1. Introduction

There are large numbers of proteins in cytomembrane related to the cell activities. Optical imaging is a powerful tool for life science studies and biomedical applications due to its high resolution and high sensitivity. However, most of the protein molecules are not fluorescent and thus need to be labeled with extraneous optical probes. To efficiently image protein molecules requires the probes to be photostable, bright, continuously emissive and as small/light as possible. Fluorescent organic dyes and quantum dots (QDs) meet the size requirement but impose other shortcomings. Fluorescent organic dyes suffer from severe photobleaching and the signal-to-noise ratio (SNR) is usually reduced by the autofluorescent background [1,2 ]. QDs have a compatible size and efficient luminescence, but they have potential cytotoxicity as well as photoblinking [3,4 ]. As an alternative, UCNPs have been considered as a new choice of luminescence probe for biological applications [1,5 ]. UCNPs utilize sequential absorption of multiple photons to produce higher energy anti-Stokes luminescence. Compared with traditional luminescence probes, UNCPs have several advantages, such as no photo-bleaching and non-blinking, having high spatial resolution and low photo-toxicity [6]. Since excited by near infrared (NIR) light, they can penetrate deeply into bio-tissue with no autofluorescence of endogenous fluorophores. All these advantages make UCNPs have remarkable applications, including optical bioimaging, high-sensitive biosensing, multimodal imaging, photodynamic therapy, gene and drug delivery and in vivo animal theranostics [2,7 ].

In order not to disturb the biomolecules of interest, it is necessary to utilize a size/mass comparable probe to label them [8]. In principle, the cell components like the cytomembrane and other biological molecules are in a nanometric size, mainly below 5 nm [9]. For example, the typical thickness of a cell membrane is around 4 nm, and the diameter of the double helix structure of DNA is only 2 nm (Fig. 1(a) ). Therefore it is best to find bright and stable probes for imaging and not too large to disrupt the protein's function. However, there is still not an appropriate UCNP existing with such a small size and high brilliance. UCNPs smaller than 10 nm have recently been synthesized, and the sub-10-nm NaLuF₄ UCNPs have already been used in in vivo imaging [10,11 ]. A 3-nm BaGdF₅-based UCNPs was successfully synthesized [12], but they haven’t applied in bioimaging. It is also difficult to control the high temperature of oil-bath process in synthesizing the most common host matrix NaYF₄ to get a uniform, ultra-small UCNP. In all, the trade-off between small size and high luminance still exists. Compared with fluoride UCNPs, Ln³⁺-doped inorganic oxide exhibit better photo-stability, chemical stability and thermal stability because of their more rigid crystalline environment and higher lattice binding energy, which make it promising as a perfect sub-5-nm molecule probe [13]. Particularly, ZrO₂ crystal has low phonon energy and a high host absorption coefficient [14]. Ultra-small (below 5 nm) and uniform ZrO₂ nanocrystals can be synthesized through a highly repeatable solvothermal procedure [13,15,16 ].

 

Fig. 1 Size comparison and 3D structure: (a) Size comparison of ZrO₂-Ln³⁺nanoparticles and GFP(green fluorescent protein), three proteins—streptavidin (SA), maltose binding protein (MBP), GNRs, gold nanorods (about 20 nm in width and 80 nm in length), General NPs, about 20 nm, and NaYF₄ NPs, which are used in this paper (about 10 nm); ZrO₂ NDs (about 4 nm) and the magnified 3-D core-shell structure.

Download Full Size | PDF

Rapid scanning microscopy is always needed in the biospecimen imaging, especially for the three-dimensional (3D) imaging. However, the luminescence lifetime of UCNPs is intrinsically long, usually more than 100 μs [17], in which circumstances a long dwelling scanning time is needed to take a clear imaging. This drawback is not beneficial for rapid scanning microscopy imaging [18]. Thanks to the concentration of doped ions in the host can be tuned, energy transfer from the sensitizer and the activator varies as the sensitizer-activator distance and provides lifetime tunability. Lu Y et. al. obtained lifetimes varied from hundreds of μs to tens of μs with the 0.2%-4% doping of Tm³⁺ [19]. Here we tuned the concentration of Er³⁺ to decrease the lifetime.

An aptamer instead of antibody was used to increase the accuracy of targeting. The complex multi-layer of primary/secondary antibody modification greatly reduces the stability and increases the steric hindrance of the nanoprobes, thus impairing endocytosis [20]. On the contrary, aptamers are of small size and easy to be modified, which can effectively reduce the steric hindrance and simplify the modification steps of nanoprobes. In this work, we proposed to design the lanthanide-doped zirconium oxide (ZrO₂-Ln³⁺) upconversion nanodots (UCNDs). To the best of our knowledge, there are only a few reports about ZrO₂-based UCNPs except for some large and nonuniform in the literatures [21–23 ]. UCNDs with high concentration of activator were synthesized to implement the rapid scanning imaging as well as the 3D imaging. A specific aptamer was modified outside the streptavidin coated UCNDs and implemented the aptamer-mediated protein-targeted imaging.

2. Materials and method

2.1 Materials

Lanthanide oxides (Y₂O₃, Yb₂O₃ and Er₂O₃) were 99.99% purity or higher and purchased from Shanghai Yuelong New Materials Co., Ltd. Lanthanide chlorides (YbCl₃.6H₂O, YCl₃.6H₂O, ErCl₃.6H₂O) (99.9% purity), 1-octadecene (ODE) (technical grade, 90%), Nitrosonium Tetrafluoroborate (NOBF₄) (95%), polyallylamine hydrochloride (PAH) and polyacrylic acid (PAA) were purchased from Sigma-Aldrich. Zirconium (IV) propoxide propanol (70wt %), sodium oleate, and NH₄F were purchased from Aladdin Industrial Corporation. Trifluoroacetic acid (99%), Oleic acid (AR), ethyl alcohol (AR), hexane (AR), acetonitrile (AR) and benzyl alcohol (AR) were purchased from Sinopharm Chemical Reagent Co., China. The streptavidin was purchased from Sangon Biotech (Shanghai) Co.,Ltd. The biotin-aptamer was synthesized by Sangon Biotech (Shanghai) Co., Ltd. The aptamer sequences are as follows: Bio−Apt1, 5′-biotin-(C6)-TTG GTG GTGGTG GTT GTG GTGGTGGTGG-3′.Deionized (DI) water was used in all the experimental procedures.

2.2 Synthesis of hydrophilic UCNDs

Monodisperse ZrO₂−Ln³⁺UCNDs (Ln = Yb³⁺, Er³⁺) were synthesized according to thesolvothermal procedure reported in the literature [16]. Here, we take the procedure of ZrO₂:20% Yb³⁺, 2% Er³⁺ as an example. Briefly, Yb₂O₃ (0.2290 mmol) and Er₂O₃ (0.0229 mmol) were mixed and then solubilized in 6 mL of 50% aqueous trifluoroacetic acid at 85°C in a three-neck round-bottom flask. After about 1.5 hours, the solution became clear and the aqueous solution was evaporated to be dry in a rotary evaporator. Subsequently 10 mL benzyl alcohol was used to dissolve the solid, following the solution was transferred into a 50 mL Teflon-lined autoclave with 0.8 mL zirconium (IV) propoxide propanol solution added in. Ultrapure argon gas was poured into the Teflon-lined autoclave to exhaust air. Then it was heated at 230°C for 24 h in the oven. After the solution cooled to RT naturally, the resulting ZrO₂−Ln³⁺ UCNDs were precipitated by adding ethanol, collected by centrifugation at 10000 rpm for 5 min, washed with ethanol several times, and finally redispersed in hexane to be used in the next step. The NaYF₄: 2% Yb³⁺ shell was synthesized through a modified method in our previous work [24].

The hydrophilic process was conducted through a modified method [25]. In a typical phase transfer process, 2 mL UCNDs hexane dispersion was first combined with 2 mL acetonitrile to form a two-phase mixture, into which an appropriate amount of NOBF₄ was added. The resulting mixture was stirred until UCNDs were transferred from the upper hexane layer to the bottom acetonitrile layer (about 20 mins). The surface modified UCNDs were then purified by centrifugation, and the precipitated UCNDs were redispersed in 2 mL deionized water.

2.3 Surface functionalization of ZrO₂−Ln³⁺ UCNDs

First the PAA layer was modified onto the UCNDs, following the PAH layer. The methods of the two layers are the same. Briefly, 200 μL of the ZrO₂-Ln³⁺ core/shell UCNDs solution was diluted to 2 mL, and then 400 μL PAA (10mg/ mL) which was dissolved in 10 mMNaCl solution and 200 μL 10 mMNaCl solution were added. Then this mixture was revolved gently for 30 minutes, after which was centrifuged at a high-speed for 5 minutes. The precipitated UCNDs were redispersed in 2 mL deionized water, and a same process was conducted to modify a PAH layer. Then the polymer-coated UCNDs were connected with the streptavidin. The process was conducted through a modified method [26]. 80 μL streptavidin (1mg·mL⁻¹) was added into the pre-prepared PAH-PAA-ZrO₂−Ln³⁺ solution, which had been previously adjusted with 0.1 M K₂CO₃ and incubated using a rotary shaker at 37°C for 120 min. The mixture was then centrifuged at 10000 rpm for 15 min. After that, the sediment was washed and redispersed in phosphate-buffered saline (PBS) buffer solution to gain SA-UCNDs.

2.4 Cell culture and treatment

The cell culture method is as the same in the previous work [24]. For UCNDs labeling, the cells were rinsed gently with pre-warmed PBS and incubated with 200 μL 200 nMBio−Aptamer in DMEM for 30 min at 37°C. After that, the cells were gently rinsed twice with PBS to remove unbound Bio−Apt, followed by replacement with a mixture of cell media and 200 μL diluted SA-PAH-PAA−UCNDs. After incubation for a certain time (60-90mins), the cells were rinsed for three times with pre-warmed PBS to remove unbound SA-PAH-PAA−UCNDs.

2.5 Measurements

The nanoparticles’ morphology was characterized by transmission electron microscopy (TEM) and X-ray diffraction(XRD).The Fourier Transform Infrared Spectroscope (FTIR) was measured by using a Thermo Nicolet 6700 device. Zeta potential measurements were performed by a NanoPlus instrument (Micromeritics) at room temperature. The absorption spectra of YbCl₃was recorded in water solution by a UV/VIS/NIR Spectrometer (Lambda 950, PerkinElmer).The emission spectra of UCNDs in Fig. 2(e) and Fig. 3(a) were recorded by a spectrograph (QE65000, Ocean Optics) in the microscope system. While the emission spectra of UCNDs in Fig. 4(a) and Fig. 5(a) were recorded by another one (HR4000, Ocean Optics). The setups of spectra and imaging measurements are illustrated in Fig. 2.

 

Fig. 2 Schematic diagram for the setup of the proposed spectroscopy, lifetime and imaging system. A tunable solid Ti: sapphire laser (Mira HP, Coherent, USA) working at 925 nm CW laser mode was utilized in this experiment. This in vitro cell UC imaging experiment was implemented in the multiphoton laser scanning microscopy system (FV10MPE-S, Olympus). Luminescence decay measurements of the emission from UCNDs were performed by modulating the excitation laser with a chopper (Stanford, model SR540). The modulation frequency was 1 KHz, time-window 1 ms for luminescence decay measurement. The measured photon counts from the PMT and the trigger signal from the chopper were recorded simultaneously by a time-correlated single photon counter (TCSPC) (NanoHarp, Picoquant).

Download Full Size | PDF

 

Fig. 3 (a) TEM of ZrO₂:20% Yb³⁺, 2% Er³⁺ (core) and histogram of the size distribution (inset). (b) EDS analysis of the core NDs. (c) TEM of ZrO₂:20% Yb³⁺, 2% Er³⁺@NaYF₄: 2% Yb³⁺ (core/shell), and histogram of the size distribution (inset at bottom right). (d) EDS analysis of the core/shell NDs. (e) Emission spectra of (I) core-shell UCNDs,(II) core UCNDs and (III) absorption spectrum of core-shell UCNDs. (f) The photos of ZrO₂−Ln³⁺ core UCNDs (upper left) and core-shell UCNDs (upper right) dispersed in hexane. The emission photos were taken by a Canon camera under a 2 W laser excitation.

Download Full Size | PDF

 

Fig. 4 (a) Emission spectra of ZrO₂−Ln³⁺ core-shell UCNDs and NaYF₄−Ln³⁺ core-shell UCNPs. (b) TEM of ZrO₂:20% Yb³⁺, 2% Er³⁺@NaYF₄: 2% Yb³⁺ and (c) TEM of NaYF₄:20%Yb³⁺, 2% Er³⁺@NaYF₄. (d) Time-course plot from a single ZrO₂−Ln³⁺ core-shell UCND with laser illumination for 10 min and (e) 1 s.

Download Full Size | PDF

 

Fig. 5 (a) Emission spectra of ZrO₂−Ln³⁺ core UCNDs with different reaction time durations. (b) XRD patterns of ZrO₂−Ln³⁺ core UCNDs with different reaction time durations. The green vertical line indicates the standard PDF value. (c) TEM of ZrO₂−Ln³⁺ core UCNDs which were heated for 48 h in oven. (d) TEM of ZrO₂−Ln³⁺ core UCNDs which were heated for 72 h in the oven.

Download Full Size | PDF

3. Results and discussion

3.1 Characterization and emission spectra of ZrO₂:20% Yb³⁺, 2% Er³⁺@NaYF₄:2% Yb³⁺ UCNDs

The sub-5-nm, bright UCNDs with the shell layout are successfully synthesized shown in Fig. 1(b). ZrO₂ instead of NaYF₄ was chosen as the host for two reasons. Firstly, the crystallization process of ZrO₂ is easily controlled and repeatable to achieve the ultra-small (2-5 nm), uniform size using the solvothermal method [15]. Secondly, the lattice phonon energy of ZrO₂ is also relatively small. The main concern on ultra-small nanoparticles would be the severe quenching effect of the surface activator ions, which will largely reduce the emission intensity. Therefore, a thin shell outside the core was designed to inhibit the quenching. Doped with 2% Yb³⁺, this active protective layer is beneficial to further enhance the emission. Capped with benzyl alcohol after the solvothermal procedure, these as-prepared UCNDs are hydrophobic and can be dispersed in several organic solvents, e.g., chloroform and cyclohexane. The size and morphology were characterized by TEM, where the UCNDs clearly exhibited a uniform spherical shape with an averaged 3.5 nm diameter (Fig. 3(a)). The corresponding composition analysis was implemented by an energy dispersive spectrometer (EDS), revealing the presence of ZrO₂ composition and the doped Yb³⁺, Er³⁺ ions (Fig. 3(b)). As shown in Fig. 5(b), XRD patterns of the core UCNDs have well-defined peaks and this measured data matched well with the cubic crystal structure of zirconium oxide (JCPDS 49-1642). The outside shell NaYF₄:2% Yb³⁺ was synthesized through a thermal decomposition procedure. The dried sample of core-shell UCNDs can be redispersed in many organic solvents to form a clear, stable colloidal solution (Fig. 3(f)). The final core-shell UCNDs in the TEM image also exhibited uniform spherical UCNDs with an average diameter of 4.5 nm. The EDS analysis verified not only the existence of core elements, but also the existence of shell composition-Na, Y, F (Fig. 3(d)).

The emission spectra of UCNDs (core and core-shell with the same concentration and dosage) were measured to be found emit efficiently. Because of the less absorption by water, the 925-nm continues wave (CW) laser was chosen as the excitation light to reduce the heating effect in the whole process [27,28 ]. It was obviously found that the emission intensity of the core-shell was an order of magnitude stronger than that of the core, as shown in Fig. 3(e). The fluorescent photos also show the similar result (Fig. 3(f)). Although the core is bright enough, the large surface-to-volume ratio and poor crystallinity produce many defects that lead too much energy quenching. The core-shell structure is one of the most effective ways to improve the upconversion efficiency. The thin shell produced in the high temperature oil environment can decrease quenching as well as improve the crystallinity. Another merit of the shell is that it is capped by an OA ligand, facilitating further functionalization. For more efficient excitation, an appropriate amount of Yb³⁺ ions were added to form an active shell [29], which further enhanced the PL intensity of the core-shell structure.

3.2 The high brilliance of ZrO₂ UCNDs

NaYF₄ was considered as a perfect host material for its relatively high efficiency [30]. However, most commonly used NaYF₄-based UCNPs are about 20-100 nm [31,32 ]. Previous reports have suggested a sub-10-nm NaYF₄-based UCNP [33]. We synthesized the NaYF₄:20% Yb³⁺, 2% Er³⁺ core and core-shell UCNPs according to the reported literature [33]. The sizes and morphologies were presented in TEM images (Fig. 4(c)). The diameter of these core-shell UCNPs is almost 9.5 nm, which is about twice as large as the diameter of core-shell UCNDs (Fig. 3(b)). Figure 4(a) clearly shows that the emission intensity of the same core-shell structures ZrO₂−Ln³⁺ is one order of magnitude stronger than that of NaYF₄−Ln³⁺ (Fig. 4(a)). This result provides compelling evidence that the efficient host material (ZrO₂) and a thin shell enhance the emission efficiency. During synthesis of bright UCNDs, we questioned whether the ultra-small size would make them susceptible to photobleaching or blinking. The photostability of UCNDs was investigated by multiphoton microscopy through CW laser illumination for 10 min, suggesting that emission is stable without bleaching and emits consecutively of the UCND (Fig. 4(d)). To further confirm the non-blinking property, the dwelling time of each scanning point was reduced to 4 μs with high temporal resolution to confirm no blinking during 1 second’ emission (Fig. 4(e)).This strong and stable upconversion emission is appealing for protein-targeted imaging as well as other forms of bioimaging like single-molecular imaging, and it can also be applied for in vivo imaging.

3.3 Optimizing the synthesis conditions

The synthesis of ZrO₂ nanoparticles relies on a solvothermal reaction at a high temperature, which is a versatile reaction system for the nonaqueous preparation of oxide nanoparticles [34]. Among all the variables, the reaction time and temperature are critical for the formation of the ultra-small and bright UCNDs. It has been reported that ZrO₂ nanoparticles could be crystallized under temperatures of 210-300°Cfor 36 hours to 72 hours [15,35 ]. And 230°Cwas commonly used in the literature [1,15 ]. As a matter of fact, a relatively long reaction time didn't lead to much difference, according to the results of XRD measurements when the reaction time was 24 hours, 48 hours, 72 hours respectively (Fig. 5(b)). The TEM images show the sizes and morphologies of UCNDs when the reaction time is 48 hours and 72 hours (Fig. 5), and they are almost the same as when the UCNDs reacted for 24 hours seen in Fig. 3(a). Meanwhile, the emission intensity also verifies negligible differences among the UCNDs by three reaction times (Fig. 5(b)). To simplify the synthesis process and improve the efficiency, we choose to use 24 hours as the reaction time with the temperatures of 230°C.

3.4 Rapid scanning imaging of UCNDs with shortened lifetime

During microscopic imaging, a tightly focused laser beam is illuminated onto the sample with high power density. Upconversion emission can be significantly enhanced by using much higher activator concentrations under a relatively high-irradiance excitation [36]. Therefore, a series concentration of Er³⁺ varying from 0.5 mol% to 20 mol% co-doped with 20 mol% Yb³⁺were synthesized. Emission spectra were gained through the microscope with a high sensitive spectrograph under high-irradiance excitation values of up to 300 kW/cm². At high power density irradiance, the 15 mol% Er³⁺nanodots generated an exceptionally bright upconversion emission which was much stronger than that of the 0.5 mol% Er³⁺nanodots. The emission was increased from 0.5 mol% to 15 mol% Er³⁺-doped core-shell UCNDs by a factor of 18 (Fig. 6(a) ). This verifies that the higher excitation irradiance can alleviate concentration quenching in upconversion luminescence to enhance emission intensity when combined with higher activator concentration. It is also obvious that the green emission (attributing to the transition of ²H_11/2→⁴I_15/2 and ⁴S_3/2→⁴I_15/2) enhances more dramatically than the red one (attributing to the transition of ⁴F_9/2→⁴I_15/2) with increasing Er³⁺ concentration. However, the emission intensity of 20 mol% Er³⁺nanodots is weaker than that of the 15 mol% Er³⁺nanodots. This is because the sensitizer (Yb³⁺) content becomes less significant at higher power when the emitter (Er³⁺) concentrations is relatively high [10]. Meanwhile, the luminescence mechanism is changed since Er³⁺ is so particular that it can activate itself when the concentration is relatively high or there is no sensitizer [10,37,38 ]. This implies that the 20 mol% Er³⁺ UCNDs are less bright than the 15 mol% Er³⁺ ones. Combined with the advantage of being ultra-small, the enhanced luminescence can contribute to protein-targeted imaging.

 

Fig. 6 (a) Emission spectra of a series of ZrO₂: 20% Yb³⁺, 2% Er³⁺@ NaYF₄:2% Yb³⁺nanodots with varied Er³⁺ concentrations from 0.5% to 20%. (b) Luminescent decay data of a series ZrO₂:Yb³⁺, Er³⁺@NaYF₄: 2% Yb³⁺ UCNDs with varying Er³⁺concentrations under the925 nm excitation. Both of the excitation irradiance was 3.0 × 10⁵ W cm⁻².

Download Full Size | PDF

The long lifetime of UCNPs is a big obstacle for rapid scanning microscopy [18], which can lead severe streaking problem in the imaging. We need a long dwelling time to get a clear image and this will greatly increase the imaging time. However, the biomolecule detection is usually time-consuming especially for the 3D imaging, thereby it is necessary to adopt measures to shorten the lifetime of UCNDs. Here we synthesized a series of concentrations (from 0.5 mol% to 20 mol%) of Er³⁺-doped UCNDs and measured the luminescence decay of each sample (Fig. 6(b)). The variation trend of the lifetime is the same with that of emission intensity when excited at a high power density. We observed from the decay curve that the 15 mol%-doped UCNDs have the shortest lifetime. After fitting and calculating, the lifetime of 15mol % Er³⁺-doped core-shell UCNDs was as short as 23.6 μs, which was much shorter than the 116.3 μsof 0.5 mol% Er³⁺-doped core-shell UCNDs (Fig. 6(b)). When the concentration of Er³⁺ increased, the distance between the Yb³⁺ and Er³⁺ became shorter and the cross relaxation of neighboring ions increased. Thus the lifetime was tuned to be shorter. However, the 20% Er³⁺-doped core-shell UCNDs had a longer lifetime than 15% ones, and this phenomenon was also caused by the different luminance mechanisms that the sensitizer content becomes less significant and the activator pumps itself when its proportion is enough high. Thus the luminescence quenching is not so severe and the lifetime becomes longer (37.1 μs) than that of 15 mol% Er³⁺ UCNDs.

The streaking phenomenon caused by the long lifetime of UCNPs has affected the scanning speed for a long time [17,39 ]. We have to apply short lifetime to acquire rapid scanning luminescence images, the merit of which was intensely reflected in the 3D imaging. The corresponding rapid scanning luminescence images were recorded (Fig. 7(a)-7(e) ). For the image of 0.5% Er³⁺-doped UCNDs, the poor lateral resolution (Fig. 7(a)), that is, the x-y plane, was caused by streaking in the scanning direction, obtained at a speed of 12.5 μs/pixel with which the shape and the size of the aggregated UCNDs could not be determined. As the Er³⁺ concentration increased from 0.5 mol% to 15 mol%, images (Fig. 7(b)-7(e)) with enhanced lateral resolution were observed because of the decreasing in streaking. The image of 15 mol% Er³⁺-doped UCNDs were clearly observed when the luminescence streaking almost vanished at a scanning speed of 12.5 μs/pixel. We analyzed in detail the streak from the emission intensity profiles obtained from the halfway line downtrend (Fig. 7(f)). The full width at half maximum (FWHM) of each curve was calculated, and these data can quantitatively demonstrate the streaking phenomenon (Fig. 7(g)). The core-shell UCNDs were fixed in the agar gel, and 100 μm depth in the z-axis direction was selected to perform the 3D imaging. At the speed of 12.5 μs/pixel, covering 512 × 512 pixels for each layer and 30 scanning layers, it took 1.638 minutes to finish the scanning of 15% Er³⁺-doped core-shell UCNDs. It can be observed that the axial resolution was also enhanced so that the 3D image is clear and distinguishable (Fig. 7(j)). But there is an obvious streaking effect in the image of the 0.5 mol% Er³⁺-doped core-shell UCNDs image (Fig. 7(h)). Clear imaging requires a 200-μs dwelling time and 26.214 minutes to complete the image. This is a long scanning time, 16 times longer than that of 15% Er³⁺-doped core-shell UCNDs (Fig. 7(i)). All these experimental results prove that a high concentration of activator ions (like 15%) has a high emission intensity as well as shorter lifetime when excited by a high power density, which make much sense in the rapid scanning imaging and also the time-consuming 3D imaging.

 

Fig. 7 (a) - (e) Multiphoton scanning luminescence images of UCNDs doped with different Er³⁺ concentrations. All the images were gained at the same dwelling time of 12.5 μs/pixel. Dashed circles indicate the points which were chosen to show the linecuts in (f). All the data were collected using a 3.0 × 10⁵ W/cm² excitation intensity. The scale bar is 150 μm. (f) Normalized intensity profile obtained from halfway line along with the long tail of a series concentration of Er³⁺-doped UCNDs. (g) 3D scanning imaging of ZrO₂:20% Yb³⁺, 0.5% Er³⁺@ NaYF₄:2% Yb³⁺ at a speed of 12.5 μs/pixel. (h) 3D scanning imaging of ZrO₂:20% Yb³⁺, 0.5% Er³⁺@ NaYF₄:2% Yb³⁺ at a speed of 200 μs/pixel. (i) 3D scanning imaging of ZrO₂:20% Yb³⁺, 15% Er³⁺@ NaYF₄:2% Yb³⁺ at a speed of 12.5 μs/pixel. These three images were all taken by 30 layers, and each layer was covering 512 × 512 pixels. The height of the 3D images is 127 μm.

Download Full Size | PDF

3.5 Aptamer-mediated proteins-targeted imaging

There are various proteins in the cell membrane including the nucleolin or prion protein (PrPC). These proteins are disease-related proteins, and the visualization of them will greatly help researchers do the relevant study. We implemented proteins-targeted imaging of the PrPc with an aptamer specifically recognizing the protein (Fig. 8 ). Compared to antibodies, aptamers have a similarly high affinity and selectivity for proteins of interest, thereby facilitating the procedure and improving the target recognition performance [36]. Before biological application, the hydrophobic OA-capped UCNDs need to be hydrophilic and biocompatible. Figure 8 shows how the aptamer links to the UCNDs.

 

Fig. 8 Schematic illustration showing the detailed process of proteins-targeted aptamer-mediated bioimaging.

Download Full Size | PDF

Consequently, the as-synthesized UCNDs were transformed to be hydrophilic by using a ligand-free method [25]. Afterwards, the nanodots were successively capped with water-soluble PAA and PAH to enhance the biocompatibility as well as to reduce toxicity [40]. Firstly the NOBF₄ was used to remove the OA capped on the UCNDs, which also shrank the core-shell structure to be coated with polymer. The surface of the UCNDs became ligand free and positively charged [25]. Then the hydrophilic UCNDs were dissolved in deionized water and coated with two layers of water-soluble polymers through electrostatic interaction. When the PAA was capped onto the surface, the UCNDs were negatively charged and this allowed for positively charged PAH to be coated outside next. After coated by polymer, the PAH-PAA-UCNDs illustrated better water solubility as well dispersibility [40]. All the surface zeta potentials were measured by a Zeta-Plus and the results are shown in Fig. 9(a) . Surface modification of the nanoparticles was further confirmed by FTIR spectroscopy (Fig. 9(b)). In the FTIR spectrum of OA-UCNPs, the peak at 3008 cm⁻¹ was attributed to the C-H stretching vibration, and peaks at approximately 2926 cm⁻¹ and 2855 cm⁻¹ were attributed to the asymmetric and symmetric stretching vibrations of methylene (CH₂) in the long alkyl chain, indicating that OA was coated on the surface of the nanoparticles. In contrast, peaks at approximately 2926 cm⁻¹ and 2855 cm⁻¹ clearly disappear in the spectrum of NOBF₄-treated UCNDs. This proved that the nanodots were free of OA. The peak at 1680 cm⁻¹ as well as the inconspicuous peak located around 745 cm⁻¹indicates the existence of an additional N-H bond, which manifests the existence of PAH.

 

Fig. 9 (a) Zeta potential of different layers of modified UCNDs, (b) FTIR spectra of (I) OA-UCNDs, (II) ligand-free UCNDs (III) PAH-UCNDs.

Download Full Size | PDF

To achieve aptamer-mediated proteins-targeted imaging, the streptavidin (SA) was attached to the surface of polymer-coated UCNDs (Fig. 8). The streptavidin can specifically recognize the biotin that was modified on the aptamer. The live cells were imaged in multiphoton microscope equipped with a 925 nm NIR laser. Figure 10(a) shows the imaging of blank cells, and no emission could be observed in the completely dark fluorescence image but the cell profile. To verify the specificity of aptamer, cells in Fig. 10(b) were first incubated with PBS and no aptamer/biotin. Thus few streptavidin-PAH-PAA-UCNDs were targeted onto the proteins and they emitted ultra-weak light from the cytomembrane. Meanwhile, other aptamer targeted cells in Fig. 10(c) were incubated with PAH-PAA-UCNDs without SA. As exhibited in the pictures just a cluster of specks of emission could be observed, indicating that a few PAH-PAA-UCNDs were adsorbed. This was because of electrostatic adherence between cancer cytomembrane and PAH-PAA-UCNDs. The bright green and red emission signal in Fig. 10(d) can be observed, agreeing well with the corresponding upconversion emission spectrum. The cells were apparently stained with yellow color from the ZrO₂: Yb³⁺/Er³⁺@NaYF₄: Yb³⁺, which indicated that the streptavidin-UCNDs specifically targeted the proteins on the cytomembrane. These results further identified that the bright and ultra-small UCNDs can be applied for bioimaging and have prospect for achieving single-molecule imaging.

 

Fig. 10 (a) –(d) In vitro cancer cell imaging using 925 nm laser excited UCNDs: images of HepG2 cells separately incubated with (a) nothing, (b) PBS pre-incubated and with SA-PAH-PAA-ZrO₂-Ln³⁺@NaYF₄:Yb³⁺, (c) Bio-Aptamer pre-incubated and then with PAH-PAA-ZrO₂-Ln³⁺ @ NaYF₄:Yb³⁺, and (d) Bio-Aptamer pre-incubated and then with SA- PAH-PAA -ZrO₂-Ln³⁺@NaYF₄:Yb³⁺. Bright field (first column), green imaging (second column), red imaging (third column) and overlay image. The scale bar is 30 μm.

Download Full Size | PDF

Conclusions

In summary, for the first time a new type of ultra-small ZrO₂-based- UCNDs was elaborately designed and successfully synthesized. The emission intensity of ZrO₂−based core-shell UCNDs is more than one order of magnitude stronger than that of 10-nm NaYF₄−Ln³⁺ core-shell UCNPs. Under a high-irradiance excitation, UCNDs with varied Er³⁺-doping concentration from 0.5 mol% to 15 mol% were synthesized. It was found that 15 mol% Er³⁺-doped UCNDs had 18 times stronger emission intensity than that of in 0.5 mol% Er³⁺-doped ones. Furthermore, higher activator doping concentration also led to a shorter lifetime (23.6 μs), and thus was applied to rapid scanning microscopy. A clear 3D image of 15 mol% Er³⁺-doped UCNDs dispersed in the agar gel was finished in 1.638 minutes, only one-sixteenth time consumption of that of 0.5 mol% Er³⁺-doped UCNDs. To achieve protein targeted-imaging, the UCNDs were designed to be biocompatible and specifically-connected by aptamertargeted to the cytomembrane proteins. Moreover, the UCNDs can also be used to image intracellular proteins through specific targeting in live or fixed cells for the small size. This method can be useful for producing efficient, uniform, ultra-small and bright UCNDs. One can apply these ultra-small and efficient UCNDs to single-molecule imaging and dynamics studies.