Monofunctional Stealth Nanoparticle for Unbiased Single Molecule

Mar 17, 2014 - Mapping Intracellular Diffusion Distribution Using Single Quantum Dot Tracking: Compartmentalized Diffusion Defined by Endoplasmic Reti...
0 downloads 9 Views 6MB Size
Download PDF
Letter pubs.acs.org/NanoLett

Monofunctional Stealth Nanoparticle for Unbiased Single Molecule Tracking Inside Living Cells Domenik Liße,† Christian P. Richter,† Christoph Drees,† Oliver Birkholz,† Changjiang You,† Enrico Rampazzo,‡ and Jacob Piehler*,† †

Department of Biology, University of Osnabrück, Osnabrück, Germany Department of Chemistry ‘‘G. Ciamician’’, University of Bologna, Bologna, Italy

‡

S Supporting Information *

ABSTRACT: On the basis of a protein cage scaffold, we have systematically explored intracellular application of nanoparticles for single molecule studies and discovered that recognition by the autophagy machinery plays a key role for rapid metabolism in the cytosol. Intracellular stealth nanoparticles were achieved by heavy surface PEGylation. By combination with a generic approach for nanoparticle monofunctionalization, efficient labeling of intracellular proteins with high fidelity was accomplished, allowing unbiased long-term tracking of proteins in the outer mitochondrial membrane. KEYWORDS: Stealth nanoparticle, cytosol, autophagy, mitochondria, monofunctionalization, single particle tracking

L

microinjected FNPs, rapid clustering in the cytosol has been observed,11,16−18 pointing toward recognition by an active metabolic machinery. Here, we aimed to systematically engineer inert FNPs to enable unbiased protein tracking on intracellular membranes with ultrahigh spatial and temporal resolution. As a scaffold, we chose the natural protein cage formed by the ferritin light chain (LCF), which by itself provides intrinsic biocompatibility by its proteinaceous surface. Moreover, these very stable and fully monodisperse protein cages offer versatile means for welldefined chemical and genetic modification of the core and surface properties, as well as monofunctionalization with the clickHTL to ensure labeling in a 1:1 stoichiometry. By systematically engineering the functional properties of these protein cages, we developed a monofunctionalized model nanoparticle exhibiting “stealth” properties within the cytosol as required for unbiased live cell single molecule studies (Figure 1). Recombinant LCF was expressed in E. coli, purified to homogeneity,19 and labeled with Cy3 maleimide via cysteine residues fused to the N-terminus of each LCF subunit (Cy3LCF; degree of labeling (DOL): 3). Both, LCF and Cy3LCF formed stable and monodisperse protein cages as confirmed by dynamic light scattering and analytical size exclusion chromatography (Supporting Information Figure S1). Moreover, its protein repelling properties were confirmed by

ocalization and tracking of individual proteins provide powerful means for unravelling complex processes that would otherwise be averaged out by ensemble measurements.1 Detecting and tracking single molecules in living cells, however, requires labeling with bright and photostable probes.2 Fluorescent nanoparticles (FNP) such as quantum dots or dye-doped nanoparticles enable long lasting imaging with ultrahigh precision and have been successfully applied for tracking individual proteins in the plasma membrane of living cells.3−8 Biological applications moreover require highly specific FNP attachment to target proteins in a defined stoichiometry without affecting its functional integrity. While a broad spectrum of modification and targeting strategies was made available for extracellular labeling with FNPs,9 these approaches mostly proved insufficient for intracellular application, which therefore remained challenging.10,11 Recently, we developed a generic approach for efficient site-specific targeting of FNPs to HaloTag12 fusion proteins in the cytoplasm of living cells, which is based on an engineered HaloTag ligand termed clickHTL.13,14 By employing this approach for tracking of single molecules in the outer mitochondrial membrane, however, we observed highly restricted mobility of individual FNP-labeled proteins not seen in case of labeling with organic dyes.15 This strong effect of intracellular FNP labeling on protein mobility, which has not been observed for cell surface labeling, could be ascribed to unspecific attractive or steric repulsive interactions with structures or compounds of the cytosolic environment. Moreover, rapid binding to target proteins was required for successful protein labeling in the cytosol of living cells, suggesting that competing interactions with cellular components play a critical role. Indeed, even for © 2014 American Chemical Society

Received: February 18, 2014 Revised: March 15, 2014 Published: March 17, 2014 2189

dx.doi.org/10.1021/nl500637a | Nano Lett. 2014, 14, 2189−2195

Nano Letters

Letter

Figure 1. Strategy of LCF surface modification and functionalization. Fluorescence dyes (here: ATTO 647N highlighted in purple) and click HTL (red) conjugated to an (EEG)3 carrier peptide (blue) were coupled to cysteine residues engineered into the N-terminus of each subunit. PEG (orange) was coupled via surface lysine residues.

Figure 2. Metabolism of LCF and engineered LCF inside living cells probed by epifluorescence microscopy. (A) Rapid aggregation of Cy3LCF injected into HeLa cells. (B) Co-localization of Cy3LCF (red channel) with transiently expressed mEGFP-LC3B (green channel). (C) Apparent degree of degradation (DOD) after 1, 5, and 16 h (black, Cy3LCF; red, Cy3LCFCOOH; green, Cy3LCFPEG750; blue, Cy3LCFPEG2k). (D) HeLa cells 16 h after microinjection of modified LCF (top, Cy3LCFCOOH; middle, Cy3LCFPEG750; bottom, Cy3LCFPEG2k). Scale bar: 5 μm in all images. Cy3

LCF immediately aggregated (Figure 2A), as indicated by the formation of bright dots on a time scale of seconds. Taking the very high colloidal stability of LCF particles into account,

monitoring the interaction of bovine serum albumin (BSA) with immobilized LCF in real time (Supporting Information Figure S2). After microinjection into HeLa cells, however, 2190

dx.doi.org/10.1021/nl500637a | Nano Lett. 2014, 14, 2189−2195

Nano Letters

Letter

microinjection small, isolated aggregates could be observed for Cy3LCF grafted with PEG750, indicating that the short PEG is still not sufficient to fully protect the particle surface. Further stabilization was achieved by grafting PEG2k, as no significant particle aggregation could be observed up to 16 h. These results provide clear evidence that PEGylation of NP surfaces at high densities is required to protect cytosolic NPs against autophagy. Because the recruitment of nanoparticles into autophagosomes is very likely mediated by interaction with cytosolic proteins, we characterized the intracellular diffusion properties of LCF before and after PEGylation by fluorescence correlation spectroscopy (FCS). For this purpose, unmodified LCF or LCFPEG2k labeled with DY-647 (DY647LCF and DY647LCFPEG2k, respectively) were microinjected at low concentration into 3T3 fibroblasts. Diffusion constants were determined from the autocorrelation curves obtained within cells and compared with the diffusion constants obtained in buffer solution (Supporting Information Figure S8). The diffusion constant of 13 μm2/s obtained for DY647LCFPEG2k in the cytosol was 5-fold lower compared to the diffusion constant of LCFPEG2k in buffer solution, a ratio that is in excellent agreement with systematic studies on intracellular diffusion carried out with indifferent polymers.27−29 In contrast, despite its lower hydrodynamic radius and similar diffusion properties determined in vitro, substantially slower diffusion of DY647LCF compared to DY647 LCFPEG2k was measured in the cytosol (Supporting Information Figure S8). The ∼10-fold decreased mobility of DY647 LCF was very likely caused by the formation of a protein corona,30 which is responsible for autophagosomal targeting. These results highlight that the very dense nanoparticle surface coating provided by PEG2k is required to achieve true stealth properties with respect to cytosolic metabolism and diffusion. Labeling of proteins with nanoparticles carrying multiple targeting moieties leads to protein cross-linking and thus to reduced mobility as well as other functional biases.13 In order to ensure specific conjugation to target proteins inside the cell in a 1:1 ratio, we therefore established monofunctionalization of AT647N LCFPEG2k for efficient targeting via the HaloTag. Both, the HaloTag and the SNAP-tag have been successfully applied for site-specifically conjugating nanoparticles with target proteins, yet high concentrations and coupling times are required due to the relatively slow reaction with their substrates.31,32 For intracellular nanoparticle targeting, these reactivities are not suitable,13 as protein concentrations are limited by the expression level in the cell and washout of nonreacted particles in not possible. We therefore further improved the reactivity of the clickHTL toward the HaloTag by an optimized linker (Supporting Information Scheme S1 and S2). The binding kinetics was probed by simultaneous total internal reflection fluorescence spectroscopy and reflectance interference detection (TIRFS/RIf) as described previously.13,33 A reaction rate constant of 105 M−1s−1 for the improved clickHTL (compound 3c) was obtained, which is about 3 times faster than for the nonoptimized one (Supporting Information Figure S9). Monofunctionalization of LCF with the improved clickHTL was implemented following a strategy initially developed for fractionation of gold nanoparticles functionalized with oligonucleotides.34 For this purpose, we employed a negatively charged carrier peptide for coupling clickHTL, thus providing means for separating LCF with different number of functional groups by anion exchange chromatography (AEX). A peptide consisting of three repeats

we suspected cellular metabolization to be responsible for this effect. Indeed, we found that Cy3LCF aggregates specifically colocalized with mEGFP-tagged LC3B (mEGFP::LC3B), a marker for autophagosomes (Figure 2B),20 indicating rapid recognition by the autophagy machinery. Importantly, the same behavior was also observed for the ferritin analog Dps (DNA protection during starvation protein; 9 nm in diameter; DOL: 3) from the Gram-positive bacterium Listeria innocua,21 which in its amino acid sequence is unrelated to human ferritin, suggesting generic recognition of intracellular nanoparticles by the autophagy machinery (Supporting Information Fμigure S3).22 To test this hypothesis, we microinjected Rhodamine B doped, core−shell PEG/silica nanoparticles (SiNPs; 12 ± 2 nm in diameter; ζ-potential: −4.3 ± 0.3 nm; Supporting Information Figure S4),23 into the same cell line, yielding particle aggregates colocalized with mEGFP::LC3B (Supporting Information Figure S3). However, the kinetics of SiNP degradation were substantially lower with respect to Cy3LCF or Cy3 Dps, suggesting that nanoparticle PEGylation interferes with recognition by the autophagy machinery. In contrast, a homogeneous distribution of purified EGFP or Cy3 labeled BSA (DOL: 1) was observed 1 h after microinjection into HeLa cells (Supporting Information Figure S5), corroborating a selective autophagy response toward nanoparticles rather than a stress-induced response caused by the microinjection procedure. In order to minimize LCF recognition by the autophagy machinery, we tested different surface modification strategies. Since the particles surface charge plays an important role for colloidal NP stability,24 the moderately negative ζ-potential of −2.2 ± 0.2 mV for the unmodified LCF was decreased by converting surface-exposed amines into carboxyl groups by reaction with succinic anhydride. Moreover, the viscoelastic surface properties of LCF were modified by attaching poly(ethylene glycol) (PEG) via lysine residues using NHS chemistry. As the surface coverage with PEGs also depends on its chain length,25,26 we tested PEG750 with an average of 13 ethylene glycol units, as well as PEG2k with an average of 42 ethylene glycol units. A decreased ζ-potential for both the carboxylated (−7.5 ± 0.1 mV) and the PEGylated (PEG750, −6.1 ± 0.3 mV; PEG2k, −6.5 ± 0.4 mV) LCF cage was obtained (Supporting Information Figure S1). The integrity of the modified LCFs was confirmed by analytical size exclusion chromatography (SEC) and dynamic light scattering, confirming a modification-dependent size increase of the LCF (Supporting Information Figure S1). Mass spectrometry of LCFPEG2k (Supporting Information Figure S6) revealed 8 PEG chains per subunit (altogether 192 chains/FNP; on average 1 chain/2.3 nm2). This result matches very well with the theoretical number of accessible amines observed in the LCF crystal structure (PDB: 3HX7). Autophagy of native and modified Cy3LCF was compared by microinjection into the cytoplasm of HeLa cells (Figure 2C,D and Supporting Information Figure S7). While the majority of nonmodified Cy3LCF was fully aggregated after 1 h and degraded after 16 h, a slightly increased persistency of carboxylated Cy3LCF was observed, indicating a decreased recruitment into autophagosomes, probably due to electrostatic repulsion. Yet, carboxylated Cy3LCF clearly aggregated between 1 to 5 h and was almost fully degraded 16 h after microinjection. In contrast, surface PEGylation substantially increased the stability of Cy3LCF and the majority of particles remained stable for at least 5 h. However, 16 h after 2191

dx.doi.org/10.1021/nl500637a | Nano Lett. 2014, 14, 2189−2195

Nano Letters

Letter

Figure 3. Purification and characterization of mono-clickHTL-functionalized LCF (mLCF). (A) AEX fractionation of LCF functionalized with clickHTL by means of a negatively charged carrier peptide (black, eluted protein fractions monitored at λ = 280 nm; blue, Gaussian fit for the quantification of fraction frequencies). (B) Binding kinetics of mLCF monitored by RIf (upper panel) and TIRFS (lower panel): (I) immobilization of His tagged HaloTag; (II) binding of 100 nM Cy3LCF monofunctionalized with clickHTL; (III) surface regeneration by imidazole (green curve). As a negative control, the same experiments carried out with 100 nM nonfunctionalized Cy3LCF is shown (red curve).

Figure 4. Specific targeting of mLCF to the actin-cytoskeleton and mitochondria. (A) Epifluorescence images of HeLa cells stably expressing Lifeact::mEGFP::HaloTag 1h after micro injection of AT647NmLCFPEG2k (green, Lifeact::mEGFP::HaloTag; red, AT647NmLCFPEG2k; yellow, overlay). (B) Time-lapse imaging of AT647NmLCFPEG2k binding to mitochondria (green, Tom20::mEGFP::HaloTag; red, AT647NmLCFPEG2k). Scale bar: 5 μm in all images. (C) Kinetics of AT647NmLCFPEG2k binding to Tom20::mEGFP::HaloTag (data points from four independent measurements were fitted using a monoexponential). The functionalization half-life (t1/2) averaged from independent experiments is shown in the inset.

H12) immobilized via its His-tag on a PEG polymer brush35 was characterized in vitro by TIRFS-RIf detection. Specific, irreversible binding of mLCF to immobilized HaloTag-H12 was observed (Figure 3B). From the binding curves, a reaction rate constant of 4 × 104 M−1 s−1 was obtained for mLCF, that is, by 50% reduced compared to clickHTL alone. Interestingly, mLCFs coated with PEG 7 5 0 (mLCF P E G 7 5 0 ), PEG 2 k (mLCFPEG2k), and PEG2k, where the clickHTL was reacted at the terminus of PEG3k as a spacer (mPEG3kLCFPEG2k), yielded similar reaction rate constants as obtained for unmodified mLCF (Supporting Information Figure S11). These results confirmed robust reactivity of clickHTL irrespective of the coating and localization of the clickHTL on the LCF surface.

of EEG was functionalized with clickHTL via a C-terminal cysteine residue (Supporting Information Scheme S3). Finally, the maleimide-activated clickHTL-peptide was coupled to cysteine residues present on the FNP surface (Supporting Information Scheme S4) and subsequently fractionated by AEX. Elution of bound LCFs through a linear salt gradient yielded 4 fractions of differently charged species (Figure 3A), corresponding to the unfunctionalized LCF as well as LCF with 1, 2, and 3 peptides attached. The elution profile was in excellent agreement with the theoretically expected binomial distribution, confirming an optimal yield of monofunctionalized LCF (Supporting Information Figure S10). The kinetics of monofunctionalized LCF (mLCF) reacting with purified HaloTag fused to a dodecahistidine-tag (HaloTag2192

dx.doi.org/10.1021/nl500637a | Nano Lett. 2014, 14, 2189−2195

Nano Letters

Letter

Figure 5. Tracking of individual AT647NmLCFPEG2k bound to Tom20 in the outer mitochondrial membrane. (A) Superresolved localization map (median localization precision: 29.4 nm) of AT647NmLCFPEG2k bound to Tom20 (red) overlaid on top of the mitochondrial network stained with Tom20::mEGFP::HaloTag (gray). (B) Selected subset (N = 6) of individual Tom20 trajectories labeled with AT647NmLCFPEG2k (median localization precision: 16.8 nm). Scale bar: 5 μm. (C) Comparison of diffusion coefficients D obtained for Tom20 labeled with TMR (left, average D: 0.247 ± 0.10 μm2/s) and AT647NmLCFPEG2k (right, average D: 0.246 ± 0.11 μm2/s) as control of monofunctionalization and stealth properties of the AT647N mLCFPEG2k label. (D) Comparison of mitochondrial radii r recovered from the transversal diffusion for Tom20 labeled with TMR (left, average r: 155 ± 53 nm) and AT647NmLCFPEG2k (right, average r: 161 ± 24 nm).

of 0.246 ± 0.11 μm2/s was obtained (Figure 5C). The distribution of diffusion coefficients was in good agreement with reference measurements carried out using HTL-TMR labeling of HaloTag::mEGFP::Tom20 indicated by the failure to reject the null hypothesis of equal distribution (p-value = 0.81, kstest2, The Mathworks MATLAB 2013a).15 The heterogeneity of observed diffusion coefficients is partly explained by the inability to precisely decompose the trajectories due to their finite observation lengths (Supporting Information Figure 15A,B) as well as the fact that mitochondria might be slightly tilted with respect to the focal plane. Further heterogeneity could stem from intrinsic mobility changes of Tom20 due to interaction. Despite these constraints, essentially unbiased tracking of proteins in intracellular membranes was achieved by labeling with AT647NmLCFPEG2k. In addition to the diffusion properties, by analysis of the transversal diffusion component of each trajectory we were able to extract local mitochondrial radii as a geometrical parameter (Figure 5D and Supporting Information Figure S14) which are in good agreement with the radii observed by electron microscopy.37 This was possible with much higher fidelity for AT647N mLCFPEG2k compared to TMR due to the longer trajectories obtained with these highly photostable nanoparticles. Taken together, we engineered a model nanoparticle and could show for the first time that specific and quantitative application of monofunctionalized nanoparticles is possible inside living cells. Rather than nonspecific nanoparticle aggregation, our results highlight that specific recognition by the autophagy machinery is a key mechanism responsible for clustering and clearance of nanoparticles in the cytoplasm. While autophagy has been frequently implicated in the downstream metabolism of endocytosed nanoparticles,38−41 direct recognition by the cytosolic autophagy machinery has not been described yet. Our results suggest that rather than

For probing specific and efficient protein labeling with mLCFPEG2k within the cytosol, we employed Lifeact, a polypeptide efficiently binding to F-actin,36 which was fused to mEGFP and the HaloTag (Lifeact::mEGFP::Halo). Upon microinjection of mLCFPEG2k, labeled with ATTO 647N (AT647NmLCFPEG2k, DOL: 11), into HeLa cells stably expressing Lifeact::mEGFP::Halo, high colocalization of AT647NmLCFPEG2k with the actin cytoskeleton was observed (Figure 4A), confirming efficient and specific reaction with the target protein. Furthermore, AT647NmLCFPEG2k was efficiently targeted to Tom20::mEGFP::Halo, which is localized in the outer mitochondrial membrane (Supporting Information Figure S12 and Figure 4A). Time lapse imaging revealed rapid binding of AT647N mLCF PEG2k to mitochondria. By quantifying the fluorescence increase specifically on mitochondria (Figure 4B,C and Supporting Information Movie 1), a functionalization time of τ ≈ 13 min was obtained. In contrast, in control experiments no colocalization could be observed by using LCF without clickHTL (Supporting Information Figure S13). These results provide clear evidence that efficient and specific targeting of AT647NmLCFPEG2k to intracellular HaloTag fusion proteins was possible using this approach. The impact of labeling HaloTag::mEGFP::Tom20 with AT647N mLCFPEG2k on its diffusion properties was evaluated by single particle tracking with AT647NmLCFPEG2k microinjected at very low concentrations. Individual particles specifically colocalizing with mitochondria (Figure 5A) could be observed without significant photobleaching. Thus, individual Tom20 diffusing along mitochondria could be readily tracked up to several hundred frames (Figure 5B and Supporting Information Movie 2). Obtained trajectories followed the shape of mitochondria, confirming high fidelity targeting even at very low nanoparticle concentrations. From fitting to the longitudinal diffusion component of each trajectory (Supporting Information Figure S14A), an average diffusion coefficient 2193

dx.doi.org/10.1021/nl500637a | Nano Lett. 2014, 14, 2189−2195

Nano Letters

Letter

(3) Michalet, X.; Pinaud, F. F.; Bentolila, L. A.; Tsay, J. M.; Doose, S.; Li, J. J.; Sundaresan, G.; Wu, A. M.; Gambhir, S. S.; Weiss, S. Science 2005, 307 (5709), 538−44. (4) Pinaud, F.; Clarke, S.; Sittner, A.; Dahan, M. Nat Methods 2010, 7 (4), 275−85. (5) Lidke, D. S.; Nagy, P.; Heintzmann, R.; Arndt-Jovin, D. J.; Post, J. N.; Grecco, H. E.; Jares-Erijman, E. A.; Jovin, T. M. Nat. Biotechnol. 2004, 22 (2), 198−203. (6) Bouzigues, C.; Morel, M.; Triller, A.; Dahan, M. Proc. Natl. Acad. Sci. U.S.A. 2007, 104 (27), 11251−6. (7) Howarth, M.; Takao, K.; Hayashi, Y.; Ting, A. Y. Proc. Natl. Acad. Sci. U.S.A. 2005, 102 (21), 7583−8. (8) Pons, T.; Mattoussi, H. Ann. Biomed. Eng. 2009, 37 (10), 1934− 1959. (9) Algar, W. R.; Prasuhn, D. E.; Stewart, M. H.; Jennings, T. L.; Blanco-Canosa, J. B.; Dawson, P. E.; Medintz, I. L. Bioconjugate Chem. 2011, 22 (5), 825−58. (10) Pierobon, P.; Cappello, G. Adv. Drug. Delivery Rev. 2012, 64 (2), 167−78. (11) Muro, E.; Fragola, A.; Pons, T.; Lequeux, N.; Ioannou, A.; Skourides, P.; Dubertret, B. Small 2012, 8 (7), 1029−37. (12) Los, G. V.; Encell, L. P.; McDougall, M. G.; Hartzell, D. D.; Karassina, N.; Zimprich, C.; Wood, M. G.; Learish, R.; Ohana, R. F.; Urh, M.; Simpson, D.; Mendez, J.; Zimmerman, K.; Otto, P.; Vidugiris, G.; Zhu, J.; Darzins, A.; Klaubert, D. H.; Bulleit, R. F.; Wood, K. V. ACS Chem. Biol. 2008, 3 (6), 373−82. (13) Lisse, D.; Wilkens, V.; You, C.; Busch, K.; Piehler, J. Angew. Chem., Int. Ed. 2011, 50 (40), 9352−5. (14) Etoc, F.; Lisse, D.; Bellaiche, Y.; Piehler, J.; Coppey, M.; Dahan, M. Nat. Nanotechnol. 2013, 8 (3), 193−8. (15) Appelhans, T.; Richter, C. P.; Wilkens, V.; Hess, S. T.; Piehler, J.; Busch, K. B. Nano Lett. 2012, 12 (2), 610−6. (16) Xu, J.; Teslaa, T.; Wu, T. H.; Chiou, P. Y.; Teitell, M. A.; Weiss, S. Nano Lett. 2012, 12 (11), 5669−72. (17) Biju, V.; Itoh, T.; Ishikawa, M. Chem. Soc. Rev. 2010, 39 (8), 3031−56. (18) Delehanty, J. B.; Mattoussi, H.; Medintz, I. L. Anal. Bioanal. Chem. 2009, 393 (4), 1091−105. (19) Kramer, R. M.; Li, C.; Carter, D. C.; Stone, M. O.; Naik, R. R. J. Am. Chem. Soc. 2004, 126 (41), 13282−6. (20) Kabeya, Y.; Mizushima, N.; Ueno, T.; Yamamoto, A.; Kirisako, T.; Noda, T.; Kominami, E.; Ohsumi, Y.; Yoshimori, T. EMBO J. 2000, 19 (21), 5720−8. (21) Ilari, A.; Stefanini, S.; Chiancone, E.; Tsernoglou, D. Nat. Struct. Biol. 2000, 7 (1), 38−43. (22) Rivera-Gil, P.; Jimenez De Aberasturi, D.; Wulf, V.; Pelaz, B.; Del Pino, P.; Zhao, Y.; De La Fuente, J. M.; Ruiz De Larramendi, I.; Rojo, T.; Liang, X.-J.; Parak, W. J. Acc. Chem. Res. 2012, 46 (3), 743− 749. (23) Bonacchi, S.; Genovese, D.; Juris, R.; Montalti, M.; Prodi, L.; Rampazzo, E.; Zaccheroni, N. Angew. Chem., Int. Ed. 2011, 50 (18), 4056−66. (24) Israelachvili, J. N. Intermolecular and Surface Forces, 3rd ed.; Academic Press: New York, 2011; p 191−499. (25) Benhabbour, S. R.; Sheardown, H.; Adronov, A. Biomaterials 2008, 29 (31), 4177−86. (26) Gref, R.; Luck, M.; Quellec, P.; Marchand, M.; Dellacherie, E.; Harnisch, S.; Blunk, T.; Muller, R. H. Colloids Surf., B 2000, 18 (3−4), 301−313. (27) Seksek, O.; Biwersi, J.; Verkman, A. S. J. Cell Biol. 1997, 138 (1), 131−42. (28) Wachsmuth, M.; Waldeck, W.; Langowski, J. J. Mol. Biol. 2000, 298 (4), 677−89. (29) Verkman, A. S. Trends Biochem. Sci. 2002, 27 (1), 27−33. (30) Monopoli, M. P.; Aberg, C.; Salvati, A.; Dawson, K. A. Nat. Nanotechnol. 2012, 7 (12), 779−86. (31) So, M. K.; Yao, H.; Rao, J. Biochem. Biophys. Res. Commun. 2008, 374 (3), 419−23.

direct trafficking from endosomes to autophagosomes, nanoparticle release into the cytosol may cause induction of autophagy responses. Surprisingly, even protein-based nanoparticles, which are physico-chemically highly stable and by definition biocompatible, are efficiently autophagized in the absence of further protection. Yet, also inorganic nanoparticles were targeted to autophagosomes, suggesting a generic intracellular mechanism for metabolism of nanomaterials at this scale. Diffusion studies indicated the formation of a protein corona on the surface of nonmodified LCF particles, which have been suggested to determine biological trafficking and metabolism.30,42,43 As the composition of this protein corona is very likely responsible for autophagosomal targeting, further studies will be required to identify the molecular components involved in particle recognition in order to design evasion strategies more systematically. However, similar as for preventing macrophage uptake of synthetic nanoparticles in organismic applications,44 dense surface PEGylation efficiently reduced nanoparticle recognition by the autophagy machinery. Thus, despite distinct mechanisms, extra- and intracellular nanoparticle clearance mechanisms seem to share common principles of surface pattern recognition, which have probably evolved as part of the antiviral defense.45 By eluding autophagy, intracellular stealth FNPs with a moderate size and surface charge were obtained, which are key requisites for unbiased single molecule studies by minimizing repulsive or attractive interactions with the cellular environment. By establishing LCF monofunctionalization with a reaction rate-enhanced clickHTL, efficient and selective labeling of proteins within the cytoplasm was achieved, enabling versatile intracellular application. Upon fulfilling all these requirements, unbiased diffusion dynamics of a membrane protein within the outer mitochondrial membrane could be observed by a ∼16 nm-sized NP. Thus, rather than the size, the surface properties of nanoparticles are the critical determinants, especially for cytosolic applications. These insights provide key guidelines for designing nanoparticles suitable for biophysical studies in the cytosol of living cells.

■

ASSOCIATED CONTENT

S Supporting Information *

Detailed description of materials and methods, supplementary figures, and movies. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Gabriele Hikade and Hella Kenneweg for technical support, Stefan Walter for mass spectrometry analyses and Sergej Korneev for synthesis of the HaloTag ligand. The cDNA for LC3B was kindly provided by Sascha Martens, Vienna. We also thank him and Fulvio Reggiori, Utrecht, for instructive discussions. This project was supported by funding from the DFG (SFB 944) and by the HFSP (RGP0005/2007).

■

REFERENCES

(1) Joo, C.; Balci, H.; Ishitsuka, Y.; Buranachai, C.; Ha, T. Annu. Rev. Biochem. 2008, 77, 51−76. (2) Lord, S. J.; Lee, H. L.; Moerner, W. E. Anal. Chem. 2010, 82 (6), 2192−203. 2194

dx.doi.org/10.1021/nl500637a | Nano Lett. 2014, 14, 2189−2195

Nano Letters

Letter

(32) Colombo, M.; Mazzucchelli, S.; Montenegro, J. M.; Galbiati, E.; Corsi, F.; Parak, W. J.; Prosperi, D. Small 2012, 8 (10), 1492−7. (33) Gavutis, M.; Lata, S.; Lamken, P.; Müller, P.; Piehler, J. Biophys. J. 2005, 88 (6), 4289−302. (34) Claridge, S. A.; Liang, H. W.; Basu, S. R.; Frechet, J. M.; Alivisatos, A. P. Nano Lett. 2008, 8 (4), 1202−6. (35) Lata, S.; Piehler, J. Anal. Chem. 2005, 77 (4), 1096−1105. (36) Riedl, J.; Crevenna, A. H.; Kessenbrock, K.; Yu, J. H.; Neukirchen, D.; Bista, M.; Bradke, F.; Jenne, D.; Holak, T. A.; Werb, Z.; Sixt, M.; Wedlich-Soldner, R. Nat. Methods 2008, 5 (7), 605−7. (37) Wilkens, V.; Kohl, W.; Busch, K. J. Cell Sci. 2013, 126 (Pt 1), 103−16. (38) Zabirnyk, O.; Yezhelyev, M.; Seleverstov, O. Autophagy 2007, 3 (3), 278−281. (39) Stern, S. T.; Zolnik, B. S.; McLeland, C. B.; Clogston, J.; Zheng, J. W.; McNeil, S. E. Toxicol. Sci. 2008, 106 (1), 140−152. (40) Ma, X.; Wu, Y.; Jin, S.; Tian, Y.; Zhang, X.; Zhao, Y.; Yu, L.; Liang, X. J. ACS Nano 2011, 5 (11), 8629−39. (41) Zhang, Y.; Zheng, F.; Yang, T.; Zhou, W.; Liu, Y.; Man, N.; Zhang, L.; Jin, N.; Dou, Q.; Zhang, Y.; Li, Z.; Wen, L. P. Nat. Mater. 2012, 11 (9), 817−26. (42) Lesniak, A.; Fenaroli, F.; Monopoli, M. P.; Aberg, C.; Dawson, K. A.; Salvati, A. ACS Nano 2012, 6 (7), 5845−57. (43) Lundqvist, M.; Stigler, J.; Elia, G.; Lynch, I.; Cedervall, T.; Dawson, K. A. Proc. Natl. Acad. Sci. U.S.A. 2008, 105 (38), 14265− 14270. (44) Perry, J. L.; Reuter, K. G.; Kai, M. P.; Herlihy, K. P.; Jones, S. W.; Luft, J. C.; Napier, M.; Bear, J. E.; DeSimone, J. M. Nano Lett. 2012, 12 (10), 5304−10. (45) Kudchodkar, S. B.; Levine, B. Rev. Med. Virol. 2009, 19 (6), 359−78.

2195

dx.doi.org/10.1021/nl500637a | Nano Lett. 2014, 14, 2189−2195