Direct Imaging of Transmembrane Dynamics of Single Nanoparticles

Mar 20, 2014 - with Darkfield Microscopy: Improved Orientation Tracking at Cell. Sidewall ... Herein, by using dual-channel polarization darkfield mic...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Direct Imaging of Transmembrane Dynamics of Single Nanoparticles with Darkfield Microscopy: Improved Orientation Tracking at Cell Sidewall Dong Xu, Yan He,* and Edward S. Yeung State Key Laboratory of Chemo/Biosensing and Chemometrics, College of Chemistry and Chemical Engineering, College of Biology, Hunan University, Changsha, Hunan 410082, P. R. China S Supporting Information *

ABSTRACT: Investigation of the cellular internalization processes of individual nanoparticles (NPs) is of great scientific interest with implications to drug delivery and NP biosafety. Herein, by using dual-channel polarization darkfield microcopy (DFM) and single gold nanorods (AuNRs) as orientation probes, we developed a method that is capable of monitoring AuNR orientation dynamics during its transmembrane process. With annular oblique illumination and a birefringent prism to split AuNR plasmonic scattering into two channels of orthogonal polarizations, the AuNR azimuth and polar angles are obtained from their intensity difference and intensity sum. By placing the focal plane of the microscope objective at the elevated cell sidewall rather than at the flat cell top, interference from cellular background is reduced and the signal-to-noise ratio of AuNR orientation sensing is improved significantly, especially for AuNRs inserting into the membrane at a large out-of-plane angle. As a result, we were able to capture the complete membrane-crossing dynamics of single AuNRs. This powerful method could be utilized to obtain valuable insights on NP endocytosis mechanisms of various cells.

I

into a zero-dimensional point, its translational and rotational degrees of freedom are decoupled; that is, the particle’s rotation does not necessitate its simultaneous lateral or axial displacement. One popular anisotropic nanoprobe for rotational and orientation sensing is the gold nanorod (AuNR), which can be approximated to a nonbleaching oscillating dipole having induced plasmonic scattering emission polarized along its longitudinal axis.20,21 Recently, with differential interference contrast microscopy (DICM) and autocorrelation analysis, Gu and co-workers obtained time-evolution of rotational speeds of functionalized single gold nanorods (AuNRs) on live cell membranes.22 Using either pattern recognition or polarization sensing, Xiao developed several darkfield microscopy (DFM)23,24 and DICM25 based methods that can be used to determine the 3D orientation angles of single AuNRs on the PM surface, but so far, only orientation dynamics of AuNRs before they entered the cell were studied. One overlooked problem in previous studies on PM−AuNR interactions is that the image plane of the sensor chip or the conjugate focal planes of the microscope was always aligned with the top or bottom membrane surface of the biological cell. This convention is needed for 2D or 3D translational tracking of single spherical NPs, whose emissions are unpolarized and uniform at all directions and can always be collected efficiently

nteractions between nanoparticles (NPs) and the plasma membrane (PM) of mammalian cells during their transmembrane endocytosis are fundamentally important to many biological and biomedical processes such as targeted drug delivery,1−4 gene therapy,5,6 and NP cytotoxicity.7 The endocytosis pathways of NPs are very complicated,8,9 involving not only the fluidic lipid bilayer and membrane proteins but also other intracellular components and extracellular structures. To get in-depth understanding of their cellular uptake mechanisms including the possible routes, intermediate states, and special events, it is crucial to monitor NP/PM interactions at the single molecule level.10,11 So far, a number of center-ofmass single particle tracking techniques based on noninvasive optical imaging has been developed,12−18 and a lot of information on the initial contact of NPs with the PM and diffusion of NPs on the PM surface has been obtained. However, the PM thickness is only ∼10 nm and far below the optical diffraction limit. Under a far-field optical microscope, an NP traversing such a shallow distance would appear laterally and axially almost static, making it unreliable to extract the NP transmembrane dynamics from variation of its centroid position. Although it is difficult to monitor the NP translational movement crossing a thin PM, the transmembrane rotational motions of some anisotropic NPs, which are also directly related to dynamic interactions between NPs and biomolecules on or near the PM,19 could be tracked in real time. This is because for a Brownian particle that can not be degenerated © 2014 American Chemical Society

Received: November 14, 2013 Accepted: March 9, 2014 Published: March 20, 2014 3397

dx.doi.org/10.1021/ac403700u | Anal. Chem. 2014, 86, 3397−3404

Analytical Chemistry

Article

the literature29 with slight modifications. In brief, 240 μL of icecold 0.010 M NaBH4 was added under stirring to a 4 mL solution containing 0.1 M CTAB and 0.00025 M HAuCl4. The obtained gold nanoparticle seed solution was stirred for another 2 min and then kept at 28 °C for at least 2 h prior to use. To 40 mL of 0.1 M CTAB solution, 825 μL of 0.02428 M HAuCl4, 1 mL of 0.004 M AgNO3, and 210 μL of 0.1 M AA were added in order. After this solution changed from bright yellow to colorless, 120 μL of the seed solution was added and the mixture was shaken for 20−30 s and then kept at 29 °C undisturbed for 4 h. Subsequently, 3 mL of the freshly prepared AuNR solution was added into a growth solution consisting of 20 mL of 0.1 M CTAB, 480 μL of 0.02428 M HAuCl4, and 185 μL of 0.1 M AA while shaking the bottle rapidly for 20 s. The solution was left undisturbed for another 2 h. The excess reagents were removed via centrifugation at 8,000 rpm followed by washing with DI water twice. For surface modification with MUTAB, 20 μL of 1 mM MUTAB was added into 2 mL of AuNR solution, and the mixture was sonicated for 2 h at room temperature. The modified AuNRs were washed by centrifugation at 8,000 rpm with DI water twice, redispersed in 2 mL of DI water, and stored at 4 °C before use. Characterization of the AuNRs were performed using a UV-1800 spectrometer (SHIMADZU Ltd. Japan), a TEM microscope (JEM1230, JEOL), and a Zetasizer Nano instrument (Malvern Instruments Ltd., U.K.). Optical Imaging. All the imaging experiments were performed using a Nikon 80i upright microscope, which was equipped with a 100 W halogen tungsten lamp, an oil immersion darkfield condenser (NA 1.20−1.43), a 40× plan fluor objective and a CoolSnap HQ2 CCD camera (Roper Scientific). A birefringent prism was placed in the detection light path to separate the scattering from single AuNRs into X and Y orthogonally polarized light. Cell Culture and Live Cell Single Particle Analysis. Cervical cancer HeLa cells were obtained from American Type Culture Collection (ATCC). The cells were cultured on a coverslip placed in a plastic Petri dish and maintained in high DMEM (Dulbecco’s Modified Eagle’s medium with high glucose, GIBCO) supplemented with 1% penicillin−streptomycin (10,000 u/mL, 10,000 μg/mL) (Invitrogen) and 10% fatal bovine serum (GIBCO) at 37 °C, 5% CO2 in a humidified atmosphere. After reaching ∼40% confluency, the coverslip was taken out of the Petri dish and inverted on top of a microscope glass slide with a 100 μL cavity filled with AuNR-containing culture medium. For high SNR imaging, the focal plane was positioned at the sidewall of cells, usually 2−3 μm away from the glass surface, and the rotational and translational motions of AuNRs were recorded at 20 frames/s. Image analysis was performed using ImageJ, MATLab, and Origin.

by an imaging objective. However, the longitudinal plasmonic scattering emission of an oscillating-dipole-like AuNR is polarized, which can only propagate at directions perpendicular to the long axis of the nanorod.20,21 Such directional emission not only forms the foundation of AuNR based orientation sensing but also means the collected signal intensity of the AuNR is orientation dependent. When an AuNR is largely parallel to the image plane, its emission would reach the image sensor efficiently and the obtained signal intensity would be high. When the AuNR is mostly vertical to the image plane, little of its emission would be collected and the observed intensity would be low. Especially in the presence of large background scattering interference from underneath intracellular components, the signal-to-noise-ratio (SNR) of the AuNR as well as the accuracy of its orientation measurement would deteriorate remarkably if the rod is tilting away from the membrane surface, yet it has been demonstrated via transmission electron microscopy (TEM) imaging and computer simulation recently that, to minimize the elastic strain in the plasma membrane, one-dimensional NPs prefer entering cells by tip recognition and rotate to high angles during the membrane invagination process.26−28 In this work, we position the focal plane of the microscope objective at the sidewall of the cell and perform orientation tracking of single AuNRs at the elevated side of the PM surface with dual-channel polarization sensing based DFM. For interaction between a Brownian AuNR in the medium solution and a biological cell of several μm in size, the AuNR’s initial contact with the PM and its subsequent rotational behaviors should be consistent either on the cell-top or at the cell sidewall. However, by creating a large angle between the membrane surface and the image plane, a majority part of the out-of-plane angle information of the AuNR relative to the PM surface is translated into the variation of the azimuth angle of the AuNR relative to the image plane. Furthermore, the background scattering noise at the thin edge of the cell is much reduced. As a result, the SNR of AuNR orientation tracking is greatly improved, especially for AuNRs inserting into the membrane vertically or with a high angle. This allowed us to obtain for the first time detailed 3D orientation dynamics of AuNRs during their entire transmembrane process. In this paper, our attention is focused on the basic principle, optical setup, and sensitivity enhancement of this single particle orientation tracking approach. The methodology that leads to identification of the subtly different dynamic states of the AuNRs during their transmembrane process and detailed descriptions of those dynamic states will be presented in a second paper.

■

EXPERIMENTAL SECTION Chemicals and Materials. Cetyltrimethylammonium bromide (CTAB) was purchased from Sinopharm Chemical (Shanghai, China). Sodium borohydride (99%), silver nitrate, L-ascorbic acid (AA), and (11-Mercaptoundecyl)-N,N,Ntrimethylammonium bromide (MUTAB) were obtained from Sigma-Aldrich. All solutions were prepared with deionized (DI) water purified with a Milli-Q system. The glass slides and coverslips (Coring) were cleaned with piraha solution (H2SO4/ H2O2 = 7:3) and then rinsed extensively with DI water prior to use. (CAUTION: “Piranha” solution reacts violently with organic materials; it must be handled with extreme care.) AuNR Preparation and Modification. The methods used to prepare the AuNR probes are similar to those described in

■

RESULTS AND DISCUSSION AuNR Preparation and Modification. The AuNRs used in the present study were prepared via a two-step seedmediated overgrowth approach. Previous studies have shown that an AuNR with aspect ratio ∼2 is closest to an oscillating dipole for the purpose of orientation sensing.24 However, scattering intensities of AuNRs prepared via the traditional seed-mediated growth method30,31 are not high enough under darkfield illumination. To get brighter AuNRs with large scattering cross sections, we first grew small AuNRs (53 × 16 nm) from 2 nm gold nanoparticle seeds and then used them as templates to prepare the desired fat AuNRs through over3398

dx.doi.org/10.1021/ac403700u | Anal. Chem. 2014, 86, 3397−3404

Analytical Chemistry

Article

Figure 1. (A) UV−vis spectra of the AuNRs before (black) and after (red) overgrowth. (B, C) TEM images of the AuNRs before (B) and after (C) overgrowth; the scale bar is 100 nm.

Figure 2. (A) Schematic diagram for cell sidewall imaging with dual-channel polarization darkfield microscopy. (B, C) Coordinate systems for cell top (B) and cell sidewall (C) observation.

techniques.24,32−35 We have demonstrated previously that, by using a birefringent prism to split the plasmonic scattering of an AuNR into two spots with polarization directions parallel and perpendicular to the optical axis of the prism, the azimuth and polar angle of the AuNR can be determined with a number of illumination schemes.36 In brief, when observed through a farfield optical microscope, the intensity of the polarized emission from an AuNR that is captured by the image detector can be described by Is ∝ K × Pa × Ps, where K is a system constant and Pa and Ps represent the excitation and detection probability function of the dipole, respectively.37 In the presence of the birefringent prism, Ps is split into two orthogonally polarized components, Psx and Psy, which can be further simplified to Psx ∝ sin2θ cos2ϕ and Psy ∝ sin2θ sin2ϕ when a low NA objective (40×, NA = 0.75) is utilized for light collection. Here, θ is the polar angle and ϕ is the azimuth angle of the AuNR. Accordingly, ϕ can be obtained from the polarization anisotropy P = (Isx − Isy)/(Isx + Isy) ≈ cos(2ϕ) regardless of the expression of Pa. To determine θ from sum intensity Is = Isx + Isy ∝ Pa sin2θ, the mathematical expression of Pa as a function of the angle

growth. The resulting AuNRs have a plasmonic spectral maximum at ∼650 nm with a mean width of 43 ± 6 nm and a mean length of 84 ± 9 nm (Figure 1) and are ∼37 times brighter than AuNRs of the same aspect ratio prepared using the conventional method. To ensure biocompatibility of the AuNRs, the CTAB layer covered on the AuNR surface is replaced with a thiolated cationic ligand, MUTAB. Zeta potential of the modified AuNRs is +28 mV but becomes −10 mV after adsorption of serum proteins. Further assays indicate that the obtained AuNRs are monodisperse, do not aggregate in the cell culture medium, and have no observable cytotoxicity within at least 12 h. They can be readily internalized by the cells (Figure S1, Supporting Information). A large amount of the AuNRs entered HeLa cells after just 0.5 h of coincubation at 37 °C. Therefore, these bright and biocompatible AuNRs are qualified as single particle orientation sensing probes. Determine AuNR Orientation with Traditional Darkfield Microscopy. Figure 2A shows the optical setup used in this study. For an AuNR treated as an oscillating dipole, its 2D or 3D orientation angles can be obtained with various 3399

dx.doi.org/10.1021/ac403700u | Anal. Chem. 2014, 86, 3397−3404

Analytical Chemistry

Article

Figure 3. Focused cell images (A, B) and calculated errors (C, D) of the azimuth angle (black) and the polar angle (red) under cell top (A, C) and cell sidewall (B, D) observation mode. The derivations of equations for error determination are provided in the Supporting Information.

θ, but in the case of cell sidewall imaging or when there is a large angle between the image plane and the PM, a matrix rotation has to be applied to obtain the values of α and β, as long as the elevation angle of the PM is known. In particular, if the PM surface is perpendicular to x-axis of the image plane, the ideal case of cell sidewall imaging α and β can be determined from sin2β = sin2θcos2ϕ and cos2βcos2α = sin2θsin2ϕ (Figure 2C). Improved AuNR Orientation Sensing at Cell Sidewall. With the availability of bright AuNR probes and the capability to retrieve their 3D orientation angles using DFM, whether or not single AuNR rotations on the PM surface can be tracked reliably is largely determined by the signal-to-noise ratio of the imaging system. Figure 3A shows an image of the top surface of a HeLa cell with adsorbed AuNRs. Since the cell top is the thickest region of the cell, Rayleigh scattering from intracellular structures is pretty strong and could cause serious interference to plasmonic scattering detection, resulting in the AuNRs appearing as blurred, barely discernible spots embedded in a huge heterogeneous background. In addition, because the background signal is everywhere and there is no obvious marker for the axial position of the PM, convincing criteria are lacking to identify whether the AuNRs are on or below the membrane surface. On the other hand, when the image was taken at the sidewall or the thin edge of the cell (Figure 3B), scattering from intracellular components is all contained within the cell boundary, which clearly defines the strike line or the PM position in the image plane. Single AuNRs approaching from the solution can be clearly visualized with high contrast on a

between the dipole axis and the direction of the illumination light must be known. Depending on the illumination scheme, the excitation probability function Pa varies. In particular, Pa equal to one could be achieved with orthogonal planar illumination.36 However, that scheme requires careful collimation and alignment of an external white light source and may not be readily implemented in most laboratorial settings. In the present study, we derived the expression of Pa for conventional darkfield microscopy, which is characterized by annular oblique illumination (see the Supporting Information).38 It turned out that this seemingly complex scheme can be decomposed into a mixture of unpolarized incident light with azimuth angles ranging from 0 to 360° and polar angles ranging from 64.5° to 90° (for oil-type darkfield condenser of NA 1.20−1.43 at glass/ water interface of refractive index n1/n2= 1.52/1.33). From the expression of Pa for unpolarized incident light with arbitrary incident angle, the time-averaged Pa for annular oblique illumination can be obtained by integration over the ranges of the incident angles. The result is Pa ∝ 0.936 − 0.404sin2θ, a function that only depends on θ. Consequently, the θ value of the AuNR can be obtained from the sum intensity of the two channels by solving the equation Isx + Isy ∝ (0.936 − 0.404sin2θ)sin2θ. Note that the obtained azimuth angle ϕ and the polar angle θ are defined relative to the image plane. For tracking rotational motions of single AuNRs on the fluidic PM, it is more intuitive to translate them into the in-plane angle α and out-of-plane angle β with respect to the PM surface. When the image plane is parallel to the PM (Figure 2B), we have α = ϕ and β = 90° − 3400

dx.doi.org/10.1021/ac403700u | Anal. Chem. 2014, 86, 3397−3404

Analytical Chemistry

Article

Figure 4. Simulated time-dependent variations of the polarization anisotropy (blue) and the sum intensity (green) of two AuNRs with large β angles under the cell top (A, C, E) and the cell sidewall (B, D, F) observation mode. (A, B) Fluctuation patterns of the β angle. (C, D) Simulated results with β values fluctuating in the range of 75−90°. (E, F) Results with β values fluctuating in the range of 60−90°.

both the cell top and the cell sidewall mode. The two AuNRs are standing on the PM surface with their β values fluctuating in the ranges of 75−90° and 60−90°, respectively. It can be seen that, under the cell top mode, these two dynamic states can not be separated because the observed sum intensities of both AuNRs are close to the noise level, but under the cell sidewall imaging mode, the two states are clearly differentiated because the acquired sum intensities are much higher than the noise. Moreover, since the out-of-plane angle is now a convolution of the azimuth angle and the polar angle, the variation of β values is exhibited not only in the sum intensity channel but also in the anisotropy channel, making it easier to identify different orientation states of the AuNRs. Monitor AuNR Orientation Variation on Cell Membrane. To demonstrate the enhanced performance of single particle orientation sensing at cell sidewall with DFM, we monitored interactions between single AuNRs and HeLa cells in real time. For the experiments, cells attached on coverglass surface were mixed with AuNR-containing culture medium under the microscope, and short movies were recorded at 20 frames/s. Due to the high viscosity of the PM and its interaction with the particles, the apparent rotational diffusion coefficients of the AuNRs on the membrane were found to be no more than several Hz (data not shown), so the temporal

low solution background, and their positions relative to the membrane can be readily determined. Indeed, the noise level observed at the cell top is at least 3 times higher than at the cell sidewall. As a result, the errors of polar and azimuth angles determined at the cell sidewall are much smaller than at the cell top (Figure 3C,D). Besides much reduced background noise, observation at cell sidewall also improved the signal intensity significantly if an AuNR is standing on the PM tip-first with a large β close to 90°. This is often the case if the AuNR is negatively charged as the PM and their initial contact area is minimized due to electrostatic repulsion. If the focal plane is positioned at the top surface of the cell, the observed AuNR polar angle θ = 90° − β would be quite small, and the obtained signal intensity would be pretty low according to equations derived above, making it difficult to distinguish variation of the β value from background noise fluctuations, but if the focal plane is put at the sidewall of the cell (assuming the membrane surface is perpendicular to xaxis of the image plane), an AuNR with β ∼ 90° would appear to be lying flat within the focal plane with both θ ∼ 90° and ϕ ∼ 0° (sin2β = sin2θcos2ϕ), leading to high signal intensity. Figure 4 showed simulated time-dependent variations of the polarization anisotropy (Isx − Isy)/(Isx + Isy) and the sum intensity Isx + Isy of two AuNRs with large β observed under 3401

dx.doi.org/10.1021/ac403700u | Anal. Chem. 2014, 86, 3397−3404

Analytical Chemistry

Article

dyes and perform optical slicing and 3D reconstruction of the cell profile,42 but that procedure is complicated and timeconsuming and could not account for local topological variation of the PM at the attachment sites of different AuNRs. In order to bypass this problem, we chose the membrane confined state of an AuNR, characterized by it being rotationally and laterally static at a PM site, as its own “internal reference”, and assumed the nanorod is perpendicular to the PM surface. For the serumprotein-coated, negatively charged AuNRs, this assumption should be valid in that, when the rotational motion of an AuNR on the cell membrane is strongly confined, it has to be firmly inserted into the cell membrane; to minimize its global energy, the most logical scenario is that the AuNR takes a 90° angle to the membrane surface. This hypothesis has been proved by TEM results and computer simulations from others.26−28 In that case, η and γ can be, respectively, determined from the time-average of the θ and ϕ values of the AuNR at its “insertion state”. For the nanorod experiencing the strongly confined situation in the second video of Movie 2, “ac403700u_si_003.avi” in the Supporting Information, the values of η and γ are calculated to be 69.3° and 71.1°, respectively. The latter is nearly the same as the γ value measured independently earlier, proving that the angle between the projection of the AuNR on the image plane and the cell strike line, βprj, is nearly 90°. Statistical analysis of 64 such strongly confined AuNRs picked from different cell images results in a histogram of βprj with only one major population narrowly distributed close to 90° (Figure S3, Supporting Information). Since the AuNRs should be randomly dispersed on the PM surface, the missing broad distribution of βprj is solid proof that the majority of those strongly confined AuNRs are perpendicular not only to the cell strike line but also to the membrane surface. Therefore, β could be obtained from sin2β = sin2θ′cos2ϕ′, where ϕ′ = ϕ − γ and θ′ = θ + 90° − η are the corrected azimuth angle and polar angle, respectively. As an example, Figure 5 shows the recovered intensities and orientation angles and the lateral displacement of centroid position of the AuNR in the third video of Movie 2, “ac403700u_si_003.avi” in the Supporting Information, which captured the exact moment of the nanorod detaching from the inner surface of the PM. It can be seen that, prior to 5.0 s, the intensities in the two channels as well as θ and ϕ of the AuNR did not alter much, the average value of β is 74.7 ± 7.2°, and the centroid position of the AuNR kept nearly unchanged, indicative the nanorod taking a large angle against the PM and still being confined by the membrane. Then, from 5.0 to 5.9 s and from 6.7 to 9.4 s, the AuNR twice swung toward the membrane surface, reaching β = 55.4 ± 3.2° and β = 39.5 ± 3.0°, respectively. This is apparently an intermediate stage where the AuNR began breaking off from the PM, which would not have been observed had only the translational motions of the AuNRs been tracked. Finally after 10.1 s, the AuNR started to rotate wildly and moved away from its membrane position via directional movement and appeared deep inside the cell doing active transportation in the next video. To our knowledge, this is the first report on direct observation of the transmembrane event of a nanoparticle approaching from the open solution. Further studies on other AuNRs and detailed analysis on their complicated transmembrane processes will be presented in another paper.

resolution is good enough to track their orientation variations. One example is shown in Movie 1, “ac403700u_si_002.avi” in the Supporting Information. It can be seen that a lot of AuNRs, appearing as close-by two-spots, are either static or flickering at various frequencies. According to the position of the cell boundary defined by the separation between the bright intracellular and dark extracellular background, whether the AuNRs are outside, on/near the membrane surface, or inside the cell can be readily discerned. Thanks to cell sidewall observation, such spatial and temporal-spatial differentiation of the AuNRs was accomplished within the same image plane. Moreover, since at least a half of the out-of-focus solution interference is now buried into the intracellular background, we can afford to use a relatively high nanorod concentration to facilitate simultaneous tracking of multiple AuNRs. In contrast, under the cell top observation mode, Rayleigh scattering backgrounds from both the heterogeneous intracellular components and the floating particles in the pericellular environment could cause strong interference in identifying the laterally static AuNRs on the PM surface and tracking their rotational motions.39 As a result, with cell sidewall imaging and no back-and-forth focal plane adjustment, we can statistically compare AuNR dynamic behaviors under different local environments and capture AuNR transmembrane events via consecutive imaging of multiple AuNRs in parallel for a long enough time. Movie 2, “ac403700u_si_003.avi” in the Supporting Information, presents four 20 s video clips of a membranecrossing AuNR taken at 4, 7, 9, and 10 min after adding the AuNR solution into the cell culture medium. It can be seen that, apparently located on the cell surface, the flickering frequency of the AuNR was fast at t = 4 min, reduced to almost zero at t = 7 min, and became fast again at t = 9 min. Before the early part of the third video, the lateral position of the AuNR remained largely unchanged, but at t = 10 min, it was doing active transportation toward inside the cell, indicating that the AuNR had penetrated into the plasma membrane. Therefore, the first 3 videos apparently correspond to the AuNR being attached to, strongly confined by, and detached from the cell membrane, respectively. Benefited from the high AuNR signal and low background, the intensities of the AuNR in every frame can be obtained by using a rolling-cycle background subtraction algorithm,40,41 and the time-dependent variations of azimuth and polar angles of the AuNR relative to the image plane can be calculated by using the equations derived earlier. The image analysis results of the AuNR in Movie 2, “ac403700u_si_003.avi” in the Supporting Information, are shown in Figure S2, Supporting Information. To obtain the out-of-plane angle information of the AuNRs relative to the PM surface during their transmembrane process, however, it is necessary to determine the orientation of the cell membrane plane with respect to the image plane. Since the connection line between the center of the two spots of each AuNR defined the y-axis and the outer boundary of the intercellular scattering defined the strike line or the intersection line between the image plane and the membrane plane, we could directly measure the angle from the cell strike line to the y-axis at the attachment site of the AuNR (denoted as γ). In the case of Movie 2, “ac403700u_si_003.avi” in the Supporting Information, for example, the γ value is found to be ∼70°. However, there is no simple way to determine the elevation angle of the membrane surface (denoted as η). One utilizable method is to stain the cell membrane with certain fluorescence 3402

dx.doi.org/10.1021/ac403700u | Anal. Chem. 2014, 86, 3397−3404

Analytical Chemistry

■

Article

ASSOCIATED CONTENT

S Supporting Information *

Additional notes and figures as mentioned in the main text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by NSFC 21127009, NSFC 91027037, NSFC 21221003, and Hunan University 985 fund.

■

REFERENCES

(1) Davis, M. E.; Chen, Z.; Shin, D. M. Nat. Rev. Drug Discovery 2008, 7, 771. (2) Slowing, I.; Viveroescoto, J.; Wu, C.; Lin, V. Adv. Drug Delivery Rev. 2008, 60, 1278. (3) Wang, A. Z.; Langer, R.; Farokhzad, O. C. Annu. Rev. Med. 2012, 63, 185. (4) Riveros, A.; Dadlani, K.; Sales, E.; Caballero, L.; Melo, F.; Kogan, M. J. J. Biomater. Tissue Eng. 2013, 3, 4. (5) Rosi, N. L.; Giljohann, D. A.; Thaxton, C. S.; Lytton-Jean, A. K. R.; Han, M. S.; Mirkin, C. A. Science 2006, 312, 1027. (6) Salem, A. K.; Searson, P. C.; Leong, K. W. Nat. Mater. 2003, 2, 668. (7) Leroueil, P. R.; Hong, S.; Mecke, A.; Baker, J. R., Jr.; Orr, B. G.; Holl, M. M. B. Acc. Chem. Res. 2007, 40, 335. (8) Doherty, G. J.; McMahon, H. T. Annu. Rev. Biochem. 2009, 78, 857. (9) Marguet, D.; Lenne, P.-F.; Rigneault, H.; He, H.-T. EMBO J. 2006, 25, 3446. (10) Yang, Y. H.; Nam, J. M. Anal. Chem. 2009, 81, 2564. (11) Chen, L. Q.; Xiao, S. J.; Hu, P. P.; Peng, L.; Ma, J.; Luo, L. F.; Li, Y. F.; Huang, C. Z. Anal. Chem. 2012, 84, 3099. (12) Jaqaman, K.; Loerke, D.; Mettlen, M.; Kuwata, H.; Grinstein, S.; Schmid, S. L.; Danuser, G. Nat. Methods 2008, 5, 695. (13) Saxton, M. J.; Jacobson, K. Annu. Rev. Biophys. Biomol. Struct. 1997, 26, 373. (14) Serge, A.; Bertaux, N.; Rigneault, H.; Marguet, D. Nat. Methods 2008, 5, 687. (15) Toprak, E.; Balci, H.; Blehm, B. H.; Selvin, P. R. Nano Lett. 2007, 7, 2043. (16) Wawrezinieck, L.; Rigneault, H.; Marguet, D.; Lenne, P.-F. Biophys. J. 2005, 89, 4029. (17) Wells, N. P.; Lessard, G. A.; Werner, J. H. Anal. Chem. 2008, 80, 9830. (18) Gu, Y.; Di, X. W.; Sun, W.; Wang, G. F.; Fang, N. Anal. Chem. 2012, 84, 4111. (19) Rong, G. X.; Wang, H. Y.; Reinhard, B. M. Nano Lett. 2010, 10, 230. (20) Sonnichsen, C.; Franzl, T.; Wilk, T.; von Plessen, G.; Feldmann, J.; Wilson, O.; Mulvaney, P. Phys. Rev. Lett. 2002, 88, 077402. (21) Sonnichsen, C.; Alivisatos, A. P. Nano Lett. 2005, 5, 301. (22) Wang, G.; Sun, W.; Luo, Y.; Fang, N. J. Am. Chem. Soc. 2010, 132, 16417. (23) Xiao, L. H.; Wei, L.; Liu, C.; He, Y.; Yeung, E. S. Angew. Chem., Int. Ed. 2012, 51, 4181. (24) Xiao, L.; Qiao, Y.; He, Y.; Yeung, E. S. Anal. Chem. 2010, 82, 5268. (25) Xiao, L.; Ha, J. W.; Wei, L.; Wang, G.; Fang, N. Angew. Chem., Int. Ed. 2012, 51, 7734.

Figure 5. Time-dependent variations of (A) two-spot intensities, (B) azimuth and polar angles, (C) out-of-plane angle, and (D) lateral displacement of the AuNR experiencing a membrane-crossing event in the t = 9 min video in Movie 2, “ac403700u_si_003.avi” in the Supporting Information. The dashed line refers to the time point when the AuNR detached from the inner surface of the cell membrane.

■

CONCLUSIONS We have established a simple but robust method to determine 3D orientation angles of individual AuNRs in real time by using dual-channel orthogonal polarization detection with conventional darkfield microscopy. With annular oblique illumination, a low NA objective, and a birefringent prism to split the plasmonic scattering from an AuNR into two spots of orthogonal polarizations, the azimuth and polar angles of the AuNR can be calculated from the relative intensity difference and the intensity sum of the two spots. By putting the image plane of the microscope at the sidewall of the cell, the detectable signal intensity of AuNRs with high angle against the membrane surface is much enhanced, the scattering interference from intracellular components is greatly reduced, the boundary of the cell can be readily defined, and the complete membrane-crossing process of an AuNR can be followed without changing the focal plane of the objective. By using the “insertion state” where the AuNR is strongly confined by the membrane as the “internal reference”, the out-of-plane angle of the AuNR on cell membrane can be determined. Consequently, we are able to obtain the time-dependent orientation variations of single AuNRs during their transmembrane endocytosis process for the first time. This improved single particle orientation tracking method could be exploited to acquire indepth knowledge on complicated interactions between nanoparticles and cell membranes for biological and biomedical studies. 3403

dx.doi.org/10.1021/ac403700u | Anal. Chem. 2014, 86, 3397−3404

Analytical Chemistry

Article

(26) Cureton, D. K.; Massol, R. H.; Saffarian, S.; Kirchhausen, T. L.; Whelan, S. P. J. PLoS Pathog. 2009, 5, DOI: 10.1371/journal.ppat.1000394. (27) Yang, K.; Ma, Y. Q. Nat. Nanotechnol. 2010, 5, 579. (28) Shi, X.; von dem Bussche, A.; Hurt, R. H.; Kane, A. B.; Gao, H. Nat. Nanotechnol. 2011, 6, 714. (29) Ni, W.; Kou, X.; Yang, Z.; Wang, J. F. ACS Nano 2008, 2, 677. (30) Nikoobakht, B.; El-Sayed, M. A. Chem. Mater. 2003, 15, 1957. (31) Sau, T. K.; Murphy, C. J. Langmuir 2004, 20, 6414. (32) Chang, W. S.; Ha, J. W.; Slaughter, L. S.; Link, S. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 2781. (33) Failla, A. V.; Qian, H.; Qian, H.; Hartschuh, A.; Meixner, A. J. Nano Lett. 2006, 6, 1374. (34) Gu, Y.; Sun, W.; Wang, G.; Fang, N. J. Am. Chem. Soc. 2011, 133, 5720. (35) Li, T.; Li, Q.; Xu, Y.; Chen, X. J.; Dai, Q. F.; Liu, H. Y.; Lan, S.; Tie, S. L.; Wu, L. J. ACS Nano 2012, 6, 1268. (36) Xiao, L. H.; Qiao, Y. X.; He, Y.; Yeung, E. S. J. Am. Chem. Soc. 2011, 133, 10638. (37) Forkey, J. N.; Quinlan, M. E.; Goldman, Y. E. Prog. Biophys. Mol. Biol. 2000, 74, 1. (38) Yang, H. J. Phys. Chem. A 2007, 111, 4987. (39) Zhou, R.; Xiong, B.; He, Y.; Yeung, E. S. Anal. Bioanal. Chem. 2011, 399, 353. (40) He, Y.; Li, H. W.; Yeung, E. S. J. Phys. Chem. B 2005, 109, 8820. (41) Fang, N.; Zhang, H.; Li, J. W.; Li, H. W.; Yeung, E. S. Anal. Chem. 2007, 79, 6047. (42) Picone, R.; Ren, X. Y.; Ivanovitch, K. D.; Clarke, J. D. W.; McKendry, R. A.; Baum, B. PLoS Biol. 2010, DOI: 10.1371/ journal.pbio.1000542.

3404

dx.doi.org/10.1021/ac403700u | Anal. Chem. 2014, 86, 3397−3404