoparticles. With regard to the ICP-OES control experiments, it was concluded: (1) there was a small but statistically significant increase within the 95% confidence limit in ICP-OES intensity when gold nanoparticles were prepared by digestion in aqua regia compared to undigested samples (Table 2), (2) any residual nanoparticle complexes adhered to the coverslips did not significantly alter the resulting ICP-OES signal (values from Table 3 compared to those in Table 4), (3) doubling the number of coverslips per sample was an effective way to double the resulting ICP-OES signal without introducing nonlinear effects in the signal (Table 4), and (4) although using a postincubation heparin wash afforded a very small yet statistically significant difference in the resulting ICP-OES signal (Table 4), the vast preponderance of complexes detected after incubation with HeLa cells were most likely internalized as opposed to merely associated with the exterior of the cells. This last conclusion is based upon the hypothesis that heparin would remove nonspecifically bound gold nanoparticles from the exterior of the cell membrane as has been reported for other molecules.23 The first examination of cellular internalization performed was designed to quantitatively examine the relationship between gold nanoparticle complex internalization and incubation time. While the internalization of gold nanoparticle complexes was observed to generally increase as incubation time increased (Figure 1), the amount of gold nanoparticle complexes detected appeared to reach a plateau around 150 min and resumed an increasing trend after 240 min. Other work24 has shown that certain cell lines under particular incubation conditions can enter a “lag phase” in their growth cycles, meaning the growth cycle of the cells may slow down for a period of time as the cells adjust to the ambient presence of a foreign agent (i.e., experimental probe or delivery vector) but resume normal (or even increased) growth rate after this adjustment period. This may help explain why the internalization of gold nanoparticle complexes by HeLa cells in Figure 1 appeared to plateau, but this phenomenon was not further probed in this study. Once a direct relationship between gold nanoparticle complex internalization by HeLa cells and incubation time was established (Figure 1), it was of interest to probe nanoparticle internalization over time as a function of the number of large T peptides per nanoparticle. Moreover, it was of interest to examine whether this internalization proceeds via an energy-dependent mechanism. To this end, several concurrent experiments were performed on cells incubated with nanoparticle complexes modified with varying amounts of large T (Figure 2). It was observed that increasing the amount of large T peptides per gold nanoparticle complex resulted in more cellular internalization irrespective of incubation time (Figure 2). Furthermore, larger amounts of gold nanoparticle complexes were internalized (23) Tyagi, M.; Rusnati, M.; Presta, M.; Giacca, M. J. Biol. Chem. 2001, 276 (5), 3254-3261. (24) Hernandez, L. D.; Hoffman, L. R.; Wolfsberg, T. G.; White, J. M. Annu. Rev. Cell Dev. Biol. 1996, 12, 627-61.
Analytical Chemistry, Vol. 79, No. 23, December 1, 2007
9157
at 37 than at 4 °C, and this disparity became greatly pronounced as the number of large T peptides per gold nanoparticle complex increased. It is unclear whether nanoparticle complexes are merely adhering to the outside of the cells at 4 °C or if they are being internalized at a much slower rate. The former option implies this internalization process proceeds through an energy-dependent mechanism, whereas the latter allows for the possibility that at least some of the gold nanoparticle complexes are able to be internalized via an energy-independent pathway. Although it is possible that cellular internalization utilizes an energy-independent pathway, the much higher uptake seen at 37 °C appears to indicate energy dependence for this process. The gold nanoparticle complexes detected at 4 °C may therefore represent complexes bound to their receptors but not internalized due to the energy constraint. In the interest of further probing the dynamics of cellular internalization of these gold nanoparticle complexes, experiments were performed using nanoparticles of varying diameters (Figure 3). These experiments showed a direct relationship between nanoparticle uptake and nanoparticle size. Furthermore, although it might have been expected that having excess large T-BSA conjugate in the growth media may have competitively inhibited cellular uptake of the nanoparticle constructs, this was not observed (Figure 4). This implies either that under these conditions the size and/or polyvalency of the large T-BSA/gold nanoparticle complexes allows for more efficient internalization versus the large T-BSA conjugates alone or that an excess of cellular receptors are present. As the observation of the internalization of the large T SV40 antigen from the pericellular environment is a new result, any cellular receptor it may be accessing and/or the number of those receptors is currently unknown. Another interesting dimension to the phenomena under scrutiny in this research is the ultimate fate of a given nanoparticle complex which has been internalized by a HeLa cell. A pulsechase experiment was performed to determine if the amount of time that internalized gold nanoparticle complexes were in HeLa cells was short (perhaps quickly shunted out of the cell via some exocytotic pathway) or if these gold nanoparticle complexes associated with a cell for an appreciable amount of time. The gold nanoparticle complexes did, in fact, remain associated with the HeLa cells in similar concentration up to 6 h after the cells were no longer exposed to the complexes (Figure 5). Experiments were designed to probe two critical aspects of any potential intracellular vector: (1) the intracellular location at which these nanoparticle complexes may reside/accumulate, and (2) cytotoxicity. Since large T is known to localize in the cell nucleus when injected directly into the cytoplasm,12 experiments were designed to quantify nuclear localization of large T-BSA/ gold nanoparticle complexes. Many cellular fractionation kits exist to facilitate the process of extracting nuclei from other cellular material. Unfortunately for the goals of this research, most of these fractionation kits utilize high-speed centrifugation as the primary method of separation of nuclei from remaining cellular organelles/ debris, and this centrifugation would certainly result in the aggregation of larger-sized gold nanoparticles resulting in an artificially large amount of gold detected in the nuclear fraction obtained in this manner. However, using a much smaller-sized nanoparticle (5 nm diameter) minimized this effect to the point 9158
Analytical Chemistry, Vol. 79, No. 23, December 1, 2007
that genuine trends in nuclear localization of nanoparticle complexes could be observed (Table 5). To quantify the effects of centrifugation on the amount of 5 nm diameter large T-BSA/gold nanoparticle complexes recovered in nuclear fractions, lysed cells were spiked with nanoparticles prior to centrifugation. A significantly higher concentration of these complexes was used to spike the samples in these control experiments than was delivered to the cells under the standard incubation conditions (Table 5). Increasing the number of large T peptides per gold nanoparticle complex did increase the percentage of complexes, which ultimately associated with the nuclear pellet solely due to centrifugation. This could be due to a stronger association between large T-BSA/gold nanoparticle complexes and the outer nuclear membrane as the number of large T peptides per particle is increased. However, when this percentage was used to correct for the number of gold nanoparticles detected per nuclear fractionsundoubtedly an overcorrectionsit was clear that centrifugation effects could not completely explain the large percentage of nanoparticles complexes observed to be associated with the nuclei. The ability of large T to function as a nuclear localization signal when attached to gold nanoparticles can be highlighted by comparing the data for BSA/gold nanoparticle complexes (no large T) and complexes carrying 150 large T peptides per nanoparticle. When corrected for centrifugation effects, 55% of the nanoparticles modified with large T were detected in the nuclear fraction, compared to 18% of the BSA-modified complexes. This result clearly demonstrates the effectiveness of large T as a nuclear localization agent, even when introduced in the pericellular environment. There is a maximum efficiency for nuclear targeting evident from the data in this study. Tables 5 and 6 show that although an increase in the number of large T peptides per nanoparticle results in an increase in the number of nanoparticles internalized, the percentage of those nanoparticles that reached the nucleus appears to have decreased slightly. Whether this is related to the increase in cytotoxicity with an increase in large T coverage, an increase in nanoparticle complex size that compromises its ability to pass through the nuclear pore complex, or some other factor is unknown. Irrespective of the origin of the trends in nuclear targeting discussed above, it is interesting to compare nuclear targeting as judged by ICP-OES with that of VEC-DIC. Using VEC-DIC, large T-BSA/gold nanoparticle complexes were previously detected in the cytoplasm of HeLa cells after 3 h but not inside the nucleus.11 It was hypothesized that the apparent failure to traverse the nuclear membrane was a result of the inability of these conjugates to escape endosomes intact. ICP-OES, in contrast, was able to detect the presence of these complexes inside fractionated HeLa cell nuclei. The only differences in the two experiments were particle size (5 nm diameter gold in ICP vs 20 nm diameter in VEC-DIC) and incubation time (6 h in ICP-OES vs 3 h in VECDIC). These disparate results may be justified by considering (i) that ICP-OES is a “ensemble average” technique and is thus less susceptible to sample heterogeneity compared to VEC-DIC, which analyzes a random (and relatively low) population of individual cells, (ii) VEC-DIC may be less sensitive than ICP-OES and thus failed to image nanoparticles in the nucleus that ICP-OES was
able to detect, and/or (iii) the larger diameter of the nanoparticles investigated by VEC-DIC slowed their transport through the nuclear pore complex. Last, the toxicity of these complexes increased with an increase in large T coverage as analyzed via FACS. As seen in Table 7, toxicity increased in direct proportion to number of large T peptides per nanoparticle. However, it should be noted that gold nanoparticle complexes containing small numbers of large T peptides exhibited negligible toxicity compared to that of complexes with no large T peptides at all (native BSA). Thus, by attaching a relatively low number of large T peptides per nanoparticle (∼30 peptides per 5 nm diameter nanoparticle), relatively high nuclear localization and low toxicity can be achieved. CONCLUSIONS The data gathered from these experiments demonstrate the efficacy of ICP-OES as an analytical tool for use in biological assays involving gold nanoparticles. ICP-OES allows for the efficient quantification of many delivery parameters (i.e., differences in internalization over time, degree of nuclear localization, etc.), which would be difficult to examine using only standard microscopic imaging techniques such as transmission electron microscopy (TEM), confocal microscopy, or VEC-DIC. It can be further deduced from the data presented that the number of large T peptides per nanoparticle was the determinant of the efficiencies of internalization and nuclear localization of these complexes. While this peptide sequence has been shown previously to enhance nuclear internalization of similar gold nanoparticle complexes when microinjected into various cell lines,11 it is noteworthy that this peptide-BSA conjugate greatly enhanced cellular internalization of large T-BSA/gold nanopar-
ticle complexes from areas external to the cells and also simultaneously maintained its inherent ability to enhance subsequent localization in the nuclei of these cells. Additionally, it has been shown in this research that the presence of excess large T-BSA conjugate during incubation of large T-BSA/gold nanoparticle complexes with HeLa cells does not seem to adversely affect the ability of the cells to internalize the gold nanoparticle complexes and that these complexes remain associated with the HeLa cells in similar concentration for up to 6 h postincubation. Last, it should be noted that although the gold nanoparticle complexes comprised of the largest concentrations of large T used in this research enhanced cellular internalization in general (and nuclear localization in specific) compared to the gold nanoparticle complexes synthesized using the smallest amounts of large T, the overall toxicity of the gold nanoparticle complexes also increased proportionally with increasing number of large T peptides in the gold nanoparticle complexes. However, the data also demonstrated that using gold nanoparticle complexes of comparatively low large T coverage enhanced cellular uptake and greatly improved the amount of gold nanoparticle complexes, which were associated with the nuclei of the cells without introducing significant toxicity. ACKNOWLEDGMENT This work was supported by Grants from The David and Lucile Packard Foundation, NSF-DMR (9900073), and NIH NCI 9819401.
Received for review July 23, 2007. Accepted September 25, 2007. AC0715524
Analytical Chemistry, Vol. 79, No. 23, December 1, 2007
9159
