Magnetic Nanoparticles
Magnetic Nanoparticles to Recover Cellular Organelles and Study the Time Resolved Nanoparticle-Cell Interactome throughout Uptake Filippo Bertoli, Gemma-Louise Davies, Marco P. Monopoli, Micheal Moloney, Yurii K. Gun’ko, Anna Salvati,* and Kenneth A. Dawson*
Nanoparticles
in contact with cells and living organisms generate quite novel interactions at the interface between the nanoparticle surface and the surrounding biological environment. However, a detailed time resolved molecular level description of the evolving interactions as nanoparticles are internalized and trafficked within the cellular environment is still missing and will certainly be required for the emerging arena of nanoparticle-cell interactions to mature. In this paper promising methodologies to map out the time resolved nanoparticle-cell interactome for nanoparticle uptake are discussed. Thus silica coated magnetite nanoparticles are presented to cells and their magnetic properties used to isolate, in a time resolved manner, the organelles containing the nanoparticles. Characterization of the recovered fractions shows that different cell compartments are isolated at different times, in agreement with imaging results on nanoparticle intracellular location. Subsequently the internalized nanoparticles can be further isolated from the recovered organelles, allowing the study of the most tightly nanoparticle-bound biomolecules, analogous to the ‘hard corona’ that so far has mostly been characterized in extracellular environments. Preliminary data on the recovered nanoparticles suggest that significant portion of the original corona (derived from the serum in which particles are presented to the cells) is preserved as nanoparticles are trafficked through the cells.
1. Introduction Nanoparticles typically interact with cells (and living organisms in general) in a fundamentally different manner than small molecules, and are taken into the cells and processed by active, energy-dependent mechanisms.[1–6] A distinct interface Mr. F. Bertoli, Dr. M. P. Monopoli, Dr. M. Moloney, Dr. A. Salvati,[+] Prof. K. A. Dawson Centre For BioNano Interactions (CBNI) School of Chemistry and Chemical Biology University College Dublin Belfield, Dublin 4, Ireland E-mail: [email protected]; [email protected] Dr. G.-L. Davies[+] School of Chemistry and CRANN Institute Trinity College Dublin Dublin 2, Ireland
between nanoparticles and the biological milieu is created by adsorption of biomolecules such as proteins and lipids from the surrounding biological environment, leading to a biomolecular layer that may be loosely divided into “hard” and “soft” corona, with (respectively) “long” and “short” typical exchange times.[1,7–11] For several nanomaterials, the lifetime Prof. Y. K. Gun’ko School of Chemistry and CRANN Institute Trinity College Dublin Dublin 2, Ireland and Saint Petersburg National Research University of Information Technologies Mechanics and Optics 197101, Saint Petersburg, Russia [+]
Current address: G.-L. D.: Department of Chemistry, University of Warwick, Gibbet Hill Road, Coventry, CV4 7AL, UK. A.S.: Department of Pharmacy, University of Groningen, Antonius Deusinglaan 1, 9713AV, Groningen, The Netherlands.
DOI: 10.1002/smll.201303841 small 2014, DOI: 10.1002/smll.201303841
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
1
full papers
F. Bertoli et al.
of the hard corona formed in plasma and serum (as well as other biofluids) has been shown to be as long as hours, certainly longer than nanoparticle uptake into cells,[8,10,12,13] suggesting that for realistic scenarios, early stage interactions and recognition of nanoparticles by the outer cell membrane can be largely mediated by this corona.[1,14–16] There is also some evidence (for few particle types) that extracellular corona proteins can be retained on the nanoparticles as they enter cells[17,18] and are ultimately trafficked to lysosomes.[2,3,19–21] However, interactions and biological recognition are key motifs that govern protein interactions throughout the whole range of intracellular trafficking and signaling pathways and it is clearly desirable to go beyond the present qualitative observations that extracellular nanoparticle corona proteins are drawn into the cell. Thus, now we will need to focus on their identification, and ultimately connect this identity to a rational molecular-scale interpretation of the processes by which nanoparticles are trafficked, and also of the processes they initiate once inside cells. A pre-requisite for this is the development of methods to allow for a time resolved determination of the biomolecular identities involved in those nanoscale interactions. We must stress here that our interest is in capturing relevant information about the endogenous trafficking and signaling processes involving the nanoparticles, rather than simply the interactions of particles with intracellular extracts, which has previously been reported.[22,23] The latter may be relevant to cases where particles enter cells in damage related processes,[23] or by another physical membrane crossing mechanism,[24] but almost universally, when nanoparticles are presented in a ‘realistic’ biological milieu, active endogenous cellular processes, most involving recognition of some sort, predominate. While in situ methods to identify and image specific nanoparticle-associated protein interactions with biomolecules of the intracellular processing machinery are highly desirable, our current knowledge is so limited that we are not yet able even to identify the relevant proteins in this interactome, as candidates for labeling. A beginning can be made by recovery, isolation and characterization of the organelles in which the nanoparticles are localized, and then isolation of the most strongly nanoparticle-associated biomolecules within those organelles. In principle this would shed light on the identity of the proteins drawn by the nanoparticles into the cell, and potentially their evolution or exchange as the nanoparticles are trafficked between different compartments. Prior time-resolved imaging studies (both live and fixed cell optical as well as TEM) in which nanoparticles are placed in contact with cells for relatively short and well defined periods, and studied after fixed elapsed times, suggested that the nanoparticles are trafficked as a ‘pulse' in which the ‘typical' nanoparticle reaches a specific organelle type at typical time periods.[3,19,25] Therefore, in this work, silica nanoparticles with a core of super paramagnetic iron oxide (SPIONs) are allowed to undergo different periods of active endocytosis in an identical manner to that used for imaging, prior to the cells being disrupted in the presence of a magnetic field[26] and isolation of nanoparticle-associated organelles. It should of course be recognized that one is averaging over various factors such as nanoparticle arrival time at
2 www.small-journal.com
the plasma membrane, and uptake into cells in different cells cycle phases,[21] thus one determines the ‘typical' situation within the cell ensemble under study. However, further separation of cells into different states – such as, for instance, cells into different cell cycle phases – is straightforward. To begin with, organelles obtained from magnetic isolation at different times have been characterized by different techniques, allowing for a time resolved mapping of the recovered subcellular compartments, in broad agreement with expectations from imaging of intact cells at same times. Then, one can further analyze the recovered organellar fractions by separating the internalized nanoparticles and biomolecules most strongly associated with them with a second magnetic extraction. This allows characterization of their evolving corona, as nanoparticles are trafficked through different subcellular compartments, yielding the evolving ‘biological identity’ of the nanoparticles within the cells. Preliminary data indicates that a significant portion of the original corona formed in serum (in which particles are presented to the cells) is preserved as nanoparticles are trafficked through the cells, suggesting that typically the early intracellular interaction of nanoparticles may be prescribed by this corona, rather than directly by the nature of the nanomaterial itself. Further variations of this approach can allow the study of recovered sub-cellular fractions, as well as rare events of nanoparticle trafficking into locations different than the endo-lysosomal pathway.[25]
2. Results and Discussion 2.1. Overview of the Magnetic Extraction Approach Used in This Study Figure 1 gives an overview of the magnetic extraction approach used in this study (Figure 1a) and describes briefly its procedure (Figure 1b-d). As discussed previously, the concept at the heart of this approach is first to extract the organelles containing the nanoparticles in order to characterize the cellular machinery interacting with them, and subsequently to perform another extraction in order to recover and isolate the internalised nanoparticles from the cellular organelles for further studies (Figure 1a). Time resolved magnetic extraction after an initial ‘pulse' exposure of cells to nanoparticles allows us to isolate at different times the different cellular compartments in which nanoparticles are trafficked, thus giving information on the pathways involved. Briefly, the basic steps required to achieve this consist of the lysis of the cells previously incubated with nanoparticles (Figure 1b), the incubation of the cell lysate on a magnet (Figure 1c) and finally the recovery of the organelles containing the nanoparticles (Figure 1d).
2.2. Nanoparticle Synthesis and Characterization Superparamagnetic magnetite nanoparticles (9.1 ± 2 nm, as shown in Supporting Table S1 and Figure 2) were prepared
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2014, DOI: 10.1002/smll.201303841
Magnetic Nanoparticles to Recover Cellular Organelles
Figure 1. a: Time resolved magnetic extraction allows to isolate the different organellar structures in which nanoparticles are located at different times following exposure to nanoparticles (in the example: early endosomes, EE, and lysosomes, Lyso). A second magnetic extraction allows the recovery of the internalized nanoparticles from the isolated organelles (see Experimental Section for details). b-d: Schematic representation of the magnetic extraction procedure. First cells are lysed mechanically by passage through a needle (b) and the lysis efficiency is checked by trypan blue staining (blue lysed cells are shown, with intracellular organelles containing nanoparticles in yellow). Then the cell lysate is incubated on a magnet to allow magnetic separation of the organelles containing the nanoparticles (c) and finally the obtained organellar fraction adhering towards the magnet (yellow) is isolated after removal of the supernatant and collected for further analysis (d).
using a traditional aqueous co-precipitation technique.[27,28] These particles were coated with silica using a modified Stöber sol-gel technique.[29,30] The silica shell thickness was varied by altering the concentration of the silica source, as detailed in the Supporting Information. Silica coated magnetite nanoparticles with sizes 50.8 ± 8 nm and 147.4 ± 17.3 nm were synthesized and are referred to throughout as 50 nm and 150 nm silica coated magnetite nanoparticles (also in Supporting Information Table S1 and Figure 2). TEM results confirmed the size difference between the two silica coated magnetite samples and showed multiple inclusions of magnetite nanoparticles (several cores per particle) (Figure 2). Overall the samples were homogenous in size, and all the nanoparticles contained magnetic cores (Figure 2a-c; see also Supporting Information Table S1). small 2014, DOI: 10.1002/smll.201303841
Figure 2. TEM images of magnetite (a), and 50 (b) and 150 (c) nm silica coated magnetite nanoparticles. Scale bar: 100 nm. d: Magnetisation curves of magnetite and silica coated magnetite nanoparticles at 20 °C, demonstrating superparamagnetism. Magnetite nanoparticles showed a saturation magnetisation value (Ms) of 41.1 Am2/kg, which is in agreement with literature values. Coating with a silica shell caused a decrease in the saturation magnetisation, which reduced further with increasing shell thickness (see also values in Supporting Table S1).
Characterization by FTIR spectroscopy, together with Raman and XRD results (Supporting Information Figure S1, S2 and S3, respectively), further confirmed that silica coated magnetite nanoparticles were obtained. Magnetisation measurements of all samples (Figure 2d and Supporting Table S1) demonstrated typical superparamagnetic characteristics, as expected. The magnetite cores had a saturation magnetisation value at room temperature (Ms) of 41.1 Am2/kg, which is in agreement with literature values.[31–33] Coating with the silica shell caused a decrease in the saturation magnetisation, which decreased further with increasing shell thickness. This resulted in a lower susceptibility for the 150 nm silica coated nanoparticles, in comparison to the 50 nm silica coated nanoparticles and the bare magnetite.
2.3. Nanoparticle Intracellular Localization A549 cells have been applied as a model cell line typically used for nanoparticle uptake studies in order to optimize the method for the recovery of the internalized nanoparticles. Prior to exposure to cells, the nanoparticle dispersions in biological media have been characterized (see Experimental Section for details and Supporting Information Table S2) and the results indicated that both samples were well dispersed also in these conditions. Cell viability measurements after
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
3
full papers
F. Bertoli et al.
Figure 3. TEM images of typical subcellular localizations of 50 (a-c) and 150 (d-f) nm silica coated magnetite nanoparticles. A549 cells were exposed to 250 µg/mL silica coated magnetite nanoparticles in cMEM for up to 24 hours. For the smaller particles, at early exposure times (20 minutes) it was possible to see nanoparticles entering the cell and localising inside early endocytic structures (a and b). After 24 hours most of the observed particles were found inside lysosomes (c). Similar locations were observed also for the 150 nm particles, which entered cells enclosed in vesicles (d) and localised inside early endosomes (e) and in the lysosomes (f). Scale bar: 200 nm.
exposure to the two nanoparticles at different concentrations indicated that no cell death was observed in the conditions applied for the study (Supporting Information Figure S4). Extensive information on how A549 cells uptake and process silica nanoparticles of sizes comparable to the silicacoated magnetite used for this work is already available[19] and allowed comparison with the results obtained here. Cells were exposed to 250 µg/mL nanoparticles, then their subcellular localization at different exposure times was determined by transmission electron microscopy (Figure 3). As it is possible to see from the images in Figure 3, most of the particles followed the endo-lysosomal pathway and reached the lysosomes as their final destination. More in detail, the smaller particles at early exposure times (20 minutes) localized inside early endocytic structures (Figure 3a-b). After 24 hours, most of the observed particles were found inside lysosomes (Figure 3c). Similar locations were observed also for the 150 nm particles, which also entered cells enclosed in vesicles (Figure 3d) and localised inside early endosomes (Figure 3e) and finally in the lysosomes (Figure 3f). However for these nanoparticles, localization in other compartments was also observed (see Supporting Information Figure S5), though such events were rare, suggesting that also in this case most of the nanoparticles followed the endo-lysosomal pathway.
2.4. Optimization and Validation of the Magnetic Extraction As a first step, the possibility of recovering the nanoparticleprotein complexes by magnetic extraction was tested in cell culture media and compared to an already established centrifugation based method (see Experimental Section for details).[34,35] The two sizes of silica coated nanoparticles were incubated in complete cell culture medium and recovered both via centrifugation and via magnetic extraction.
4 www.small-journal.com
The proteins most tightly bound on the nanoparticle surface were recovered and separated by SDS-PAGE (Figure 4a-b). The results showed that the pattern of corona proteins in the samples obtained with the two different methods was similar, and that the recovered protein content increased at increasing nanoparticle concentration. This overall suggested that the magnetic isolation procedure was successful at the tested nanoparticle concentrations and allowed isolation of corona complexes with same protein patterns as in samples isolated via previously developed methods based on centrifugation. (Note that the 150 nm coated silica nanoparticle samples showed a lower level of recovery in comparison to the 50 nm samples, and this is probably due to the lower number of nanoparticles (same mass concentrations were applied) and lower magnetic susceptibility, as shown in Figure 2d). A549 cells were then exposed to 250 µg/mL nanoparticles for 24 hours in order to ensure that high doses of nanoparticles were internalized, and allow optimization of the extraction method. Then the cells were harvested and their membrane lysed as described in details in the Experimental Section; finally the organelle mixtures were put in contact with a magnet in order to separate the organelles in which nanoparticles were located. The isolated sample was then characterized by Western Blot against the lysosomal protein LAMP1 in order to confirm TEM observations and verify that the recovered organelles were indeed enriched in lysosomes. Lysosomes were also recovered using a classic isolation method by density gradient[36] and the lysosomal enrichment obtained by the two different methods compared (see Figure 4c and Supporting Information Figure S6-S7 for further characterization of the two isolation methods). Densitometry analysis showed that the fractions isolated by magnetic extraction had LAMP1 levels comparable to those obtained by the classic density gradient isolation method (also in Figure 4, see Experimental Section for details).
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2014, DOI: 10.1002/smll.201303841
Magnetic Nanoparticles to Recover Cellular Organelles
Figure 4. Optimization and characterisation of the magnetic extraction. SDS-PAGE gels of hard corona proteins recovered from 50 (a) and 150 nm (b) silica coated magnetite after incubation in cMEM (37 °C for 20 minutes as carried out in experiments with cells). Nanoparticle concentration was 500 (lane 1), 250 (lane 2) and 125 (lane 3) µg/ml. Nanoparticles were recovered using an already established centrifugation based method (Centr) and by magnetic extraction (Magn, also marked with a star (*)) (see Experimental Section for details). The same pattern of proteins could be recovered by both methods, suggesting that succesfull magnetic isolation of nanoparticle-corona complexes was possible. c: Western blot and corresponding densitometry results of the lysosomal enriched fractions isolated by density gradient (Gr) and the organellar fractions obtained by magnetic extraction (Magn). Roughly 40 million A549 cells were exposed to 250 µg/mL 50 or 150 nm silica nanoparticles in cMEM for 24 hours prior to fractionation and extraction as described in the Experimental Section. The same total amount of proteins was loaded for each sample. The densitometry results (after normalization performed as described in the Experimental Section) indicated that levels of lysosomal marker LAMP1 obtained in the fractions collected by magnetic extraction were comparable with that obtained by gradient. d: Classification of the subcellular localisation of the proteins identified by mass spectrometry in the fraction isolated by magnetic extraction from cells treated with 50 nm silica coated magnetite for 24 h (250 µg/ml). 522 different proteins were identified and classified (using Goest software) according to their subcellular localisation. Results clearly confirmed the presence of proteins of organelles of the endo-lysosomal pathway (44.3% of total proteins); signals of different serum proteins were also identified (consistent with previous findings that the serum protein corona associated with the nanoparticles is retained, at least in part, as they enter cells),[17] as well as different proteins from ER/Golgi and the cytoplasm. e-g: TEM images of the organellar fractions isolated by magnetic extraction from cells treated for 24 h with 250 µg/mL 150 nm silica confirmed presence of nanoparticles in the majority of the isolated organelles.
Mass spectrometry was also used to identify the total protein composition (See Supporting Information Figure S8 and Table S3). Goest software analysis was used to classify the subcellular localization of the identified proteins (Figure 4d). For this purpose only the most abundant proteins (i.e. spectral count of two or more) were considered. Markers of the endocytic pathway were the most abundant (44.3%), while the presence of others (nuclei, mitochondria, peroxisomes) was low (less than 5% each). It is worth commenting on issues of cross-contamination during the extraction process. Obviously one advantage of the magnetic approach is that isolated assemblies can be held and washed to differing degrees, eliminating the most loosely bound excess components. Naturally, contaminants of high affinity remain. However, one should be cautious about simply associating proteins usually identified with other organelles as contaminants, and it is likely that there is significant overlap between proteins in different organelles, many associated with transport and transfer between compartments.[37] The presence of proteins from cytoplasm and ER/Golgi can be explained by similar overlap, although it could also be linked to rare events of localisation of the nanoparticles in such compartments, as discussed above (Supporting Information Figure S5). Finally TEM imaging of the isolated fractions confirmed the presence of nanoparticles inside the majority of the isolated organelles (Figure 4e-g). small 2014, DOI: 10.1002/smll.201303841
2.5. Time Resolved Magnetic Extraction of the Organelles Containing Nanoparticles After optimization and validation of the magnetic extraction method, in order to follow in more details the location of the nanoparticles as they trafficked in different cell compartments, a different time resolved approach was used, consisting of a short exposure to the nanoparticles (20 minute, 250 µg/mL) followed by incubation in nanoparticle-free medium for increasing times (20 minutes and 24 hours, after the initial 20 minute exposure). The fractions isolated in this way were then tested against different markers of the endolysosomal pathway by Western Blot (Figure 5; see also Supporting Information Figure S9 for the complete gel lanes). The results clearly showed that in the fraction recovered after a short time in nanoparticle-free medium (20 minutes), markers of typical organelles of the early stages of endocytosis were present (early endosomes), while lysosomal markers were absent. On the contrary, in the fraction isolated after 24 hour incubation in nanoparticle-free medium, the opposite was obtained, thus the isolated fraction did not contain early endocytic markers, while lysosomal proteins were present. This was in strong agreement with the results on nanoparticle localization obtained by electron microscopy for corresponding timescales (Figure 5a-b) and suggested that this method can be used to carefully follow in a time-resolved
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
5
full papers
F. Bertoli et al.
Figure 5. Time resolved magnetic extraction. a-b: TEM images of A549 cells exposed to 250 µg/mL 50 nm silica coated magnetite nanoparticles in cMEM for 20 minutes, followed by 20 minutes (a) or 24 hours (b) in nanoparticle-free medium. The results showed that after 20 minutes (a) the majority of the internalized nanoparticles localized inside early endocytic compartments, while after 24 hours (b) the majority of the internalized nanoparticles localized inside lysosomes. c: Western blot of organelle fractions isolated by magnetic extraction after 20 minute exposure to nanoparticles and further growth in nanoparticle-free medium (for 20 minutes and 24 hours as in the images in panels a and b, respectively) (See Experimental Section for details). The same total amount of proteins was loaded for each sample. In the early fraction (20 minutes) different proteins of organelles of the early endocytic pathway were present (Caveolin 1, clathrin heavy chain, and the early endosomal protein EEA1) while the lysosomal marker LAMP1 was absent. In the late fraction (24 hours) the lysosomal marker LAMP1 was present while no signal could be detected for the markers of the early stage of endocytosis. This was in agreement with TEM observations (panels a and b). d: Silver stained 1D SDS-PAGE of the corona proteins isolated from 50 nm silica coated nanoparticles recovered from different subcellular organelles. Cells were exposed for 20 minutes to 250 µg/mL 50 nm silica coated magnetite nanoparticles in cMEM, followed by particle removal and further growth for different times in nanoparticle-free medium (as detailed above the gel). The recovered organelle fractions were lysed mechanically and the internalised nanoparticles were collected with a second magnetic extraction as explained in the Experimental Section. The results showed an increase in the concentration of recovered proteins with time, but overall the pattern of corona proteins remained similar, even if the nanoparticles localized into different cellular compartments (as shown by TEM in panels a-b and Figure 3).
way nanoparticle intracellular trafficking, even after very short exposure to the nanoparticles. The presence of markers of two different uptake pathways like clathrin and caveolin at early exposure times (20 minutes) may suggest the possible involvement of different uptake mechanisms for these nanoparticles; however more studies are needed to exclude possible contamination and carefully address this point, though the approach is overall well placed to support such investigation. As a final step, the nanoparticles contained in the recovered organelles were isolated by a second mechanical lysis (this time of the organellar membrane) and magnetic extraction. After washing the recovered particles to isolate the
6 www.small-journal.com
hard corona nanoparticle complexes (as detailed in the Experimental Section) 1D SDS-PAGE showed that at different times after particle removal, the pattern of the proteins strongly associated to the internalized nanoparticles remained relatively constant, as the nanoparticles were trafficked in different sub-cellular compartments along the endo-lysosomal pathway (as determined by TEM and shown in Figure 2 and 5). This result, although preliminary, suggested that for these particles there was little exchange of the original extracellular hard corona proteins associated to the nanoparticles with the proteins of the different organelles in which the nanoparticles were trafficked. The fact that the concentration of recovered nanoparticles seemed to increase with time (thus the higher protein content in the gel lanes in Figure 5d) was possibly due to a higher efficiency of recovery of the organelles as more nanoparticles accumulated from early endosomes into the lysosomes. BCA results on the protein concentration of the recovered fractions (Supporting Information Figure S10) also confirmed that the fractions recovered after longer times (8 h and 12 h) were richer in protein concentration than the ones recovered after shorter chase times. However, this could be explained by both higher efficiency of recovery and higher protein concentration in the organelles recovered at long chase times (the lysosomes).
3. Conclusions
Here magnetic properties of silica coated magnetite nanoparticles have been used to successfully recover the subcellular fractions in which the nanoparticles were contained. We have demonstrated that this approach can be applied in a time resolved way in order to study organelle – nanoparticle interactions and to follow the intracellular trafficking of the nanoparticles. This allowed us to characterize and identify the subcellular compartments in which the silica coated magnetite nanoparticles were found, confirming the results obtained by electron and optical microscopy which showed accumulation along the endo-lysosomal pathway, as also previously reported for simple silica nanoparticles. The characterization of the isolated fractions suggested that these particles may enter the cells through several different endocytic mechanisms. Analysis of the proteins most firmly bound to the particles within those organelles suggested that much of the serum derived corona is preserved through uptake.
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2014, DOI: 10.1002/smll.201303841
Magnetic Nanoparticles to Recover Cellular Organelles
The whole approach of isolating candidates for proteins and other biomolecules relevant to nanoparticle uptake is quite a different undertaking to the study of endogenous uptake mechanisms of known proteins. Firstly, the initial nanoparticle serum interface replaces any concept of target protein being taken up, then the possibility of a dynamical exchange of these throughout the process could lead to differences in detailed endocytosis processes. Various improvements and developments will be required, but the basic concept of using time resolved extraction of magnetic cored particles with relevant materials surfaces seems a promising direction to study these questions.
4. Experimental Section Cell Culture: Human lung carcinoma epithelial A549 cells (passage 1–30 after defrosting from liquid nitrogen; original batches from ATCC, item number CCL-185, at passage number 82) were cultured at 37 °C in 5% CO2 in Minimum Essential Medium (MEM, with additional L-Glutamine) supplemented with 10% Foetal Bovine Serum (FBS, Gibco), 1% penicillin/streptomycin (Invitrogen), and 1% MEM non-essential amino acids (HyClone) (complete medium, cMEM). Cells were confirmed to be mycoplasma negative using the MycoAlert kit (Lonza Inc. Allendale, NJ) and were tested monthly. Exposure of Cells to Nanoparticles: Cells were plated at a density of 2.5 × 105 cells in a 6 cm diameter plate and at a density of 6×106 in a T175 flask (Grenier) for electron microscopy and magnetic extraction experiments, respectively. After seeding, cells were allowed to adhere for 24 hours before exposure to the nanoparticle dispersions at the appropriate concentrations (in the case of T175 flask, cells were allowed to adhere and grown for 48 hours before the experiment). Nanoparticle dispersions were prepared by diluting the concentrated nanoparticle stock solutions into the complete medium used for cell culture at room temperature, immediately prior to exposure to cells. Cells were incubated with nanoparticles for the required times, depending on the experiment, then the particle containing medium was discarded. To follow the intracellular trafficking of the internalized nanoparticles, after exposure to cells for 20 minutes (250 µg/mL), the nanoparticle dispersion was discarded and after 3 washes with DPBS, fresh cMEM without nanoparticles was added to the cells, which were further grown for increasing times (chase time). This allows to follow in a time resolved manner the location of the internalised nanoparticles as they are trafficked to different intracellular compartments. The medium was finally removed and the samples were washed thrice with DPBS and cells harvested after incubation with trypsin for 3 minutes. Time Resolved Magnetic Extraction of Nanoparticle-Containing Organellar Fractions and Internalised Nanoparticles: After exposure to 50 and 150 nm silica coated magnetite nanoparticles (250 µg/mL in complete medium cMEM) for the required time, A549 cells were harvested with trypsin and centrifuged at 1500 rpm at room temperature. The cells were then resuspended in Lysis Buffer (5 mM Tris base 1 mM EDTA with protease inhibitors, from Mini Complete protease inhibitor, Roche) and the cell membrane lysed mechanically with multiple passages (15–20) into a G 22 needle on a 1 mL syringe. The percentage of cells with broken cell membrane was evaluated on an optical microscope after small 2014, DOI: 10.1002/smll.201303841
staining with Trypan Blue: the mechanical lysis was performed until most of the cells (around 80%) were stained by the dye. Prior to organelle isolation, the cell lysate was spun at 1500 for 3 minutes at 4 °C in order to pellet and discard unbroken cells. To recover the organelles containing the magnetic particles, the cells lysate was then incubated on a magnet (a F999S-N42 Neodymium Magnet acquired from first4magnets.uk; see Supporting Information Figure S11 for magnet geometry) overnight at 4 °C. The supernatant was then discarded, while the pellet was resuspended in 500 μL of lysis buffer containing protease inhibitors (Roche MiniPrep). In order to recover the nanoparticles from the isolated organelles, the same method was used to perform a second magnetic extraction, after the lysis of the organellar memebrane (performed following the same method). The magnetic extraction is illustrated in Figure 1. Isolation of Lysosomes via Centrifugation in a Density Gradient: In order to allow comparison with the fractions obtained by magnetic extraction and validate the method, lysosomes were isolated using a Lysosome Isolation Kit (Sigma).[36] After lysis of the cell membrane performed as described above, in order to recover the first the crude lysosomal fraction (CLF), the cell lysates were centrifuged two times. These centrifugations ensure the separation of eventual intact cells, large cell debris and nuclei (first centrifugation: 10 minutes at 1000 g; the CLF is contained in the supernatant), and also allow to separate the CLF from other smaller cellular compartments such as endoplasmic reticulum and cytoplasm (second centrifugation: 20 minutes at 20 000 g; the CLF is contained in the pellet). In this way the final CLF contains mainly lysosomes. The CLF was then resuspended in a solution containing 19% Optiprep Density Gradient Medium to a final protein concentration of 0.5–1.0 mg-protein/mL. The lysosomes were then further purified with a separation by density gradient centrifugation (150 000 g for 4 hours at 4 °C) in order to obtain 6 fractions, separated through the gradient according to their density. Western blot of LAMP1 was used to confirm lysosomal enrichment between the 3rd and 4th fractions, as per instructions (Supporting Information Figure S6). Protein Corona Isolation: In order to isolate hard corona nanoparticle complexes, nanoparticles were incubated in the MEM medium used for cell culture, supplemented with 10% fetal bovine serum (complete medium, cMEM) at different concentrations (500, 250, 125 µg/mL NPs) for 20 minutes at 37 °C. The samples were then centrifuged for 20 minutes at 20 krcf at 4 °C to pellet the particle-protein complexes and separate them from the supernatant. The pellet was resuspended in 500 μL of PBS and centrifuged again in the same way for a total of three washing-steps, before resuspension to the desired concentration in PBS. This treatment allowed us to remove the proteins with low affinity for the nanoparticle surface and isolate hard corona nanoparticle complexes.[34] The same protocol was applied to isolate the hard corona nanoparticles from the extracted subcellular organelles after a second magnetic extraction. BCA Assay: In order to determine the protein concentration in the isolated organelle fractions, the BCA assay (BCA Protein Assay Reagent, Pierce) was performed according to Manufacturer’s instructions. Briefly the samples were incubated with the BCA protein assay reagents in a 96 well plate for 30 minutes at 37 °C and with 1% Triton X to ensure the lysis of the isolated organelles; the
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
7
full papers
F. Bertoli et al.
absorbance at 562 nm was then read on a VarioSkan Flash plate reader (Thermo, USA). SDS-PAGE and Western Blotting: SDS-PAGE was used to separate and characterize the proteins contained in the isolated coronananoparticle complexes and in the recovered organelle fractions. Samples were diluted in protein loading buffer (62.5 mM Tris-HCL pH 6.8, 2% (w/v) SDS, 10% glycerol, 0.04M DTT and 0.01% (w/v) bromophenol blue), and incubated for 5 minutes at 100 °C. After necessary dilutions in PBS to normalize the protein content among different samples recovered, an equal sample volume (thus same total protein concentration) was loaded for each well in a 10% gel polyacrylamide gel. This was performed on all gels except for the gels present in Figure 4a-b and in Figure 5d, in which equal sample volume (without protein normalization) was loaded. Gel electrophoresis was performed at 120V, 400 mA for about 60 minutes, until the proteins neared the end of the gel. To detect the separated proteins, the gels were stained for 1 hour in Coomassie blue staining (50% methanol, 10% acetic acid, 2.5% (w/v) brilliant blue) and de-stained overnight in 50% methanol, 10% acetic acid. When higher sensitivity of detection was required, gels were stained using the silver staining Daiichi kit according to protocol given by the supplier. Gels were scanned using a Biorad GS-800 calibrated densitometer scanner. For the western blotting procedure, after separation of the recovered proteins by SDS-PAGE, the proteins were transferred to a nitro cellulose membrane at 100 mV for 1 h using a wet transfer method. Then, the membrane was incubated with a 2% BSA blocking solution in PBS TWEEN 0,01% for 1 h at room temperature. The membrane was then incubated with the antibody of interest (LAMP1, Clathrin heavy chain, Caveolin 1, EEA1; monoclonal, mouse, from Abcam) in blocking solution (1:4000; 1:2000 depending on the antibody used) overnight at 4 °C. Finally, the membrane was washed 4–5 times with PBS TWEEN 0.01% and incubated with the secondary antibody (anti mouse HRP, Abcam) at a dilution of 1:4000 in blocking solution for 1 h. The membrane was then washed with PBS TWEEN 0.01% for 4 times and incubated with the substrate solution for the chemiluminescent reaction (ECL PIERCE) for 1 minute and finally developed on X-Ray film in a dark room. Mass Spectrometry: In order to identify the recovered proteins by mass spectrometry analysis, after separation by SDS-PAGE performed as described above, the bands of interest were excised from the gel and digested in-gel with trypsin (porcine trypsine, Promega), according to the method of Shevchenko et al.[38] The resulting peptide mixtures were resuspended in 0.1% formic acid and analyzed by electrospray liquid chromatography mass spectrometry (LC MS/MS) using an HPLC (Surveyor, ThermoFinnigan, CA) interfaced with an LTQ Orbitrap (ThermoFinnigan, CA). Chromatography buffer solutions (Buffer A, 0.1% formic acid; Buffer B, 100% acetonitrile and 0.1% formic acid) were run using a 72 minute gradient. A 150 μL/min flow rate was used at the electrospray source. Spectra were analysed with Bioworks Browser 3.3.1 SP1 (ThermoFisher Scientific) using Sequest Uniprot/Swiss-Prot database (www.expasy.org). An exclusion filter was applied to reduce false positives, where peptides with P
Magnetic Nanoparticles to Recover Cellular Organelles and Study the Time Resolved Nanoparticle-Cell Interactome throughout Uptake Filippo Bertoli, Gemma-Louise Davies, Marco P. Monopoli, Micheal Moloney, Yurii K. Gun’ko, Anna Salvati,* and Kenneth A. Dawson*
Nanoparticles
in contact with cells and living organisms generate quite novel interactions at the interface between the nanoparticle surface and the surrounding biological environment. However, a detailed time resolved molecular level description of the evolving interactions as nanoparticles are internalized and trafficked within the cellular environment is still missing and will certainly be required for the emerging arena of nanoparticle-cell interactions to mature. In this paper promising methodologies to map out the time resolved nanoparticle-cell interactome for nanoparticle uptake are discussed. Thus silica coated magnetite nanoparticles are presented to cells and their magnetic properties used to isolate, in a time resolved manner, the organelles containing the nanoparticles. Characterization of the recovered fractions shows that different cell compartments are isolated at different times, in agreement with imaging results on nanoparticle intracellular location. Subsequently the internalized nanoparticles can be further isolated from the recovered organelles, allowing the study of the most tightly nanoparticle-bound biomolecules, analogous to the ‘hard corona’ that so far has mostly been characterized in extracellular environments. Preliminary data on the recovered nanoparticles suggest that significant portion of the original corona (derived from the serum in which particles are presented to the cells) is preserved as nanoparticles are trafficked through the cells.
1. Introduction Nanoparticles typically interact with cells (and living organisms in general) in a fundamentally different manner than small molecules, and are taken into the cells and processed by active, energy-dependent mechanisms.[1–6] A distinct interface Mr. F. Bertoli, Dr. M. P. Monopoli, Dr. M. Moloney, Dr. A. Salvati,[+] Prof. K. A. Dawson Centre For BioNano Interactions (CBNI) School of Chemistry and Chemical Biology University College Dublin Belfield, Dublin 4, Ireland E-mail: [email protected]; [email protected] Dr. G.-L. Davies[+] School of Chemistry and CRANN Institute Trinity College Dublin Dublin 2, Ireland
between nanoparticles and the biological milieu is created by adsorption of biomolecules such as proteins and lipids from the surrounding biological environment, leading to a biomolecular layer that may be loosely divided into “hard” and “soft” corona, with (respectively) “long” and “short” typical exchange times.[1,7–11] For several nanomaterials, the lifetime Prof. Y. K. Gun’ko School of Chemistry and CRANN Institute Trinity College Dublin Dublin 2, Ireland and Saint Petersburg National Research University of Information Technologies Mechanics and Optics 197101, Saint Petersburg, Russia [+]
Current address: G.-L. D.: Department of Chemistry, University of Warwick, Gibbet Hill Road, Coventry, CV4 7AL, UK. A.S.: Department of Pharmacy, University of Groningen, Antonius Deusinglaan 1, 9713AV, Groningen, The Netherlands.
DOI: 10.1002/smll.201303841 small 2014, DOI: 10.1002/smll.201303841
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
1
full papers
F. Bertoli et al.
of the hard corona formed in plasma and serum (as well as other biofluids) has been shown to be as long as hours, certainly longer than nanoparticle uptake into cells,[8,10,12,13] suggesting that for realistic scenarios, early stage interactions and recognition of nanoparticles by the outer cell membrane can be largely mediated by this corona.[1,14–16] There is also some evidence (for few particle types) that extracellular corona proteins can be retained on the nanoparticles as they enter cells[17,18] and are ultimately trafficked to lysosomes.[2,3,19–21] However, interactions and biological recognition are key motifs that govern protein interactions throughout the whole range of intracellular trafficking and signaling pathways and it is clearly desirable to go beyond the present qualitative observations that extracellular nanoparticle corona proteins are drawn into the cell. Thus, now we will need to focus on their identification, and ultimately connect this identity to a rational molecular-scale interpretation of the processes by which nanoparticles are trafficked, and also of the processes they initiate once inside cells. A pre-requisite for this is the development of methods to allow for a time resolved determination of the biomolecular identities involved in those nanoscale interactions. We must stress here that our interest is in capturing relevant information about the endogenous trafficking and signaling processes involving the nanoparticles, rather than simply the interactions of particles with intracellular extracts, which has previously been reported.[22,23] The latter may be relevant to cases where particles enter cells in damage related processes,[23] or by another physical membrane crossing mechanism,[24] but almost universally, when nanoparticles are presented in a ‘realistic’ biological milieu, active endogenous cellular processes, most involving recognition of some sort, predominate. While in situ methods to identify and image specific nanoparticle-associated protein interactions with biomolecules of the intracellular processing machinery are highly desirable, our current knowledge is so limited that we are not yet able even to identify the relevant proteins in this interactome, as candidates for labeling. A beginning can be made by recovery, isolation and characterization of the organelles in which the nanoparticles are localized, and then isolation of the most strongly nanoparticle-associated biomolecules within those organelles. In principle this would shed light on the identity of the proteins drawn by the nanoparticles into the cell, and potentially their evolution or exchange as the nanoparticles are trafficked between different compartments. Prior time-resolved imaging studies (both live and fixed cell optical as well as TEM) in which nanoparticles are placed in contact with cells for relatively short and well defined periods, and studied after fixed elapsed times, suggested that the nanoparticles are trafficked as a ‘pulse' in which the ‘typical' nanoparticle reaches a specific organelle type at typical time periods.[3,19,25] Therefore, in this work, silica nanoparticles with a core of super paramagnetic iron oxide (SPIONs) are allowed to undergo different periods of active endocytosis in an identical manner to that used for imaging, prior to the cells being disrupted in the presence of a magnetic field[26] and isolation of nanoparticle-associated organelles. It should of course be recognized that one is averaging over various factors such as nanoparticle arrival time at
2 www.small-journal.com
the plasma membrane, and uptake into cells in different cells cycle phases,[21] thus one determines the ‘typical' situation within the cell ensemble under study. However, further separation of cells into different states – such as, for instance, cells into different cell cycle phases – is straightforward. To begin with, organelles obtained from magnetic isolation at different times have been characterized by different techniques, allowing for a time resolved mapping of the recovered subcellular compartments, in broad agreement with expectations from imaging of intact cells at same times. Then, one can further analyze the recovered organellar fractions by separating the internalized nanoparticles and biomolecules most strongly associated with them with a second magnetic extraction. This allows characterization of their evolving corona, as nanoparticles are trafficked through different subcellular compartments, yielding the evolving ‘biological identity’ of the nanoparticles within the cells. Preliminary data indicates that a significant portion of the original corona formed in serum (in which particles are presented to the cells) is preserved as nanoparticles are trafficked through the cells, suggesting that typically the early intracellular interaction of nanoparticles may be prescribed by this corona, rather than directly by the nature of the nanomaterial itself. Further variations of this approach can allow the study of recovered sub-cellular fractions, as well as rare events of nanoparticle trafficking into locations different than the endo-lysosomal pathway.[25]
2. Results and Discussion 2.1. Overview of the Magnetic Extraction Approach Used in This Study Figure 1 gives an overview of the magnetic extraction approach used in this study (Figure 1a) and describes briefly its procedure (Figure 1b-d). As discussed previously, the concept at the heart of this approach is first to extract the organelles containing the nanoparticles in order to characterize the cellular machinery interacting with them, and subsequently to perform another extraction in order to recover and isolate the internalised nanoparticles from the cellular organelles for further studies (Figure 1a). Time resolved magnetic extraction after an initial ‘pulse' exposure of cells to nanoparticles allows us to isolate at different times the different cellular compartments in which nanoparticles are trafficked, thus giving information on the pathways involved. Briefly, the basic steps required to achieve this consist of the lysis of the cells previously incubated with nanoparticles (Figure 1b), the incubation of the cell lysate on a magnet (Figure 1c) and finally the recovery of the organelles containing the nanoparticles (Figure 1d).
2.2. Nanoparticle Synthesis and Characterization Superparamagnetic magnetite nanoparticles (9.1 ± 2 nm, as shown in Supporting Table S1 and Figure 2) were prepared
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2014, DOI: 10.1002/smll.201303841
Magnetic Nanoparticles to Recover Cellular Organelles
Figure 1. a: Time resolved magnetic extraction allows to isolate the different organellar structures in which nanoparticles are located at different times following exposure to nanoparticles (in the example: early endosomes, EE, and lysosomes, Lyso). A second magnetic extraction allows the recovery of the internalized nanoparticles from the isolated organelles (see Experimental Section for details). b-d: Schematic representation of the magnetic extraction procedure. First cells are lysed mechanically by passage through a needle (b) and the lysis efficiency is checked by trypan blue staining (blue lysed cells are shown, with intracellular organelles containing nanoparticles in yellow). Then the cell lysate is incubated on a magnet to allow magnetic separation of the organelles containing the nanoparticles (c) and finally the obtained organellar fraction adhering towards the magnet (yellow) is isolated after removal of the supernatant and collected for further analysis (d).
using a traditional aqueous co-precipitation technique.[27,28] These particles were coated with silica using a modified Stöber sol-gel technique.[29,30] The silica shell thickness was varied by altering the concentration of the silica source, as detailed in the Supporting Information. Silica coated magnetite nanoparticles with sizes 50.8 ± 8 nm and 147.4 ± 17.3 nm were synthesized and are referred to throughout as 50 nm and 150 nm silica coated magnetite nanoparticles (also in Supporting Information Table S1 and Figure 2). TEM results confirmed the size difference between the two silica coated magnetite samples and showed multiple inclusions of magnetite nanoparticles (several cores per particle) (Figure 2). Overall the samples were homogenous in size, and all the nanoparticles contained magnetic cores (Figure 2a-c; see also Supporting Information Table S1). small 2014, DOI: 10.1002/smll.201303841
Figure 2. TEM images of magnetite (a), and 50 (b) and 150 (c) nm silica coated magnetite nanoparticles. Scale bar: 100 nm. d: Magnetisation curves of magnetite and silica coated magnetite nanoparticles at 20 °C, demonstrating superparamagnetism. Magnetite nanoparticles showed a saturation magnetisation value (Ms) of 41.1 Am2/kg, which is in agreement with literature values. Coating with a silica shell caused a decrease in the saturation magnetisation, which reduced further with increasing shell thickness (see also values in Supporting Table S1).
Characterization by FTIR spectroscopy, together with Raman and XRD results (Supporting Information Figure S1, S2 and S3, respectively), further confirmed that silica coated magnetite nanoparticles were obtained. Magnetisation measurements of all samples (Figure 2d and Supporting Table S1) demonstrated typical superparamagnetic characteristics, as expected. The magnetite cores had a saturation magnetisation value at room temperature (Ms) of 41.1 Am2/kg, which is in agreement with literature values.[31–33] Coating with the silica shell caused a decrease in the saturation magnetisation, which decreased further with increasing shell thickness. This resulted in a lower susceptibility for the 150 nm silica coated nanoparticles, in comparison to the 50 nm silica coated nanoparticles and the bare magnetite.
2.3. Nanoparticle Intracellular Localization A549 cells have been applied as a model cell line typically used for nanoparticle uptake studies in order to optimize the method for the recovery of the internalized nanoparticles. Prior to exposure to cells, the nanoparticle dispersions in biological media have been characterized (see Experimental Section for details and Supporting Information Table S2) and the results indicated that both samples were well dispersed also in these conditions. Cell viability measurements after
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
3
full papers
F. Bertoli et al.
Figure 3. TEM images of typical subcellular localizations of 50 (a-c) and 150 (d-f) nm silica coated magnetite nanoparticles. A549 cells were exposed to 250 µg/mL silica coated magnetite nanoparticles in cMEM for up to 24 hours. For the smaller particles, at early exposure times (20 minutes) it was possible to see nanoparticles entering the cell and localising inside early endocytic structures (a and b). After 24 hours most of the observed particles were found inside lysosomes (c). Similar locations were observed also for the 150 nm particles, which entered cells enclosed in vesicles (d) and localised inside early endosomes (e) and in the lysosomes (f). Scale bar: 200 nm.
exposure to the two nanoparticles at different concentrations indicated that no cell death was observed in the conditions applied for the study (Supporting Information Figure S4). Extensive information on how A549 cells uptake and process silica nanoparticles of sizes comparable to the silicacoated magnetite used for this work is already available[19] and allowed comparison with the results obtained here. Cells were exposed to 250 µg/mL nanoparticles, then their subcellular localization at different exposure times was determined by transmission electron microscopy (Figure 3). As it is possible to see from the images in Figure 3, most of the particles followed the endo-lysosomal pathway and reached the lysosomes as their final destination. More in detail, the smaller particles at early exposure times (20 minutes) localized inside early endocytic structures (Figure 3a-b). After 24 hours, most of the observed particles were found inside lysosomes (Figure 3c). Similar locations were observed also for the 150 nm particles, which also entered cells enclosed in vesicles (Figure 3d) and localised inside early endosomes (Figure 3e) and finally in the lysosomes (Figure 3f). However for these nanoparticles, localization in other compartments was also observed (see Supporting Information Figure S5), though such events were rare, suggesting that also in this case most of the nanoparticles followed the endo-lysosomal pathway.
2.4. Optimization and Validation of the Magnetic Extraction As a first step, the possibility of recovering the nanoparticleprotein complexes by magnetic extraction was tested in cell culture media and compared to an already established centrifugation based method (see Experimental Section for details).[34,35] The two sizes of silica coated nanoparticles were incubated in complete cell culture medium and recovered both via centrifugation and via magnetic extraction.
4 www.small-journal.com
The proteins most tightly bound on the nanoparticle surface were recovered and separated by SDS-PAGE (Figure 4a-b). The results showed that the pattern of corona proteins in the samples obtained with the two different methods was similar, and that the recovered protein content increased at increasing nanoparticle concentration. This overall suggested that the magnetic isolation procedure was successful at the tested nanoparticle concentrations and allowed isolation of corona complexes with same protein patterns as in samples isolated via previously developed methods based on centrifugation. (Note that the 150 nm coated silica nanoparticle samples showed a lower level of recovery in comparison to the 50 nm samples, and this is probably due to the lower number of nanoparticles (same mass concentrations were applied) and lower magnetic susceptibility, as shown in Figure 2d). A549 cells were then exposed to 250 µg/mL nanoparticles for 24 hours in order to ensure that high doses of nanoparticles were internalized, and allow optimization of the extraction method. Then the cells were harvested and their membrane lysed as described in details in the Experimental Section; finally the organelle mixtures were put in contact with a magnet in order to separate the organelles in which nanoparticles were located. The isolated sample was then characterized by Western Blot against the lysosomal protein LAMP1 in order to confirm TEM observations and verify that the recovered organelles were indeed enriched in lysosomes. Lysosomes were also recovered using a classic isolation method by density gradient[36] and the lysosomal enrichment obtained by the two different methods compared (see Figure 4c and Supporting Information Figure S6-S7 for further characterization of the two isolation methods). Densitometry analysis showed that the fractions isolated by magnetic extraction had LAMP1 levels comparable to those obtained by the classic density gradient isolation method (also in Figure 4, see Experimental Section for details).
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2014, DOI: 10.1002/smll.201303841
Magnetic Nanoparticles to Recover Cellular Organelles
Figure 4. Optimization and characterisation of the magnetic extraction. SDS-PAGE gels of hard corona proteins recovered from 50 (a) and 150 nm (b) silica coated magnetite after incubation in cMEM (37 °C for 20 minutes as carried out in experiments with cells). Nanoparticle concentration was 500 (lane 1), 250 (lane 2) and 125 (lane 3) µg/ml. Nanoparticles were recovered using an already established centrifugation based method (Centr) and by magnetic extraction (Magn, also marked with a star (*)) (see Experimental Section for details). The same pattern of proteins could be recovered by both methods, suggesting that succesfull magnetic isolation of nanoparticle-corona complexes was possible. c: Western blot and corresponding densitometry results of the lysosomal enriched fractions isolated by density gradient (Gr) and the organellar fractions obtained by magnetic extraction (Magn). Roughly 40 million A549 cells were exposed to 250 µg/mL 50 or 150 nm silica nanoparticles in cMEM for 24 hours prior to fractionation and extraction as described in the Experimental Section. The same total amount of proteins was loaded for each sample. The densitometry results (after normalization performed as described in the Experimental Section) indicated that levels of lysosomal marker LAMP1 obtained in the fractions collected by magnetic extraction were comparable with that obtained by gradient. d: Classification of the subcellular localisation of the proteins identified by mass spectrometry in the fraction isolated by magnetic extraction from cells treated with 50 nm silica coated magnetite for 24 h (250 µg/ml). 522 different proteins were identified and classified (using Goest software) according to their subcellular localisation. Results clearly confirmed the presence of proteins of organelles of the endo-lysosomal pathway (44.3% of total proteins); signals of different serum proteins were also identified (consistent with previous findings that the serum protein corona associated with the nanoparticles is retained, at least in part, as they enter cells),[17] as well as different proteins from ER/Golgi and the cytoplasm. e-g: TEM images of the organellar fractions isolated by magnetic extraction from cells treated for 24 h with 250 µg/mL 150 nm silica confirmed presence of nanoparticles in the majority of the isolated organelles.
Mass spectrometry was also used to identify the total protein composition (See Supporting Information Figure S8 and Table S3). Goest software analysis was used to classify the subcellular localization of the identified proteins (Figure 4d). For this purpose only the most abundant proteins (i.e. spectral count of two or more) were considered. Markers of the endocytic pathway were the most abundant (44.3%), while the presence of others (nuclei, mitochondria, peroxisomes) was low (less than 5% each). It is worth commenting on issues of cross-contamination during the extraction process. Obviously one advantage of the magnetic approach is that isolated assemblies can be held and washed to differing degrees, eliminating the most loosely bound excess components. Naturally, contaminants of high affinity remain. However, one should be cautious about simply associating proteins usually identified with other organelles as contaminants, and it is likely that there is significant overlap between proteins in different organelles, many associated with transport and transfer between compartments.[37] The presence of proteins from cytoplasm and ER/Golgi can be explained by similar overlap, although it could also be linked to rare events of localisation of the nanoparticles in such compartments, as discussed above (Supporting Information Figure S5). Finally TEM imaging of the isolated fractions confirmed the presence of nanoparticles inside the majority of the isolated organelles (Figure 4e-g). small 2014, DOI: 10.1002/smll.201303841
2.5. Time Resolved Magnetic Extraction of the Organelles Containing Nanoparticles After optimization and validation of the magnetic extraction method, in order to follow in more details the location of the nanoparticles as they trafficked in different cell compartments, a different time resolved approach was used, consisting of a short exposure to the nanoparticles (20 minute, 250 µg/mL) followed by incubation in nanoparticle-free medium for increasing times (20 minutes and 24 hours, after the initial 20 minute exposure). The fractions isolated in this way were then tested against different markers of the endolysosomal pathway by Western Blot (Figure 5; see also Supporting Information Figure S9 for the complete gel lanes). The results clearly showed that in the fraction recovered after a short time in nanoparticle-free medium (20 minutes), markers of typical organelles of the early stages of endocytosis were present (early endosomes), while lysosomal markers were absent. On the contrary, in the fraction isolated after 24 hour incubation in nanoparticle-free medium, the opposite was obtained, thus the isolated fraction did not contain early endocytic markers, while lysosomal proteins were present. This was in strong agreement with the results on nanoparticle localization obtained by electron microscopy for corresponding timescales (Figure 5a-b) and suggested that this method can be used to carefully follow in a time-resolved
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
5
full papers
F. Bertoli et al.
Figure 5. Time resolved magnetic extraction. a-b: TEM images of A549 cells exposed to 250 µg/mL 50 nm silica coated magnetite nanoparticles in cMEM for 20 minutes, followed by 20 minutes (a) or 24 hours (b) in nanoparticle-free medium. The results showed that after 20 minutes (a) the majority of the internalized nanoparticles localized inside early endocytic compartments, while after 24 hours (b) the majority of the internalized nanoparticles localized inside lysosomes. c: Western blot of organelle fractions isolated by magnetic extraction after 20 minute exposure to nanoparticles and further growth in nanoparticle-free medium (for 20 minutes and 24 hours as in the images in panels a and b, respectively) (See Experimental Section for details). The same total amount of proteins was loaded for each sample. In the early fraction (20 minutes) different proteins of organelles of the early endocytic pathway were present (Caveolin 1, clathrin heavy chain, and the early endosomal protein EEA1) while the lysosomal marker LAMP1 was absent. In the late fraction (24 hours) the lysosomal marker LAMP1 was present while no signal could be detected for the markers of the early stage of endocytosis. This was in agreement with TEM observations (panels a and b). d: Silver stained 1D SDS-PAGE of the corona proteins isolated from 50 nm silica coated nanoparticles recovered from different subcellular organelles. Cells were exposed for 20 minutes to 250 µg/mL 50 nm silica coated magnetite nanoparticles in cMEM, followed by particle removal and further growth for different times in nanoparticle-free medium (as detailed above the gel). The recovered organelle fractions were lysed mechanically and the internalised nanoparticles were collected with a second magnetic extraction as explained in the Experimental Section. The results showed an increase in the concentration of recovered proteins with time, but overall the pattern of corona proteins remained similar, even if the nanoparticles localized into different cellular compartments (as shown by TEM in panels a-b and Figure 3).
way nanoparticle intracellular trafficking, even after very short exposure to the nanoparticles. The presence of markers of two different uptake pathways like clathrin and caveolin at early exposure times (20 minutes) may suggest the possible involvement of different uptake mechanisms for these nanoparticles; however more studies are needed to exclude possible contamination and carefully address this point, though the approach is overall well placed to support such investigation. As a final step, the nanoparticles contained in the recovered organelles were isolated by a second mechanical lysis (this time of the organellar membrane) and magnetic extraction. After washing the recovered particles to isolate the
6 www.small-journal.com
hard corona nanoparticle complexes (as detailed in the Experimental Section) 1D SDS-PAGE showed that at different times after particle removal, the pattern of the proteins strongly associated to the internalized nanoparticles remained relatively constant, as the nanoparticles were trafficked in different sub-cellular compartments along the endo-lysosomal pathway (as determined by TEM and shown in Figure 2 and 5). This result, although preliminary, suggested that for these particles there was little exchange of the original extracellular hard corona proteins associated to the nanoparticles with the proteins of the different organelles in which the nanoparticles were trafficked. The fact that the concentration of recovered nanoparticles seemed to increase with time (thus the higher protein content in the gel lanes in Figure 5d) was possibly due to a higher efficiency of recovery of the organelles as more nanoparticles accumulated from early endosomes into the lysosomes. BCA results on the protein concentration of the recovered fractions (Supporting Information Figure S10) also confirmed that the fractions recovered after longer times (8 h and 12 h) were richer in protein concentration than the ones recovered after shorter chase times. However, this could be explained by both higher efficiency of recovery and higher protein concentration in the organelles recovered at long chase times (the lysosomes).
3. Conclusions
Here magnetic properties of silica coated magnetite nanoparticles have been used to successfully recover the subcellular fractions in which the nanoparticles were contained. We have demonstrated that this approach can be applied in a time resolved way in order to study organelle – nanoparticle interactions and to follow the intracellular trafficking of the nanoparticles. This allowed us to characterize and identify the subcellular compartments in which the silica coated magnetite nanoparticles were found, confirming the results obtained by electron and optical microscopy which showed accumulation along the endo-lysosomal pathway, as also previously reported for simple silica nanoparticles. The characterization of the isolated fractions suggested that these particles may enter the cells through several different endocytic mechanisms. Analysis of the proteins most firmly bound to the particles within those organelles suggested that much of the serum derived corona is preserved through uptake.
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2014, DOI: 10.1002/smll.201303841
Magnetic Nanoparticles to Recover Cellular Organelles
The whole approach of isolating candidates for proteins and other biomolecules relevant to nanoparticle uptake is quite a different undertaking to the study of endogenous uptake mechanisms of known proteins. Firstly, the initial nanoparticle serum interface replaces any concept of target protein being taken up, then the possibility of a dynamical exchange of these throughout the process could lead to differences in detailed endocytosis processes. Various improvements and developments will be required, but the basic concept of using time resolved extraction of magnetic cored particles with relevant materials surfaces seems a promising direction to study these questions.
4. Experimental Section Cell Culture: Human lung carcinoma epithelial A549 cells (passage 1–30 after defrosting from liquid nitrogen; original batches from ATCC, item number CCL-185, at passage number 82) were cultured at 37 °C in 5% CO2 in Minimum Essential Medium (MEM, with additional L-Glutamine) supplemented with 10% Foetal Bovine Serum (FBS, Gibco), 1% penicillin/streptomycin (Invitrogen), and 1% MEM non-essential amino acids (HyClone) (complete medium, cMEM). Cells were confirmed to be mycoplasma negative using the MycoAlert kit (Lonza Inc. Allendale, NJ) and were tested monthly. Exposure of Cells to Nanoparticles: Cells were plated at a density of 2.5 × 105 cells in a 6 cm diameter plate and at a density of 6×106 in a T175 flask (Grenier) for electron microscopy and magnetic extraction experiments, respectively. After seeding, cells were allowed to adhere for 24 hours before exposure to the nanoparticle dispersions at the appropriate concentrations (in the case of T175 flask, cells were allowed to adhere and grown for 48 hours before the experiment). Nanoparticle dispersions were prepared by diluting the concentrated nanoparticle stock solutions into the complete medium used for cell culture at room temperature, immediately prior to exposure to cells. Cells were incubated with nanoparticles for the required times, depending on the experiment, then the particle containing medium was discarded. To follow the intracellular trafficking of the internalized nanoparticles, after exposure to cells for 20 minutes (250 µg/mL), the nanoparticle dispersion was discarded and after 3 washes with DPBS, fresh cMEM without nanoparticles was added to the cells, which were further grown for increasing times (chase time). This allows to follow in a time resolved manner the location of the internalised nanoparticles as they are trafficked to different intracellular compartments. The medium was finally removed and the samples were washed thrice with DPBS and cells harvested after incubation with trypsin for 3 minutes. Time Resolved Magnetic Extraction of Nanoparticle-Containing Organellar Fractions and Internalised Nanoparticles: After exposure to 50 and 150 nm silica coated magnetite nanoparticles (250 µg/mL in complete medium cMEM) for the required time, A549 cells were harvested with trypsin and centrifuged at 1500 rpm at room temperature. The cells were then resuspended in Lysis Buffer (5 mM Tris base 1 mM EDTA with protease inhibitors, from Mini Complete protease inhibitor, Roche) and the cell membrane lysed mechanically with multiple passages (15–20) into a G 22 needle on a 1 mL syringe. The percentage of cells with broken cell membrane was evaluated on an optical microscope after small 2014, DOI: 10.1002/smll.201303841
staining with Trypan Blue: the mechanical lysis was performed until most of the cells (around 80%) were stained by the dye. Prior to organelle isolation, the cell lysate was spun at 1500 for 3 minutes at 4 °C in order to pellet and discard unbroken cells. To recover the organelles containing the magnetic particles, the cells lysate was then incubated on a magnet (a F999S-N42 Neodymium Magnet acquired from first4magnets.uk; see Supporting Information Figure S11 for magnet geometry) overnight at 4 °C. The supernatant was then discarded, while the pellet was resuspended in 500 μL of lysis buffer containing protease inhibitors (Roche MiniPrep). In order to recover the nanoparticles from the isolated organelles, the same method was used to perform a second magnetic extraction, after the lysis of the organellar memebrane (performed following the same method). The magnetic extraction is illustrated in Figure 1. Isolation of Lysosomes via Centrifugation in a Density Gradient: In order to allow comparison with the fractions obtained by magnetic extraction and validate the method, lysosomes were isolated using a Lysosome Isolation Kit (Sigma).[36] After lysis of the cell membrane performed as described above, in order to recover the first the crude lysosomal fraction (CLF), the cell lysates were centrifuged two times. These centrifugations ensure the separation of eventual intact cells, large cell debris and nuclei (first centrifugation: 10 minutes at 1000 g; the CLF is contained in the supernatant), and also allow to separate the CLF from other smaller cellular compartments such as endoplasmic reticulum and cytoplasm (second centrifugation: 20 minutes at 20 000 g; the CLF is contained in the pellet). In this way the final CLF contains mainly lysosomes. The CLF was then resuspended in a solution containing 19% Optiprep Density Gradient Medium to a final protein concentration of 0.5–1.0 mg-protein/mL. The lysosomes were then further purified with a separation by density gradient centrifugation (150 000 g for 4 hours at 4 °C) in order to obtain 6 fractions, separated through the gradient according to their density. Western blot of LAMP1 was used to confirm lysosomal enrichment between the 3rd and 4th fractions, as per instructions (Supporting Information Figure S6). Protein Corona Isolation: In order to isolate hard corona nanoparticle complexes, nanoparticles were incubated in the MEM medium used for cell culture, supplemented with 10% fetal bovine serum (complete medium, cMEM) at different concentrations (500, 250, 125 µg/mL NPs) for 20 minutes at 37 °C. The samples were then centrifuged for 20 minutes at 20 krcf at 4 °C to pellet the particle-protein complexes and separate them from the supernatant. The pellet was resuspended in 500 μL of PBS and centrifuged again in the same way for a total of three washing-steps, before resuspension to the desired concentration in PBS. This treatment allowed us to remove the proteins with low affinity for the nanoparticle surface and isolate hard corona nanoparticle complexes.[34] The same protocol was applied to isolate the hard corona nanoparticles from the extracted subcellular organelles after a second magnetic extraction. BCA Assay: In order to determine the protein concentration in the isolated organelle fractions, the BCA assay (BCA Protein Assay Reagent, Pierce) was performed according to Manufacturer’s instructions. Briefly the samples were incubated with the BCA protein assay reagents in a 96 well plate for 30 minutes at 37 °C and with 1% Triton X to ensure the lysis of the isolated organelles; the
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
7
full papers
F. Bertoli et al.
absorbance at 562 nm was then read on a VarioSkan Flash plate reader (Thermo, USA). SDS-PAGE and Western Blotting: SDS-PAGE was used to separate and characterize the proteins contained in the isolated coronananoparticle complexes and in the recovered organelle fractions. Samples were diluted in protein loading buffer (62.5 mM Tris-HCL pH 6.8, 2% (w/v) SDS, 10% glycerol, 0.04M DTT and 0.01% (w/v) bromophenol blue), and incubated for 5 minutes at 100 °C. After necessary dilutions in PBS to normalize the protein content among different samples recovered, an equal sample volume (thus same total protein concentration) was loaded for each well in a 10% gel polyacrylamide gel. This was performed on all gels except for the gels present in Figure 4a-b and in Figure 5d, in which equal sample volume (without protein normalization) was loaded. Gel electrophoresis was performed at 120V, 400 mA for about 60 minutes, until the proteins neared the end of the gel. To detect the separated proteins, the gels were stained for 1 hour in Coomassie blue staining (50% methanol, 10% acetic acid, 2.5% (w/v) brilliant blue) and de-stained overnight in 50% methanol, 10% acetic acid. When higher sensitivity of detection was required, gels were stained using the silver staining Daiichi kit according to protocol given by the supplier. Gels were scanned using a Biorad GS-800 calibrated densitometer scanner. For the western blotting procedure, after separation of the recovered proteins by SDS-PAGE, the proteins were transferred to a nitro cellulose membrane at 100 mV for 1 h using a wet transfer method. Then, the membrane was incubated with a 2% BSA blocking solution in PBS TWEEN 0,01% for 1 h at room temperature. The membrane was then incubated with the antibody of interest (LAMP1, Clathrin heavy chain, Caveolin 1, EEA1; monoclonal, mouse, from Abcam) in blocking solution (1:4000; 1:2000 depending on the antibody used) overnight at 4 °C. Finally, the membrane was washed 4–5 times with PBS TWEEN 0.01% and incubated with the secondary antibody (anti mouse HRP, Abcam) at a dilution of 1:4000 in blocking solution for 1 h. The membrane was then washed with PBS TWEEN 0.01% for 4 times and incubated with the substrate solution for the chemiluminescent reaction (ECL PIERCE) for 1 minute and finally developed on X-Ray film in a dark room. Mass Spectrometry: In order to identify the recovered proteins by mass spectrometry analysis, after separation by SDS-PAGE performed as described above, the bands of interest were excised from the gel and digested in-gel with trypsin (porcine trypsine, Promega), according to the method of Shevchenko et al.[38] The resulting peptide mixtures were resuspended in 0.1% formic acid and analyzed by electrospray liquid chromatography mass spectrometry (LC MS/MS) using an HPLC (Surveyor, ThermoFinnigan, CA) interfaced with an LTQ Orbitrap (ThermoFinnigan, CA). Chromatography buffer solutions (Buffer A, 0.1% formic acid; Buffer B, 100% acetonitrile and 0.1% formic acid) were run using a 72 minute gradient. A 150 μL/min flow rate was used at the electrospray source. Spectra were analysed with Bioworks Browser 3.3.1 SP1 (ThermoFisher Scientific) using Sequest Uniprot/Swiss-Prot database (www.expasy.org). An exclusion filter was applied to reduce false positives, where peptides with P
