Long-term live-cell microscopy with labeled nanobodies delivered by laser-induced photoporation

Abstract

Fluorescence microscopy is the method of choice for studying intracellular dynamics. However, its success depends on the availability of specific and stable markers. A prominent example of markers that are rapidly gaining interest are nanobodies (Nbs, ~ 15 kDa), which can be functionalized with bright and photostable organic fluorophores. Due to their relatively small size and high specificity, Nbs offer great potential for high-quality long-term subcellular imaging, but suffer from the fact that they cannot spontaneously cross the plasma membrane of live cells. We have recently discovered that laser-induced photoporation is well suited to deliver extrinsic labels to living cells without compromising their viability. Being a laser-based technology, it is readily compatible with light microscopy and the typical cell recipients used for that. Spurred by these promising initial results, we demonstrate here for the first time successful long-term imaging of specific subcellular structures with labeled nanobodies in living cells. We illustrate this using Nbs that target GFP/YFP-protein constructs accessible in the cytoplasm, actin-bundling protein Fascin, and the histone H2A/H2B heterodimers. With an efficiency of more than 80% labeled cells and minimal toxicity (~ 2%), photoporation proved to be an excellent intracellular delivery method for Nbs. Time-lapse microscopy revealed that cell division rate and migration remained unaffected, confirming excellent cell viability and functionality. We conclude that laser-induced photoporation labeled Nbs can be easily delivered into living cells, laying the foundation for further development of a broad range of Nbs with intracellular targets as a toolbox for long-term live-cell microscopy.

1. Introduction

Studying intracellular dynamic processes is of great value for understanding the biological mechanisms that govern living cells. Long-term microscopy imaging of living cells labeled with fluorescent stains is crucial in that regard [1]. The fluorescent labels should be bright and photostable with the ability to target subcellular structures with high specificity. Even though fluorescent proteins (FPs) have proven tremendously useful for live-cell microscopy in the past 2 decades, they come with some drawbacks, such as the potential for an altered cell phenotype due to protein overexpression, as well as their relatively low brightness, photostability and limited spectral range [2]. Instead, extrinsic fluorescent labels, such as organic dyes or quantum dots (QDs), can be designed to have better photostability and brightness, are available in a wide range of colors, and can be coupled to targeting moieties for specific labeling of subcellular structures [3–5].

Unfortunately, many of the existing high-quality extrinsic labels cannot spontaneously permeate through the cell membrane of living cells, especially after conjugation to a suitable targeting ligand. For instance, although organic dye functionalized antibodies have been used successfully for many years in fixed and permeabilized cells, these cannot be used for the intracellular labeling of living cells as they cannot permeate through the cell membrane [5, 6]. While there are ongoing efforts to engineer cell-permeable organic labels for different applications [7, 8], the use of a biocompatible method to deliver already existing high-quality labels in living cells would be a game-changer for the live-cell imaging field.

Intracellular delivery methods have been frequently investigated for this purpose, often adapted from the drug delivery or cell transfection field. One approach is the use of chemical transfection agents, which typically associate with the labels to form nanoparticles that can be internalized into cells by endocytosis [9, 10]. Unfortunately, none of the currently available transfection agents is 100% effective in letting their cargo escape from the endosomes, meaning that at best a confounding staining is obtained of endosomes on the one hand and the subcellular structure of interest on the other hand. Methods that allow direct delivery into the cytosol are, therefore, preferred. A first example is the pore-forming bacterial toxin streptolysin O (SLO) [11], which was used to deliver external fluorophores into the cytoplasm, ranging from small organic dyes to labeled antibodies. The formed pores, however, are limited to molecules of 250 kDa and the method is difficult to optimize for different cell types because it requires a delicate balance of the treatment time to keep the pores open without causing too much toxicity. Microfluidic cell squeezing is another example [12, 13], for which labeling of subcellular structures with labeled nanobodies and small cell impermeable dyes was demonstrated. However, it requires cells to be in suspension, limiting its practical use for imaging which is often performed on adherent cells. Another example is the photothermal nanoblade, which was used to deliver QDs into selected individual living cells [14]. As it is essentially based on the principle of micro-injection, it can deliver a wide range of compounds into cells but the method has very limited throughput. Thus, there is still a need for a fast, efficient, and universal permeabilization method for living cell labeling, preferably compatible with standard growth conditions and recipients for microscopy (e.g. glass slides, petri dishes, and multi-well plates).

Laser-induced photoporation is an upcoming new intracellular delivery method which fulfills these requirements [15, 16]. In recent work we gave a proof of concept that a range of cell impermeable probes could be easily delivered into the cytoplasm of living cells and stain their target structure [17]. In particular, we gave an example of staining an intracellular target protein (vimentin) with a labeled nanobody (Nb). Nbs are fragments derived from the antibody (Ab) heavy chain with the ability to specifically bind to antigens. Being only ~ 15 kDa, Nbs are ten times smaller than conventional Abs, making them quite interesting for the labeling of subcellular targets [18, 19], especially for superresolution imaging [20]. Spurred by this positive initial result, here we further explored the suitability of various Nbs delivered by photoporation for long-term time-lapse microscopy imaging of subcellular structures in living cells [17]. In particular, we demonstrate successful long-term imaging of living HeLa cells with three different nanobodies. Two were commercially obtained, one Nb targeting to the GFP and one Nb targeting to the histone labeled with ATTO647N and ATTO488, respectively. The third one was an in-house developed Alexa Fluor® 488 labeled fascin Nb for targeting the fascin protein in filopodia and microspikes. Time-lapse confocal microscopy images were recorded over 24 h, showing cell division, migration and the dynamics of the respective subcellular structures. Even after 72 h, the labeling quality remained excellent. Combined with the cell division time, cell migration and mitochondrial morphology remaining unaffected, these results convincingly demonstrate for the first time the possibility of using labeled Nbs for high-quality long-term microscopy imaging of living cells.

2. Results

2.1. Intracellular delivery of Nbs by laser-induced photoporation

Nbs will be delivered into living cells by laser-induced photoporation, which makes use of photothermal NP that after attachment to the cell membrane can temporarily increase its permeability upon laser irradiation. Here we used grapheme quantum dot (GQD) as photothermal sensitizers, which we have demonstrated before that they are very well suited for this purpose [17]. To improve their colloidal stability and produce more uniform photothermal effects, GQDs were functionalized with polyethylene glycol (PEG) (Fig. S1(a) in the Electronic Supplementary Material (ESM)). PEGylation did not substantially alter the ultraviolet–visible (UV–VIS) absorption spectra (Fig. S1(b) in the ESM). Atomic Force Microscope (AFM), showed that QGD had a thickness of ~ 4 nm and a lateral diameter of 20 to 30 nm (Figs. S1(c) and S1(d) in the ESM). This was further confirmed by dynamic light scattering (DLS), which showed a hydrodynamic size of around 20 nm as well (Fig. S1(e) in the ESM). While the size of GQD and GQD-PEG were virtually the same, a significant increase in zeta potential from −35 to −12.4 mV was found after PEGylation, indicative of successful PEGylation (Fig. S1(e) in the ESM).

For photoporation, GQDs-PEG were dispersed in the cell medium of cells grown on a microscopy compatible substrate and incubated with the cells for 30 min to allow them to interact with the cell membrane, as illustrated in Fig. 1. Next, unbound GQD-PEG is washed and new cell medium to which 40 μg/mL Nb in DPBS is added as well. The sample is then scanned through the pulsed laser beam (7 ns, λ = 561 nm) in such a way that essentially a single laser pulse is applied at every location. The laser pulse fluence at the sample was 2.6 J/cm², which is twice the VNB formation threshold of GQD-PEG. This ensures that virtually all GQD-PEG nanoparticles will effectively form VNBs, whose physical force will generate small transient pores in the cell membrane. While the pores typically reseal in less than one minute, it gives the Nbs sufficient time to diffuse from the cell culture medium into the cytoplasm and stain their target structure [16, 21].

Figure 1.

Schematic illustration of living cell labeling by laser-induced photoporation. Cells are first cultured on a microscopy compatible substrate and incubated with photothermal pegylated graphene quantum dots (GQD-PEG) for 30 min to allow them to interact with the cell membrane. Next, the fluorescent probe of interest is added to the cell culture medium and the cells are irradiated with pulsed laser light. VNBs are formed around the GQD-PEG whose physical force forms transient holes in the cell membrane through which the fluorescent probes can diffuse into the cells. After washing and adding fresh cell culture medium, the cells are labeled and ready for imaging.

The cytotoxicity of GQD-PEG and laser-induced photoporation were studied using the CellTiter-Glo® assay. It demonstrated that GQD-PEG by itself did not induce any noticeable cytotoxicity, while laser irradiation and VNB formation induced toxicity to the cells in a concentration-dependent manner, as is to be expected (Fig. S2 in the ESM). Since a commonly used rule of thumb is to select conditions with > 80% cell viability, we selected a concentration of 5.1×10⁹ nanoparticles/mL GQD-PEG for all further Nb delivery experiments.

2.2. GFP Nb enhanced cell labeling for long-term microscopy imaging of mitochondrial dynamics

As a first example, we selected an anti-GFP Nb that can target proteins fused with GFP (or YFP). The commercial anti-GFP Nb (GFP Nb in short) that we selected was labeled with ATTO647N, and can be used to mitigate the limited brightness and photostability of GFP, which is especially useful for long-term imaging. At the same time, it retains the benefit of being able to use genetic engineering for labeling specific proteins, especially since Nb labeling technology is still fairly new and only a limited range of Nb with the intracellular target are currently available. The target that we chose is mitofusin, a mitochondrial protein that mediates the fusion of mitochondria. This target is interesting from a validation point of view since mitochondria become fragmented if the labeling method would interfere with its functionality.

HeLa cells were first transfected with YFP-coupled mitofusin2 (Mfn2) encoding plasmid DNA (pDNA) by nucleofection as illustrated in Fig. 2(a). Note that GFP Nb is not only able to bind to GFP but to YFP as well [22]. 24 h after transfection, the GFP Nb was delivered into living cells by photoporation. After washing, cells were brought to the microscope for imaging. As a control, Mfn2 YFP labeled HeLa cells were incubated with GFP Nb for 2 min (which is the same time as the photoporation process) to check for spontaneous uptake without laser treatment. As expected, the Nb were not able to enter the cytoplasm spontaneously (Fig. 2(b)). Confocal microscopy images revealed strong colocalization between the YFP (displayed in green) and Nb (magenta) signal (Figs. 2(b) and 2(c)). Colocalization was further quantified as the Pearson’s R value, Li’s intensity correlation quotient (ICQ) and Costes P value [23–25]. For perfect colocalization, the Pearson’s R value approaches 1, while this is 0.5 for the ICQ value. A Costes P value of 1 means that none of the randomized images had better correlation, while anything lower than 0.95 means that the colocalization in the real images is not likely to be better than random chance. Based on the Pearson’s P value, cells labeled with both Mfn2-YFP and GFP Nb exhibited significantly more colocalization than control cells with only one of both labels (Fig. 2(d)), of which some exemplary images together with the Li’s ICQ value and Costes P value are shown in Fig. S3 in the ESM.

Figure 2.

Labeling living Hela cells with GFP Nb targeted to Mfn2-YFP. (a) Schematic representation of the experimental set-up. HeLa cells are first transiently transfected with Mfn2-YFP after which they are additionally labeled with GFP Nb by photoporation. (b) Representative confocal microscopy images of living HeLa cells showing the YFP signal (green) and the Nb signal (magenta). Mfn2-YFP expressed cells are incubated with GFP Nb as control (first row) and are co-labeled with GFP Nb delivered by photoporation (middle and last rows). The scale bar is 50 μm. The scale bar is 50 μm. (c) Fluorescence intensity profile along the dashed line in (b). (d) Quantification of colocalization with Pearson’s R value based on 10 cells. Two-way ANOVA was used to compare the groups; ****P < 0.0001.

Next, long-term live-cell imaging (24 h with 20 min time intervals) was performed of GFP Nb labeled cells. As can be seen in Fig. 3 and Movie ESM1, imaging cells over 24 h was very well possible, with frequent cell divisions being observed. The white and red arrows in Fig. 3 indicate two subsequent cell divisions, showing that the GFP Nb label is nicely redistributed over the daughter cells. While cell division is a first hallmark sign of good cell viability, we investigated if the Nb labeling did not interfere with the mitochondrial fusion/fission dynamics.

Figure 3.

Time-lapse microscopy imaging of Mfn2-YFP HeLa cells labeled with GFP Nb by photoporation. Confocal images were recorded of HeLa cells every 20 min for 24 h. Selected image frames are shown in which two cells are dividing subsequently, as indicated by the white and yellow arrowheads. The scale bar is 50 μm.

Mitochondrial morphology was quantified from confocal microscopy images of cells labeled with only Mfn2-YFP and cells labeled with both Mfn2-YFP and GFP Nb at t = 0 h, t = 12 h, and t = 24 h. The ImageJ plugin Mitochondria Analyzer was used to quantify multiple morphological parameters [26]. The analysis was performed on the YFP images in both cases because the Nb signal is much brighter and may introduce a bias in the analysis. Four different morphological parameters were determined, including the mean area (Fig. 4(a)), the form factor (Fig. 4(b)), the number of branches (Fig. 4(c)) and the branch length (Fig. 4(d)). Statistical analysis revealed that there is no significant difference for any of the parameters at any of the time points after GFP Nb staining by photoporation. This confirms that neither the photoporation procedure itself nor the GFP Nb staining, has a noticeable effect on mitochondria and can be used for reliable long-term imaging of mitochondrial dynamics.

Figure 4.

Quantification of mitochondrial morphology with and without GFP Nb staining by photoporation. (a)–(d) The morphology of mitochondria was quantified in HeLa cells expressing Mfn2-YFP with (Mfn2-YFP2) and without (Mfn2-YFP1) GFP Nb co-staining at three different time points (0, 12, and 24 h) using the ImageJ plugin Mitochodria Analyzer. The quantified parameters are (a) mean area, (b) form factor, (c) branches per mitochondria, and (d) branch length. Quantification was performed on YFP images from 15 cells in each case. Two-way ANOVA was used to compare the groups; ns = not significant.

2.3. Fascin Nb labeled living cells for long-term microscopy imaging

Based on the positive results with the GFP Nb targeted to the mitochondria, which are readily accessible in the cytoplasm, we proceeded with trying to target a finer intracellular structure. Therefore, as a second case study, we labeled filopodia and microspikes in HeLa cells with an in-house developed Alexa Fluor® 488 labeled Fascin Nb (Fascin Nb in short), again delivered by photoporation. Filopodia are thin cell protrusions containing actin and protruding from the plasma membrane. Playing an important role in cell adhesion, migration and invasion, they receive increasing attention as a promising prognostic biomarker of metastatic disease [27, 28]. During photoporation, alongside the Fascin Nb we co-delivered (cell impermeable) phalloidin Alexa Fluor®647 into the cells to label the F-actin. Fascin Nb (green) and phalloidin (magenta) colocalized well in the filopodia (Fig. 5). This is clearly visible from the fluorescence profiles in Fig. 5(b). A three-dimensional (3D) confocal image is shown in Fig. S4 in the ESM as well, showing the distribution of microspikes around the cell surface (Movie ESM2) [29]. In addition to normal confocal microscopy, we also used Airyscan superresolution imaging and total internal reflection fluorescence (TIRF) microscopy to better resolve the fine structured filopodia and microspikes (Fig. 5(c)). TIRF microscopy imaging showed with a high contrast of the structures near the bottom of the cells where they attach to the cell culture substrate. The dynamics of the finger-like filopodia can be clearly seen in the TIRF time-lapse recordings (Movie ESM3).

Figure 5.

Fluorescence microscopy imaging of living HeLa cells labeled with Fascin Nb (green) and phalloidin Alexa Fluor® 647 (magenta). (a) Confocal microscopy images of HeLa cells labeled with Fascin Nb (left), phalloidin Alexa Fluor® 647 (middle) and the merged image (right). (b) Fluorescence intensity profiles are plotted along the dashed lines across the cell in (a). (c) Different microscopy images of HeLa cells labeled with fascin Nb. From left to right: confocal microscopy, Airyscan superresolution imaging, and TIRF microscopy. The scale bar is 10 μm.

Next, 24 h time-lapse confocal microscopy was performed at 25 min intervals (Fig. 6 and Movie ESM4). Upon cell division, Fascin-Nb was redistributed equally into the daughter cells and labeling the filopodia and microspikes. Examples of dividing cells can be seen in Figs. 6(a) (yellow arrows) and 6(b) (cyan arrows). Interestingly, one can see sharp thin and finger-like filopodia along the plasma membrane in the direction of migration, forming the initial adhesion sites (magenta arrows, Fig. 6(b)) [30]. This is expected because the filopodia are “scouting” the microenvironment as cells migrate [31]. On the opposite side of the cell, long and thin filopodia can be seen which communicates to other cells through the cell-cell junctions (magenta arrows). It is the second confirmation that living cells labeled with Nb delivered by photoporation can be imaged over prolonged periods of time.

Figure 6.

Time-lapse microscopy imaging of HeLa cells in which filopodia are stained with Fascin Nb by photoporaion. (a) A selected panel of time-lapse images show cell division during long-term imaging, as indicated by the yellow arrows. (b) A selected panel of time-lapse images shows the dynamics of filopodia during cell migration. For the cell in the lower-left corner, the direction of movement is indicated by the white arrow. Finger-like small filopodia are formed along the cell membrane in the direction of movement, as indicated by the yellow arrows. Long filopodia on the other side of the cell connect to neighboring cells for intercellular communication, as indicated by magenta arrows. The cell indicated by the cyan arrows in here also shows cell division during the long-term imaging. The scale bars are 50 μm.

Since filopodia are involved in cell migration, we additionally performed a scratch wound healing assay to verify that cell migration was not affected by labeling with Fascin Nb [32]. Untreated cells were compared to cells treated by photoporation without Nb and cells labeled with Fascin Nb by photoporation. In all three cases, the scratch wound healed after 12 h (Fig. 7(a)). The wound area were quantified as the normalized percentage area over time (Fig. 7(b)). The slopes of the lines fitted to those three data sets were not significantly different, confirming that labeling with fascin Nb did not affect cell migration.

Figure 7.

Scratch wound healing of HeLa cells labeled with Fascin Nb by photoporation. (a) Representative confocal microscopy images of HeLa cells are shown at t = 0 h, t = 4 h, t = 8 h, and t = 12 h. From top to down bottom are transmission images of untreated cells (upper row), transmission images of cells after photoporation without Nb (middle row), and fluorescent images of cells labeled with Fascin Nb by photoporation (lower row). The wound area is delineated by the yellow lines. The scale bars are 20 μm. (b) Quantification of the wound area over time. The area is normalized to 100% at t = 0 h in each condition. A linear function (dashed line) is fitted to the experimental data (exp). Error bars are the standard deviation from 3 independent experiments.

2.4. Histone Nb intracellular labeling for long-term microscopy imaging

As a third case study, we photoporated HeLa cells with Alexa Fluor® 488 conjugated Nbs that target histone H2A/H2B heterodimers (Histone Nb in short). Even though photoporation delivers the Nb into the cytoplasm, due to their relatively small size (15 kDa), it is expected that the Histone Nb could reach the nucleus by passive diffusion through the nucleus pore complexes (NPC) which have a cut-off size of approximately 60 kDa [33]. The confocal microscopy images in Figs. 8(a) and 8(b) indeed confirm that HeLa cell nuclei could effectively be labeled with Histone Nb after photoporation. Cells were additionally labeled with the living cell nucleus stain Hoechst, showing excellent colocalization with the Histone Nb with 86% of the Hoechst labeled cells being positive for the Histone Nb (Fig. 8(c)). After photoporation, cells were furthermore incubated with TO-PRO-3, a DNA intercalating dye that will only enter dead cells. The image analysis showed that only 2% of the cells were positive for the TO-PRO-3 stain, confirming excellent cell viability after photoporation (Fig. 8(c)).

Figure 8.

Living HeLa cells labeled with Histone Nb and co-stained with Hoechst and TO-PRO-3. Histone Nb was delivered into HeLa cells by photoporation, after which cells were stained with Hoechst (live cell nuclear stain) and TO-PRO-3 (dead cell stain). (a) Representative confocal images showing the various stains with low magnification and a large field of view (scale bar = 200 μm). The dashed rectangle indicates the area that is shown with higher magnification in (b) (scale bar = 50 μm). (c) Histone Nb delivery efficiency and cell toxicity induced by photoporation based on imaging analysis.

To further underscore the need for a more performant Nb delivery technology like photoporation, we performed a comparison with a commercial transfection agent (PULSin) that is marketed for efficient cytosolic delivery of proteins and peptides, including Abs, which are even much larger than the Nbs used in this study. Following the manufacturer’s instructions, the Nb was first mixed with the PULSin solution and then added into the cells, as schematically shown in Fig. S5(a) in the ESM. After 4 h incubation, cells were washed and supplemented with fresh cell medium, after which they were ready for imaging. While many cells exhibited green fluorescence from the Histone Nb, the intracellular staining pattern was very inhomogeneous (Fig. S5(b) in the ESM). Crucially, only very few cells showed nucleus Nb staining, while most of the signal was associated with a dotted pattern in the cytoplasm, indicating that most of the PULSin-Nb complexes were still present in endosomes. As already pointed out in the Introduction, this is a known problem of transfection agents [34, 35]. In addition, we also observed toxicity effects, since the Histone Nb positive cells were also positive for TO-PRO-3. This corroborates the current need for new efficient and non-toxic technologies, such as photoporation, for delivering cell-impermeable labels into living cells.

Having established that cells can be efficiently and safely labeled with Histone Nb by photoporation, long-term imaging was subsequently performed with time-lapse confocal microscopy. Cells were imaged for 24 h with images being recorded every 20 min. As can be appreciated from the representative Movie ESM5, HeLa cells stained with Histone Nb appeared active and healthy during the entire period. Since the HeLa cell division cycle is ~ 24 h, cell mitosis was frequently observed as well, which is a hallmark of healthy viable cells. An example is shown in Fig. 9 where two cells are undergoing mitosis consecutively, as indicated by the white and yellow arrows. Importantly, the Histone Nb was always transferred to both daughter cells, whose nuclei remained positive for the staining.

Figure 9.

Time-lapse microscopy imaging of HeLa cells labeled with Histone Nb by photoporation. Confocal microscopy images were acquired of HeLa cells every 20 min for 24 h. Selected image frames are shown in which two cells are dividing subsequently, as indicated by the white and yellow arrowheads. The scale bar is 50 μm.

To explore whether the Nb labeling is stable in living cells for even longer times, confocal microscopy imaging was performed up to 72 h. In this experiment, HeLa cells were trypsinized and reseeded every day to compensate for cell density. We visibly observed that Histone Nb signal gradually decreased over time (Figs. 10(a) and 10(b)), which was confirmed by intensity measurements (Fig. 10(c)) and consequently leads to a decrease in number of labeled cells as such (Fig. 10(d)). This decrease is due to dilution of the Nb labels over daughter cells upon sequential cell divisions, which is unavoidable for externally added dyes. Note that this decrease is not due to cell death or photobleaching because an exponential fit to the data in Fig. 4(c) revealed a decay time τ = 23.7 h, which is precisely the expected 24 h cell division time of HeLa cells and another confirmation shows that neither the photoporation nor the labeling affects the cell’s natural functioning [36].

Figure 10.

Long-term microscopy imaging of Histone Nb labeled HeLa cells. (a) HeLa cells labeled with Histone Nb by photoporation were imaged at t = 0, 24, 48 and 72 h. The contrast in the 48 and 72 h images was two-fold enhanced to visually compensate for a decrease in Nb staining due to label dilution by cell division. The scale bare is 200 μm. (b) Higher magnification images of the areas indicated with a dashed rectangle in a. Scale bar is 50 μm. (c) The normalized mean fluorescence intensity (nMFI) is shown in function of time of cells labeled with fluorescent Histone Nb by photoporation. The green line represents the exponential fit (Exp fit). (d) Quantification of the normalized percentage of positive cells over time.

3. Discussion

Imaging intracellular dynamics helps to deepen our understanding of cell functioning. It requires high-quality labels that allow imaging specific subcellular structures with high contrast for extended time periods. Nanobodies, conjugated to high-quality fluorescent labels, are a new class of targeted probes that fit these requirements. Due to their size, however, they require suitable delivery technology to get them across the plasma membrane of living cells. This delivery problem was circumvented by Rothbauer et al. who used genetically expressed fusion constructs of Nbs with fluorescent proteins, also known as chromobodies [18]. They imaged cells expressing chromobodies for 10 h to follow mitotic events, giving the very first proof that Nbs may be used for long-term imaging of specifically stained structures in living cells. However, being genetically expressed fusion constructs with fluorescent proteins, chromobodies still suffer from limited photostability, spectral range and brightness, apart from the risk of inducing overexpression artifacts. Furthermore, the FPs are even bigger (> 25 kDa) compared with the Nb itself that can impede the intracellular labeling [37]. Instead, Herce et al. designed cell-permeable Nbs by combining them with cell-penetrating peptides [38]. They showed that Nbs could enter living cells and bind to the target antigens. Even though they did not perform extended time-lapse imaging, they showed that the label was still present after 24 h. While an elegant approach, it is clear that every different type of Nb would need to be specifically designed to become cell-permeable without affecting its binding affinity. A more general approach was presented by Klein et al., who delivered labeled Nbs into living cells by microfluidic cell squeezing [37]. This fairly new intracellular delivery technology offers tremendously high throughput in the order of a million cells per second with good cell viability. They showed that 20 h after cell squeezing the Nb labeling was still present. However, being a microfluidic technique, it requires cells to be in suspension, which complicates matters considering that most microscopy work is done on adherent cells.

Recently we demonstrated that laser-induced photoporation with GQD is well-suited for the intracellular delivery of cell-impermeable labels, including Nbs [17]. Conveniently, photoporation is compatible with cells grown in common imaging chambers and combines excellent cell viability with high labeling efficiencies. Building forth on those initial proof-of-concept experiments, here we demonstrate that fluorescently labeled Nbs can be reliably used for specific staining of subcellular structures, allowing long-term imaging for at least as long as 72 h. As a first commercially available Nb, we labeled Mfn2-YFP expressing cells with a Nb that can bind to YFP. The use of such a Nb targeted to fluorescent proteins is of interest to enhance the labeling intensity and photostability as compared to the fluorescent protein alone. It is also useful in cases where fusion constructs with the protein of interest are already available while a specific Nb is not. We found excellent colocalization of YFP with the Nb delivered by photoporation, allowing observing mitochondrial dynamics over a time-span of 24 h. Importantly, Nb labeling by photoporation did not disturb the mitochondria morphology which is a good indication that it is suitable to study mitochondrial dynamics over extended periods of time. As a second labeling case, we used an in-house developed Fascin Nb that is of interest to stain filopodia and microspikes. In filopodia, the fascin Nb colocalized well with actin stained by phalloidin, which was co-delivered by photoporation. Long-term time-lapse imaging nicely revealed that finger-like filopodia guided cell motion, while long filamentous filopodia on the other side of the cell remained in contact with neighboring cells, likely for intercellular communication [39]. To demonstrate that dye conjugated Nbs are suitable labels for super-resolution microscopy as well, we additionally performed Airyscan and TIRF microscopy, showing in great detail and with enhanced contrast the filopodia that are formed between the bottom of the cell and the underlying substrate. Importantly, a scratch wound healing assay demonstrated that neither the photoporation process nor the Fascin Nb labeling interfered with cell migration. As a third and final example, we tested a commercial Nb targeted to the histones in the nucleus. Thanks to its small size (15 kDa, ~ 2 nm), the Histone Nb could passively diffuse through the NPC and reach the nucleus. More than 80% of the cells were positive for the Histone Nb after photoporation with virtually no signs of acute toxicity (~ 2% To-Pro-3 positive cells). The reason why not 100% were successfully labeled is that we selected conditions (GQD-PEG concentration and laser fluence) to have > 80% cell viability after photoporation, which naturally means there is a trade-off with labeling efficiency. This is inherent to any delivery technology, which always induces some cytotoxicity since there is a manipulation of the cell involved. Excellent cell viability of cells was further confirmed by confocal time-lapse microscopy imaging over 24 h, showing frequent cell divisions as a hallmark for the well-being of cells. As expected for an extrinsic label, the Nb signal gradually diminished over time due to dilution over subsequent cell generations. Still, after 72 h (~ 3 cell division cycles) the labeling quality remained very good, indicating that the Histone Nb did not lose its target binding capacity.

Here we have used GQD-PEG as photothermal sensitizers during photoporation. Thanks to the outstanding photothermal property of graphene-based nanomaterials, water vapor bubbles can be formed around the GQD-PEG after pulsed laser irradiation to mechanically disrupt the cell membrane. Although currently there is not a direct way to measure the laser-induced local temperature of the GQD-PEG nanoparticles, this observational fact shows that temperatures above the water spinodal temperature must be reached [40]. This seems quite plausible since also in a previous report it was already demonstrated that temperatures of 400–500 °C could be reached in solid graphene oxide with millisecond pulsed laser irradiation [41]. For nanosecond laser pulses this can even easily go beyond 500 °C.[42]

Taken together, it is clear that photoporation is well-suited to deliver labeled nanobodies into the cytoplasm, where they can bind to their cytosolic or nuclear target proteins. We have demonstrated this possibility of HeLa cells in the present study, as it is one of the most common and best-characterized cell lines. Since photoporation was successfully carried out on many other cell types before [16, 43–45], it is expected that intracellular delivery of Nbs by laser-induced photoporation can be readily transferred to other cell types as well. At present, however, relatively few (intracellular) proteins have thus far been a target for nanobody development and much work remains to be done in this area to increase the number of available intracellularly targeted nanobodies. The advent of technologies like photoporation can instigate this development process and we argue that nanobodies, as high affinity and specific binders targeting enzymatic and non-enzymatic (structural) proteins alike, will no doubt reveal new insights into protein function in living cells, thus circumventing the requirement for expression modulation, which may perturb endogenous equilibria.

4. Conclusion

In conclusion, we have demonstrated that labeled Nbs can be reliably used for long-term microscopic observations of subcellular structures in living cells. Photoporation proves to be a well-suited method to deliver the Nbs into living cells with high efficiency and low toxicity, while offering the convenience of being readily compatible with common cell recipients. With the intracellular delivery issue being solved, future work can focus on expanding the available range of Nbs with intracellular target for high-quality and high-resolution microscopy.

5. Materials and methods

5.1. Cell culture

HeLa cells (ATCC® CCL-2™) were cultured in DMEM/F-12 (Gibco-Invitrogen) supplemented with 10% heat-inactivated fetal bovine serum (FBS, Biowest), 2 mM glutamine (Gibco-Invitrogen), and 100 U/mL penicillin/streptomycin (Gibco-Invitrogen). Cells were passaged using Dulbecco's Phosphate-Buffered Saline (DPBS) (Gibco-Invitrogen) and trypsin- ethylenediaminetetraacetic acid (EDTA) (0.25%, Gibco-Invitrogen). HeLa cells were cultivated in a humidified tissue culture incubator at 37 °C and 5% CO2. All cell culture products were purchased from Life Technology unless specifically stated otherwise.

5.2. GQDs and GQD-PEG synthesis method

The synthesis of GQDs was performed according to a previously reported procedure [17]. In brief, graphene oxide (GO) was synthesized by following a modified Hummer’s method as indicated in reference. Reduced graphene oxide (rGO) was synthesized by the treatment of GO with hydrazine monohydrate. To synthesize small GQD nanoparticles, 100 mg of rGO powder was dispersed in 100 mL of 30% H2O2 and ultrasonicated for 30 min. The obtained uniformly dispersed solution was kept refluxing for 12 h at 60 °C. The resulting solution was filtered and GQDs separated from porous reduced graphene. The obtained GQD suspension was further dialyzed to remove excess H₂O₂ and to separate rGO from small-sized GQDs. The purified GQDs were dissolved in deionized (DI) water to a final concentration of 1 mg/mL.

10 mg N-(3-Dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride (EDC, Sigma, Belgium) and 11.3 mg N-Hydroxysulfosuccinimide sodium salt (Sulfo-NHS, Sigma, Belgium) were added into 10 mL of the newly synthesized GQDs. The solution was stirred at room temperature for 30 min. Afterwards, 20 mg mPEG-amine (2 kD MW, Creative PEGWorks, US) was added under water bath sonication (BRANSON 2510) at room temperature for 2 h. 10 μL of mercaptoethanol was added into the mixture to quench the reaction. The solution was then transferred into the dialysis cassette 2000 MWCO (Thermo Scientific, Belgium) and dialyzed in DI water for 2 days.

5.3. UV–Vis spectra, DLS and NanoSight measurements

UV–Vis absorption spectra were measured by a NanoDrop 2000c Spectrophotometer (ThermoFisher Scientific). Hydrodynamic size and zeta potential were determined by DLS on a Zetasizer Nano ZS (Malvern, UK). The particle concentration of GQD-PEG was measured by Nanoparticle Tracking Analysis (NanoSight LM10, Malvern, UK) in reflection mode.

5.4. AFM measurement

GQDs dispersions (10 μL) were deposited onto muscovite mica (V-1 grade from SPI Supplies) by drop casting and dried at room temperature before imaging. The sample was imaged with a Bruker Dimension 3100 Model AFM (Veeco, Santa Barbara, CA). Rectangular single beam silicon cantilevers (Nanosensors PPP-NCHR) with tip radius and stiff ness of ~ 7 nm and ~ 40 N/m, respectively were used. All AFM images were acquired in tapping mode.

5.5. Cell viability assay by CellTiter Glo®

Hela cells were incubated with different concentrations of GQD-PEG for 30 min. Afterward, cells were washed with DPBS to remove the unbound NPs, followed by laser treatment. After a washing step, cells were put back in the incubator to allow for an initial recovery for 2 h. Before the cytotoxicity assay, the cell medium was replaced by 100 μL fresh CCM. 100 μL CellTiter Glo® solution (CellTiter-Glo® Luminescent Cell Viability Assay, Promega, Belgium) was added into each well. The plate was put on a shaker at 100 rpm for 10 min. The luminescence was measured by GloMax (GloMax® 96 Microplate Luminometer, Promega). All data is expressed as the mean ± SD (n = 3).

5.6. Nb information and fluorophore labeling

The Histone Nb (Histone-Label Atto488) and GFP Nb (GFP-Booster Atto647) are purchased from ChromoTek (Germany). Fascin Nb (FASNb2 in this work [46]) was site-specifically labeled with Alexa Fluor® 488 (AF488) using a click chemistry based procedure that will be described in more detail elsewhere (Hebbrecht et al., in preparation). Briefly, the Fascin Nb was produced and purified by IMAC as described before [47–53]. Afterward the CuAAC reaction was performed. The reaction mixture of 20–30 mM pAzF containing nanobody, 40–60 mM alkyne-AF488, 0.1 mM CuSO4, 1 mM THTPA and 7.5 mM sodium ascorbate and 10 mM HEPES was incubated for 60 min at 33 °C. Afterward, an EDTA buffer (20 mM HEPES, 500 μM EDTA, pH 7.4) was added into the mixture, followed by a purification step with an Amicon Ultra-0.5 Centrifugal Filter Unit (Sigma-Aldrich, St. Louis, MO, USA) and a PD Spintrap™ G-25 (Sigma-Aldrich).

5.7. Nucleofection

To transfect the cells with Mfn2 pDNA, HeLa cells were trypsinized and counted as described in the cell culture part. The nucleofection solution (SE Cell Line 4D-NucleofectorTM X Kit S, Lonza, Belgium) was prepared as the manufacturer indicates. 200000 cells were suspended in 20 μL nucleofection solution to which 0.6 μg Mfn2 pDNA was added. The mixture was then transferred into one well of the NucleocuvetteTM strip. The strip with closed lid was placed into the retainer of the 4D Nucleofector TM X unit (Lonza, Belgium). Nucleofection was performed with a program of “HeLa” recommended by the manufacturer for the transfection of HeLa cells. After nucleofection, the strip was removed from the retainer. 80 μL pre-warmed CCM was added into the well in the strip and mixed with the cell suspension. Cells were then transferred and cultured in the μ-slide or μ-dish (for next step of photoporation) for 24 h.

5.8. Nb intracellular delivery by laser-induced photoporation

A μ-slide (μ-Slide Angiogenesis, ibidi, Belgium) was seeded with 5,000 HeLa cells per well one day in advance for short-term microscopy imaging. For long-term microscopy imaging, HeLa cells were seeded at a density of 5,000 cells/well in an inserted 4 well μ-dish (micro-Insert 4 Well in μ-Dish 35 mm, high, ibidi, Belgium). Before laser treatment, cells were incubated with GQD-PEG suspended in CCM for 30 min. The labeled Nb was dispersed in DPBS at a concentration of around 40 μg/mL. 10 μL Nb solution was added into each well (both for μ-slide and inserted 4 well μ-dish) before the laser treatment. A 7-ns pulsed laser tuned to a wavelength of 561 nm (Opolette HE 355 LD, OPOTEK Inc., USA) was applied to excite the graphene-based materials. As the photoporation laser beam has a diameter of 150 μm, a scanning procedure was used to treat all cells within each well. The sample was scanned through the photoporation laser beam (20 Hz pulse frequency) using an electronic microscope stage (HLD117, Prior Scientific, USA). The scanning speed was 2.1 mm/s, and the distance between subsequent lines was 0.1 mm to ensure that each cell received at least one single laser pulse. The total scanning time per well was 2 min for μ-slide and 30 s for μ-dish. Afterward, the fluorophore solution was removed, and the cells were gently washed with DPBS and supplemented with fresh CCM. For cells in the μ-dish, the inserted well was removed by tweezers and 3 mL CCM was added into the dish for long-term microscopy imaging.

Next, cells in the μ-slide for short-term imaging were stained with Hoechst (nuclear stain, 33342, Life Technologies, Belgium) and TO-PROTM-3 (cell viability staining, O-PRO™-3 Iodide (642/661), Life Technologies, Belgium) diluted in CCM to a final concentration of 1 μg/mL and 0 .5 μM, respectively. 50 μL solution was added into each well. After incubation for 10 min at 37 °C, cells were washed once and supplemented with fresh CCM. After that, the cells are ready for imaging.

5.9. Nb intracellular delivery with PULSin transfection reagent

The commercial transfection reagent PULSin (Polyplus Transfection, France) was used for comparison, as it is marketed by the manufacturer for the intracellular labeling of living cells with antibodies. The Nb delivery protocol was based on the one provided by the manufacturer. In brief, HeLa cells were seeded in the μ-slide as described before. 2 μL Histone Nb (0.5–1 μg/μL, chromotek, Germany) was diluted in 100 μL HEPES buffer in a centrifugation tube to get a concentration of 20 mM. The tube was vortexed gently and centrifuged down. 5 μL Pulsin reagent was added to the tube and vortexed. The mixture was incubated at room temperature for 15 min. The prepared cells in the μ-slide were first washed with DPBS and then filled with 40 μL of opti-MEM (Gibco, Begium) per well. 10 μL of the mixture was added into each well. After 4 h incubation at 37 °C, the cells were washed once and the solution was replaced with CCM. Hoechst and TO-PROTM-3 staining were performed after the washing step as described above.

5.10. Confocal microscopy imaging and time-lapse microscopy imaging

Laser scanning confocal images were recorded on a Nikon C2 confocal laser scanning microscope or on a Nikon A1R confocal laser scanning microscope with a 60×/1.4 Plan Apo VC Oil immersion objective (CFI Plan Apo VC, Nikon). The following lasers were used for excitation: a 405 nm continuous wave laser (Melles Griot 56ICS/S2695) for Hoechst; a 488 nm continuous wave laser (Coherent Sapphire) for ATTO 488 and Alexa Fluor® 488; a 640 nm continuous wave laser (Melles Griot 56ICS/S2695) for Alexa Fluor® 647.

Time-lapse recordings were performed on a spinning disk confocal microscope (Nikon eclipse Ti-e inverted microscope, Nikon, Japan) equipped with an MLC 400 B laser box (Agilent technologies), a Yokogawa CSU-22 Spinning Disk scanner (Andor Technology, UK) and an iXon ultra EMCCD camera (Andor Technology, UK). HeLa cells were imaged in a stage-top cell incubator (37 °C with 5% CO2 supplied, Tokai Hit) for 24 h with a time interval of 20 or 25 min using a 60×/1.4 oil immersion objective lens (CFI Plan Apo VC 60 × oil, Nikon, Japan).

For long-term microscopy imaging up to 72 h, the HeLa cells were imaged after photoporation, recorded as t = 0 on the spinning disk confocal microscope. Then the cells were trypsinized and seeded with the same initial number. HeLa cells were imaged again at t = 24 h. Then the cells were trypsinized and seeded for another time to image at t = 48 h. A third time tripsining and seeding were performed for imaging at t = 72 h.

5.11. Colocalization analysis

To quantify the degree of colocalization between Mfn2 YFP and GFP Nb, the FIJI plugin Coloc2 was used [54]. First, images were processed by a background subtraction. Further analysis was performed on individual ROI’s within the cell. Mitochondrial regions were selected based on a manual intensity threshold. In these mitochondrial regions the pearson/Li coefficient was calculated with Coloc2 using Costes regression threshold.

5.12. Mitochondria morphology analysis

The mitochondria morphology was quantified with the FIJI plugin Mitochondria Analyzer [26]. First, a single cell was selected manually and the 2D threshold was applied to the image. The mean local threshold method was used with adjusting the block size and C-value. After setting the threshold, the 2D analysis was applied to acquire the morphology parameters.

5.13. Scratch wound healing assay

HeLa cells were seeded in the μ-slide for 5000 per well one day in advance. Photoporation without Fascin Nb and with Fascin Nb were performed as described before. After photoporation process, cells were washed with DPBS and supplemented with fresh cell medium. A 20–200 μL tip was used to scratch in the wells of untreated cells, photoporation without Nb cells, and Fascin Nb labeled cells. Confocal microscopy imaging was acquired on the spinning disk microscope with a 20× objective (Plan Fluor, Nikon, Japan) at t = 0 h, t = 4 h, t = 8 h, and t = 12 h at the same location.

The imaging analysis was processed with FIJI ImageJ. The scratch area was gated manually with the freehand selections. Then the area was counted by Measure in ImageJ.

5.14. Fluorescence intensity quantification of Histone Nb labeled cells

FIJI ImageJ was used to quantify the number of positive cells and the fluorescence intensity of Histone Nb labeled cells. 13 to 18 cells were quantified per data point of normalized mean fluorescence intensity (nMFI). The following exponential equation was used to obtain the cell division time τ:

$$\text{nMFI}=2^{\wedge }\left(-\left(t/\tau \right)\right)$$ (1)

where t is imaging time in this work. The number of positive cells was quantified from 64 images taken with a 60×/1.4 Plan Apo VC Oil immersion objective (CFI Plan Apo VC, Nikon).

5.15. Airyscan superresolution imaging

Images were collected on an LSM880 Airyscan (Carl Zeiss) system with a 63× DIC M27 objective (PlanApo NA: 1.4, oil immersion, Carl Zeiss) using the operating software ZEN blue 2.3. Alexa Fluor® 488 was excited with the 488 line of a multi Argon laser. The filter set BP 420-480 + BP 495-550 were used in conjunction with the Airyscan detector in the fast mode. A pixel reassignment algorithm and a 2D Wiener °Clter were carried out post-acquisition.

5.16. TIRF microscopy imaging

A total of 150,000 HeLa cells were seeded in a high glass bottom 35-mm μ-Dish (Ibidi) 24 h before imaging. Images were collected on a Zeiss Observer 1.1 microscope (Carl Zeiss) with a 100× TIRF oil immersion objective (PlanApo, NA: 1.46, Carl Zeiss). Alexa Fluor® 488 was visualized with a 488 nm diode laser at an incident angle of −70°, which allows selective excitation of molecules within ~ 80 nm of the cover glass. The filter set 77 HE GFP/mRFP/Alexa 633 (Carl Zeiss) was used in conjunction with an EMCCD Image MX2 camera (Hamamatsu). Image acquisition was performed at an exposure time of 200 msec and an EM gain of 100.

Supplementary Material

Supplementary material ( further details of the characterization of GQD, cell viabil ity of photoporation with GQD PEG, image analysis of colocalization of Mfn2 YFP with GFP Nb, 3D projection of Fasin Nb labeled HeLa cell, control experiment of PULSin® transfection agent used for the intracellular delivery of Histone Nb, and description of supplemented movies ) is available in the online version of this article at https://doi.org/10.1007/s12274-020-2633-z.