Monitoring the endocytosis of bovine serum albumin based on the fluorescence lifetime of small squaraine dye in living cells

: Bovine serum albumin (BSA) has a wide range of physiological functions involving the binding, transportation, and delivery of fatty acids, porphyrins, bilirubin, steroids, etc. In the present study, we prepared a small squaraine dye (SD), which can selectively detect BSA using ﬂuorescence lifetime imaging microscopy (FLIM), to monitor the endocytosis of BSA in live cultured cells in real time. This approach revealed that BSA uptake is concentration-dependent in living cells. Furthermore, we used paclitaxel (PTX), a chemotherapeutic drug, to inﬂuence the endocytosis of BSA in living cells. The results demonstrated that the endocytic rate was clearly reduced after pretreatment with 0.4 µM PTX for 2 h. The present study demonstrates the potential value of using the ﬂuorescence lifetime of SD to detect BSA concentration and study the physiological mechanism of chemotherapeutic drugs.


Introduction
Serum albumins (SA), which are the major proteins of circulatory system, act as critical transport proteins for delivery of fatty acids, amino acids and drug molecules [1][2][3][4]. Patrick et al. reported that the overexpression of the gp60 receptor and SPARC protein in cancer cells contributed to SA retention in the tumor microenvironment [5]. In addition, accumulation of intravascular SA in the tumor vicinity can be attributed to the increased permeability and retention effect of the leaky abnormal blood vessels of tumor [6]. Basically, SA are regarded as the important transporter between intracellular and extracellular microenvironment. Bovine SA (BSA) has been extensively used in many researches because of the structural homology with human SA [7][8][9]. Several studies have reported the detection of BSA using optical sensors [10][11][12][13]. However, all these researches were performed in solution or plasma. To the best of our knowledge, there are few reports on the endocytosis of BSA in single living cells during a chemotherapeutic drug treatment [14,15], which is a complex dynamic process associated with physiological functions of cancer cells, including dynamic changes of proteins concentration, metabolism and apoptosis [16][17][18][19][20].
Fluorescence lifetime imaging microscopy (FLIM) has been widely used to measure the fluorescence lifetime at each pixel across a sample in biomedical applications [21]. In conventional fluorescence intensity detection, emission signal are strongly affected by multifarious environmental influences, such as fluctuations of light source power and detector voltage, autofluorescence of some tissue samples, concentration of probes and fluorescence resonance energy transfer, which results in the difficulty in quantifying the intensity-based data. However, the fluorescence lifetime indicates the local physicochemical microenvironment around the fluorophore [22][23][24][25][26], which was measured as the rate of emission decay. Therefore, FLIM can provide the contrast mechanism images even if the fluorescence intensities are similar, which makes FLIM as a powerful tool to research physiological processes in living cells.
Previously, we have synthesized a small squaraine dye (SD), which exhibited excellent twophoton emission when interacting with BSA [27]. In the present study, we demonstrated that this SD can selectively detect BSA on the basis of the changes in the proportion of a relatively longer lifetime component. It can be used to monitor the endocytosis of BSA in living ovarian cancer cells (OVCAR-3) in real time. Upon adding different concentrations of BSA into the cell culture medium, the proportion of this relatively longer lifetime component (interaction of BSA with SD) increased gradually and tended to be constant within 0.5 h. Subsequently, we used 0.4 µM paclitaxel (PTX, a chemotherapeutic drug) to pretreat OVCAR-3 cells for 2 h and found that the endocytosis of BSA was clearly reduced. Through this study, we provide a method to monitor the endocytosis of BSA in single living cells using FLIM, which expands the toolbox for biological proteinology research.

Synthesis of SD
The synthesis of target compound 1 was a relatively simple two-step reaction (Fig. 1). Our strategy for the synthesis of SD 1 started with the quaternization of the free base of 2,3,3trimethylindolenine with an excess of ethyl iodide in acetonitrile. Dimethylation of squaric acid 3 with ethanol and triethylorthoformate followed by reaction with ethylindolium 4 generated target compound 1 with 66% yield.

Confocal laser scanning microscopy (CLSM)
OVCAR-3 cells seeded in the glass dishes were stained with the synthesized SD and Mito-G for 20 min and then washed with PBS. Fluorescence images were captured using a confocal laser scanning microscope (A1R MP, Nikon, Japan) equipped with an oil-immersion 60× objective (NA = 1.4, Nikon, Japan). Mito-G, Rho-123 and BSA-Alexa Fluor 488 fluorescence was excited at 488 nm, and the emission was collected from 500 to 550 nm. SD fluorescence was excited using a 561 nm laser, and the emission was collected from 570 to 620 nm, which matched well with the spectra range of SD that we have previously reported [27]. Colocalization analyses of two fluorescence channels were conducted using the Nikon software NIS-Elements AR.

FLIM
When the cells reached 70% confluence, 2 µM SD was used to stain the cells for 20 min. After being washed with PBS for 3 times, the cells were incubated in RPMI-1640 media without Phenol Red and FBS (Gibco, Invitrogen, NY, USA) during imaging. FLIM images were captured using a Nikon microscope connected to a confocal scanning FLIM system (DCS-120, Becker & Hickl GmbH, Germany) and a picosecond (pulse width: 6 ps) supercontinuum laser (WL-SC-400-4, Fianium, UK). The excitation wavelength was set at 570 nm, and the emission for FLIM imaging was collected between 580 and 650 nm. The 40-MHz pulsed laser worked in an on-off modulated mode, which allowed FLIM in the laser-off period to be obtained through a single-photon counting module (SPC150, Becker & Hickl GmbH, Germany). The cells were maintained at 37°C and 5% CO 2 in an environmental chamber (H301, Okolab, Italy) during measurements.
The data were recorded in the "first in, first out image" mode of the SPC150 module. The images were scanned and recorded at a resolution of 256 time channels. The fluorescence lifetime at each pixel of a 256 × 256 image was calculated, with components of biexponential fitting expressed as follows: where τ 1 and τ 2 are fluorescence lifetimes of free and bound SD molecules, respectively, and a 1 and a 2 denote the amplitudes of two corresponding components. The final a 2 images were constructed using the Becker & Hickl SPCImage software (version 7.3).

Fluorescence lifetime responses of SD to BSA in solution
To establish the role of SD in the selective detection of BSA, we first investigated the fluorescence lifetime response of SD (3 µM) in the presence of other biomolecules such as ATP, NADH, FAD, GSH, and transferrin in that order at the same concentration (3 µM). As shown in Fig. 2(a), these biomolecules could not change the fluorescence decay curve of SD, thus indicating that SD is insensitive to these biomolecules, except to BSA. Subsequently, we studied the relationship between the fluorescence lifetime of SD and the concentration of BSA. As shown in Fig. 2(b), with BSA titration, the fluorescence decay curves changed from a single lifetime component (0 µM) to double lifetime components (from 0.09 to 1.35 µM) and then to another single lifetime component (1.8 µM), which indicated environmental changes surrounding the squaraine chromophore because the electrostatic and hydrophobic interactions between SD and BSA resulted in reducing the aggregation of the squaraine moiety [28], as shown in Fig. 2(c). By fitting the fluorescence decay curves (addition of 0 and 1.8 µM BSA) to a monoexponential model, we obtained τ 1 and τ 2 , which were 204 and 2651 ps, respectively. We fitted the other fluorescence decay curves to a double exponential model, while fixing the two florescence lifetime values, and obtained a linear relationship between the proportion of the long lifetime component a 2 and the concentration of BSA [ Fig. 2(d)].

Subcellular localization of SD in living OVCAR-3 cells
To examine the intracellular distribution of SD, the cells were stained with SD and Mito-G then imaged with CLSM. As shown in Fig. 3(a), colocalization analyses indicated that the Pearson coefficient (r) values of the fluorescence images of SD (red) and Mito-G (green) were 0.95 when the cells were stained with 0.3 µM SD and 0.69 when the cells were stained with 3 µM SD. The boxplot shows that the average r value decreased from 0.93 to 0.68 for the cells stained with concentrations of SD ranging from 0.3 to 3 µM [ Fig. 3(b)]. These results demonstrated that SD can target mitochondria well at a low concentration, whereas it also exists in the cytoplasm at a high concentration. To evaluate how SD molecules target to mitochondria, we observed the fluorescence images of the cells stained with Mito-G, Rho-123 and SD, respectively [ Fig. 3(c)]. The fluorescence were obviously faded in the cells stained with SD after fixing just as Rho-123, which demonstrated that SD molecules attach to mitochondria by electrostatic adsorption.

Endocytosis of BSA in living cells
BSA is considered as a transport protein that plays an important role in cell metabolism. Therefore, the process of BSA endocytosis is closely related to the physiological state of the cells. We firstly examined whether the presence of intracellular SD can affect the BSA endocytosis behavior (Fig. 4). The results showed similar rising tendencies of the average fluorescence intensity in the cells with and without the presence of SD, demonstrating that the presence of SD has no effect on the BSA endocytosis. Subsequently, we investigated the endocytosis of BSA in OVCAR-3 cells by monitoring the changes in a 2 values after the addition of BSA, to reflect the concentration and time of BSA uptake and to characterize the physiological processes in the cells. As shown in Fig. 5, we implemented time-series FLIM measurements of the cells stained with 3 µM SD. The fluorescence lifetime data of the cells fitted well to the double exponential model, while fixing the long lifetime value to 2651 ps, and we obtained the a 2 images [ Fig. 5(a)]. In the control group, we added PBS into the dishes containing the cells and found that the a 2 value was constant from 0 to 30 min, which implied that the concentration of BSA was relatively stable in the cells during the measurements. However, upon the addition of BSA at 2, 4, and 6 µM, respectively, the a 2 values increased gradually at different rates, thereby demonstrating that the endocytic rate of BSA depends on the concentration of BSA. In particular, the a 2 values tended to reach saturation and become constant 8 min after the addition of 6 µM BSA, which is possibly caused by the equilibrium of BSA uptake in the cells. The statistical results of at least 100 cells are presented in Fig. 5(b).

The influence of the endocytosis of BSA in single cells under PTX treatment
PTX can stabilize microtubules and induce apoptosis, and it is widely used chemotherapeutic agent in several types of cancers [29,30]. To examine whether PTX treatment can affect the endocytosis of BSA in living cells, SD-stained cells were pretreated with 0.4 and 2 µM PTX for 2 h. After adding 2 µM BSA, the time-series measurements of fluorescence lifetime were obtained, and the corresponding a 2 images are shown in Fig. 6(a). For the control group, the endocytic rate of BSA was similar to the result in Fig. 5(b, red line). However, the increase in a 2 values was significantly lower in the cells treated with 0.4 µM PTX than in the control group, indicating that the endocytosis of BSA was inhibited. For the cells treated with 2 µM PTX, the morphological changes (shrinking and rounding) during imaging indicated that the cells had undergone apoptosis, whereas the a 2 values were scarcely increased. The statistical results from 15 cells are presented in Fig. 6(b).
The cells stained with the low SD dose (0.3 µM) show that almost no free intracellular SD is present (Fig. 3). We also monitored the fluorescence lifetime of the cells stained with 0.3 µM SD (Fig. 7). Almost all the intracellular SD molecules were bound to the BSA uptake during cell culture (RPMI-1640 media supplemented with 10% FBS), which resulted in the FLIM data were fitted well with a monoexponential model and exhibited the long fluorescence lifetime (∼2609 ps) of bound SD molecules (τ 2 ). We observed no significant changes in τ 2 values after 2 µM BSA addition [ Fig. 7(b)], which implied that the only free SD molecules can be bound to endocytosed BSA in the cells stained with the higher SD dose (Figs. 5 and 6). SD molecules can enter the cells rapidly because of the positive charge of its 3,3-dimethyl-ethylindole group. However, a small number of SD molecules remain attached to the cell membrane, which results in the larger a 2 values on the cell membrane than on the cytoplasm [white arrow in Fig. 6(a)], thus indicating that there are some free intracellular SD molecules before the addition of BSA. We also found that the a 2 values on the cell membrane were relatively stable [ Fig. 6(a)], whereas these values on the cytoplasm clearly increased in the control group upon the addition of BSA, which implied that the normal cells endocytose BSA quickly and that the free intracellular SD molecules were bound.
The a 2 values represent the proportion of SD molecules bound to BSA. Before the addition of BSA into the glass-dishes containing the cells stained by 2 µM SD, the a 2 values at 0 min were dependent on the amount of BSA uptake from the media during cell culture, and exhibited a slight difference between Fig. 5 and Fig. 6. However, with the addition of BSA at the same concentration (2 µM), the a 2 values were increased similarly [the second row in Fig. 5(a) and the first row in Fig. 6(a)], which indicated that the slight difference of the amount of BSA uptake during cell culture did not affect the free intracellular SD molecules to bind endocytosed BSA after subsequent BSA addition. Previous reports support the finding that PTX affects the metabolism of cells by disrupting membrane trafficking and affecting the microtubulin carrier [31][32][33][34]. In the present study, the endocytic rate and concentration of BSA decreased significantly in the cells pretreated with PTX, thus indicating that PTX damaged the signaling pathway of endocytosis and then inhibited BSA uptake even in the apoptotic cells in which membrane permeability had changed [imaging of the cells pretreated with 2 µM PTX at 30 min; Fig. 6(a)]. The results of the present study demonstrated that FLIM can be used to study the endocytosis of BSA in living cells on the basis of the changes in the fluorescence lifetime of SD.

Conclusion
In summary, the present study demonstrated that the SD we synthesized can selectively detect BSA in solution and living cells using FLIM. Upon the addition of BSA at different concentrations, the rates of BSA uptake differed, which can be reflected by monitoring the changes in the proportion of the long lifetime component. We also found that drug-induced apoptosis can influence endocytosis in living cells. We proved that the method used in the present study can be applied to monitor the endocytosis of proteins and explain some biochemical processes in living cells.

Disclosures
The authors declare no conflicts of interest.