Membrane dynamics—including mechanical stiffness and exchange of lipid molecules (membrane fluidity)—has a large impact on cell function and may be an important factor for pathogenesis of, e.g., cancer1 or Alzheimer’s disease.2 An essential component of lipid bilayers is cholesterol, which determines its fluidity and rigidity. Membrane dynamics has so far been deduced from measurements of fluorescence polarization3 or fluorescence recovery after photobleaching (FRAP)4 using specific membrane markers. In addition, polarity-sensitive probes, e.g., 6-dodecanoyl-2-dimethylamino naphthalene (laurdan), whose electronic excitation energy is different in polar and nonpolar environments,5, 6 proved to be useful for membrane studies. Once incorporated into cell membranes, the fluorescence spectrum of laurdan shows a shift toward longer wavelengths when the laurdan molecules come into contact with adjacent water molecules of the cytoplasm, e.g., when membrane stiffness decreases or when a phase transition from the tightly packed gel phase to the liquid crystalline phase of membrane lipids occurs. Fluorescence maxima around and have been observed, so far, and the generalized polarization7 with corresponding to the fluorescence intensity at and to that at .
In addition to fluorescence spectra, measurements of fluorescence anisotropy can give valuable information about membrane dynamics according toand corresponding to the fluorescence intensities with an electrical field vector parallel and perpendicular to the plane of incidence. The faster a molecule rotates between optical excitation and fluorescence, the lower is the expected value . This value can be determined from steady state as well as from time-resolved fluorescence measurements. In the latter case, decays exponentially after short laser pulse excitation according to , the limiting value (observed at times that are long in comparison with the fluorescence lifetime and expressing that the rotational motion of a fluorophore may be hindered in one direction), and the rotational correlation time which is a direct measure of membrane viscosity.3
While conventional wide-field microscopy permits simultaneous measurements of different cell membranes—including the plasma membrane and various intracellular membranes—the plasma membrane can be examined selectively upon total internal reflection (TIR) of a laser beam8, 9 propagating through a medium of refractive index (e.g., glass substrate) and meeting a second medium of refractive index (e.g., cell). Despite being totally reflected at all angles of incidence that are greater than a critical angle , the incident beam establishes an evanescent electromagnetic field that penetrates into the second medium and decays exponentially with the distance from the interface. According to the relationbetween about and more than are attained depending on the wavelength and the angle of incidence .
The present paper reports on fluorescence microscopy of human U373-MG glioblastoma cells incubated with laurdan and presents spectral images and fluorescence anisotropy images upon variation of temperature as well as of the intracellular amount of cholesterol. Changes of cholesterol in cell membranes may affect protein interactions,10 e.g., in pathogenesis of Alzheimer’s disease,11, 12 and may have some influence on drug resistance13 as well as on cell aging.14 Plasma membranes illuminated selectively by TIR illumination are distinguished from intracellular membranes assessed by epi-illumination of whole cells.
Materials and Methods
U373-MG human glioblastoma cells were obtained from the European Collection of Cell Cultures (ECACC No. 89081403). Cells were routinely grown in RPMI 1640 medium supplemented with 10% fetal calf serum and antibiotics at and 5% . The hydrophobic membrane marker 6-dodecanoyl-2-dimethylamino naphthalene (laurdan) was obtained from Molecular Probes (Göttingen, Germany) and prepared as a Stock solution in ethanol. Water-soluble cholesterol and methyl- -cyclodextrine were obtained from Sigma (Munich, Germany). After seeding of , cells were grown on microscope object slides for prior to rinsing with Earl’s balanced salt solution (EBSS) and incubation with laurdan or coincubation with laurdan and or coincubation with laurdan and a :cholesterol complex diluted in culture medium without serum. Cholesterol depletion after application of as well as cholesterol enrichment after application of a :cholesterol complex are well documented in the literature.15 After incubation for in each case, the intracellular amount of cholesterol was either reduced to about 50% or enriched to about 200% ( :cholesterol complex), as determined with an Amplex Red Cholesterol Assay Kit (Molecular Probes) and visualized by coupling of cholesterol to the antibiotic filipin (Sigma). After rinsing with EBSS, cells were measured in an open aluminum chamber at variable temperatures ranging from using a water immersion objective lens. This custom-made chamber (filled with a layer of EBSS) contained a Peltier element for heating or cooling, as well as a calibrated thermocouple for temperature measurements in close vicinity to the measured part of the samples.
As depicted in Fig. 1 , a diode laser with high repetition pulses (LDH 400 with driver PDL 800-B, Picoquant, Berlin, Germany; wavelength: ; pulse energy: ; pulse duration: ; repetition rate: ; average power: ) was adapted to a fluorescence microscope (Axioplan 1, Carl Zeiss, Jena, Germany) using either a multimode fiber for epi-illumination of whole cells or a single-mode fiber-optic system (kineFlex-p-3-S-395, Point Source, Southampton, UK) together with a custom-made condenser unit16 for illumination of the plasma membrane and adjacent cellular sites under total internal reflection (TIR). Commonly, an angle of incidence was used, permitting a penetration depth of the evanescent wave of about , if refractive indices (glass substrate) and (cytoplasm) are assumed. For anisotropy measurements, the electrical field vector was always polarized perpendicular to the plane of incidence. Fluorescence images were recorded by an electron multiplying (EM)–CCD camera with Peltier cooling and a sensitivity below (DV887DC, ANDOR Technology, Belfast, UK)17 in combination with appropriate spectral filters (bandpass filters for and used sequentially for spectral imaging or long-pass filter for used for anisotropy imaging). For anisotropy imaging, two polarized fluorescence images (with an electrical field vector parallel and perpendicular to the plane of incidence) were generated by a polarization beamsplitter and detected simultaneously on different segments of the EM-CCD camera (Dual View, Optical Insights, Tucson, Arizona). Slight differences of sensitivity within the two detection paths were corrected by determination of the so-called G factor, i.e., the factor of fluorescence intensities measured within the two paths when in both cases the polarization of emission and excitation light was parallel. For additional measurements of fluorescence spectra, a custom-made polychromator together with an image-intensifying detection unit (IMD 4562, Hamamatsu Photonics, Ichino-Cho, Japan) was fixed on top of the microscope,18 whereas for measurements of the rotational correlation time, fluorescence decay kinetics was recorded with a time-gated image-intensifying camera (Picostar HR 12; LaVision, Göttingen, Germany) at a temporal resolution of (Ref. 19).
The fluorescence spectrum of U373-MG glioblastoma cells incubated with laurdan is depicted in Fig. 2 together with the molecular structure of laurdan. All spectra are related to single cells either without further treatment (lower curve) or on enrichment of cholesterol (to about 200% of its natural amount; middle curve) or on cholesterol depletion (to about 50% of its natural amount; upper curve). While absolute fluorescence intensities varied from cell to cell up to a factor of about 2, the intensity ratio of the fluorescence bands around and increased with the amount of cholesterol. This finding can be described by an increase of the generalized polarization (GP) with increasing membrane stiffness according to Eq. 1.
GP values determined from fluorescence images recorded within both spectral bands are depicted in Fig. 3 as a function of the intracellular amount of cholesterol on illumination of whole cells [(a): epi-illumination] and plasma membranes [(c): TIR illumination]. Corresponding images of fluorescence intensity [(b), (d), recorded at ] as well as a color scale for the GP values have been added. This figure proves the increasing GP values with increasing amounts of cholesterol and shows that GPs are always higher for the plasma membranes than for inner parts of the cells. This holds for all temperatures in the range , although GP decreases continuously with increasing temperature (data not shown).
Fluorescence Anisotropy Imaging
Polarized measurements of laurdan fluorescence in U373-MG glioblastoma cells always showed higher fluorescence intensity within the plane of the incident electrical field vector than perpendicular to this plane. In addition, time-resolved measurements after picosecond laser pulse excitation showed a fluorescence decrease with a lifetime around (Ref. 19) for both polarizations as well as some delayed increase for the polarization perpendicular to the plane of incidence.
Images of steady state anisotropy determined according to Eq. 2 are depicted in Fig. 4 for whole-cell illumination and in Fig. 5 for TIR illumination as a function of temperature and intracellular amount of cholesterol. These figures clearly show increasing anisotropies with increasing amounts of cholesterol and decreasing temperatures. In addition, anisotropy values are always higher for the plasma membranes (visualized as cell edges on epi-illumination and cell surfaces on TIR illumination) than for inner parts of the cells including the nuclear membrane. A similar result is obtained if instead of the fluorescence anisotropy , the rotational correlation time is determined from fluorescence decay kinetics, as depicted in Fig. 6 . Again, decreases with temperature and increases with the amount of cholesterol (not shown) due to decreasing membrane viscosity in the first case and increasing viscosity in the second case. Obviously, the most pronounced change of occurs in the temperature range between and .
Several parameters characterizing membrane dynamics of living cells have been reported. While the generalized polarization (GP) of laurdan characterizes mechanical stiffness, the fluorescence anisotropy and the rotational correlation time seem to characterize membrane viscosity or (inverse to viscosity) membrane fluidity. Obviously, similar information can be deduced from and . GP on one hand and as well as on the other hand show a pronounced decrease with temperature and an increase with the intracellular amount of cholesterol. This proves an inverse behavior of mechanical stiffness and fluidity of cell membranes. Experiments of staining cholesterol with fluorescent filipin showed that on cholesterol depletion with methyl- -cyclodextrine , this depletion occurred primarily from inner parts of the cells, whereas comparably high cholesterol amounts were maintained in the plasma membranes.20 This explains their higher stiffness and lower fluidity in comparison with intracellular membranes.
A decrease of membrane stiffness with increasing temperature has been well documented for artificial membranes,5, 6, 7 but only in a few cases for living cells.21 While in artificial membranes, this decrease occurred in a small temperature interval (typically, around ), it occurred more gradually in a temperature range between about and for living cells. This is in good agreement with our own experiments and may be explained by the fact that several different membranes are expected to interfere.
It is of long-term interest to know how cholesterol may affect intermolecular interactions in cell membranes and whether it might favor or prevent pathogenesis of pathological processes. Just recently, we could show that enrichment as well as depletion of cholesterol could prevent interaction of amyloid precursor protein (APP) and -secretase,12 which favors the production of amyloid peptide in the pathogenesis of Alzheimer’s disease. In addition, an influence of cholesterol on cell aging is well described in the literature14 and verified by our own experiments (not shown): Aging cells of subcultures 38–42 (i.e., cells split 38–42 times) were stiffer than cells of the subcultures 18–24 generally used for our experiments.
While fluorescence spectra5, 6, 7, 21, 22 and anisotropy22 measurements of laurdan have been well documented in the literature, only few imaging data of living cells are available so far.23, 24 Therefore, it was a main purpose of this paper to present spectral and anisotropy images of single cells and to compare laurdan fluorescence arising from plasma membranes with that arising from whole cells (and dominated by intracellular membranes). Switching of the illumination mode between total internal reflection (TIR) and epi-illumination is performed easily by two different illumination paths and a sliding mirror, as depicted in Fig. 1. TIR illumination with a hemispherical or hemicylindrical glass prism (Fig. 1) has further advantages in comparison with commercial TIR devices using high aperture objective lenses: easy variation of the penetration depth of the evanescent electromagnetic field by variation of the angle of incidence, and use of any kind of objective lens, even of rather low numerical aperture. This is essential for selection of larger object fields and for maintaining the polarization of light in fluorescence anisotropy measurements.
Imaging experiments profit to a large extent from the availability of highly sensitive EM-CCD cameras with excellent signal-to-noise ratio. Microscopic resolution in spectral and anisotropy imaging, however, is commonly limited by the Abbé criterion to about but can be improved to a few nanometers in the axial direction by variable-angle TIR microscopy.16 So, microdomains of different membrane stiffness or fluidity can be visualized down to diameters around , whereas cholesterol-enriched lipid rafts of only a few nanometers in size cannot be resolved.25 The excellent axial resolution, however, permits us to measure cholesterol-dependent cellular adhesions on a substrate in a nanometer scale, and loosening of those adhesions on depletion of cholesterol—finally leading to detachment of whole cells—has previously been well documented.26
The authors thank C. Hintze (Hochschule Aalen) for cell cultivation as well as P. Kruse and E. Winkler (ILM Ulm) for their support in quantitative measurements of cholesterol. Research was funded by the Bundesministerium für Bildung und Forschung (BMBF) and the Land Baden-Württemberg.