Abstract
The p53 protein is a major regulator of cell cycle progression and apoptosis. We used a p53-enhanced green fluorescent protein (EGFP) construct for transfections into human breast cancer (MCF-7) cells. Cells expressing p53-EGFP showed an increased apoptotic index compared to cells transfected with EGFP alone. Interestingly, apoptotic cells showed localization of p53-EGFP to both nuclei and cytoplasm, whereas non-apoptotic cells usually only showed nuclear localization of p53-EGFP. This result is in agreement with the hypothesis that p53 induces apoptosis by interaction with both nuclear and cytoplasmic targets. Transfected p53-deficient osteosarcoma cells were used for immunofluorescence quantitation. The intensity of immunofluorescence for either p53 or EGFP showed excellent linear correlation to the EGFP autofluorescence, proving that measurements of immunofluorescence intensities can be used for determining endogenous protein levels.
M
However, correlations of the type described involve more or less artificial models or make use of biochemical measurements on large populations of cells. A better approach would be to directly correlate staining intensities of individual cells to their content of antigens. Since the cloning of green fluorescent protein from the jellyfish Aequorea, directed mutations have produced genes encoding proteins that show autofluorescence in the blue to red range of the spectrum (reviewed by Kendall and Badminton 1998; Tsien 1998). These genes can be fused to genes coding for normal cell proteins and expressed in vivo under the control of appropriate promoters. Cells transfected with such constructs express fluorescent chimeric proteins that often show subcellular distributions and functions identical to those of the normal endogenous proteins (Kendall and Badminton 1998; Tsien 1998). Importantly, the autofluorescence of these proteins can be used to determine their concentrations (Niswender et al. 1995; Albano et al. 1998).
It occurred to us that fluorescent proteins might constitute useful internal standards for indirect immunofluorescence. We therefore transfected human Saos-2 cells, which lack the p53 gene, with a construct encoding a p53-enhanced green fluorescent protein (p53-EGFP) chimera. In double-stained cells, the endogenous EGFP fluorescence was measured and correlated to indirect immunofluorescence measurements of the EGFP and p53 parts of the chimeric protein. The results show that quantitation of endogenous proteins by indirect immunofluorescence is feasible. Moreover, the results also show that, similar to wild-type p53, the p53-EGFP chimeric protein induces apoptosis in transfected cells.
Materials and Methods
Cell Lines and Transfections
Saos-2 cells were cultured in Dulbecco's minimum essential medium (MEM) with glutamax-1 (Life Technologies; Lexington, KY) and 10% fetal calf serum (FCS). MCF-7 cells were cultured in Eagle's MEM with glutamax-1 and 10% FCS. Both media were fortified with 100 U/ml penicillin and 100 μg/ml streptomycin. For transfections with clonfectin (Clontech; Palo Alto, CA), Saos-2 or MCF-7 cells were grown on sterile slides in 90-mm culture dishes. When they reached about 70% confluency, cells were transfected with 1.5 μg/ml of either a p53-EGFP vector (pp53-EGFP; Clontech) or a control vector encoding only EGFP (pEGFP-C1; Clontech) using 1.5 μg/ml clonfectin for 8 hr (as recommended by the manufacturer). Both vectors were under control of the human cytomegalovirus (CMV) immediate-early promoter. Twenty-four hours after transfection, cells were fixed in 3.7% paraformaldehyde (Merck; Darmstadt, Germany) in 0.1 M sodium phosphate buffer, pH 7.4, for 10 min at 4C and were then permeabilized with 1% Triton X-100 for 10 min.
Immunofluorescence and TUNEL Staining
Cells were either mounted directly in antifade medium or underwent indirect immunofluorescence staining using either a monoclonal p53 (IgG2b) antibody (5 μg/ml) (Dako; Glostrup, Denmark) or a monoclonal anti-green fluorescent protein (GFP) (IgG2a) antibody (2–5 μg/ml) (Clontech) crossreacting with EGFP. Specimens to be used for quantitation were double stained with p53 and GFP antibodies using biotin-conjugated goat anti-mouse IgG2b antibody (10 μg/ml) (Southern Biotechnology Associates; Birmingham, AL) followed by aminomethyl-coumarin (AMCA)-labeled streptavidin (20 μg/ml) (Vector Laboratories; Burlingame, CA) and Texas red-labeled goat-anti mouse IgG2a antibody (Southern Biotechnology Associates) for detection.
Some specimens were also reacted with the TUNEL method (Gavrieli et al. 1992) for detection of fragmented DNA in apoptotic cells. Cells were permeabilized in 0.5% Triton X-100 in 0.05 M PBS for 1 hr at room temperature (RT) and washed in PBS four times for 2 min. A mixture of 100 U/ml terminal deoxynucleotidyl transferase (TdT), 0.5 μM digoxigenin-dUTP (Boehringer Mannheim; Mannheim, Germany) and 0.5 μM d(AGC)TP (Applied Biosystems; Foster City, CA) in TdT buffer (0.2 M sodium cacodylate, 2.5 mM cobalt chloride, 0.1 mM dithiothreitol, pH 6.6) was then applied for 2 hr at 37C. The reaction was stopped with 0.3 M sodium chloride, 0.03 M sodium citrate for 15 min at RT, followed by washing in PBS. Cells were subsequently reacted with rhodamine-labeled goat anti-digoxigenin (Boehringer Mannheim) for 1 hr at RT, followed by washing in PBS. TUNEL staining was combined with staining of DNA using bisbenzimide (Hoechst 33258; Sigma, St Louis, MO) for confirming the presence of condensed DNA in TUNEL-positive cells (Cao et al. 2000).
Quantitations
Cells transfected with p53-EGFP and double stained for p53 and EGFP were examined in an inverted Nikon fluorescence microscope equipped with a Photometrics cooled CCD camera or in a Leica DMRXA microscope fitted with a Leica Q-FISH system. Transfected cells were located using excitation in blue (488-nm) light and images were sequentially acquired using selective FITC, AMCA, and Texas red filter blocks. Image acquisition time was fixed by the software and the time from cell localization to initation of image capture was less than 5 sec. Captured images were analyzed using either the Image-Pro Plus program (Media Cybernetics; Silver Spring, MA) or the Leica Q-WIN software. Nuclei were selected by tracing and measured. An area outside the cells was similarly traced and measured and the value subtracted from the nuclear fluorescence values. In addition, specimens containing cells transfected with either p53-EGFP or with EGFP alone and re-stained with the TUNEL method and bisbenzimide were coded and the number of transfected cells that contained condensed DNA and were TUNEL-positive was determined by counting.
Results
Human breast cancer (MCF-7) cells, which express a normal p53 gene, were transfected with a p53-EGFP construct. The transfections resulted in the appearance of the green fluorescent chimeric protein in about 5% of the cells. In most cells the green fluorescence was confined to the nuclei (excluding nucleoli), but occasional cells also displayed cytoplasmic green fluorescence. Control cells transfected with EGFP alone expressed both nuclear and cytoplasmic fluorescence. Immunocytochemical double stainings of p53-EGFP-transfected cells showed that the green fluorescent cells also were immunopositive for both the p53 and the EGFP portion of the chimeric protein (Figure 1). Occasional non-transfected cells also stained for p53, reflecting the endogenous production of this protein in MCF-7 cells.
TUNEL and bisbenzimide staining showed that many of the p53-EGFP-transfected cells were apoptotic. Interestingly, all of these cells showed the presence of p53 in both the nucleus and cytoplasm (Figure 1), whereas non-apoptotic cells showed localization of p53 either to the nucleus only or to both the nucleus and cytoplasm. Counting of TUNEL-positive cells showed that cells transfected with p53-EGFP had a much higher apoptotic index than control cells transfected with EGFP alone. Thus, whereas 34.7 ± 10.9% of the p53-EGFP-transfected cells were apoptotic, only 14.3 ± 3.5% of cells transfected with EGFP alone were apoptotic. However, the apoptotic index of the EGFP-transfected control cells was higher than that of untransfected cells (1.1 ± 0.6%), indicating that transfection with EGFP alone also induces apoptosis.
(
Because MCF-7 cells express endogenous p53 they were not suitable for quantitative immunofluorescence measurements of the expression of transfected EGFP-p53. For this part of the experiment we selected Saos-2 cells, which do not express endogenous p53. Transfection of such cells with the p53-EGFP construct induced green fluorescence in nuclei of scattered cells. In most experiments about 5% of the cells were transfected. The intensity of the EGFP fluorescence varied among cells. Immunofluorescent double staining for p53 and EGFP revealed detectable fluorescence only in cells that also expressed green EGFP autofluorescence. Most fluorescence occurred over nuclei (excluding nucleoli) but additional cytoplasmic fluorescence was also observed in some cells. Measurements of indirect immunofluorescence intensities for p53 and EGFP over individual nuclei showed that the intensities correlated well with each other and with the green EGFP autofluorescence (Figure 2). Although highest fluorescence intensities were recorded with Texas red, all transfected cells were detected with all three methods. The GFP antibody was tested at dilutions of 2 and 5 μg/ml, both of which resulted in maximal fluorescence values, indicating that saturating antibody concentrations had been reached.
Correlations between EGFP immunofluorescence, p53 immunofluorescence, and EGFP-p53 autofluorescence in nuclei of transfected Saos-2 cells that have been double stained for p53 and EGFP immunofluorescence as in Figures 1A-1C. Correlation coefficients (R2) are given in the respective graphs.
Discussion
Our results show that transfection of cultured neo-plastic cells with a p53-EGFP construct results in nuclear targeting of the chimeric protein and in an increased apoptotic index. Together, these results strongly indicate that the chimera functions like wild-type p53. The induction of apoptosis agrees with previously published findings using transfections with p53 constructs not containing EGFP (Ko and Prives 1996; Schuler et al. 2000). Expression of p53 is normally induced in response to DNA damage caused by, e.g., ionizing radiation, nitric oxide, or chemotherapeutic agents (Levine 1997). In the absence of these agents, basal expression of p53 is relatively low in MCF-7 cells, as shown by stainings for endogenous p53 in non-transfected cells. Together, these data show that high levels of p53 expression are associated with induction of apoptosis in MCF-7 cells. Interestingly, we observed that transfected apoptotic cells always contained p53-EGFP in both nuclei and cytoplasm. This could reflect leakiness of the nuclear membrane in apoptotic cells or massive overexpression of EGFP-p53. Alternatively, our observations may indicate that both nuclear and cytoplasmic targets must be reached by p53 before apoptosis is commenced. This possibility is supported by findings showing that p53 not only affects transcriptional targets but also interacts with mitochondria (Marchenko et al. 2000) and can promote apoptosis and caspase activation by inducing release of cytochrome C from these organelles (Schuler et al. 2000).
Our data support the belief that EGFP-p53 constitutes a good marker for endogenous p53 trafficking in cells and may be useful for elucidating nuclear and cytoplasmic targets for this protein. The observation that EGFP expression alone increases apoptosis concurs with previous data on the toxicity of overexpressed EGFP (cf. Tsien 1998). Although the use of control transfections with EGFP alone corrects for over-interpretations, we are currently testing alternative fluorescent protein variants.
The central role of p53 in apoptotic regulation makes it of great interest to quantitate p53 in individual cells. To study this aspect we used Saos-2 cells, which lack endogenous p53 expression. Our results show that p53 quantitation by immunofluorescence is indeed possible and indicate that the approach used may also be helpful for quantitating other proteins. It should be emphasized that transfected cells had to be located in fluorescent light because there was no other means to identify them. This approach is associated with the risk for photobleaching of the fluorophores, and we therefore performed cell measurements in as quick and standardized a way as possible. Although the excellent correlations obtained show that this approach is possible, we strongly advocate the use of cell identification by phase or differential interference contrast optics before measurements, whenever possible.
Footnotes
Acknowledgments
Supported by the Danish Medical Research Council, the Danish Agricultural and Veterinary Research Council, and the Danish Cancer Society.