OUP user menu

Organelle association visualized by three dimensional ultrastructural imaging of the yeast cell

Andreas Perktold, Bernd Zechmann, Günther Daum, Günther Zellnig
DOI: http://dx.doi.org/10.1111/j.1567-1364.2007.00226.x 629-638 First published online: 1 June 2007


This study was aimed at a better understanding of organelle organization in the yeast Saccharomyces cerevisiae with special emphasis on the interaction and physical association of organelles. For this purpose, a computer aided method was employed to generate three-dimensional ultrastructural reconstructions of chemically and cryofixed yeast cells. This approach showed at a high level of resolution that yeast cells were densely packed with organelles that had a strong tendency to associate at a distance of <30 nm. The methods employed here also allowed us to measure the total surface area and volume of organelles, the number of associations between organelles, and the ratio of associations between organelles per surface area. In general, the degree of organelle associations was found to be much higher in chemically fixed cells than in cryofixed cells, with endoplasmic reticulum/plasma membrane, endoplasmic reticulum/mitochondria and lipid particles/nuclei being the most prominent pairs of associated fractions. In cryofixed cells, similar preferences for organelle association were seen, although at lower frequency. The occurrence of specific organelle associations is believed to be important for intracellular translocation and communication. Membrane contact as a possible means of interorganelle transport of cellular components, especially of lipids, is discussed.

Key words
  • Saccharomyces cerevisiae
  • three-dimensional reconstruction
  • cell ultrastructure
  • organelle association
  • mitochondria
  • endoplasmic reticulum


Organelles of all eukaryotic cells communicate with each other for various reasons. Firstly, transfer of metabolites between compartments is required to maintain the cellular balance of pathways distributed among different organelles. As an example, steps of the glyoxylate cycle are distributed between mitochondria and peroxisomes. Secondly, transfer of components is needed for the exchange of information between subcellular fractions. Prominent examples of this process are the transport of mRNA from the nucleus to the ribosomes and various routes of signal transduction. Finally, interorganelle transfer of components such as proteins or lipids is a prerequisite to maintain structure and function of subcellular compartments. As most organelles are not autonomous with respect to their molecular equipment, the aspect of organelle communication deserves our special attention.

Components can migrate between organelles by diffusion, with the aid of helper proteins, by vesicle flux or through membrane contact. Small molecules such as metabolites or ions are mainly transported as monomers with or without protein catalysis, whereas larger molecules, e.g. polypeptides, migrate in most cases through receptor mediated processes or by vesicle flux. The most prominent example with that respect is protein secretion, which requires a complex machinery to translocate proteins from the interior of the cell to the cell periphery. Intracellular translocation of lipophilic components is a specific task, because for biophysical reasons and as a result of their hydrophobic properties, lipids cannot cross aqueous compartments without appropriate auxiliary components.

Phenomena of intracellular migration of molecules have been investigated with various types of eukaryotic cells. Several components involved in these processes were studied at the molecular level, and some of the mechanisms are understood already in some detail. Less attention has been paid to the structural arrangement of organelles in connection to transport and communication processes. This aspect appears to be specifically important for the assembly of lipids into organelle membranes. Increasing recent evidence suggested that membrane contact may be the most relevant mechanism for lipid migration between organelles. This mechanism was proposed to govern the import of lipids from the endoplasmic reticulum (ER) into mitochondria in various experimental systems (for reviews see Daum & Vance, 1997; Nebauer., 2003; Voelker, 2004, 2005; Rosenberger & Daum, 2005), but also for the supply of lipids from internal membranes to the plasma membrane (Pichler., 2001; Baumann., 2005; Schnabl., 2005; Sullivan., 2006).

To visualize membrane association and contact which may be a prerequisite for translocation processes and/or related cellular events we have used transmission electron microscopy (TEM) for three-dimensional (3D) imaging of the yeast Saccharomyces cerevisiae. TEM of ultrathin serial sections of cells was originally used both for morphometric measurements and evaluation of section profiles to obtain a better view of organelles and cellular structures (Preuss., 1992; Bähler., 1993; Mulholland., 1994; Sato & Yano, 1994; Bennett & Laurie, 1995; Gaffal., 1995; Manella., 1997; Singh., 1998; Yamaguchi., 2002). Although various yeast species have already been examined by ultrathin sectioning and computer assisted 3D reconstructions, these studies were mostly restricted to the visualization of the heterogeneity of mitochondrial structure (Yamaguchi., 2003) or single selected cell structures (Osumi, 1998; Winey., 2005). Recently, the 3D ultrastructure including basic quantitative data of the organelle structure of the pathogenic yeast Exophiala dermatitidis was investigated by ultrathin serial sectioning (Biswas., 2003; Yamaguchi., 2003). 3D images of the internal structural organization of Saccharomyces cerevisiae cells were also obtained by X-ray tomography (Larabell & Le Gros, 2004) and using dual beam electron microscopy (Heymann., 2006). The advantage of these methods is the combination of fairly high resolution with a short time of data collection. However, the resolution of these methods is limited to c. 60 nm. Therefore internal membrane or organelle associations with distances of <30 nm cannot be identified. The reason for the paucity of such data may be that 3D reconstruction of cells, organelles and fine structures based on electron microscopic inspection of ultrathin serial sections is still laborious and time consuming.

The method of 3D image reconstruction of yeast cells employed in this study is similar to procedures reported earlier (Stevens, 1977; Osumi, 1998; Zellnig., 2004). The advantage of our method is that distinct structures of interest can be shown separately while other elements like organelles or their substructures are faded out. Selected structures can be visualized from various angles, yielding detailed information about association and arrangement of organelles. For comparative reasons and to minimize misinterpretation due to fixation artifacts, both chemical fixation and cryofixation of yeast cells (Biswas., 2003; Konomi., 2003; Feron., 2005; Winey., 2005; Binns., 2006) were used in this study to obtain detailed data about organelle organization and membrane association in yeast.

Materials and methods

Strain and culture conditions

The haploid wild-type strain S. cerevisiae W303 was used throughout this study. Cells were grown on YPD medium containing 1% yeast extract, 2% peptone and 2% glucose to the late logarithmic phase, harvested by centrifugation and subjected to fixation as described below.

Fixation of cells and sectioning for electron microscopy

Electron microscopic investigations were performed with chemically fixed and high pressure frozen/freeze-substituted material. For chemical fixation, cells were treated with 4% paraformaldehyde/5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.0) and 1 mM CaCl2 for 90 min at room temperature. Then, cells were washed in buffer with 1 mM CaCl2 for 1 h and incubated for 1 h with a 2% aqueous solution of KMnO4. After washing for 30 min, samples were dehydrated in a graded series of ethanol (50–100%, with en bloc staining in 2% uranylacetate in 70% ethanol overnight) and gradually infiltrated with increasing concentrations of Spurr resin (30%, 50%, 70% and 100%) mixed with ethanol for a minimum of 3 h of each step. Samples were finally embedded in pure, fresh Spurr resin and polymerized at 60°C for 48 h.

To prepare high pressure frozen (HPF) and freeze-substituted cells, pellets of yeast were high pressure frozen in the Leica EM Pact (Leica Microsystems, Vienna, Austria). Platelets containing frozen samples were then transferred and stored under liquid nitrogen conditions in transfer boxes or prechilled specimen baskets. For freeze substitution (FS), specimens were transferred into precooled cryogenic vials (Corning Incorporated) filled with FS-medium consisting of 2% osmium tetroxide in anhydrous acetone containing 0.2% uranyl acetate. FS was carried out at −80°C (72 h), −65°C (24 h), −30°C (24 h), 0°C (12 h) and room temperature (1 h). After rinsing the samples twice in anhydrous acetone for 15 min, they were infiltrated by solutions containing acetone and Agar 100 epoxy resin (mixtures 2 : 1, 1 : 1, 1 : 2) and pure epoxy resin for at least 3 h at room temperature. Embedded samples were then polymerized in pure, fresh Agar 100 epoxy resin for 48 h at 60°C.

Series of 80-nm ultrathin sections were cut with an ultramicrotome (Reichert Ultracut S) and transferred to single slot grids because their transmission area is high enough to show a complete ribbon of serial ultrathin sections without any disturbing bars. The ultrathin sections were stained with lead citrate and viewed with a Philips CM 10 transmission electron microscope (TEM). Section thickness restricts resolution in z direction to 80 nm.

Computer aided 3D reconstruction and determination of surface areas and volumes

TEM micrographs of the serial sectioned yeast cells were digitized by a scanner (Epson 4990 Photo) connected with a Pentium IV computer (2.8 GHz, 1 GB RAM) and imported as TIFF files. The digitized micrographs were pixel images which were transferred into vectorgraphics by tracing selected cell structures (organelles) semi-automatically using a computer program (Corel Trace) or by hand (Corel Draw). For 3D reconstructions, vectorgraphics were aligned by centering specific structures. This step required visibility of two section edges on selected micrographs, which allowed general centering of images. If necessary, centering was corrected by a one-upon-another arrangement of selected ultrastructures of successive sections. 3D reconstructions were created by the program carrara studio (Softline). Complete cells and organelles were measured and three-dimensionally reconstructed. It is noteworthy that this method led to images of reconstructed, real existing cell structures and not to models based on statistical possibilities.

The circumference and area of sectioned organelles was measured by the computer program optimas 6.5 (Bio Scan) and data were exported to the software program excel. The total surface area and volume of organelles were calculated by the sum of their circumferences/areas multiplied by the section thickness. Organelle associations were determined on the TEM micrographs by counting the number of associations when two selected cell compartments (e.g. mitochondria and ER) were at a distance of <30 nm from each other. The counting was performed on every single section of the section series and resulted in the total number of associations of two compartments within the whole cell. As the number of associations may depend on the abundance and size of the organelles, data were also presented as number of associations per 10 μm2 organelle surface area.


General structural aspects

Conventional ultrathin sections of yeast cells were characterized by sectioned profiles of the major organelles such as nucleus, mitochondria, vacuoles, endoplasmic reticulum (ER) and lipid particles (Fig. 1a and c). It is, however, impossible to judge the frequency of organelles and their spatial relation from 2D images of the cell, or to assign separate profiles of the same structure to a single or more organelles of the same kind, e.g. vacuolar profiles in Fig. 1a. In contrast, 3D reconstruction of cells provides information about the number, arrangement and association of different organelles and subcellular structures. Reconstructed images of yeast demonstrated the dense organization of different organelles (Fig. 1b and d). The large number of different structures and organelles in the cell, however, made it almost impossible to obtain a distinct view of individual structural elements. An optical cut of the cell (Fig. 1b and d) or fading out selected structures (Fig. 2) facilitated identification of structural details. When cell wall and plasma membrane were electronically subtracted from the image, a dense network of peripheral ER became visible, especially in chemically fixed cells (Fig. 2a). This subcellular structure is often referred to as cortical ER, which is structurally different from the perinuclear/nuclear ER (Prinz., 2000; Voeltz., 2002; Estrada., 2005). The occurrence of ER fractions in proximity to the yeast plasma membrane has been described before by Preuss. (1992) and Prinz. (2000). It has to be noted, however, that the peripheral ER did not completely shield internal organelles from the plasma membrane because it was not a strict continuum (Fig. 2a). In cryofixed cells (Fig. 2b) association of the ER with the plasma membrane was less pronounced than in chemically fixed cells. It appeared that chemical fixation of yeast cells caused changes in the cellular distribution of the ER resulting in a delocalization towards the cell periphery. Nevertheless, associations of the ER with the plasma membrane could also be observed in cryofixed cells.


Transmission electron micrographs of ultrathin sections (a, c; bar=1 μm) and corresponding 3D reconstructions of serial sections (b, d) of a chemically fixed (a, b) and cryofixed (c, d) yeast cell. (a, c) 2D section profiles of some cell compartments. (b, d) the complexity of the cell organelles is demonstrated by means of a 3D view. CW, cell wall; ER, endoplasmic reticulum; LP, lipid particle; M, mitochondrion; N, nucleus; V, vacuole.


3D reconstructions of internal structures of a chemically fixed (a, c, e) and cryofixed (b, d, f) yeast cell showing selected organelles after fading out of distinct cell structures. Square=1 μm2. (a, b) The cell wall and the plasma membrane are faded out and the peripheral ER system (green) becomes clearly visible in the chemically fixed cell. (c, d) Arrangement and association of the internal ER (green) and mitochondria (blue) after fading out of all other cell structures. (e, f) Arrangement of the nucleus (red) and vacuoles (purple) after fading out of all other cell structures and rotation of the image in (f). The fragmentation of the vacuole in the chemically fixed cell (e) is clearly visible. ER, endoplasmic reticulum; M, mitochondrion; N, nucleus; V, vacuole.

Besides the peripheral network, the ER was also found in the interior of the cell where it came in contact with other organelles. The latter observation became most evident when other cell organelles were faded out (Figs. 2b–d). A prominent type of organelle association occurred between the ER and mitochondria. Physical interaction of these two organelles from the yeast has been described before (Gaigg., 1995; Achleitner., 1999). Mitochondria formed a compact cell compartment consisting of 16–35 single organelles, which were at least in some cases found in close vicinity to the ER (Fig. 2c and d).

The ER was also associated with lipid particles (Fig. 2b). The physiological relevance of this observation has not yet been proven. Studies from our laboratory (Leber., 1998; Athenstaedt., 1999) and from others (Lum & Wright, 1995) had led to the assumption that lipid particles of the yeast may be derived from the ER (for a recent review see Athenstaedt & Daum, 2006). This hypothesis is in line with the model of oil body biogenesis in plant cells (Napier., 1996; Galili., 1998; Murphy & Vance, 1999; Zweytick., 2000; Hills & Roscoe, 2006). Contact between yeast lipid particles and the ER as shown in this study confirmed a possible relationship between these two compartments. Lipid particles of the yeast were often also found to be attached to the vacuole (Fig. 2a and b). This observation is in line with cell fractionation studies (Zinser., 1991; Leber., 1994), which showed that during sucrose density gradient centrifugation crude vacuoles float to the top of the gradient due to their attachment to lipid particles (Uchida., 1988). Only upon treatment with EDTA is this association destroyed; vacuoles can then be sedimented like ‘regular’ organelles, whereas purified lipid particles still float as a result of their low density.

The largest organelles of the yeast were vacuole and nucleus, which were always found in close vicinity (Fig. 2e and f). 3D reconstructions revealed that in chemically fixed cells deformation of the two organelles had occurred (Fig. 2e). In cryofixed cells, the compact structured core of the cell was protected and smoother (Fig. 2f) than in chemically fixed cells. This picture most likely showed the true structural organization of nucleus and vacuole in the yeast cell due to the cryo-preservation of the yeast cell in milliseconds. Changes of the nuclear and vacuolar shape in chemically fixed cells could mainly be attributed to the strongly reduced volume of both organelles under these conditions, whereas the surface area was only affected for vacuoles, as described below in more detail.

Association of organelles

3D reconstructions provided valuable information regarding the detailed structure of organelles from various angles. These images allowed us to detect associations between organelles, but also to quantify association rates. Such calculations may be highly relevant for the evaluation of transport or exchange rates of certain substances between different compartments. We considered a distance of <30 nm between two organelles as an association event due to the expected size of ‘contact proteins’ attached to or associated with the membrane surface in the range of 4–45 nm (Achleitner., 1999). Such proteins may facilitate interaction between organelles as suggested by Corazzi and coworkers (Rakowska., 1994; Camici & Corazzi, 1997; Corazzi., 1998), who described a microsomal protein with potential fusogenic properties.

Associations between organelles of the yeast are numerous and may therefore be relevant for interorganelle communication which bypasses the cytoplasmic compartment (Binns., 2006). Some organelles exhibit a specifically high tendency to associate with each other. Based on 3D imaging of yeast cells described in this study the frequency of association between subcellular compartments was documented by counting membrane contacts per series of cell sections (Table 1). Due to different surface areas of organelles, however, the number of associations per defined organelle surface area was regarded as a more relevant measure for the affinity between two compartments (Table 2). Both types of calculation were performed for chemically fixed and cryofixed yeast cells. It has to be mentioned that all data presented in this work were obtained from cells grown on glucose. Other carbon sources may influence the cell ultrastructure, the abundance of organelles and, consequently, association rates between subcellular compartments. Such observations, which suggested changes of the cellular organization of yeast cells cultivated on fatty acids as a carbon source resulting in the induction of peroxisomes, were discussed in a previous paper from our laboratories (Achleitner., 1999). More recent investigations (our own unpublished results) confirmed this view.

View this table:

Number of total associations of organelles (distance between two organelles less than 30 nm) in yeast cells grown on YPD counted on all serial sections of the cell

No. of associations in chemically/cryofixed cells
  • Data presented for chemically fixed/cryofixed cells.

  • ER, endoplasmic reticulum; LP, lipid particle; M, mitochondrion; N, nucleus; PM, plasma membrane; V, vacuole.

View this table:

Total organelle surface area (μm2) and organelle volume (μm3) per cell and corresponding number of associations (distance between two organelles less than 30 nm) per 10 μm2 organelle surface area

Organelle data for chemically/cryofixed cells
No. of associations per 10 μm2organelle surface area
  • Example of a calculation: 28.5 associations were calculated per 10 μm2 PM with 10 μm2 ER in a chemically fixed cell. Data presented for chemically/cryofixed cells.

  • ER, endoplasmic reticulum; LP, lipid particle; M, mitochondrion; N, nucleus; PM, plasma membrane; V, vacuole.

Numerical data from cells treated with the two different methods of fixation were in some cases dramatically different. The most prominent example was the calculated volume of nuclei and vacuoles from the two types of fixed cells. Whereas cryofixation led to a calculated volume of both organelles which roughly corresponds to a globular shape of the organelle, chemical fixation dramatically decreased the volume values. Obviously, the latter treatment led to massive shrinking of these large organelles. As can be seen from Tables 1 and 2, however, the method of fixation not only affected the shape of cell organelles as a result of alterations in surface area and volume, but also the number of organelle associations. Generally speaking, associations between organelles were found to be more numerous in chemically fixed cells than in cryofixed cells. In chemically fixed cells the highest number of associations related to the surface area was found between the organelle pairs lipid particles/nucleus, lipid particles/vacuole, ER/plasma membrane and vacuole/nucleus. This calculation roughly reflected the relative number of associations between organelles (see Table 1) with the exception of ER/plasma membrane and ER/mitochondria, whose single associations appeared to be more frequent. With cryofixed cells the relative number of ER/mitochondria associations (Table 1) was similar to that of ER/plasma membrane. Other associations which were considered numerous in chemically fixed cells were less pronounced in cryofixed cells. The calculated ratios of associations per organelle surface area (Table 2) demonstrated high association rates between lipid particles/mitochondria, ER/mitochondria and lipid particles/nucleus.


Contact between organelles has been recognized during the last decade as important prerequisite for a number of cellular processes. This surface contact of subcellular compartments does obviously not occur randomly, but appears to be a directed event. Thus, it is not surprising that different organelles associate with different efficiency as described in this study. The molecular basis of organelle association may be catalysis of fusogenic proteins, as has already been suggested (Rakowska., 1994; Camici & Corazzi, 1997; Corazzi., 1998). On the other hand, the occurrence of ‘contact blockers’ which may prevent association events between subcellular compartments has to be taken into account. It will be a task for the future to identify and characterize such components, which may be highly important for communication and biochemical interaction between organelles.

To visualize association between yeast organelles we have chosen the method of 3D reconstruction of electron microscopic images derived from serial ultrathin sections, which allowed structural investigations at the required level of resolution. Whereas the method of computational reconstruction was well applicable for this purpose, the technique of cell fixation turned out to be most critical. The choice of the fixation method either by chemical treatment or by cryotechnique strongly affected the shape of certain organelles as well as the occurrence of associations between subcellular compartments. Although the general message of these two experimental approaches appears to be the same, namely that certain organelles such as mitochondria and the ER, or the ER and the plasma membrane show a clear preference for association, quantitative differences must not be ignored. It appears that in chemically fixed cells swelling or shrinkage of some organelles occurs, resulting in a much higher number of associations than in cryofixed cells. The organelles most strongly affected by the method of fixation are the ER, the nucleus and the vacuole as shown by calculation of their contacts to other organelles, their surface area and their volume. Thus, although both fixation methods may be valuable for specific purposes, quantitative results should be interpreted with caution, and possible effects of the fixation procedure leading to osmotic imbalance inside the cell have to be taken into account. Altogether, cryofixation seems to yield a better and conserved view of the native structure of the yeast than chemical fixation.

Visualization of organelle association does not, of course, explain its biological role or relevance. It has been postulated that contact between subcellular compartments may be highly important for various cell dynamic processes. One of the best studied organelle interactions is that of the ER and mitochondria; this has also been the most prominent association found in the present study. The possible role of this contact in lipid migration between the two organelles was reported before (Gaigg., 1995; Daum & Vance, 1997; Achleitner., 1999; Voeltz., 2002; Levine, 2004). A subfraction of the ER associated with mitochondria from mammalian (Vance, 1990, 1991; Rusinol., 1994) and yeast cells (Gaigg., 1995; Achleitner., 1999) named MAM (mitochondria associated membrane) was isolated and characterized. The MAM fraction of the yeast was shown to be distinct from the bulk ER, especially because of its high capacity to synthesize phospholipids. Experiments with reconstituted isolated organelles (Gaigg., 1995; Achleitner., 1999) or permeabilized cells (Achleitner., 1995) demonstrated that contact between MAM and mitochondria was a prerequisite for lipid transport between the two compartments. Schumacher. (2002) showed that a yeast strain mutated in MET30, which encodes a substrate recognition subunit of a multiprotein SCF (Skip1/Cullin/F box protein components) ubiquitin ligase, was affected in phosphatidylserine transport between the two organelles. It has also been shown that inactivation of Met4p by Met30p-mediated ubiquitination (Kaiser., 1998), alleviated the inhibition of phosphatidylserine transport, possibly by disturbing membrane contact.

Another function that was ascribed to ER/mitochondria contact, although not with the yeast, was Ca2+ signaling and regulation (Rizzuto, 1998; Csordas., 1999; Rutter & Rizzuto, 2000; Filippin., 2003; Levine, 2004; Rutter, 2006). This process may be facilitated by enrichment of IP3-activated receptors at the ER/mitochondria interfaces (Rizzuto., 2004). Ca2+-ions released from mammalian ER in response to hormonal stimulation might be preferentially transferred into the mitochondrial matrix, causing the local activation of ATP synthesis. The increase in mitochondrial Ca2+ might trigger a release of apoptosis-activating substrates and modify the activity of ER-located Ca2+ release channels and thus the dynamics of cytosolic Ca2+ oscillations.

Contact of the ER with organelles different from mitochondria, as also documented in this study, may be of equal importance for short range nonvesicular intracellular trafficking (Levine, 2004). A subfraction of the yeast ER associated with the plasma membrane might fulfill such functions (Pichler., 2001). Similar to MAM, this portion of the ER exhibited a high capacity to synthesize lipids. An involvement in direct lipid transport to the plasma membrane similar to the MAM/mitochondria system was postulated but not yet proven at the experimental level. Transport of ergosterol from the ER to the plasma membrane in yeast by a nonvesicular mechanism as shown recently by Baumann. (2005) may occur through such membrane contact zones. Intimate plasma membrane/ER interactions have also been discussed as potentially essential factors in the capacitative Ca2+ entry in nonexcitable mammalian cells (Putney, 1999; Yao., 1999; Patterson., 1999). It was suggested that refilling of Ca2+ stores of the ER upon Ca2+ release to the cytoplasm requires direct interaction of the plasma membrane with the ER (Putney., 2001).

In summary, the reason for the juxtaposition of the ER to other organelles in the cell appears to be that all organelles need components that are made or stored in the ER. For mitochondria, which have no connection with the ER via vesicular trafficking, direct transfer appears to be especially important. In the case of the plasma membrane, traffic via membrane contact may provide a selective alternative to vesicle flux.

Another prominent organelle association is the nucleus–vacuole junction (NV junction) which has been studied in S. cerevisiae in some detail (for reviews see Voeltz., 2002; Kvam & Goldfarb, 2006). NV junctions are created by specific binding interactions between the vacuole membrane protein Vac8p and the nuclear membrane protein Nvj1p (Pan., 2000) forming a stabile complex. NV junctions mediate piecemeal microautophagy of the nucleus (PMN), an autophagic process that targets portions of the nucleus for degradation in the vacuole (Roberts., 2003). At NV junctions two proteins involved in lipid metabolism are accumulated, namely Tsc13p and Osh1p (Levine & Munro, 2001; Kohlwein., 2001; Kvam & Goldfarb, 2004). Both proteins are supposed to alter the local lipid composition of the nuclear and vacuole membranes, thereby facilitating the biogenesis of PMN vesicles (Kvam & Goldfarb, 2006). Osh1p belongs to the yeast Osh family, which shows some homology to mammalian oxysterol-binding protein (OSBP) (Levine & Munro, 2001). OSBP and OSBP-related proteins are cytoplasmic lipid-binding proteins and might specifically mediate lipid exchange at membrane contact sites (Olkkonen & Levine, 2004).

In conclusion, examples of membrane contact discussed in this study demonstrate the possible physiological relevance of organelle associations. Visualization and quantification of these associations on an ultrastructural level are therefore important for the understanding and evaluation of various cellular processes.


This study was financially supported by the Fonds zur Förderung der wissenschaftlichen Forschung in Österreich (project 17321 to GD, and project P15374-B03 to GZ).


View Abstract