The Lysosomotropic Agent Monodansylcadaverine Also Acts as a Solvent Polarity Probe

Abstract

The autofluorescent substance monodansylcadaverine has recently been reported as a specific in vivo marker for autophagic vacuoles. However, the mechanism for this specific labeling remained unclear. Our results reveal that the common model of ion trapping in acidic compartments cannot completely account for the observed autophagic vacuole staining. Because autophagic vacuoles are characterized by myelin-like membrane inclusions, we tested whether this lipid-rich environment is responsible for the staining properties of monodansylcadaverine. In in vitro experiments using either liposomes or solvents of different polarity, monodansylcadaverine showed an increased relative fluorescence intensity in a hydrophobic environment as well as a Stokes shift dependent on the solvent polarity. To test the effect of autophagic vacuoles or autophagic vacuole lipids on monodansylcadaverine fluorescence, we isolated autophagic vacuoles and purified autophagic vacuole lipids depleted of proteins. Entire autophagic vacuoles and autophagic vacuole lipids had the same effect on monodansylcadaverine fluorescence properties, suggesting lipids as the responsible component. Our results suggest that the in vivo fluorescence properties of monodansylcadaverine do not depend exclusively on accumulation in acidic compartments by ion trapping but also on an effective interaction of this molecule with autophagic vacuole membrane lipids.

A variety of reagents have been described for in vivo labeling of acidic cellular compartments (Anderson and Orci 1988). The autofluorescent substance monodansylcadaverine (MDC) has recently been shown to stain autophagic vacuoles (AVs), which are part of the lysosomal compartment (Biederbick et al. 1995). AVs are organelles in which cellular components such as mitochondria, peroxisomes, and cytoplasmic constituents are degraded (Dunn 1994). Although there is still some debate about the biogenesis of autophagic vacuoles, most experimental evidence supports the idea that these structures originate from ribosome free cisternae of the endoplasmic reticulum (Dunn 1990a). These membranes first sequester the intracellular constituents being addressed for degradation. Then the membranes of the cisternae fuse and constitute a closed vacuole lined by a double membrane (Dunn 1990b). Degradation occurs after their fusion with primary lysosomes. A morphological hallmark of the vacuoles that then develop is the accumulation of membrane whirls filling the AV lumen (Papadopoulos and Pfeiffer 1987). The origin, composition, and fate of this membranous material have not been described in detail thus far.

Like other lysosomal compartments, AVs possess an acidic pH which is generated by a V-type H⁺-ATPase (Al-Awqati 1986; van Dyke 1996). Weak bases capable of crossing biological membranes are believed to be concentrated in acidic compartments by protonation (DeDuve et al. 1974). This ion trap mechanism has been proposed for a variety of so-called lysosomotropic substances such as N-(3-[(2,4-dinitrophenyl)-amino]-propyl)-N-(3-aminopropyl-methylamine)dihydrochloride (DAMP) (Anderson et al. 1984), primaquine, and chloroquine (DeDuve et al. 1974). At higher concentrations, lysosomotropic substances disturb the pH gradient between the acidic compartment and the cytosol, leading to impaired functioning of these organelles. In particular, a decreased degradative capacity, e.g., for proteins, has been documented for a variety of monoamines, diamines, amino alcohols, and other amino compounds (Seglen and Gordon 1980).

The autofluorescent substance MDC is also a weak base. In addition to its ability to label AVs in vivo (Biederbick et al. 1995), MDC is effective as a substrate or inhibitor for the protein crosslinking enzyme transglutaminase (Lorand et al. 1979), as a potential inhibitor of endocytosis (Davies et al. 1980, 1984), and as an stimulator of synthesis of phospholipid-like phosphatidylinositols (Gracia Gil et al. 1984).

The fluorescence properties of MDC in solution are dependent on the polarity of the solvent (Narayanan and Balaram 1976). The effect of solvent polarity on fluorescence properties has been intensively investigated for 6-propionyl-2-(dimethyl-amino)naphthalene (PRODAN), which is structurally similar to MDC, but not for MDC itself (Weber and Farris 1979).

It was assumed that the capability of MDC to label AVs in vivo depends also on an ion trap mechanism (Biederbick et al. 1995). Here we show that only a part of the total MDC is concentrated in acidic compartments by ion trapping, using the human pancreatic adenocarcinoma cell line PaTu 8902 (Elsässer et al. 1993). Furthermore, an interaction of MDC with membrane lipids appears to be another mechanism underlying the observed in vivo and in vitro fluorescence properties of MDC.

Materials and Methods

Materials

Phosopholipon 90 was obtained from Nattermann-Phospholipid (Cologne, Germany). Phospholipon 90 is lecithin purified from soybean (USP XXIII) and contains 93 ± 3% phosphatidylcholine, 3 ± 3% lysophosphatidylcholine, and a minimum of 0.1%. d,I-α-tocopherol. The fatty acid composition is 12 ± 2% palmitic acid (16:0), 3 ± 1% stearic acid (18:0), 10 ± 3% oleic acid (18:1), 66 ± 5% linoleic acid (18:2), and 5 ± 2% linolenic acid (18:3). Acridine orange was purchased from Serva (Heidelberg, Germany). Destruxin B was prepared as described elsewhere (Muori et al. 1994). All other chemicals used were obtained from Sigma (Diesenhofen, Germany).

Fluorescence Measurement and Scanning

All experiments were performed using the Fluorescence Photometer SFM 25 (Kontron Instruments; Zurich, Switzerland). For fluorescence measurement of MDC excitation wavelength was set to 335 nm and emission wavelength to 525 nm, unless otherwise indicated. Emission scanning was performed from 600 nm to 400 nm with an excitation wavelength of 335 nm and a scan speed of 100 nm/min. Excitation scans were performed in the same way, with emission wavelength set to 525 nm and scanning range from 450 nm to 200 nm. Amplification by high voltage setting was chosen individually in each experiment to record data within the linear scale of the photometer between relative fluorescence values (RF) of 10 and 160. Therefore, RF can differ in similar experiments.

Cell Culture Experiments

All experiments were performed using the cell line PaTu 8902, which was established from a human primary pancreatic adenocarcinoma (Elsässer et al. 1993). Cells were cultured in 10 ml Dulbecco's modified Eagle's medium (DMEM) containing 5% horse serum and 5% fetal calf serum until confluence on 55-cm² Falcon plastic dishes (Becton-Dickinson; Heidelberg, Germany). To determine uptake, MDC was added to the medium at a final concentration of 0.1 mM and cells were incubated for the indicated time periods at 37C, washed three times with ice-cold PBS, scraped, and resuspended in 1 ml 10 mM Tris-HCl, pH 8.0. To destroy intracellular pH gradients, cells were incubated for 10 min before and during further incubation with 50 mM ammonium chloride. In pulse-chase experiments, cells were incubated for 1 hr with MDC as described before, washed with PBS, and recultured in 10 ml medium containing 10 μM of destruxin B. Controls were chased with medium only. Cells were harvested as described above. All cell suspensions were passed seven times through a Microlance 3 needle (Becton-Dickinson) before quantitation of MDC fluorescence. To equalize cell numbers among different samples within one experiment, the relative amount of DNA was determined. One μl of a 1 mg/ml Hoechst Dye 33258 (Polysciences; Warrington, PA) aqueous stock solution was added to each sample, which was incubated for at least 10 min and then measured at an excitation wavelength of 365 nm and an emission wavelength of 460 nm. MDC fluorescence was adjusted to equal DNA content.

Fluorescence Microscopy

Staining of PaTu 8902 cells with MDC and subsequent fixation of the cells were described elsewhere (Biederbick et al. 1995). Cells were analyzed using a Zeiss inverted fluorescence microscope IM35 equipped with a long distance X40 lens and filter sets 01 and 15. Acridine orange staining was performed as described elsewhere (Muori et al. 1994) and was analyzed with a confocal laser scanning microscope LSM 410 (Zeiss; Goettingen, Germany).

Preparation of Liposomes

For preparation of multilamellar vesicles, 30 mg of Phospholipon 90 was dissolved in 2 ml dichlormethane. The solution was dried under vacuum. The dried lipids were kept under vacuum for at least 45 min and then suspended in 2 ml 10 mM Tris-HCl, pH 8.0. These preliposomes were sonicated with a Labsonic U (B. Braun; Melsungen, Germany). After sonification, vesicles were centrifuged at 10,000 times; g for 10 min at 4C and the supernatant was used for further investigations. The remaining supernatant was used only when a small pellet was visible and no changes in the phospholipid concentration could be observed according to Stewart (1980).

Purification and Analysis of Autophagic Vacuoles

PaTu 8902 cells were grown until confluence. Cells from eight 55-cm² Falcon plastic dishes (Becton-Dickinson) were homogenized as described before (Biederbick et al. 1995). The postnuclear supernatant was centrifuged at 105,000 times; g for 1 hr at 4C. The pellet was resuspended in 400 μl 10 mM Tris-HCl, pH 8.0, and separated on a sucrose step gradient [40%, 27.5%, 20%, and 10% (w/w)] at 80,000 times; g for 5.5 hr. Organelles from the 10-20% sucrose step of four gradients were collected and again centrifuged at 105,000 times; g for 1 hr at 4C. Organelles were characterized either by acid phosphatase biochemical analysis or by electron microscopy as described previously (Biederbick et al. 1995). For preparation of lipids, AVs were extracted with chloroform/methanol according to Bligh-Dyer (New 1992) and phospholipid concentration was measured according to Stewart (1980). The phospholipid concentration in each sample was adjusted to 2 μg/μl.

Results

In a previous study we showed by cell fractionation and ultrastructural studies that the autofluorescent substance MDC is capable of staining autophagic vacuoles in vivo (Biederbick et al. 1995). However, the mechanism of this specific staining behavior remained unclear and is also poorly understood for other similar substances with a specific affinity for acidic compartments.

One common and widespread hypothesis for the mechanism underlying this specific staining behavior is that an ion trapping mechanism is responsible for the accumulation of various substances in the acidic environment of lysosomal compartments (DeDuve et al. 1974). To test this for MDC, we destroyed the acidic pH by applying ammonium chloride before cells were stained with MDC. The effect of ammonium chloride on the acidic pH in lysosomes was tested with acridine orange (Figures 1e and 1f; Muori et al. 1994). As shown in Figure 1a, this treatment led to a reduction of MDC uptake. In a complementary experiment, cells were incubated with MDC under standard conditions, then chased with MDC-free medium and subsequently incubated with or without destruxin B, an inhibitor of the vacuolar H⁺-ATPase (Muori et al. 1994). Cells treated with destruxin B also lost more MDC than the untreated control cells (Figure 1d), in a comparable amount to that observed in the ammonium chloride experiment. When fluorescent microscopy was applied to the MDC-treated cells, the cytoplasm of these cells in both experiments was blurred by MDC when the pH gradient between lysosomal vacuoles and the cytoplasm was disturbed (Figures 1b, and 1c), indicating that an acidic pH is indeed responsible for retaining MDC in lysosomal vacuoles. However, MDC-positive structures but not acridine orange-positive structures could be observed even after extended incubation times with destruxin or ammonium chloride. This indicates that MDC labeling of vacuolar structures does not completely depend on ion trapping, as seen for acridine orange.

Figure 1

Uptake and release of MDC after disturbance of the acidic pH of lysosomes with either ammonium chloride or the H⁺-V-ATPase inhibitor destruxin B. Uptake of MDC in PaTu 8902 cells was reduced by 50 mM ammonium chloride compared with untreated cells (a). This reduction was observed for every time point indicated. When ammonium chloride-treated vital cells were analyzed by fluorescence microscopy 30 min after MDC was added, diffuse cytoplasmic and vacuolar staining occurred (c), whereas in untreated vital cells MDC-positive vacuoles were sharply delineated and no fluorescence could be detected in the cytoplasm (b). Bars = 27 μm. When cells were pulse-labeled with MDC for 1 hr and the MDC release was subsequently measured with or without 10 μM destruxin B, greater loss of MDC in destruxin B-treated cells was observed (d). To control the disturbance of the acidic pH of lysosomes, cells were treated in a separate experiment with acridine orange for 10 minutes. (e) Untreated cells; (f) cells treated with 50 mM ammonium chloride for 30 min. ▀, control; ○, ammonium chloride- or destruxin B-treated. Bars = 25 μm.

As mentioned above, isolated MDC-positive organelles are autophagic vacuoles (Biederbick et al. 1995) and are part of the lysosomal system. A hallmark of these vacuoles is the presence of myelin-like membrane whirls filling the vacuole lumen (Papadopoulos and Pfeiffer 1987). We suspected that this lipid-rich environment might be responsible for the observed specific fluorescence of these organelles. To test this possibility, we measured the fluorescence of 5 μM MDC for its dependence on increasing amounts of liposomes. As shown in Figure 2a, the fluorescence of MDC and hence the relative fluorescence intensity increased with the concentration of lipid added to the solution (up to ~10 μM lipid), indicating that the interaction of MDC with lipids alters the autofluorescent properties of MDC. Furthermore, we observed not only an increase of fluorescence but also a blue shift in the maximal value of the emission wavelength, indicating that MDC acts as a solvent polarity probe (Figure 2b). Similar increases in fluorescence were observed for liposomes prepared from different lipids (not shown). In addition, Tween-20 above the critical micelle concentration (CMC = 0.065%) could increase fluorescence of MDC in a dose-dependent manner (not shown).

Similar fluorescence behavior had been shown for the structurally related solvent polarity probe 6-propionyl-2-(dimethyl-amino)naphthalene (PRODAN) in response to bovine serum albumin (BSA) (Weber and Farris 1979). Therefore, we analyzed the influence of this protein on the fluorescence properties of MDC. Increasing amounts (up to 300 μM) of BSA were added to 5 μM MDC. BSA increased MDC fluorescence in a dose-dependent manner (Figure 2c) and could induce a blue shift comparable to that revealed with liposomes (Figure 2d). However, a 50-fold higher molar concentration of BSA was needed compared to lipids. For other proteins, such as casein or dimethylcasein, hardly any effect was observed at concentrations up to 100 μM (not shown). To confirm that MDC is a solvent polarity probe, its fluorescence properties were compared with those of the solvent polarity probe PRODAN. Therefore, the excitation and emission maximal wavelength of MDC in solvents of different polarity was measured. Figure 3 shows the Stokes shift [(excitation maximum wavelength minus emission maximum wavelength) in wave numbers] in dependence of the orientational polarizibility Δf of the solvents according to Lippert (Lippert 1957; Weber and Farris 1979). MDC Stokes shift showed a correlation to solvent polarity equal to that the solvent polarity probe PRODAN (Weber and Farris 1979). From these data, Δf values and thus polarity of a certain environment can be obtained by measuring the Stokes shift of MDC fluorescence (see below). MDC exhibited an increased relative fluorescence intensity under the influence of hydrophobic molecules as well as a Stokes shift dependent on the polarity of the solvent, characterizing the MDC molecule as a solvent polarity probe.

Figure 2

The relative fluorescence intensity and the emission wavelength maximum of MDC fluorescence are altered by liposomes (a,b) and bovine serum albumin (BSA) (c,d). A liposome stock solution (15 mg/ml) was made from a commercially available lipid preparation (Phospholipon 90) and further diluted to the concentrations indicated. Note the different molar concentrations of lipid and BSA needed to increase relative fluorescence intensity. Liposomes shifted the maximal emission wavelength from 525 nm to 498 nm and BSA to 492 nm.

These results suggest that, within the cell, a specific interaction of MDC with membrane lipids might be responsible for AV staining. To test this, we purified AVs on a sucrose step gradient. Organelles floating on 20% sucrose, containing the lysosomal marker acid phosphatase (data not shown; Biederbick et al. 1995), were used for further investigations. Figure 4b shows the purity of the AVs in this fraction by ultrastructural analysis. Equal volumes of prepared AVs, with a phospholipid concentration of 2 μg/μl, were either used for lipid extraction according to Bligh and Dyer (New 1992) or sedimented in a 105,000 times; g centrifugation step. Total AVs and AV lipids extracted from the same amount of entire vacuoles were dissolved in 750 μl 10 mM Tris-HCl (pH 8.0) containing 0.1% Tween-20. Figure 4a shows the influence of equal volumes of each preparation on the fluorescence of 5 MDC. Entire AVs and AV lipids had the same effect on MDC fluorescence, suggesting that lipids were responsible for the increase in relative fluorescence intensity, leading to staining of AVs in vivo. To prove the interaction of MDC with a lipophilic environment in vivo, we measured the Stokes shift in intact cells (Δv = 9280 cm⁻¹; Δf = 0.310), in isolated AVs (Δv = 8960 cm⁻¹; Δf = 0.303), and liposomes (Δv = 9070 cm⁻¹; Δf = 0.307). The obtained Δf values are almost identical and are similar to those obtained for methanol and ethanol (Figure 3).

Figure 3

Characterization of MDC as a solvent polarity probe. The excitation and emission spectra of MDC were recorded in solvents of different polarity indicated by numbers: 1, benzene; 2, chloroform; 3, dimethyformamide; 4, acetone; 5, ethanol; 6, methanol; 7, water. Stokes shift Δv = emission maximum wavelength (cm⁻¹) - excitation maximal wavelength (cm⁻¹). The values of the orientational polarizibility Δf of the different solvents are from Weber and Farris (1979). The orientational polarizibility was calculated according to the dipole interaction theory of Lippert (1957): Δf = ((n²-1)/(n² + 1))-((D-1)/(D + 1)), with n = refractive index and D = dielectric constant of the solvent.

Discussion

The term “lysosomotropic agent” was introduced by DeDuve and co-workers (1974) to designate substances that are taken up selectively into lysosomes. This definition leaves open the chemical nature of a lysosomotropic substance and the mechanism of its uptake. However, almost all agents with lysosomotropic properties based on cell permeation rather than endocytotic uptake belong to the class of weak bases. This has led to the assumption that the specific uptake of lysosomotropic substances into lysosomes depends on the acidic pH of these compartments, ion-trapping weak bases (DeDuve et al. 1974; Weissmann 1979). In this model, the most suitable weak bases are those with a pK around 8.

The autofluorescent substance MDC has been used as a lysosomotropic agent to study the function of lysosomes in different cell types (Verhoef and Sharma 1983; Vandenbroucke-Grauls et al. 1984). Recently, MDC has been shown to label autophagic vacuoles (AVs), a subpopulation of vacuoles belonging to the lysosomal compartment (Biederbick et al. 1995). However, although MDC is a weak base, the ion trapping model is not sufficient to explain this specificity. Our biochemical and morphological data show that only a certain amount of MDC is released from or taken up by AVs, respectively, when the pH gradient between the cytosol and the lysosome is disturbed either by ammonium chloride or the vacuolar H⁺-ATPase inhibitor destruxin B (Figure 1). In contrast, under the same conditions the lysosomotropic agent acridine orange showed no accumulation in lysosomes. Finally, vacuolar structures were also stained with MDC in paraformaldehyde-fixed cells, in which no proton gradients are preserved and ion trapping cannot occur (not shown). This suggests another mechanism for the specific fluorescent labeling of AVs by MDC. Interestingly, some lysosomotropic agents are described in the literature that do not belong to the class of weak bases but which are neutral molecules (Allison and Young 1969; Brown et al. 1992), so that they also cannot be retained in acidic compartments according to an ion trapping mechanism. Consequently, we tried to clarify the role of other structural AV components such as lipids and proteins for their interaction with MDC.

Figure 4

Increase in relative MDC fluorescence intensity using autophagic vacuole lipids ▀ or total autophagic vacuoles ○. Autophagic vacuoles were prepared by density gradient centrifugation from homogenized PaTu 8902 cells and were characterized by biochemical analysis (not shown) and by electron microscopy (b). Bar = 0.5 μm. The volumes indicated on the x-axis contained equal amounts of phospholipids lipids compared to total vacuoles (2 μg/μl). Isolated autophagic vacuole lipids revealed the same dose-dependent increase in relative fluorescence intensity as entire autophagic vacuoles (a).

AVs are characterized by a high content of lipids originating from membrane material regularly filling the internal space of these vacuoles as myelin-like whirls (Papadopoulos and Pfeiffer 1987). It has been shown for a variety of substances classified as solvent polarity probes that the fluorescent properties can strongly depend on the hydrophobicity of their environment (Slavik 1994). Typical properties of solvent polarity probes are their increased fluorescence quantum yield and a Stokes shift dependent on the polarity of the environment. In our experiments, we could show that MDC fluorescence was enhanced by increasing concentrations of the detergent Tween-20 or liposomes made from phospholipon 90 (Figure 2a). This confirms the description of MDC as a fluorescent probe for solvent polarity (Narayanan and Balaram 1976), similar to the intensively studied substances 1,8-anilinonaphthalenesulfonate (ANS) and its derivatives (Slavik 1982) or 6-propionyl-2-(dimethyl-amino)naphthalene (PRODAN) (Weber and Farris 1979), the latter being structurally more related to MDC than ANS. When the Stokes shift of MDC was measured in solutions of different polarity, the orientational polarizibility Δf correlated with the Stokes shift according to the dipole interaction theory of Lippert (1957), further confirming MDC as a solvent polarity probe. Moreover, as shown for ANS (Slavik 1982) and PRODAN (Weber and Farris 1979), MDC also exhibited a Stokes shift in the presence of liposomes (Δv = 9070 cm⁻¹) correlating with a Δf of 0.307, which corresponds to Δf values of ethanol. Finally, as mentioned in Results, a similar Stokes shift was also seen with isolated autophagic vacuoles and with a cell suspension stained with MDC in vivo. However, this effect could also have been induced by AV proteins. It has been shown that BSA acts similarly on PRODAN as lipids do, enhancing relative fluoresence intensity and blue-shifting the emission maximal wavelength (Weber and Farris 1979). A similar effect was observed for dansylglycine and human serum albumin (Chignell 1973). Our experiments show that this is also true for MDC, with a Stokes shift of Δv = 8827 cm⁻¹. Increase in relative fluorescence intensity depended on the amount of BSA added to a constant amount of MDC, and this was saturable with an 60-fold molar excess of BSA (Figure 2c). This finding appears to be in contradiction to data published earlier for MDC (Narayanan and Balaram 1976), in which an interaction of MDC with BSA could not be observed. However, in this report 60 μM MDC was incubated with only 3 μM BSA, a concentration that also gave no detectable increase in relative fluorescence intensity in our experiments (Figure 2c). The BSA effect occurred only at a 50-fold higher molar concentration than the effect lipids have on MDC fluorescence properties. Furthermore, isolated AV lipids had the same effect on Stokes shift and increased relative fluorescence intensity as entire isolated AVs (Figure 4a), indicating that AV proteins are not responsible for altering the fluorescent properties of MDC. Whether the observed effect is due exclusively to the interaction of lipid molecules with MDC or whether there is also a concentration effect of MDC in the lipid phase cannot be answered by the data presented. The partition coefficient for MDC between an aqueous phase and octanol was found to be 1:1 (not shown), which does not imply a higher affinity of MDC for a hydrophobic environment. However, because the lipids of the AVs are organized in regular bilayers with defined hydrophobic and hydrophilic domains, it is possible that the partition of a substance between two phases of opposite hydrophobicity does not properly reflect the interaction of a molecule with a lipid bilayer.

The question remains of why MDC stains only AVs and not the plasma membrane or other endomembranes. One reason could be the high concentration of membrane material in these organelles in which MDC molecules might be concentrated and its relative fluorescence intensity is increased in a small volume, so that the sensitivity of a fluorescent microscope is sufficient for MDC detection, whereas in other membranes the amount of emitted fluorescent light is below the threshold of fluorescent microscopic sensitivity. When cells were fractionated after MDC incubation, a weak fluorescent label could also be detected by the more sensitive fluorescent spectrophotometry in membranes other than in AVs (not shown).

Alternatively, the lipid composition of AV membranes or the presence of one AV-specific lipid with a high affinity for MDC binding could also account for the specific in vivo MDC staining of AVs. It has recently been shown that lysobisphosphatidic acid (LBPA), the antigen of the autoimmune disease antiphospholipid syndrome (Kobayashi et al. 1998), is specific for late endosomes with multivesicular structure, being similar to the structure of AVs. This demonstrates that in addition to proteins, lipid molecules also can occur as specific components of membranes, and new insights into lipid sorting in the vesicular transport pathway using a variety of fluorescent lipid derivatives have been appreciated only recently (Menon 1998). However, the role of individual lipid molecules in the integration and interaction of polarity-sensitive molecules such as MDC in the membranes of organelles like AVs remains to be elucidated. Our data indicate that MDC accumulates in intracellular vacuoles via both ion trapping and an interaction with lipid molecules, preferentially of those contained in multilamellar bodies. Such organelles were originally classified as autophagosomes by electron microscopy. However, further analyses have revealed that multilamellar bodies can occur under different physiological conditions in a variety of cell types (Schmitz and Müller 1991). Examples are Type II pneumocytes of the bronchiolar epithelium, which store components of the surfactant as multilamellar bodies, or keratinocytes in the upper layer of the skin epithelium, which contain multilamellar bodies called keratinosomes. In addition, under certain physiological or pathophysiological conditions, multilamellar bodies are created during impaired lipid metabolism or as a result of enhanced membrane turnover. In all multilamellar bodies, lysosomal marker proteins have been found. Thus, MDC is incorporated into multilamellar bodies by both an ion trapping mechanism and the interaction with membrane lipids functioning as a solvent polarity probe. The origin of the stored lipids cannot be discriminated by MDC staining.

In conclusion, we provide evidence that the lysosomotropic agent monodansylcadaverine (MDC), which preferentially labels autophagic vacuoles, is not exclusively effective as a lysosomotropic agent according to an ion trapping mechanism that depends on the proton gradient between the cytosol and the lysosomal compartment. Instead, MDC, which might be effectively incorporated into at least some cell membranes because of its amphipathic structure, also functions as a solvent polarity probe, exhibiting a Stokes shift and an increased relative fluorescence intensity in a hydrophobic environment. This property depends on the in teraction of MDC with lipid molecules. Whether the specificity of MDC staining of AVs depends only on the high content of lipids in these organelles, or whether MDC can interact specifically with lipid molecules unique for AVs, remains to be elucidated.