Real-Time Molecular Imaging throughout the Entire Cell Cycle by

Sep 14, 2012 - Laser Dynamics Laboratory, School of Chemistry and Biochemistry, Georgia Institute of ... Moustafa R. K. Ali , Yue Wu , Tiegang Han , X...
0 downloads 0 Views 2MB Size
Download PDF
Letter pubs.acs.org/NanoLett

Real-Time Molecular Imaging throughout the Entire Cell Cycle by Targeted Plasmonic-Enhanced Rayleigh/Raman Spectroscopy Bin Kang,†,‡ Lauren A. Austin,‡ and Mostafa A. El-Sayed* Laser Dynamics Laboratory, School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, Georgia 30332-040, United States S Supporting Information *

ABSTRACT: Due to their strong enhancement of scattered light, plasmonic nanoparticles have been utilized for various biological and medical applications. Here, we describe a new technique, Targeted Plasmonic-Enhanced Single-Cell Rayleigh/Raman Spectroscopy, to monitor the molecular changes of any cell-component, such as the nucleus, during the different phases of its full cell cycle by simultaneously recording its Rayleigh images and Raman vibration spectra in real-time. The analysis of the observed Raman DNA and protein peaks allowed the different phases of the cell cycle to be identified. This technique could be used for disease diagnostics and potentially improve our understanding of the molecular mechanisms of cellular functions such as division, death, signaling, and drug action. KEYWORDS: Plasmonic nanoparticles, Raman/Rayleigh spectroscopy, cell cycle

T

nucleus, the cytoplasm, cellular membrane, etc.) during the full cell cycle through the use of targeted gold nanospheres (AuNS) or other plasmonic nanoparticles. In this work, we target the nucleus of human oral squamous carcinoma (HSC-3) cells. Our experimental setup combines dark-field Rayleigh scattering imaging optics, a confocal micro-Raman instrument, a home-built live cell chamber, and a data analysis system (Figure 1a). This arrangement enables the simultaneous acquisition of Rayleigh scattering cell images and spectra, Raman mapping images, and Raman spectra from any component within single living cells (Figure 1b). Dark-field Rayleigh scattering images were taken to locate the intracellular position of the plasmonic AuNSs while Rayleigh scattering spectra were recorded from the single living cell. The high-NA objective lens allowed a laser spot size on the sample of approximately 1 μm with the focal plate depth being tunable by the control of the confocal pinhole size. Due to the rejection of Raman scattering signal from the regions outside the focus plate by the confocal optics, we were able to obtain high signal-tonoise ratios for single cell spectroscopy as well as highresolution spectra-mapping images. For targeting, AuNSs with different organelle targeting peptides were used. Poly(ethyelene glycol) stabilized AuNSs with a diameter of 24 nm were functionalized with an RGD and a nuclear localzing signal (NLS) peptide that allows for selective delivery of the nanoparticles to the nucleus of the cell

he development of modern molecular and cell biology has provided valuable techniques such as fluorescence microscopy, confocal microscopy, and flow cytometry to monitor cellular functions and the expression level of specific molecules.1−6 Recently, making use of the strong plasmonic enhancement of light scattering by gold or silver nanoparticles, nanotechnology has improved the ability to detect and sense molecular information in single living cells7−10 and help improve medical diagnosis using dark field microscopy.11 Using plasmonically enhanced Rayleigh scattering in dark field and targeting techniques, our group has been able to follow the response of the cancer cells whose nuclei have been targeted with gold or silver nanoparticles during their entire cell cycle through cell division.12−15 While concentrations below 0.4 nM of nuclear-targeted gold nanospheres were not able to cause cell death, the nuclear-targeted silver particles exhibited cell death at concentrations as low as 0.1 nM. Raman scattering provides information on molecular vibrations and thus on molecular identity. Recently, several groups have used Raman spectroscopy to discriminate between cell cycle phases and live and dead cells as well as for in vivo tumor detection.16−24 All of these methods offered an excellent, detailed view of a specific cellular function or molecule. In the present work we present a technique that is potentially able to acquire both structural and molecular information in real-time in the different phases of a single cell during its full cycle. Here, we present a new technique, Targeted Plasmonically Enhanced Single Cell Raman/Rayleigh Spectroscopy (TPESCRR). This technique enables cell imaging and molecular monitoring of any targeted cellular component (i.e., the © 2012 American Chemical Society

Received: July 25, 2012 Revised: September 4, 2012 Published: September 14, 2012 5369

dx.doi.org/10.1021/nl3027586 | Nano Lett. 2012, 12, 5369−5375

Nano Letters

Letter

Figure 1. Experimental design of recording the plasmonically enhanced Rayleigh and Raman spectra and images for targeted AuNSs during the full cell cycle. (a) The setup of plasmonically enhanced single-cell Rayleigh/Raman nanospectroscopy; (b) the dark field Rayleigh scattering images (top left), Rayleigh plasmon scattering spectrum (top right), Raman scattering images (bottom left), and surface-enhanced Raman spectrum (bottom right) from a single living cell recorded from the setup. The Rayleigh spectrum (b) shows the surface plasmon spectrum of the targeted-AuNSs and suggests that their plasmonic field is comprised of two components, one ∼600 nm from single nanoparticles and a stronger broader band at ∼700 nm from AuNS aggregates.

Figure 2. Plasmonic-enhancement of Rayleigh and Raman scattering signals from targeted-AuNSs. (a) Rayleigh scattering dark field images of the nanospheres targeted to the nucleus of a HSC-3 cell; (b) Raman mapping of the same cell shown in a; (c) the Raman spectra from different locations within cell, specifically at the targeted nucleus (blue) and away from the nucleus (red); (d) the mapping of the Raman peaks shown in c. Raman mapping gives the locations where molecular vibrations are detected, and due to extensive coverage this might suggest that the plasmonic field covers the entire nucleus (most likely from the aggregates).

(Figure S1 and Table S1 of the Supporting Information).12−14,25 The surface plasmon resonance (SPR) properties of the AuNSs are exploited to enhance the Rayleigh scattering, showing bright scattering signals in the dark-field images, which aids in the determination of their intracellular location (Figure 2a). Once the particles have been internalized by the cells (∼24 h incubation), the SPR scattering preak of the AuNSs shifts

from 524 nm to the strong peak at 620 nm and a much stronger aggregation peak at ∼700 nm (Figure S1, Figure 1b). A red laser (785 nm) was used as the exciting light. Under the Raman spectra mapping mode, we acquired Raman spectrograph images at different positions within the cell (Figure 2b), which clearly shows the signal enhancement at the nucleus due to the use of AuNSs targeting. With a 1 s acquisition time, the Raman 5370

dx.doi.org/10.1021/nl3027586 | Nano Lett. 2012, 12, 5369−5375

Nano Letters

Letter

Figure 3. Real-time dark-field images and spectra of the targeted nucleus during the complete cycle of a single HSC-3 cell: (a) the cell cycle; (b) the plasmonically enhanced Rayleigh images; (c) the plasmonically enhanced Raman spectra. (Also see Supplementary Figure S2 and S6 for individual spectra collected at each time point in three independent experiments and the Raman spectrum of the nuclear-targeted AuNSs used in this work.)

spectra of the cytoplasm (without AuNSs) showed a flat spectra with no features; however, the cell nucleus (with AuNSs) exhibited a strong, well-resolved Raman spectrum due to the calculated ∼103−104 AuNS enhancement factor (Figure 2c). The Raman spectra mapping mode also allowed spacial mapping of each individual Raman band (Figure 2d). The mapping images showed little difference in the signal distribution suggesting that the surface plasmon field of the aggregated AuNSs was relatively homogeneous. Cell Nucleus Imaging and Its Enhanced Molecular Raman Spectra during the Cell Cycle. Figure 3 shows the plasmonically enhanced Rayleigh scattering images (Figure 3b) and enhanced Raman spectra (Figure 3c) in real-time during the full cell cycle (Figure 3a) of HSC-3 cells treated with targeted-AuNSs that were synchronized in the G1 phase by serum deprevation and released into fresh medium.26−28 Darkfield images using enhanced Rayleigh and Raman spectra were recorded at various time points after release until cell division occurred. The acquired Raman spectra was averaged from three independent experiments, and for each experiment, spectra from 10 individual cells were taken to provide an average value within each experimental run (Figure S2, Supporting Information). In our experiment, cells were able to pass through the cell phases in the order from G1-S-G2-M, with increasing release time (Figure 3a). Dark-field images were

obtained at each time point to monitor the cellular community as well as to locate the position of the AuNSs (Figure 3b). The plasmonically enhanced Raman spectra of the cell nucleus were collected by focusing the laser beam on the cell nucleus (Figure 3c). Distinct changes with time in the spectrum are observed during the cycle. Most can be correlated with the phase changes as determined from cytometry (using the amount of DNA that characterizes each phase). During the first 8 h, the Raman spectra do not display any significant changes; however, at hours 9 and 10, the bands around 511 cm−1, 658 cm−1, 844 cm−1, and 1322 cm−1, which correspond to the −S− S− vibration, guanine, sugar−phosphate and adenine, and guanine and RNA, respectively,8,21,23,29 begin to decrease and remain relatively weak from 10 to 20 h. After cell division and the cell enters into the G1 phase, these peaks regain their intensity. The bands between 1400 and 1600 cm−1 (protein C−H, amide I band) increase significantly from 9 to 10 h through 20−21 h and then decrease roughly back to their original intensities at hours 22−24. This corresponds to the cells’ transition from the M to the G1 phase to the cell cycle in the daughter cell. It is important to note that at hours 23 and 24, when the cells have progressed into the G1 phase of the daughter, the Raman spectra resemble those in the G1 phase at hours 3 and 4 of the parent. As cells in phases undergo different 5371

dx.doi.org/10.1021/nl3027586 | Nano Lett. 2012, 12, 5369−5375

Nano Letters

Letter

Figure 4. Fit of the time dependence of the Raman vibrations of molecules appearing in different times of the cell cycle by time-resolved Raman spectroscopy to the time course of the different phases of the cell cycle as determined from flow cytometry. Flow cytometry results identifying the different phases: G1, S, G2/M with time of release (a,d,g), Raman spectroscopy (b,e,h), and the overlay comparisons (c, f,i).

phases of the complete cell cycle, we followed the cell cycle by flow cytometry. Cells were synchronized in the G1 phase, and cell cycle phase analysis was conducted at different time points of release using DNA staining via propidium iodide (PI). The flow cytometry results clearly show the dynamics of the cell cycle from G1 phase through S and G2/M and returning to G1 (Figure 5a, Figure S3−4 of the Supporting Information). After the cells were released into fresh medium, a recovery time of ∼5 h was required. The G1 phase showed its maximum at ∼9 h, and as it decreased at around 12 h, the percentages of cells in the S phase, which started at 8−9 h, showed a rapid increase at ∼12 h. At hour 18, the S phase became the dominat phase, and the conversion into the G2/M phase followed. Complete passage through G2/M and the new cell cycle occurred after ∼25 h, as indicated by the reappearance of the G1 phase. To correlate the trends of the Raman peak intensities seen in Figure 3c with the progression through the different phases of the cell cycle, we determined the intensity of the strongest bands as a function of release time, and the phase-related dynamics were assigned according to our flow cytometry results (Figure 4b, e, and h). We found that the cell cycle dynamics recorded from flow cytometry had a time delay of ∼4−5 h when compared to the results of Raman spectroscopy. This delay is most likely due to the fact that flow cytometry requires the analysis of at least 10 000 cells, while Raman spectroscopy follows single-cells. After accounting for the time delay, the

biochemical processes such as DNA replication, chromosome condensation/separation, and protein synthesis, the expression level of the relative proteins and chemical compounds during these phases most likely are also different. Here, the phasedependent Raman spectra during the entire cell cycle, which will be discussed in more detail in the following section, has the potential to offer new insights and better resolution into the molecular mechanisms of the functions and biochemical events in single living cells as a function of cell perturbation, for example, as a result of diseases or by the addition of drugs. Since our group has previously demonstrated that nucleartargeted AuNSs with a concentration of 0.1 nM did not induce cell death, we utilized an even lower concentration of 0.05 nM to avoid cell cycle perturbations. In addition, we compared the cell cycle obtained from flow cytometry with and without 0.05 nM gold nanoparticles. The resulting time-dependence dynamics of each phase, as determined by flow cytometry, are shown in Figure 4a, d, and g. As indicated above, the use of 0.05 nM AuNSs in the present experiments did not not seem to disrupt the cell cycle transitions (Figure S4, Table S2 in the Supporting Information).14 It should also be noted that the spectra obtained from the live single cells differs from that of the spectra obtained from the conjugated AuNS (Figure S5). To ensure G1 synchronization was successful and to demonstrate the relationship between the dynamic changes in the Raman spectra with the progression through the different 5372

dx.doi.org/10.1021/nl3027586 | Nano Lett. 2012, 12, 5369−5375

Nano Letters

Letter

Figure 5. Correlation between cell cycle phase progressions as recorded with flow cytometry (a) and Targeted Plasmonic-Enhanced Single-Cell Rayleigh/Raman Spectroscopy (b). Dashed blue lines indicate Raman peaks associated with the G1 phase, while peaks associated with the S and G2/ M phases are shown by green and orange dashed lines, respectively.

dynamics of 511, 658, 845, and 1322 cm−1 show very similar trends with the G1 phase dynamics obtained from flow cytometry (Figure 4c), and the peaks located at 1541 and 1600 cm−1 show similar trends with the S phase (Figure 4f). The Raman band at 1273 cm−1 assigned to the G2/M phase did not fit the flow cytometry dynamics for this phase as tightly as those for G1 and S phases (Figure 4i). This discrepancy could be due to the shorter time period of the G2/M phase when compared to the G1 and S phases, which could induce a larger relative difference as well as the increased time from synchronization release. Additionally, cytometry follows the changes in DNA, and the 1273 cm−1 band could be an enhanced vibration of non-DNA-coupled molecules. The Cell Cycle as Observed through Its Molecular Vibrations. During the different phases of the cell cycle many different molecules with varying functions are synthesized, changing configurations and relative orientation in the cell in preparation for cell division and cell replication. The average amount of DNA in each phase is different thus using DNA as a marker; flow cytometry is able to determine the percentage of cells in the different phases and to mark the time in the cell cycle at which each phase begins and ends. Figure 5a gives the percentage of cells in the different phases at different times during the cell cycle. Since our results in Figure S4 of the Supporting Information showed that presence of 0.05 nM

AuNSs at the nucleus does not change the time progression of the phases in the cell cycle, one can correlate the changes in the time-resolved vibration ranges of molecules in the nucleus during the cell cycle to the changes in the different phases of the cell cycle. Below, we will first examine the time dependence of the different phases (Figure 5a) and use this and the observed changes in the time-resolved vibration spectrum to assign the different spectral regions of molecules in the nucleus to the different cycle phases. From Figure 5a several important observations should be noted. (1) Initially the percentage of cells in the G1 phase remains relatively constant; however around 12−13 h there is an apparent decrease in conjunction with a sharp increase in the percentage of cells in the S phase. (2) The rise of cells in the S phase begins at 7−8 h, shows an apparent sharp increase at hours 12−13 due to the decrease in the percentage of cells in the G1 phase, and reaches its maximum at 19−20 h. (3) In correspondence with the decreasing of the percentage of cells in the S phase, at hours 17−18 the percentage of cells in the G2/M phase suddenly increases and reaches its maximum at hours 23−24. (4) During hours 24−27 it can be concluded from the dark field images of the cell in Figure 3b that the cell division (M-phase) has occurred resulting in the formation of two daughter cells and beginning of G1 phase dominance. 5373

dx.doi.org/10.1021/nl3027586 | Nano Lett. 2012, 12, 5369−5375

Nano Letters

■

Once the most important observations were identified from cell cycle analysis, we then began to correlate the time dependence of the phases discussed above with the time dependence of the vibration frequency ranges in the enhanced Raman spectrum. Since there is a large overlap between the many vibration bands observed that could not be resolved due to the low dispersion of the spectrometer used, a qualitative comparison was carried out, and the following was inferred by the data obtained. (1) When analyzing the observed bands in the Raman spectra (Figure 4b) from 3 to 7 h an assignment to the G1 phase can be made. (2) At the seventh hour, vibration bands in the frequency ranged between 658 and 844 cm−1 and above 1400 cm−1 begin to appear as cells in the S phase start to increase in number. As discussed earlier, most of the bands between 658 and 844 cm−1 are assigned to DNA vibrations and the bands having frequencies above 1400 cm−1 are assigned to protein vibrations. This is consistent with the biochemical processes that occur during the S-phase. In this phase, the number of chromosome strands double, and each strand is composed of DNA and protein molecules.30 (3) It should be mentioned that the band composition in these two regions mentioned above changes in a complex manner and stops changing in the DNA region but continues to change in the protein region at the 17th hour. This change corresponds to the percentage increase of cells in the G2 phase. (4) During hours 22−24, the band profile in the 1322 and 1500 cm−1 region becomes distinguishable and correlates to cell division (Figure 3b), which occurs in the M phase. (5) Once cell division has occurred, as confirmed by the dark field Rayleigh images (Figure 3b), the spectra become very similar to that of the G1 phase at time 3 h and resulting the daughter cells beginning their progression through a new cell cycle. In conclusion, we have demonstrated that Targeted Plasmonic-Enhanced Single-Cell Rayleigh/Raman Spectroscopy can simultaneously provide structural and molecular information from a single-cell over the complete cell cycle. Although, at this early stage, we are still unable to clearly assign all the collected Raman bands to specific molecular vibrations, this work can, in principle, provide new insights in the studies of cellular functions and processes, such as cell division, singling, cell disease diagnostics, and cell death mechanisms, in single cell studies. Furthermore, in the future, this technique has the potential of developing a more quantitative molecular technique in cytometry. We are currently investigating the possibility of using this technique to distinguish between the vibration spectra of healthy and sick (i.e., cancer) cells throughout the cell cycle, with plasmonic nanoparticles targeted to various cell positions (i.e., cell nucleus, cytoplasm, cell membrane, etc.) and hope to detect the molecular (genetic) signs of the early stages of cancers.

■

Letter

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: 404-894-0292. Fax: 404-894-0294. Present Address †

College of Material Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, P. R. China.

Author Contributions ‡

These authors contributed equally.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors would like to thank Megan A. Mackey for taking TEM micrographs of the synthesized gold nanoparticles. M.A.E. would like to thank the Georgia Institute of Technology.

■

REFERENCES

(1) Hoebe, R. A.; Van Oven, C. H.; Gadella, T. W. J.; Dhonukshe, P. B.; Van Noorden, C. J. F.; Manders, E. M. M. Nat. Biotechnol. 2007, 25, 249. (2) Rieder, C. L.; Khodjakov, A. Science 2003, 300, 91. (3) Stephens, D. J.; Allan, V. J. Science 2003, 300, 82. (4) Tang, D.; Kang, R.; Cheh, C. W.; Livesey, K. M.; Liang, X.; Schapiro, N. E.; Benschop, R.; Sparvero, L. J.; Amoscato, A. A.; Tracey, K. J.; Zeh, H. J.; Lotze, M. T. Oncogene 2010, 29, 5299. (5) Chen, J.; Young, S. M.; Allen, C.; Seeber, A.; Péli-Gulli, M.-P.; Panchaud, N.; Waller, A.; Ursu, O.; Yao, T.; Golden, J. E.; Strouse, J. J.; Carter, M. B.; Kang, H.; Bologa, C. G.; Foutz, T. D.; Edwards, B. S.; Peterson, B. R.; Aubé, J.; Werner-Washburne, M.; Loewith, R. J.; De Virgilio, C.; Sklar, L. A. ACS Chem. Biol. 2012, 7, 715. (6) van Manen, H.-J.; Kraan, Y. M.; Roos, D.; Otto, C. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 10159. (7) Kneipp, K.; Haka, A. S.; Kneipp, H.; Badizadegan, K.; Yoshizawa, N.; Boone, C.; Shafer-Peltier, K. E.; Motz, J. T.; Dasari, R. R.; Feld, M. S. Appl. Spectrosc. 2002, 56, 150. (8) Xie, W.; Wang, L.; Zhang, Y.; Su, L.; Shen, A.; Tan, J.; Hu, J. Bioconjugate Chem. 2009, 20, 768. (9) Li, J. F.; Huang, Y. F.; Ding, Y.; Yang, Z. L.; Li, S. B.; Zhou, X. S.; Fan, F. R.; Zhang, W.; Zhou, Z. Y.; WuDe, Y.; Ren, B.; Wang, Z. L.; Tian, Z. Q. Nature 2010, 464, 392. (10) Aslan, K.; Lakowicz, J. R.; Geddes, C. D. Curr. Opin. Chem. Biol. 2005, 9, 538. (11) El-Sayed, I. H.; Huang, X.; El-Sayed, M. A. Nano Lett. 2005, 5, 829. (12) Austin, L. A.; Kang, B.; Yen, C.-W.; El-Sayed, M. A. Bioconjugate Chem. 2011, 22, 2324. (13) Austin, L. A.; Kang, B.; Yen, C.-W.; El-Sayed, M. A. J. Am. Chem. Soc. 2011, 133, 17594. (14) Kang, B.; Mackey, M. A.; El-Sayed, M. A. J. Am. Chem. Soc. 2010, 132, 1517. (15) Qian, W.; Huang, X.; Kang, B.; El-Sayed, M. A. J. Biomed. Opt. 2010, 15. (16) Xie, Y.; Yin, T.; Wiegraebe, W.; He, X. C.; Miller, D.; Stark, D.; Perko, K.; Alexander, R.; Schwartz, J.; Grindley, J. C.; Park, J.; Haug, J. S.; Wunderlich, J. P.; Li, H.; Zhang, S.; Johnson, T.; Feldman, R. A.; Li, L. Nature 2009, 457, 97. (17) Okada, M.; Smith, N. I.; Palonpon, A. F.; Endo, H.; Kawata, S.; Sodeoka, M.; Fujita, K. Proc. Natl. Acad. Sci. U.S.A. 2012, 109, 28−32. (18) Pliss, A.; Kuzmin, A. N.; Kachynski, A. V.; Prasad, P. N. Biophys. J. 2010, 99, 3483. (19) Swain, R. J.; Jell, G.; Stevens, M. M. J. Cell. Biochem. 2008, 104, 1427. (20) Xie, C.; Dinno, M. A.; Li, Y.-q. Opt. Lett. 2002, 27, 249.

ASSOCIATED CONTENT

S Supporting Information *

Methods, experimental procedures, TEM, UV−vis, and DLS analysis of the synthesized NLS/RGD-AuNS are provided; the Raman spectra of the full cycle from three independent trials over 24 h each to show the degree of the reproducibility of the Raman results; flow cytometry cell cycle histograms; cell cycle phase profiles; cell synchronization analysis; and comparative Raman spectra of NLS/RGD-AuNS and that of the full cycle to show that none of the observed Raman bands are those of the capping material. This material is available free of charge via the Internet at http://pubs.acs.org. 5374

dx.doi.org/10.1021/nl3027586 | Nano Lett. 2012, 12, 5369−5375

Nano Letters

Letter

(21) Kneipp, K.; Kneipp, H.; Kneipp, J. Acc. Chem. Res. 2006, 39, 443. (22) Lui, H.; Zhao, J.; McLean, D. I.; Zeng, H. Cancer Res. 2012, 72, 2491−2500. (23) Oshima, Y.; Takenaka, T.; Sato, H.; Furihata, C. Proc. SPIE 2010, 7560, 756008. (24) Qian, X.; Peng, X.-H.; Ansari, D. O.; Yin-Goen, Q.; Chen, G. Z.; Shin, D. M.; Yang, L.; Young, A. N.; Wang, M. D.; Nie, S. Nat. Biotechnol. 2008, 26, 83. (25) Oyelere, A. K.; Chen, P. C.; Huang, X.; El-Sayed, I. H.; El-Sayed, M. A. Bioconjugate Chem. 2007, 18, 1490. (26) Bonnesen, C.; Nelander, G.-M.; Hansen, B.; Jensen, P.; Krabbe, J.; Jensen, M.; Hegelund, A.; Svendsen, J.; Oleksiewicz, M. Cell. Biol. Toxicol. 2010, 26, 293. (27) Kues, W. A.; Anger, M.; Carnwath, J. W.; Paul, D.; Motlik, J.; Niemann, H. Biol. Reprod. 2000, 62, 412. (28) Zhang, E.; Li, X.; Zhang, S.; Chen, L.; Zheng, X. Biochem. Biophys. Res. Commun. 2005, 333, 1171. (29) Cao, Y. C.; Jin, R.; Mirkin, C. A. Science 2002, 297, 1536. (30) Alberts, B.; Johnson, A.; Lewis, J.; Raff, M.; Roberts, K.; Walter, P. Molecular Biology of the Cell, 4th ed.; Garland Science: New York, 2002.

5375

dx.doi.org/10.1021/nl3027586 | Nano Lett. 2012, 12, 5369−5375