Chapter 17 - Analysis of Cholesterol Trafficking with Fluorescent Probes

https://doi.org/10.1016/B978-0-12-386487-1.00017-1Get rights and content

Abstract

Cholesterol plays an important role in determining the biophysical properties of biological membranes, and its concentration is tightly controlled by homeostatic processes. The intracellular transport of cholesterol among organelles is a key part of the homeostatic mechanism, but sterol transport processes are not well understood. Fluorescence microscopy is a valuable tool for studying intracellular transport processes, but this method can be challenging for lipid molecules because addition of a fluorophore may alter the properties of the molecule greatly. We discuss the use of fluorescent molecules that can bind to cholesterol to reveal its distribution in cells. We also discuss the use of intrinsically fluorescent sterols that closely mimic cholesterol, as well as some minimally modified fluorophore-labeled sterols. Methods for imaging these sterols by conventional fluorescence microscopy and by multiphoton microscopy are described. Some label-free methods for imaging cholesterol itself are also discussed briefly.

Introduction

Cholesterol is an essential component of mammalian cell membranes, and it plays an important role in determining the biophysical characteristics of these membranes (Mesmin and Maxfield, 2009, Wüstner, 2009). The concentration of cholesterol varies greatly among organelles with levels around 30% of the lipid molecules in the plasma membrane and about 5% in the endoplasmic reticulum. Cholesterol is transported among organelles by a mixture of vesicular and non-vesicular transport processes, and the mechanisms regulating this transport are only partially understood. Cholesterol transport in cells can be studied by following radiolabeled cholesterol, but this requires stringent purification of organelles under conditions in which the sterol does not redistribute. Contamination with a small fraction of membranes with high cholesterol content can lead to significant errors in measurements of cholesterol content in organelles such as the endoplasmic reticulum.

Fluorescence microscopy has been a powerful tool for studying the intracellular transport of proteins. A difficulty in studying transport of cholesterol (and other lipids) using microscopy is that coupling to a fluorophore can dramatically change the properties of the molecule and its interactions with other components of the membrane. Recently, two types of approaches for studying fluorescent sterols have been implemented. One approach has been to use an added fluorophore selected to minimize the perturbation as compared to cholesterol itself. Bora-diaza-indacene (BODIPY)-cholesterol (Fig. 1) has been used as a cholesterol probe in model membranes and in trafficking studies in living cells (Hölttä-Vuori et al., 2008), (Shaw et al., 2006), (Chiantia et al., 2008), (Ariola et al., 2009), (Wüstner et al., 2011a). The second approach takes advantage of intrinsically fluorescent sterols (Fig. 1) (Hao et al., 2002), (Yeagle et al., 1982), (Fischer et al., 1974), (Wüstner, 2010), (McIntosh et al., 2008), (Mukherjee et al., 1998), including dehydroergosterol (DHE; a natural sterol found in yeast) and cholestatrienol (CTL; a synthetic sterol that is close to cholesterol in its structure and in many of its biophysical properties). Both of these fluorescent sterols have two additional double bonds in the steroid rings, which creates the fluorophore. Fluorescent sterols can be added to cells either by delivery to the plasma membrane or incorporated into reconstituted lipoproteins. The movement of the sterols through the cells can then be observed directly, or they can be used for techniques such as fluorescence recovery after photobleaching to determine transport kinetics (Hao et al., 2002, Wüstner et al., 2005, Wüstner et al., 2002).

In this chapter, we discuss the relative merits of these fluorescent sterols, and we discuss the methodology for their use. We also discuss some recent developments in the label-free detection of sterols in cells. We begin by discussing the use of fluorescent sterol-binding molecules that are useful for analyzing the distribution of cholesterol in fixed cells or tissues.

Sterol-binding natural products have been used to localize cholesterol in cells, and their use has been described in detail (Gimpl, 2010). Filipin is a naturally fluorescent polyene antibiotic that binds to cholesterol but not to esterified sterols. Thus, it is useful for detecting free (i.e., unesterified) cholesterol in biological membranes. Filipin fluorescence is observed with UV excitation around 360 nm and emission around 480 nm. Filipin binding perturbs the bilayer structure, so filipin cannot be used on living cells. Beyond the free 3’-OH group, the exact basis for filipin's specificity is not fully understood. There are some important practical considerations in the use of filipin. Stock solutions in DMSO must be rigorously dried with Molecular Sieves to remove residual water. Filipin is rapidly photobleached with the UV light intensity available in most fluorescence microscopes. However, with a good camera, we have been able to attenuate the incident light using neutral density filters to 1–10% of the full brightness and obtain good images that photobleach slowly. Excessive photobleaching without the use of a neutral density filter is one of the greatest problems in reproducibility of results using filipin. Since the basis for filipin binding is not fully understood, it must be considered that there can be interfering effects. Nevertheless, we have been able to correlate brightness of filipin labeling with chemically measured cholesterol levels in cells in which cholesterol content was altered by incubation with methyl-β-cyclodextrin (MβCD) or with cholesterol:MβCD complexes (Qin et al., 2006). Filipin treatment causes dimpling of the membrane that can be observed by electron microscopy (Orci et al., 1981). It should be noted, however, that filipin deformation of the membrane can be affected by membrane:protein interactions (Steer et al., 1984). A recent paper discussed the use of filipin to label brain sections from mice with lysosomal storage disorders (Arthur et al., in press). It was found that in addition to cholesterol, filipin was also labeling the GM1 ganglioside. This points to the importance of verifying the validity of sterol-labeling reagents in each experimental system. In addition, particular care should be taken in cells that express large amounts of GM1.

Pore-forming cytolysins that bind to sterols have also been used to visualize cholesterol in cells. These polypeptide bacterial toxins bind to cholesterol in membranes and self-associate to form pores in the bilayer. For fluorescence microscopy the toxins can be labeled with a fluorescent dye and imaged with filter sets appropriate for the dye. Some investigators have preferred the use of these toxins, including perfringolysin-O or a biotinylated derivative of this called BCθ-toxin, because the fluorescent dyes (or the labeled avidin for BCθ-toxin) are more photostable than filipin. One issue is that the perfringolysin-O binding to membranes increases very nonlinearly as cholesterol content increases (Sokolov and Radhakrishnan, 2010). It appears that the binding is preferentially to cholesterol molecules with a high chemical activity coefficient. The cholesterol content at which the rapid increase in binding is observed depends on the lipid content of the membrane. For ER lipids, this occurs at about 5% cholesterol, but in other membranes binding is seen above 20% cholesterol (Gimpl, 2010).

Filipin labeling has been used for many years for diagnosis of Niemann-Pick Disease type C (NPC), an inherited disorder leading to cholesterol accumulation in lysosomal storage organelles (LSOs), which are modified late endosomes and lysosomes (Peake and Vance, 2010). Recently, filipin labeling of the cholesterol accumulation in LSOs has been used for high throughput automated microscopy screening of compounds that might reduce the cholesterol accumulation (Pipalia et al., 2006, Pipalia et al., 2011). In this application, imaging is essential because the total cholesterol level in the NPC mutant cells does not change greatly. Nevertheless, differences caused by the mutation are easily quantified by microscopy because the filipin-labeled LSOs accumulate in the peri-nuclear region, while the rest of the cell actually has lower cholesterol levels than normal cells (Pipalia et al., 2006).

Cells are plated in 384-well plates, treated with various concentrations of test compounds, fixed with 1.5% paraformaldehyde, and labeled with 50 μg/mL filipin (from a 25 mg/mL stock in dry DMSO) in phosphate buffered saline (PBS) for 45 min. The cells are then rinsed and imaged on an automated microscopy system (ImageXpressMicro from Molecular Devices or equivalent) using UV excitation and a dry 10× objective. Automated image analysis is used to quantify the fluorescence power per cell in areas that are above a threshold brightness corresponding to the brightness in LSOs of untreated cells. Further details are provided elsewhere (Pipalia et al., 2006, Pipalia et al., 2011). Screens based on this procedure have identified compounds that reduce the cholesterol levels in NPC mutant cells by various mechanisms (Pipalia et al., 2011), (Rosenbaum et al., 2009), (Rujoi et al., 2010), (Rosenbaum and Maxfield, 2011).

Valid fluorescent analogs of cholesterol should reproduce its biophysical interactions with lipids in the bilayer as closely as possible. Cholesterol has a small headgroup (the 3′-β-hydroxyl group) that can create hydrogen bonds to other lipid headgroups in the interfacial region of the bilayer, while its hydrophobic steroid ring system and isooctyl side chain are buried in the hydrophobic region of the membrane (Ohvo-Rekila et al., 2002), (Nagle and Tristram-Nagle, 2000), (Hofsass et al., 2003). Cholesterol molecules are shielded under the phospho- and sphingolipid headgroups, thereby minimizing contact with water molecules. This shielding, together with attractive van der Waals interactions with neighboring acyl chains, causes lipid membranes to have a reduced area in the presence of cholesterol. This property is described as cholesterol's condensing effect (Hofsass et al., 2003), (Lindahl and Edholm, 2000), (Henriksen et al., 2006).

When cholesterol is added to a membrane consisting of phospholipids with saturated acyl chains, (e.g., dipalmitoylphosphatidylcholine) below the phase transition temperature of the phospholipid, the sterol will perturb the high structural order in the liquid-crystalline state (Vist and Davis, 1990). In contrast, when added to membranes in the liquid-disordered (ld) phase, cholesterol has an ordering effect, and it can induce a liquid-ordered (lo) phase at high sterol mole fractions (approx. above 30 mol% -- that is, 30% of the lipid molecules) (Ipsen et al., 1987). The lo phase is characterized by high lateral lipid mobility, straightened acyl chains and tight lipid packing. The lo phase has attracted interest, due to similar properties observed in cellular membranes (Ahmed et al., 1997), (Mukherjee and Maxfield, 2004), (Munro, 2003). Despite this resemblance, it has to be emphasized that the lo phase is strictly defined only for simple two and three-component lipid mixtures at thermodynamic equilibrium. Partitioning of fluorescent sterols between lo and ld phase as well as their potential to induce the lo phase can be used as criteria to assess their potential as cholesterol mimic (Wüstner et al., 2011b, Garvik et al., 2008; L. Solanko, D. Wüstner et al., manuscript in preparation).

Both intrinsically fluorescent sterols, DHE and CTL, contain three conjugated double bonds in the steroid ring system giving these probes their slight fluorescence in the near UV-region of the spectrum. Since they contain an identical chromophore, the photo-physical properties of DHE and CTL are nearly identical (Smutzer et al., 1986), (Rogers et al., 1979), (Chong and Thompson, 1986), (Yeagle et al., 1990). Both sterols have excitation and emission maxima in membranes of λex = 320 nm and λem = 370–400 nm (Fischer et al., 1974, Rogers et al., 1979, Schroeder et al., 1988, Hyslop et al., 1990). The fluorescence lifetime of both sterols is short, around τf = 0.3–0.8 ns in various organic solvents and lipid membranes at room temperature (Smutzer et al., 1986, Chong and Thompson, 1986, Hyslop et al., 1990, Fischer et al., 1985a). Low extinction coefficient (ɛ ≈ 11,000 M-1·cm-1) and quantum yield (Φf = 0.04 in ethanol) results in low fluorescence brightness of these sterols. The environmental sensitivity of DHE and CTL is low since the molecular dipole does not change greatly during electronic transition from the ground S0, state to the excited S1 state (Chong and Thompson, 1986, Yeagle et al., 1990, Hyslop et al., 1990, Fischer et al., 1985a). In dimyristoylphosphatidylcholine liposomes, the fluorescence lifetime, τf, of DHE decreases in a sigmoidal fashion from τf = 2 ns at 10 °C to τf = 0.5 ns at 42 °C (Chong and Thompson, 1986). It has been reported that DHE and CTL self-quench at increasing mole fractions in phosphatidylcholine model membranes, as indicated by a drop in fluorescence intensity (Schroeder et al., 1988, Schroeder et al., 1987). Since no change in fluorescence lifetime was observed in these studies, it is likely that the proposed self-quenching is caused by static quenching (Lakowitz, 2006). By measuring fluorescence intensity of DHE in giant unilamellar vesicles (GUVs), we found no evidence for self-quenching but instead a linear relationship between DHE's emission and the mole fraction in the membranes (Garvik et al., 2008). Thus, quenching effects, which could complicate analysis of sterol distribution by fluorescence microscopy of DHE or CTL are unlikely, but systematic studies with varying lipid composition might be necessary to rule out any sterol self-quenching at concentrations used for studies in cells. Fluorescence anisotropy and quantum yield of DHE are very temperature-sensitive and much lower in membranes above than below the phase transition temperature (Smutzer et al., 1986, Chong and Thompson, 1986, Fischer et al., 1985a). Accordingly, under live cell imaging conditions the fluorescence brightness of both sterols should be further reduced. Another unfortunate property of DHE and CTL, at least for cellular imaging, is the high photobleaching propensity of these probes. How this can be avoided or used to advantage for microscopic investigation will be discussed later in this chapter.

Due to their minimal chemical alterations, DHE and CTL resemble the natural sterols ergosterol and cholesterol very closely. DHE differs from ergosterol only in one double bond, while CTL contains two more double bonds in the ring system than cholesterol. Accordingly, DHE is a naturally occurring sterol in yeast and red sponges, and it is the ideal mimic of ergosterol. CTL is the closest analog of cholesterol (Rogers et al., 1979, Wüstner, 2007a). The close resemblance of DHE and CTL to ergosterol and cholesterol is reflected by their biophysical properties in model membranes. DHE can, like ergosterol and cholesterol, induce the lo phase in ternary lipid mixtures and partitions with high preference into that phase (Garvik et al., 2008). CTL also prefers the lo over the ld phase in GUV membranes (Baumgart et al., 2007). CTL has a very similar potential to order acyl chains in membranes as cholesterol even at high sterol mole fraction, while DHE is slightly less efficient (Scheidt et al., 2003). Similarly, DHE (like ergosterol, but in contrast to cholesterol) shows a concentration saturation effect in various biophysical effects on lipid membranes. That means that bilayer properties such as bending elasticity and ordering of phospholipid acyl chains depend on sterol mole fraction only up to about 10 mol% DHE in the membranes (Henriksen et al., 2006, Garvik et al., 2008, Scheidt et al., 2003, Henriksen et al., 2004). In contrast, acyl chain ordering was found to depend linearly on cholesterol and CTL concentrations in the membrane, even at high sterol mole fractions (Henriksen et al., 2006, Scheidt et al., 2003). The less flexible side chain of DHE and ergosterol as compared to CTL and cholesterol is likely responsible for this difference. We found that DHE and CTL have very similar intracellular trafficking itineraries to cholesterol, suggesting that the minor differences in biophysical properties of the analogs are not determining their overall transport in mammalian cells (Wüstner and Færgeman, 2008a), (Hartwig Petersen et al., 2008), (Mondal et al., 2009). Interestingly, in a study in yeast cells, we observed different metabolism of DHE and ergosterol compared to cholesterol, indicating that DHE is the most suitable sterol analog in organisms containing ergosterol (Kohut et al., 2011).

Recently, cholesterol analogs containing a BODIPY fluorophore have been synthesized by Bittman and co-workers (Li et al., 2006). The BODIPY moiety is particularly suitable for fluorescent lipid probes since it is relatively non-polar, allowing for insertion of the analogs into the hydrophobic interior of lipid membranes (Marks et al., 2008). The BODIPY dye is electrically neutral, has a low environmental sensitivity, low Stokes shift (typically, λex = 505 nm and λem = 515 nm), high extinction coefficient, and high quantum yield (Φf ≈ 0.9 in organic solvents) (Bergström et al., 2002). Several BODIPY dyes exhibit a red-shifted excited-state dimer (excimer) at high concentration in membranes (Chen et al., 1997), (Puri et al., 2001), (Pagano et al., 1991). This has been used to determine lateral clustering and sorting of these lipid probes in cells by quantitative fluorescence microscopy (Puri et al., 2001), (Puri et al., 1999), (Sharma et al., 2003), (Sharma et al., 2004). BODIPY-cholesterol (BChol) with the dye at carbon 24 of the sterol side chain is a promising new cholesterol analog, which can supplement the well-established intrinsically fluorescent probes DHE and CTL (Hölttä-Vuori et al., 2008), (Shaw et al., 2006), (Chiantia et al., 2008), (Ariola et al., 2009), (Wüstner et al., 2011a). BChol partitions preferentially into the lo phase compared to the ld phase in various ternary lipid mixtures (Shaw et al., 2006, Ariola et al., 2009), though a direct comparison with DHE revealed that the latter sterol has an even higher preference for the lo phase (Wüstner et al., 2011a). Since Bchol is more than 500-fold brighter than DHE, it can be used at very low concentrations (about 0.1–0.5 mol% of lipids) (Hölttä-Vuori et al., 2008, Wüstner et al., 2011b). In contrast, visualization of DHE in cells may require replacing up to 5% of the sterols (Wüstner et al., 2011b, Wüstner et al., 2002). BChol has been used to investigate cholesterol mobility in the lo and ld phase in GUV's made of dioleoylphosphatidylcholine, egg yolk sphingomyelin, and cholesterol (Ariola et al., 2009). The BODIPY-moiety of this BChol is oriented perpendicular to the bilayer normal (and the acyl chains), as measured by two-photon fluorescence polarization (Ariola et al., 2009; L. Solanko, D. Wüstner et al., manuscript in preparation), and the strength of this orientation is enhanced in the presence of cholesterol in the membranes (L. Solanko, D. Wüstner et al., manuscript in preparation). BChol self-quenches at concentrations above 3 mol% in lipid membranes, but it does not form red-shifted excimers (Wüstner et al., 2011a). Instead, dark ground state dimers seem to be responsible for the self-quenching effect, as reported for some other BODIPY-tagged probes (Wüstner et al., 2011b, Bergström et al., 2002, Qin et al., 2005). Molecular dynamics simulations of BChol indicate that it has a higher molecular tilt compared to cholesterol (Hölttä-Vuori et al., 2008). This can affect the local lipid structure surrounding the probe and might contribute to its lower partition into the lo phase compared to DHE (Hölttä-Vuori et al., 2008, Wüstner et al., 2011b). The molecular tilt has been proposed to be a main determinant of the ability of a sterol to order phospholipid acyl chains and condense lipid bilayers (Aittoniemi et al., 2006). For the intrinsically fluorescent sterols, DHE and CTL, the order parameter, S, which is inversely related to the molecular tilt, could be measured by fluorescence polarization spectroscopy, since the transition dipole of these sterols lies along the long molecular axis (Chong and Thompson, 1986, Fischer et al., 1985a). The order parameter was found to be only a little lower for DHE compared to cholesterol (S = 0.67 for DHE and S = 0.85 for cholesterol in dimyristoylphosphatidylcholine at 37 °C)(Fischer et al., 1985a, Fischer et al., 1985b, Marsan et al., 1999).

Cholesterol tagged with a 7-nitrobenz-2-oxa-1,3-diazole (NBD)-group at carbon 22 or carbon 25 has been used in model membrane and cellular trafficking studies in yeast and in mammalian cells (Shaw et al., 2006, Mukherjee et al., 1998, Scheidt et al., 2003, Craig et al., 1981, Reiner et al., 2005). A problem with these probes is their up-side down orientation in model membranes compared to cholesterol and intrinsically fluorescent sterols, as well as their low ordering capacity and partitioning into the ld phase in ternary model membranes (Scheidt et al., 2003, Loura et al., 2001). NBD-cholesterol with the fluorophore at carbon 25 has been shown to be mistargeted in cells to mitochondria (Mukherjee et al., 1998). Dansyl-cholesterol is another fluorescent cholesterol analog used in cellular studies (Wiegand et al., 2003, Huang et al., 2010). The Dansyl moiety was linked to carbon 6 of the steroid ring system, and recent fluorescence studies found that the Dansyl-group of this sterol is localized on average 1.56 nm from the bilayer center (Shrivastava et al., 2009). This should significantly affect the lipid acyl chain packing in proximity of this probe when inserted into membranes. Partitioning of Dansyl-cholesterol between lo and ld phases in model membranes has not been reported. Quantitative studies of intracellular sterol distribution based on fluorescence of NBD- and Dansyl-cholesterol are also hampered by the high environmental sensitivity of the attached fluorophores. Accurate measurement of sterol distribution requires that emitted fluorescence is proportional to probe concentration, and this is not likely for these cholesterol probes (Benson et al., 1985). Further details about NBD- and Dansyl-cholesterol can be found elsewhere (Gimpl, 2010, Wüstner, 2007a).

Both DHE and CTL are suitable for studies in living cells. DHE is widely available from commercial sources, although concerns have been expressed about the purity of this material, which may vary depending on the supplier and the lot number (McIntosh et al., 2008). CTL is not available commercially, but the method for synthesis has been published (Fischer et al., 1984, Fellmann et al., 1994). Like many lipids, these fluorescent sterols are subject to oxidation, so they should be protected from exposure to air (e.g., by purging solvents with argon). They are also sensitive to light and should be stored in the dark. The purity can be checked by HPLC (McIntosh et al., 2008). Of particular concern, oxidized DHE or CTL may affect the structure of lipid bilayers.

Several methods have been used to incorporate DHE or CTL into cells. For simplicity, we will describe methods for DHE, but the same methods would apply for CTL. The simplest method is to inject DHE in an ethanolic stock solution into the culture medium. The DHE is very poorly soluble in water; some of it will adsorb to serum proteins, but most will form microcrystals. These microcrystals may be taken up by the cells and slowly dissolved to allow the DHE to distribute into cell membranes. In our experience, this procedure results in very heterogeneous labeling of cells and incomplete breakup of the microcrystals. A much better procedure involves preparation of DHE complexes with MβCD, which solubilizes the sterol and allows it to be rapidly exchanged into the plasma membrane (Box 2). Sterol: MβCD complexes form at a 1:2 ratio (Breslow and Zhang, 1996), but these complexes can dissociate rapidly. Thus, it is necessary to maintain an excess of the MβCD in order to store sterol: MβCD complexes without precipitation of the DHE.

In normal cell physiology, cholesterol is delivered to cells as cholesteryl esters in the core of lipoproteins. DHE-oleate esters can be synthesized and incorporated into the core of reconstituted LDL (Wüstner et al., 2005). Established procedures for reconstituting LDL can be modified to incorporate DHE-oleate into the core of LDL or acetylated-LDL (Krieger, 1986). We have used this method to incorporate a mixture of DHE-oleate and cholesteryl-oleate into acetylated-LDL at a molar ratio of 1:10, using 6 mg of total neutral lipid per 1.9 mg of AcLDL protein (Wüstner et al., 2005).

The LDL can be taken into cells by receptor-mediated endocytosis via the LDL receptor. For uptake by macrophages, the reconstituted LDL can be acetylated, and the acetylated-LDL is taken up via scavenger receptor type A. The esters are hydrolyzed in late endosomes and lysosomes, releasing the unesterified DHE. We have verified that DHE esters are hydrolyzed, transported out of the digestive organelles, and distributed among cellular membranes (Wüstner et al., 2005). DHE and CTL are also effective substrates for the esterifying enzyme, ACAT, which forms steryl esters that are incorporated into cytoplasmic lipid droplets, and these can be visualized in cells (Wüstner et al., 2005).

Although in this chapter we emphasize studies of mammalian cells in culture, Caenorhabditis elegans can be labeled with DHE by dietary feeding (Matyash et al., 2001, Wüstner et al., 2010). The yeast, Saccharomyces cerevisiae, can also take up DHE when grown under hypoxic conditions (Georgiev et al., 2011), and the distribution and transport of the DHE can be analyzed by microscopy. For yeast, DHE is a naturally formed sterol, and it is also a natural component of the diet of Caenorhabditis elegans.

Photobleaching of DHE can be a problem when several frames of the same field of view are acquired (e.g., in time-lapse imaging or when determining the effect of a drug or fluorescence quencher). Photobleaching of DHE can be used to advantage, however, for dynamic measurements of sterol transport. After selectively destroying DHE fluorescence by illumination of only a small region with closed field aperture on a wide field microscope, fluorescence recovery of DHE into that region was measured over time with the field aperture opened to image the whole cell (Hao et al., 2002, Wüstner et al., 2005, Wüstner et al., 2002, Wüstner, 2007b). To quantify fluorescence recovery, DHE intensity in the bleached region was normalized to total cell intensity, thereby correcting for fluorescence loss during repeated acquisition.

Transbilayer asymmetry is a general feature of most lipids in the plasma membrane and many other organelles. This asymmetry has important consequences for membrane physical properties and cell signaling. Although cholesterol is a major lipid in these membranes, its transbilayer distribution is not well understood. Fluorescent sterols such as DHE and CTL can be used with fluorescence quenchers that do not cross the bilayer to determine the sterol asymmetry in membranes (McIntosh et al., 2008, Mondal et al., 2009, Hale and Schroeder, 1982). In most such studies, it has been found that the abundance of DHE is greater on the cytoplasmic leaflet of the plasma membrane than on the exofacial leaflet. Using imaging methods described here, it is also possible to examine the DHE transbilayer distribution in organelles. When the membrane impermeant quencher, trinitrobenzene sulfonic acid, was microinjected into DHE-labeled cells, about 60% of the fluorescence in both the endocytic recycling compartment and the plasma membrane was quenched (Hartwig Petersen et al., 2008). When the trinitrobenzene sulfonic acid was added outside the cell, only 20–30% of the fluorescence was quenched.

DHE-oleate that is delivered to lysosomes in the core of LDL particles is hydrolyzed, presumably by lysosomal acid lipase, the lysosomal enzyme that hydrolyzes cholesteryl esters (Wüstner et al., 2005). DHE is also esterified by the endoplasmic reticulum enzyme, ACAT, and stored in lipid droplets. Esterification of DHE can be verified by extracting cellular neutral lipids and using HPLC to quantify free and esterified DHE (Hölttä-Vuori et al., 2008, Wüstner et al., 2005). Incorporation into lipid droplets can be measured by fluorescence microscopy by determining the co-localization of DHE with a lipid droplet vital stain such as Nile Red or Lipidtox (Wüstner et al., 2011b, Wüstner et al., 2005, Wüstner and Færgeman, 2008b). While DHE mimics many aspects of cholesterol in cells, it is not recognized by the regulatory proteins, SCAP and Insig, in the endoplasmic reticulum (A. Radhakrishnan, personal communication).

Excitation of DHE requires UV wavelengths around 330 nm, and the emission is around 400 nm. Both of these wavelengths impose special requirements as compared to typical epi-fluorescence microscopy. Most fluorescence microscopes contain glass elements that block transmission of light below 340 nm, and this greatly diminishes the excitation of DHE unless modifications are made. The most important modification is in the lamp housing. Conventional collecting lenses are a few cm thick, and they completely block transmission at 330 nm. Most research grade microscopes can be modified to incorporate UV-transmitting collecting lenses. An issue with single element collecting lenses is that the focal points for UV and visible wavelengths are very different, so optimal Köhler illumination cannot be obtained for the visible and UV simultaneously. We have obtained a multi-lens collector from Leica that corrects illumination adequately for wavelengths from 335 to 630 nm. If fiber optical systems are used for excitation, it is essential to verify that they have high transmittance at 330 nm.

Dichroic filter cubes for DHE imaging are available from various suppliers. We have used 335 nm (20 nm bandpass) excitation filter/365 nm longpass dichromatic filter/405 nm (40 nm bandpass) emission filter from Chroma (Brattleboro, VT). A microscope objective must also be chosen that has significant transmission below 340 nm. Finally, various glass elements in the optical path must be checked for transmission at 330 nm and replaced if necessary. For example, some infrared-blocking (heat) filters have poor transmission at 330 nm and must be replaced.

Because DHE fluorescence is weak and in the near UV, it places special requirements on the detector. The sensitivity of many CCD cameras falls off sharply around 400 nm, but detectors are available with coatings that extend sensitivity into the near UV. An important development is the availability of near-UV sensitive electron multiplying CCD (EMCCD) cameras. These cameras can provide a large increase in signal-to-noise ratio for detecting DHE as compared to previous versions of commercially available cameras. Using these near-UV sensitive EMCCD cameras, excitation light intensity can be reduced by 80–90% while still obtaining high quality images with less than 1 s exposure time. This allows repeated observations of the same cell for time course studies or for obtaining images at multiple focal positions.

Despite the advances in imaging technology described above, reliable detection of sub-cellular DHE and CTL distribution remains a challenge due to the weak fluorescence of these sterol probes. One potential problem of wide field imaging is a significant contribution of out-of-focus light (Swedlow, 2007). Image restoration methods based on deconvolution can improve both axial and lateral resolution in wide field fluorescence images (Swedlow, 2007), (Young, 1989), (Sibarita, 2005). The performance of any deconvolution algorithm is limited by spherical aberration, image noise and by photobleaching of the fluorophore as multiple images are obtained at different focal planes (Sibarita, 2005, Markham and Conchello, 2001). We have shown that iterative deconvolution in combination with photobleaching correction can significantly improve the performance of DHE imaging (see Fig. 2) (Wüstner and Færgeman, 2008b).

Chromatic aberration caused by the wavelength dependence of the microscope optical glass elements is a potential problem in multicolor imaging with DHE and organelle markers. While good research objectives are normally corrected for chromatic aberration in the visible range of the spectrum, they fail to collect UV light at the same focal position as red and green light. This can be corrected for by acquiring z-stacks of multicolor fluorescent beads, measuring the axial and lateral off-set between the UV channel and red or green channels and correcting for these off-sets in a post-processing step (Wüstner and Færgeman, 2008b). Software has been developed that performs a nonlinear regression of various decay models to photobleaching-induced intensity loss on a pixel-by-pixel basis (Wüstner et al., 2010). This analysis revealed that DHE bleaches homogenously throughout cells, in stark contrast to other fluorescently labeled sterols, for example NBD-cholesterol (Benson et al., 1985, Wüstner et al., 2010). Moreover, DHE bleaching follows a simple mono-exponential decay, and the measured decay rate is comparable in model and cellular membranes and quite independent of DHE intensity (Wüstner et al., 2010). Accordingly, bleach-rate imaging can be used as a fingerprint to detect DHE in the presence of other spectrally indistinguishable fluorophores. For example, we showed that bleach-rate imaging of DHE allows for selective detection of sterol enriched tissue in Caenorhabditis elegans despite the presence of significant autofluorescence (Wüstner et al., 2010, Wüstner and Sage, 2010).

As an alternative to conventional epifluorescence microscopy, intrinsically fluorescent sterols can be monitored in cells using multiphoton (MP) excitation. In this imaging modality, almost simultaneous absorption of two or more photons causes an electronic transition in the fluorophore from the ground to the first excited state. The absorbed photons have approximately one-half (for two-photon excitation) or one-third of the energy (for three-photon excitation) compared to the one-photon excitation process. Thus, as a rough rule of thumb, excitation of a fluorescent probe by two- or three-photon microscopy can be stimulated using twice or three times the wavelength used for one-photon excitation. For example, DHE is excited around 320 nm by a one-photon process, and it can be excited with 920 nm light for MP microscopy (McIntosh et al., 2008), (Wüstner et al., 2010), (McIntosh et al., 2003), (Frolov et al., 2000), (Zhang et al., 2005), (Wüstner et al., 2011b). The longer excitation wavelength has several advantages for UV microscopy including less cytotoxicity and photobleaching, deeper specimen penetration and less light scattering. Apart from these advantages, the key benefit of using MP microscopy is its intrinsic sectioning capability since multiphoton events only occur in the focal plane (Wüstner, 2010), (Zipfel et al., 2003), (Masters and So, 2008), (Xu and Zipfel, 2008), (Schrader et al., 1997).

Schroeder, Webb, Gratton and colleagues, pioneered MP excitation microscopy of DHE (Frolov et al., 2000). Using an excitation wavelength of λex = 920–930 nm, these authors visualized DHE in L-cell fibroblasts and found the sterol mainly in the plasma membrane and in lipid droplets (LDs) (Frolov et al., 2000). The published images were, however, of low quality (i.e., low signal-to-noise ratio, SNR), and no direct comparison to non-stained cells was made. The extremely low signal of DHE in MP microscopy is mostly a consequence of the low propensity of DHE for multiphoton excitation (Wüstner, 2010, Wüstner et al., 2011a). Accordingly, high laser powers are required to obtain the necessary photon density at the focal point. We compared wide field and MP microscopy of DHE directly in the same samples and found that both methods provide similar information about sterol distribution (Wüstner et al., 2011b). Average laser powers in the range of 40–70 mW were required for MP excitation of DHE at λex = 920 nm in CHO and HepG2 cells, and even then the images of DHE-stained cells were of low SNR. For comparison, two-photon excitation of enhanced green fluorescent protein at the same excitation wavelength required a laser power of only 0.3 mW (Wüstner et al., 2011b). One could also try to excite DHE more efficiently by a two-photon process using excitation wavelengths ≤700 nm. Our attempts to do that in cells, however, were unsuccessful due to overwhelming autofluorescence. Photobleaching of DHE is restricted to the focal plane in MP microscopy. Accordingly, we were able to acquire many frames (typically between 50 and 100) without significant intensity loss in MP microscopy of DHE. Although individual frames were of low SNR, averaging many frames increased the SNR dramatically. The frame-averaged MP image of DHE has a higher lateral resolution than a corresponding single-frame image of the same specimen acquired on an epifluorescence microscope.

As has been reported for other fluorescent lipids, we did not find evidence for lateral domains of DHE in the plasma membrane of living cells by deconvolution of wide field microscopy images or by MP microscopy (Wüstner and Færgeman, 2008a, Wüstner, 2007b, Wüstner et al., 2011a). By both methods, fine cell protrusions in various cell types and sterol-containing nanotubes could be observed by microscopy of DHE. The need for frame averaging dramatically lowers the time-resolution of MP microscopy of DHE. While one image of DHE stained cells with good SNR can be acquired in less than 0.5 s by widefield microscopy, a comparable image consists of about 30 averaged frames in MP microscopy, which requires between 1 and 2 min total acquisition time (depending on the pixel-dwell time) (Wüstner et al., 2011b). Accordingly, dynamic events occurring on a shorter time scale cannot be monitored and will result in motion-blurring. Despite the slow acquisition, this strategy has been used to record z-sections of DHE labeled CHO cells, and the sterol distribution in three dimensions could be observed (Wüstner et al., 2011b).

In MP microscopy of DHE, a photomultiplier tube operating in photon-counting mode is used as the detector. In photon-counting the main source of noise in detection is the statistical uncertainty in the number of detected photons in a fixed time interval (Yazdanfar and So, 2008). Computational progress in denoising routines allowed us to post-process individual frames of MP sequences of DHE stained cells achieving dramatic improvement in image quality (F. Lund, D. Wüstner et al., manuscript in preparation; Luisier, 2010). After denoising MP microscopy images, we could follow the dynamics of individual vesicles containing DHE in living CHO cells with a time-resolution of 4 s (F. Lund, D. Wüstner et al., manuscript in preparation). We found similar types of vesicle dynamics by wide field microscopy of DHE stained HepG2 and J774 cells (Wüstner et al., 2002, Wüstner and Færgeman, 2008a). Since intracellular vesicles must contain significant amounts of DHE (about 3--5 mol%) to be detectable by either wide field or MP microscopy, both imaging modalities will only catch dynamics of sterol-enriched vesicles.

As discussed earlier, BChol is the only fluorescent cholesterol analogue with an extrinsic fluorescence moiety that partitions preferentially into the lo phase in model membranes (Ariola et al., 2009, Wüstner et al., 2011b). Thus, it fulfills an essential requirement of any suitable fluorescent analog of cholesterol. Co-labeling of M19 cells, a partial sterol-auxotroph Chinese hamster ovary (CHO) cell-line, with 3H-cholesterol and BChol followed by sucrose density fractionation indicated a comparable distribution of cholesterol and BChol among cell membranes of varying equilibrium density (Hölttä-Vuori et al., 2008). Efflux of BChol from CHO cells to extracellular acceptors like BSA or apoA1 is significantly higher than that of cholesterol (Hölttä-Vuori et al., 2008). This is likely a consequence of the less efficient packing of phospholipid acyl chains around the BODIPY-moiety, as suggested by molecular dynamics simulations (Hölttä-Vuori et al., 2008) and by studies in model membranes (Shaw et al., 2006, Ariola et al., 2009, Wüstner et al., 2011b). BChol becomes esterified by Raw264.7 macrophages after incubation with AcLDL, though with a lower efficiency than cholesterol (Hölttä-Vuori et al., 2008). By directly comparing transport of DHE and BChol, we found that both sterols are targeted to the endocytic recycling compartment (ERC), a major cellular sterol pool, with identical kinetics in baby hamster kidney (BHK) and in HeLa cells (Wüstner et al., 2011a). Uptake of DHE and BChol from the cell surface was strongly reduced in BHK cells overexpressing a dominant-negative clathrin heavy chain, and co-internalization of BChol with fluorescent transferrin, a marker for this uptake pathway, could be demonstrated (Wüstner et al., 2011a). These observations suggest that some plasma membrane sterol is internalized by clathrin-dependent endocytosis in these cells. By fluorescence lifetime imaging, we demonstrated that BChol's emission is constant throughout the cell, such that the fluorescence of this probe is a reliable measure of sterol concentration (Wüstner et al., 2011a).

BChol has also been used to study sterol trafficking in CHO cells with defective Niemann-Pick C1 protein and to determine the role of oxysterol-binding protein-related proteins (ORPs) in cholesterol transport (Hölttä-Vuori et al., 2008, Jansen et al., 2011). In the latter study, an impact of ORPs on transport of BChol from the plasma membrane to the ER and to LDs has been suggested using HeLa cells treated with oleic acid to stimulate droplet formation (Jansen et al., 2011). Importantly, we found that BChol is preferentially targeted to LDs in oleic-acid treated HeLa and BHK cells compared to DHE (Wüstner et al., 2011a). Low free cholesterol content of oleic-acid induced LDs in HeLa cells was recently confirmed by filipin staining by another group (Du et al., 2011). Thus, the preferred targeting of BChol to LDs compared to cholesterol or DHE indicates that the BODIPY-fluorophore affects the intracellular transport of BChol in cells with elevated fat content (Wüstner et al., 2011a).

Based on our experience, we suggest using BChol exclusively for analysis of endocytic sterol trafficking, for example between plasma membrane and ERC. Moreover, we recommend performing co-localization studies of BChol with DHE by multi-color wide field microscopy, and only under conditions in which the probe distributions coincide, can one use the much better fluorescence properties of BChol for studies of sterol transport (Wüstner et al., 2011a). This is exemplified in Fig. 3, where cells were labeled with BChol using similar protocols as described for DHE in Box 2, and imaged on a multiphoton microscope. The intrinsic sectioning capability of this technique combined with negligible bleaching propensity outside the focal region enabled us to follow vesicles containing BChol over several hundred frames (F. Lund, D. Wüstner et al., manuscript in preparation). In this particular example, two dimly fluorescent vesicles form close to the brightly stained plasma membrane and fuse over a time course of a few minutes (Fig. 3B). If DHE would have been used instead of BChol, such vesicles would be hard to detect. Due to the much higher molecular brightness of BChol, we can detect it in intracellular membranes even after several rounds of vesicle fusion and fission (Lund et al., manuscript in preparation). In summary, BChol is useful as supplement to imaging of intrinsically fluorescent sterols in cells. In addition to its much better photo-physical properties, fluorescence imaging of BChol can be combined with electron microscopy studies making use of the diaminobenzidine reaction (Rohrl et al., 2011). Upon illumination of BChol in glutaraldehyde or formaldehyde fixed cells, reactive oxygen species cause photo-oxidation of diaminobenzidine, thereby creating an electron-dense precipitate in sub-cellular regions where BChol is located (Rohrl et al., 2011, Ellinger et al., 2010). By this method, BChol was found in tubular endosomes, multivesicular bodies and the trans-Golgi network in human HepG2 hepatoma cells. The same approach has been used to follow the intracellular fate of BChol-oleate incorporated into high density lipoprotein (Rohrl et al., 2011). A potential problem is the necessary fixation step, which, depending on the fixative and experimental conditions, can alter the intracellular sterol distribution.

Ideally, one would like to visualize intracellular sterol distribution without the need for introducing labeled sterol probes. Two approaches for this that are being developed are reviewed here.

Imaging by mass spectrometry has been used to determine the steady state distribution of cholesterol and other membrane lipids based on chemically specific ionization patterns. Secondary ion mass spectrometry (SIMS) allows for detecting cholesterol and other biomolecules based on characteristic fragmentation patterns. Matrix-enhanced SIMS and time-of-flight SIMS were used to detect phospholipids and cholesterol in neuroblastoma cells and in tissue slices and to visualize cholesterol in cell attachment sites (Altelaar et al., 2006, Nygren and Malmberg, 2004). Chemical mapping is typically limited by the low signal intensity from various biomolecules. Fragmentation of cholesterol in SIMS imaging typically provides several peaks in the mass range of the fragment ion m/z=366–370 which can be used to visualize and quantify cellular cholesterol. Recently, a characteristic fragment of cholesterol of mass-charge ratio m/z= 147 has been used as a fingerprint to visualize cholesterol accumulation in J774 murine macrophages (Piehowski et al., 2008). J774 cells pre-treated with MβCD-cholesterol (Zidovetzki and Levitan, 2007) had about two times more cholesterol in the plasma membrane than control cells (Piehowski et al., 2008, Ostrowski et al., 2007). For cellular cholesterol imaging by mass spectrometry, two major limitations need to be overcome; SIMS can be used only in freeze-dried samples, and the method has a spatial resolution approximating the cell diameter (i.e., at present no sub-cellular sterol distribution can be revealed). The technique might currently be useful for determining sterol distribution in model membranes and lipid monolayers, since domain sizes in these systems can be quite large (Sostarecz et al., 2004), (Zheng et al., 2007), (Boxer et al., 2009), (Anderton et al., 2011).

Raman scattering is due to inelastic scattering of light, which results in frequency shifts compared to the incident light. Scattered light of lower frequency (ωS =ωpωR) or increased frequency (ωAS =ωp + ωR) compared to the incident light are called Stokes and anti-Stokes components, respectively (Potma and Xie, 2008). The frequencies ωR are caused by molecular vibrations, which can be used to identify and characterize (bio)molecules. Since the Raman signal is very weak, application of this technique to bioimaging is severely limited. Nevertheless, the method combined with confocal detection has been used to characterize cholesterol esters in HeLa cells and macrophage foam cells (van Manen and Otto, 2009). Moreover, Raman spectroscopy has been used as diagnostic tool to detect breast cancer, and combined Raman and fluorescence microscopy were applied to characterize atherosclerotic plaques (Haka et al., 2009, Haka et al., 2011).

A major technical improvement is coherent anti-Stokes Raman scattering (CARS). In CARS, two coherent light beams of frequency ω1 and ω2 are used to drive a Raman scattering of frequency ωR, which results in scattering signals about 105-fold stronger than spontaneous Raman scattering (Potma and Xie, 2008, Cheng et al., 2002). The coherent addition of anti-Stokes Raman scattering radiation from many molecular oscillators in the focal volume is responsible for the sensitivity of the technique. This paved the way for biomedical applications of CARS, and this technique has been used for analysis of neutral lipids (i.e., cholesteryl and triacylglycerol esters) in cytoplasmic lipid droplets of mammalian cells (Nan et al., 2003), Drosophila melanogaster (Chien et al., 2011) and Caenorhabditis elegans (Yen et al., 2010) with chemical specificity. Neutral lipids are particularly suitable for CARS analysis due to the abundance of C-H stretching vibrations (Nan et al., 2003). Characteristic vibrations of saturated versus unsaturated fatty acyl chains allow monitoring of lipid remodeling during nutritional or genetic perturbations (Yen et al., 2010, Mörck et al., 2009). Another advantage of CARS is its ease combination with nonlinear fluorescence microscopy. For example in C. elegans, fat storage organelles have been visualized by CARS, while multiphoton fluorescence microscopy provided complementary information on autofluorescent gut granules in the same samples (Yen et al., 2010). Quantitative CARS microscopy revealed heterogeneity in acyl chain saturation of LDs in individual adipocytes (Rinia et al., 2008). Despite the improvements in CARS, vibrational microscopy is limited to cellular regions with large stores of similar lipid species (e.g., lipid droplets), and CARS remains largely confined to detection of neutral lipids due the contribution of the C-H stretching and the C=C vibrations to the CARS signal. CARS microscopy has great potential to unravel the physico-chemical basis of membrane phase separation, with CARS combined with fluorescence microscopy to visualize cholesterol-mediated lipid demixing in supported membranes (Li et al., 2005). Finally, CARS microscopy has been combined with other nonlinear optical imaging methods, like two-photon excitation and sum-frequency generation to monitor atherosclerotic lesions and plaque formation in aortas of ApoE-deficient mice (Wang et al., 2009, Lim et al., 2010).

Section snippets

Concluding remarks

There has been considerable progress over the past decade in the ability to observe the distribution of sterols in cells and to measure their transport between organelles using optical microscopy methods. CTL and DHE are reasonably good analogs of cholesterol and it seems unlikely that fluorescent analogs that are better cholesterol analogs will be developed. However, these molecules are not identical to cholesterol and this caveat must always accompany studies using them in mammalian cells.

Acknowledgments

DW acknowledges funding by grants of the Lundbeck Foundation, the Danish Research Agency Forskningsstyrelsen, Forskningsrådet for Natur og Univers (FNU), and the Danish Research Agency Forskningsstyrelsen, Forskningsrådet for Sundhed og sygdom (FSS). FRM received support from NIH Grant R37-DK27083.

References (126)

  • M. Hao et al.

    Vesicular and non-vesicular sterol transport in living cells. The endocytic recycling compartment is a major sterol storage organelle

    J. Biol. Chem.

    (2002)
  • J. Henriksen et al.

    Universal behavior of membranes with sterols

    Biophys. J.

    (2006)
  • H. Huang et al.

    Use of dansyl-cholestanol as a probe of cholesterol behavior in membranes of living cells

    J. Lipid Res.

    (2010)
  • P. Kohut et al.

    The role of ABC proteins Aus1p and Pdr11p in the uptake of external sterols in yeast: Dehydroergosterol fluorescence study

    Biochem. Biophys. Res. Commun.

    (2011)
  • M. Krieger

    Reconstitution of the hydrophobic core of low-denstiy lipoprotein

    Methods Enzymol.

    (1986)
  • L. Li et al.

    Quantitative coherent anti-stokes Raman scattering imaging of lipid distribution in coexisting domains

    Biophys. J.

    (2005)
  • R.S. Lim et al.

    Multimodal CARS microscopy determination of the impact of diet on macrophage infiltration and lipid accumulation on plaque formation in ApoE-deficient mice

    J. Lipid Res.

    (2010)
  • L.M.S. Loura et al.

    Exclusion of a cholesterol analog from the cholesterol-rich phase in model membranes

    Biochim. Biophys. Acta

    (2001)
  • M.P. Marsan et al.

    Cholesterol orientation and dynamics in dimyristoylphosphatidylcholine bilayers: a solid state deuterium NMR analysis

    Biophys. J.

    (1999)
  • A.L. McIntosh et al.

    Fluorescence and multiphoton imaging resolve unique structural forms of sterol in membranes of living cells

    J. Biol. Chem.

    (2003)
  • B. Mesmin et al.

    Intracellular sterol dynamics

    Biochim. Biophys. Acta

    (2009)
  • S. Mukherjee et al.

    Cholesterol distribution in living cells: fluorescence imaging using dehydroergosterol as a fluorescent cholesterol analog

    Biophys. J.

    (1998)
  • S. Munro

    Lipid rafts: elusive or illusive?

    Cell

    (2003)
  • J.F. Nagle et al.

    Structure of lipid bilayers

    Biochim. Biophys. Acta

    (2000)
  • X. Nan et al.

    Vibrational imaging of lipid droplets in live fibroblast cells with coherent anti-Stokes Raman scattering microscopy

    J. Lipid Res.

    (2003)
  • H. Ohvo-Rekila et al.

    Cholesterol interactions with phospholipids in membranes

    Prog. Lipid Res.

    (2002)
  • K.B. Peake et al.

    Defective cholesterol trafficking in Niemann-Pick C-deficient cells

    FEBS Lett.

    (2010)
  • N.H. Pipalia et al.

    Automated microscopy screening for compounds that partially revert cholesterol accumulation in Niemann-Pick C cells

    J. Lipid Res.

    (2006)
  • H.A. Rinia et al.

    Quantitative label-free imaging of lipid composition and packing of individual cellular lipid droplets using multiplex CARS microscopy

    Biophys. J.

    (2008)
  • J. Rogers et al.

    The organisation of cholesterol and ergosterol in lipid bilayers based on studies using non-perturbing fluorescent sterol probes

    Biochim. Biophys. Acta

    (1979)
  • A.I. Rosenbaum et al.

    Chemical screen to reduce sterol accumulation in Niemann-Pick C disease cells identifies novel lysosomal acid lipase inhibitors

    Biochim. Biophys. Acta

    (2009)
  • H.A. Scheidt et al.

    The potential of fluorescent and spin-labeled steroid analogs to mimic natural cholesterol

    J. Biol. Chem.

    (2003)
  • F. Schroeder et al.

    Fluorescence properties of cholestatrienol in phosphatidylcholine bilayer vesicles

    Biophys. Chem.

    (1988)
  • D.K. Sharma et al.

    Glycosphingolipids internalized via caveolar-related endocytosis rapidly merge with the clathrin pathway in early endosomes and form microdomains for recycling

    J. Biol. Chem.

    (2003)
  • J.E. Shaw et al.

    Correlated fluorescence-atomic force microscopy of membrane domains: structure of fluorescence probes determines lipid localization

    Biophys. J.

    (2006)
  • G. Smutzer et al.

    Physical properties of the fluorescent sterol probe dehydroergosterol

    Biochim. Biophys. Acta

    (1986)
  • A. Sokolov et al.

    Accessibility of cholesterol in endoplasmic reticulum membranes and activation of SREBP-2 switch abruptly at a common cholesterol threshold

    J. Biol. Chem.

    (2010)
  • J.R. Swedlow

    Quantitative fluorescence microscopy and image deconvolution

    Methods Cell Biol.

    (2007)
  • S.N. Ahmed et al.

    On the origin of sphingolipid/cholesterol-rich detergent-insoluble cell membranes: physiological concentrations of cholesterol and sphingolipid induce formation of a detergent-insoluble, liquid-ordered lipid phase in model membranes

    Biochemistry

    (1997)
  • J. Aittoniemi et al.

    Tilt: major factor in sterols’ ordering capability in membranes

    J. Phys. Chem. B

    (2006)
  • A.F. Altelaar et al.

    Gold-enhanced biomolecular surface imaging of cells and tissue by SIMS and MALDI mass spectrometry

    Anal. Chem.

    (2006)
  • Arthur, J.R., Heinecke, K.A., and Seyfried, T.N. Filipin recognizes both GM1 and cholesterol in GM1 gangliosidosis...
  • D.M. Benson et al.

    Digital imaging fluorescence microscopy: spatial heterogeneity of photobleaching rate constants in individual cells

    J. Cell Biol.

    (1985)
  • F. Bergström et al.

    Dimers of Dipyrrometheneboron difluoride (BODIPY) with light spectroscopic applications in chemistry and biology

    J. Am. Chem. Soc.

    (2002)
  • S.G. Boxer et al.

    Advances in imaging secondary ion mass spectrometry for biological samples

    Annu. Rev. Biophys.

    (2009)
  • R. Breslow et al.

    Cholesterol recognition and binding by cyclodextrin dimers

    J. Am. Chem. Soc.

    (1996)
  • C.H. Chien et al.

    Label-free imaging of Drosophila in vivo by coherent anti-Stokes Raman scattering and two-photon excitation autofluorescence microscopy

    J. Biomed. Optics

    (2011)
  • Craig, I.F., Via, D.P., Mantulin, W.W., Pownall, H.J., Gotto, A.M.J., and L.C.S. (1981). Low density lipoproteins...
  • X. Du et al.

    A role for oxysterol-binding protein-related protein 5 in endosomal cholesterol trafficking

    J. Cell Biol.

    (2011)
  • Fischer RT, Stephenson FA, Shafiee A, and F., S. (1974). Delta 5,7,9(11)-Cholestatrien-3 beta-ol: a fluorescent...
  • Cited by (0)

    View full text