Lipid droplets lighting up: Insights from live microscopy

FEBS Letters 584 (2010) 2168–2175

journal homepage: www.FEBSLetters.org

Review

Lipid droplets lighting up: Insights from live microscopy Margarete Digel, Robert Ehehalt, Joachim Füllekrug * Molecular Cell Biology Laboratory Internal Medicine IV, Im Neuenheimer Feld 345, University of Heidelberg, D-69120 Heidelberg, Germany

a r t i c l e

i n f o

Article history: Received 1 March 2010 Revised 19 March 2010 Accepted 22 March 2010 Available online 30 March 2010 Edited by Wilhelm Just

a b s t r a c t Lipid droplets emerge as important intracellular organelles relevant for lipid homeostasis and the pathophysiology of metabolic diseases. Here, we present a personal view on the current knowledge about the biogenesis of mammalian cytoplasmic lipid droplets, with a focus on microscopy and especially live imaging. We also discuss difficulties related to the lipid droplet proteome, contentious views on lipid droplet growth, and last but not least the evidence for the heterogeneity of lipid droplets within a single cell. We conclude with an outline of the most important future challenges. Ó 2010 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

Keywords: Lipid droplet Live microscopy Biogenesis Targeting

1. Introduction Cytosolic lipid droplets (LDs) are intracellular storage organelles occurring in most if not all cell types including those of plants and microorganisms. Historically considered as inert and static aggregates of neutral lipids (‘‘great balls of fat” [1]), this view has changed dramatically in recent years so that LDs are now regarded as dynamic subcellular organelles [2–5]. Concomitantly, LDs have been recognized as highly relevant for widespread and serious human diseases like diabetes type II, hepatosteatosis and atherosclerosis. More than 10 years after the seminal review on lipid droplet biogenesis by Murphy and Vance [6], many fundamental questions regarding LD biology remain unanswered. While the number of papers on lipid droplets certainly exploded during the last decade, the amount of decisive findings seems not to have increased proportionally. That may also be the reason why cell biology textbooks appear still reluctant to cover LDs more comprehensively. LDs are of globular shape, and their size ranges from 0.1 to 2 lm in most mammalian cells, with the notable exception of adipocytes (up to 150 lm). They are composed of a hydrophobic core containing neutral lipids, mainly triglycerides (TG) and cholesteryl esters. The surrounding membrane consists of a phospholipid monolayer which is a unique feature for a cytoplasmic organelle. Structural proteins of the perilipin family are specific for LDs and are estimated to cover about 15% of the surface [7]. Lipid metabolizing en-

* Corresponding author. Fax: +49 6221 565399. E-mail address: [email protected] (J. Füllekrug).

zymes are the second major class of proteins associated with LDs [8,9]. LDs have an obvious function as storage depots; neutral lipids may be mobilized for the generation of energy by b-oxidation, or for the synthesis of membrane lipids and signaling molecules. Less evident is the regulatory function on cellular and whole-organism lipid homeostasis. Knockout mice lacking LD-associated proteins have revealed striking metabolic changes: For instance, perilipin ko mice showed increased basal lipolysis and were resistant to both diet induced and genetically determined obesity [10]. Fsp27 ko mice showed elevated mitochondrial biogenesis in white adipose tissue and were protected from insulin resistance and diet induced obesity [11]. Apart from their metabolic role, LDs are also exploited as transient storage vehicles for certain proteins which may be released or degraded at later time points. This unexpected function (the LD sequestration hypothesis) has been eloquently summarized recently [12]. De novo formation of lipid droplets is tightly connected to the endoplasmic reticulum (ER) but the molecular mechanism remains unsolved. Triglycerides synthesized by ER-localized enzymes would only be dissolved to a small extent in the phospholipid bilayer (up to 3% in model systems [13]). Beyond this, phase separation occurs and the neutral lipids may accumulate between the two leaflets of the ER membrane. Perilipin family proteins or still unknown adaptors may then shape this fluid ‘‘lens” of lipid into a small bud, eventually giving rise to a globular lipid droplet. This model elegantly explains the origin of the phospholipid monolayer surrounding the neutral lipid core but is not supported by unequivocal evidence to date [2,3,6]. Alternative models have been proposed which however are based on even less data so far [5,14,15].

0014-5793/$36.00 Ó 2010 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2010.03.035

2169

M. Digel et al. / FEBS Letters 584 (2010) 2168–2175

Formation of lipid droplets is a widely conserved process among cells of different origin [16]. It appears unlikely that any single gene would be essential for the biogenesis of LDs. Still, genome wide RNAi screens identified many genes regulating the amount and morphology of lipid droplets [17,18]. Prominent among these were enzymes of phospholipid synthesis suggesting that the surrounding monolayer membrane has a critical influence on the size distribution of LDs. More surprising was the discovery that the COPI coatomer complex long known for its role in the secretory pathway plays an important role in LD morphology and lipolysis; this is covered in this issue by Beller et al. The identification of sorting signals for proteins specifically localized to the LD surface is just beginning. Not even an idea has been formulated so far regarding what kind of machinery might be involved in recognizing the few targeting motifs identified, let alone how these molecules are actually guided to LDs from the cytosol or the ER. A decade ago, the most important contributions to the field of lipid droplet biology from different model systems could be summarized into one comprehensive review (even including lipoproteins; [16]). This is no longer possible. In this mini-review we focus on the biogenesis of mammalian LDs with an emphasis on the contributions from microscopy, especially from live imaging. 2. Lipid droplet protein composition The first proteins to be recognized as constitutively associated with lipid droplets were members of the PAT family (perilipin, adipophilin, TIP47), now renamed to perilipin family proteins (perilipins 1–3; [19]); perilipins probably serve as a dynamic scaffold, regulating formation, growth and lipolysis of LDs. A bit surprising was the discovery that enzymes of lipid metabolism are also abundantly located at the LD surface (e.g. [8,20]), even if most of them are not as exclusively associated with LDs as the perilipins. Please refer to Table 1 for abbreviations and nomenclature of lipid droplet associated proteins discussed in this review. Without doubt, proteomic studies have been a great contribution to the characterization of the lipid droplet organelle (recently summarized [21]). Mass spectrometry technology has seen an awesome development in recent years and has been pushed to ever higher sensitivity. However it seems appropriate to advise some caution when drawing conclusions from these data. The foremost concern when looking at the long lists of putative LD-associ-

ated proteins is of course contamination of the LD fractions. First, during homogenization hydrophobic patches will be very likely getting exposed leading to unspecific capture of unrelated proteins. Since it is well known that even luminal ER proteins which are enclosed by double membranes are released during homogenization, this can be expected to be an even bigger problem for LDs. Second, although the abbreviation ‘‘LAM” for LD-associated membrane did not come into widespread use [22], numerous labs using a variety of approaches have documented close association of LDs with other organelles, especially the ER. It seems not surprising at all that some of these attachments are not resolved by biochemical fractionation procedures. For instance, calnexin is a very well characterized, abundant integral transmembrane protein of the endoplasmic reticulum and has served in countless studies as a marker protein for this compartment. To us, detection of calnexin in LD preparations would indicate contamination by ER membranes. Third, regarding the logic in assuming that a protein has a higher chance of being indeed an LD-associated protein because it was found in several independent studies could be misleading. A contaminant will always remain a contaminant. Finally, lack of quantification is an inherent difficulty of mass spectrometry. It is often impossible to even get only a rough idea on how abundant the identified proteins are. Two protein families may exemplify the difficulties: About twenty members of the rab family of small GTPases have been documented by proteomics of LD preparations, but when assessed on an individual basis by microscopy only rab18 turned out to be strongly LD associated so far [22,23]. It cannot be excluded that other rabs also play a significant role on LDs (e.g. [24]) but this awaits more in vivo evidence. Similarly, although five mammalian long chain acyl-CoA synthetases (ACSL) have been listed as LD associated by proteomics, our own work on the subcellular targeting of these enzymes could only confirm the presence of ACSL3 [9,25] on lipid droplets [26]. Concluding, the LD proteome needs confirmation, preferably by high-resolution microscopy and other independent methods. In spite of all this, the repeated documentation of ER proteins (both membrane and luminal soluble proteins) by LD proteomics has legitimately sparked innovative ideas on the biogenesis of lipid droplets. The bicelle formation model [15] predicts the presence of ‘‘wrinkles” on the LD surface corresponding to small areas of bilayer membrane which would harbor transmembrane proteins like calnexin. The vesicular budding model [5] involves transient

Table 1 Abbreviations and protein nomenclature. All lipid droplet associated proteins mentioned in this review are listed with their original designation and the most common alternatives. A weblink to the entries for the human homologues within the UniProt database (http://www.uniprot.org/) is provided as a comprehensive reference. References are biased towards analysis of targeting signals rather than first report of LD association. Abbr.

Protein names; selected alternatives

UniProt

Ref.

17b-HSD 11 17b-HSD 13 AAM-B ACSL3 ADP ALDI ATGL Caveolin cav3DGV Core CYB5R3 DGAT2 fsp27 LSDP5 Perilipin Rab18 Seipin TIP47 UBXD8

17b-Hydroxysteroid dehydrogenase type 11 17b-Hydroxysteroid dehydrogenase type 13 Methyltransferase-like protein 7A Long-chain acyl-CoA synthetase 3, FACL3 Adipophilin, perilipin 2, ADRP (adipose differentiation-related protein) Associated with lipid droplet protein 1, methyltransferase-like protein 7B Adipose triglyceride lipase, desnutrin Caveolin-1 Truncated version of caveolin-3 HCV core protein of isolate HC-J6 Cytochrome b5 reductase 3 Diacylglycerol O-acyltransferase 2 Fat-specific protein FSP27 homolog, Cidec Lipid storage droplet protein 5, perilipin 5 Perilipin 1 Ras-related protein rab-18 Bernardinelli-Seip congenital lipodystrophy type 2 protein 47 kDa Mannose 6-phosphate receptor-binding protein, perilipin 3 FAS-associated factor 2, UBX domain-containing protein 8

Q8NBQ5 Q7Z5P4 Q9H8H3 O95573 Q99541 Q6UX53 Q96AD5 Q03135 P56539 P26660 P00387 Q96PD7 Q96AQ7 Q00G26 O60240 Q9NP72 Q96G97 O60664 Q96CS3

[45] [45] [46] [9] [88] [52] [39] [51] [48] [41] [46] [59] [11] [89] [33] [22] [64] [29] [53]

2170

M. Digel et al. / FEBS Letters 584 (2010) 2168–2175

formation of a bilayer vesicle with an aqueous lumen, taking resident ER proteins with it. Later on, these vesicles would acquire progressively more neutral lipids which may eventually squeeze out the ER remnants. It will be very interesting to follow future papers putting these hypotheses rigorously to the test.

3. Lipid droplet biogenesis The biogenesis of lipid droplets may be defined as comprising the initial accumulation of neutral lipids at the ER followed by the recruitment of specific proteins which together with continued lipid synthesis leads eventually to a globular LD. For the sake of clarity we include here data obtained with more mature LDs as well, assuming that the targeting principles are essentially the same. 3.1. Initial formation of lipid droplets Nascent LDs need to reach a certain size before they can be faithfully identified. It is estimated that the primordial lipid droplet could have a diameter of only 25 nm ([27]; discussed in [3]). Maturing from this to lm-sized LDs implies enormous changes in membrane curvature and therefore surface phospholipid and protein composition; virtually nothing is known. A structure corresponding topologically to a neutral lipid globule between the two membrane leaflets of the ER has been reported recently [28]. These lipid droplets (surrounded by adipophilin (ADP) on one side and apolipoprotein B on the other) are much larger than nascent LDs, but are still considered to provide support for the prevailing model of LD biogenesis. 3.2. Lipid droplet protein attachment and targeting Identifying lipid droplet targeting sequences has proven tedious in some cases, and no universal motif has been identified. Owing to the phospholipid monolayer of lipid droplets, a classic a-helical transmembrane domain is too long to fit well, and the surrounding polar amino acids will be repelled by the hydrophobic lipid core anyway; polytopic membrane proteins are excluded as well. That leaves amphipathic helices (e.g. hepatitis C virus core protein), lipid anchors (e.g. rab18) and hairpin-like topologies (e.g. caveolin) as the principal modes of direct membrane attachment. It is difficult to see how these would confer specific binding to LD membranes, so additional features are expected to be present. 3.2.1. Perilipin family proteins All perilipin family proteins contain 11 amino acid repeats in the N-terminal region which have the potential to form amphipathic helices. For TIP47, this region was sufficient for partial localization to LDs [29]. In another study however, the N-terminus was found to be dispensable for LD localization; a four helix bundle with structural similarity to apolipoprotein E [30] and a putative hydrophobic cleft contained within are also discussed as targeting determinants [31]. Perilipin requires three separate hydrophobic sequences in the central domain for LD attachment [32,33]; structural information is not available. ADP is stably localized only to LDs but degraded in the cytosol. FRAP (fluorescence recovery after photobleaching) is a powerful application of live microscopy which allows the measurement of diffusional mobility [34]. ADP and LSDP5 were recovered fast after photobleaching of CHO cells as would be expected for proteins exchanging between cytosol and LDs [35]. However, perilipin itself was slow to recover; the mechanistic reason for this is not clear. ADP was slowly recovered in an earlier study using human hepatoma Huh7 cells [36], suggesting there could be cell type specificity. For a more detailed description of perilipin family proteins see recent reviews [37,38].

3.2.2. Atgl Deficiencies in the gene of adipose triglyceride lipase (ATGL) cause neutral lipid storage disease with myopathy. Naturally occurring mutations affecting the C-terminus retain catalytic activity but do not associate with lipid droplets [39]; the sorting signal has not been mapped further so far. Targeting of ATGL to LDs is dependent on COPI and COPII coatomer proteins, suggesting that ATGL could be a protein which is originally inserted into the ER and then moves to LDs [40]. Mobility as assessed by FRAP is similar to perilipin on the longer time scale. 3.2.3. Hepatitis C virus core protein Association of the core protein of hepatitis C virus (HCV) with LDs was reported many years ago [41] but only recently it was recognized that lipid droplets are crucial for the production of infectious HCV virus particles [42]. Fifty amino acids of the core protein comprising two amphipathic a-helices are necessary and sufficient for LD association [43]. Remarkably, point mutations within this region which affect virus assembly also change the mobility (measured by FRAP) [44]; it remains as yet unclear if the mobility of the core protein is also mechanistically relevant for the targeting to LDs. 3.2.4. N-terminal hydrophobic anchoring In contrast to the maybe a bit confusing variety of sorting signals for the perilipin family proteins, just the N-terminal 35 amino acids of 17b-HSD 11 and 13 were required for localization to LDs [45]. Similarly, the first thirty amino acids of the putative methyltransferase AAM-B, the closely related ALDI and the cytochrome CYB5R3 were reported to be sufficient for LD association albeit with varying efficiencies [46]. Common to all these proteins is a stretch of hydrophobic residues which appears to be distinct from the classical N-terminal signal sequence of the secretory pathway, and consequently is not cleaved off by the signal peptidase at the ER. In addition to the N-terminal hydrophobic region, basic amino acids have been shown to be required for efficient sorting to LDs (for ALDI and a truncated caveolin mutant; [47]). Caveolin molecules strongly associate with LDs under certain conditions (e.g. liver regeneration, exposure to brefeldin A; [48– 51]). A truncated version of caveolin-3 (cav3DGV-GFP) localizes to the ER in lipid starved cells, but already two minutes after exogenous addition of fatty acids, caveolin appeared in small dots corresponding to tiny lipid droplets. Live imaging suggests that caveolin moves from the ER directly to LDs [51]. The same behavior was recently shown for ALDI [52] and UBXD8 [53] and is independent of ongoing protein synthesis. Although this localization shift is highly suggestive of a membrane continuity between ER and small LDs (and does not require the COPII budding machinery of the ER [53]), direct proof is not available. Rather on the contrary, larger LDs appear not to be connected to each other, neither directly nor using the ER as a conduit: FRAP of caveolin and ALDI is on the time scale of several minutes to one hour [52,54]. 3.3. Is there continuity between the ER and lipid droplets? In organelles with a continuous membrane system like the Golgi apparatus or the ER, membrane protein associated fluorescence recovers within seconds [34,55]. In line with this, all three perilipin family proteins showed fast recovery within a single lipid droplet, proving that these proteins move freely within the LD monolayer [35]. Summarizing the available FRAP studies, larger lipid droplets appear physically disconnected from the ER membrane system. Membrane transport between different organelles (e.g. between the Golgi apparatus and the plasma membrane [56], or Golgi–ER cycling [57]) involves time consuming cargo collection and transport steps. The observed recovery times for constitutively

M. Digel et al. / FEBS Letters 584 (2010) 2168–2175

membrane associated LD proteins are consistent with replenishment from a separate organelle. 3.4. Is there lipid mediated sorting? Regardless whether proteins move from the cytosol or the ER to the LD surface, the most intriguing question to us is – how is the LD membrane recognized by all the LD proteins which do not penetrate beyond the monolayer? A proteinaceous master receptor appears unlikely, and has to be sorted itself anyway without any help from other proteins. One hint is provided by recent work suggesting that the perilipin family protein TIP47 is targeted to diacylglyceride-rich membranes [58]. Diacylglycerides are the immediate biochemical precursors for the TG contained within LDs, and are themselves also present in lipid droplet fractions [59]. Another consideration is that most of the current LD cartoons may be an over-simplification when it comes to the interface between the phospholipid monolayer and the neutral lipid core. The core which is usually depicted as a uniform static disk is in reality a quite energetic if not messy assembly of many different molecules jumping in all directions. It is to be expected that triglycerides and cholesteryl esters intercalate dynamically into the surrounding monolayer of amphipathic lipids all the time even if the steady-state percentage is low (see figure, ‘‘counting”). However if the difference to standard ER membranes is sufficiently high to create a lipid microdomain is as yet uncertain. In the light of lateral membrane lipid heterogeneity gaining ever more evidence [60], lipid-mediated sorting of proteins to LDs or TG-rich ER membranes seems a relevant hypothesis to be tested.

2171

against a prominent role of fusion events for the growth of LDs (e.g. [66]). 4.2. Lipid droplets on the move In many model systems, the most interesting question might not be where and how the LDs are moving but why they are moving at all. Early on, microtubules were shown to allow bidirectional movement of lipid droplets dependent on the dynactin complex (observed by DIC microscopy [67]). Later use of GFP fusion proteins labeled LDs specifically and revealed two main modes of movement: Short range oscillatory fluctuations of LDs are observed frequently whereas long range saltatory motions (which are microtubule-dependent) are only sporadic events [23,51,52]. See Supplement #1 for a vivid example of LD movements. The origin of small and fast entities noted occasionally is not certain and could also correspond to ER derived vesicles. Interestingly, in one study on steroidogenic cells imaged by live CARS (coherent anti-Stokes Raman scattering) microscopy the trafficking activity depended very much on the shape of the cells: Flat cells firmly attached to the substratum showed only few movements as opposed to rounded up cells suggesting that the actin cytoskeleton could impair LD movements [68]. CARS microscopy is a recently developed technique which allows the direct monitoring of unlabeled triglycerides in living cells [69]. This method even permits imaging of living animals as exemplified for adipocytes in nerve tissue [70]. For a more detailed introduction to LD trafficking, please resort to [71]. 4.3. Autonomous growth by LD localized biosynthesis

4. Lipid droplet growth Lipid droplets usually keep growing after reaching a globular shape if cellular metabolism continues to deliver neutral lipids. How this is achieved is discussed intensely, with some labs favoring fusion between highly mobile LDs while others prefer to see the lipid droplet as an autonomous organelle capable of local lipid synthesis by enzymes at the LD surface. 4.1. Fusion of LDs FRAP data argue against a physical connection between LDs and ER, posing the problem how fast growing LDs acquire additional neutral lipids. Lipid carrier proteins remain a possibility but currently there is no strong evidence for this. In principle, LDs may grow by fusing with each other or by progressively incorporating more neutral lipids. With few exceptions (e.g. [61]), fusion of lipid droplets has been observed only rarely [4,54,62], and there is considerable skepticism in the field if fusions occur fast enough to account solely for LD growth. Still, RNAi knockdown of the LDassociated SNARE fusion machinery does decrease the size and fusion activity of lipid droplets [63]. Recently the yeast homologue of human seipin (mutant forms cause lipodystrophy) has been identified by two independent screens as a key protein regulating LD morphology and fusion activity [64,65]. In vitro evidence suggests that no cytosolic proteins are needed for the enhanced fusion activity; the concomitantly observed changes in phospholipid composition (from long chain to medium chain fatty acids) [65] could be behind this but direct evidence is not available. It remains to be seen how comparable the situation in mammalian cells will turn out. Efficient fusion would require high LD motility, which has generally been connected to microtubules. Depolymerisation would inhibit most fusion events; but it seems that incorporation of newly synthesized triglycerides is only slightly less efficient, arguing

Local synthesis of neutral lipids at the LD surface seems a very simple and straightforward way to account for growth. This idea is supported by two recent studies reporting the in vitro synthesis of TG from biochemically isolated lipid droplet fractions [59,72]. In line with this, an enzyme catalyzing the final step of triglyceride synthesis (DGAT2) has been shown to be partially localized to LDs [59]. However, some difficulties remain. In the same studies, TG synthesis continues at heavy membranes leaving still the question how this proportion of neutral lipids is going to be moved to LDs. As discussed above, contamination of LDs by ER membranes could also be a problem, and there are conflicting data on the localization of DGAT2 [73]. Regarding the dynamic growth and shrinking of LDs, a very interesting question is how the phospholipid monolayer is being adjusted concomitantly [3]. Again, the conceptually easiest mechanism would be local synthesis and degradation. The rate-limiting enzyme of phosphatidylcholine biosynthesis has been shown to be not only localized to LDs but also to be functionally relevant [17], but many open questions remain. Synthesis of TG from diacylglycerol requires activated fatty acids which are synthesized through esterification with coenzyme A by the housekeeping enzymes of the acyl-CoA synthetase family [74,75]. ACSL3 was identified as the second most abundant protein on lipid droplets prepared from human hepatoma cells (Huh7) [9]. Unlike adipophilin, ACSL3 showed a bimodal distribution and was also present in heavy membranes presumably corresponding to the ER. While it is tempting to speculate that ACSL3 provides fatty acyl-CoAs for the local synthesis of TG and phospholipids, another or even more important function could be the re-capture of fatty acids liberated by basal lipolysis. If ACSL enzymes are inhibited by Triacsin C, fatty acids are rapidly lost from LDs suggesting there is considerable turnover of TG at steady-state. Other acyl-CoA synthetases like ER-localized FATP4/ACSVL4 are likely also contributing to TG synthesis albeit indirectly by providing activated fatty

2172

M. Digel et al. / FEBS Letters 584 (2010) 2168–2175

acids. Overexpression of FATP4 drives fluorescent fatty acids efficiently towards lipid droplets ([76]; M.D. and J.F., unpublished).

5. Lipid droplet subpopulations In many cell types, LDs show a wide variety of different sizes even when extracellular conditions remain constant. It is a tempting speculation that LDs might specialize within a single cell to increase their overall efficiency. It will be extremely interesting to see how this might be regulated. 5.1. LD heterogeneity defined by proteins At first sight LDs appear to constitute a quite homogenous population, at least under steady-state conditions. However when two LD-associated proteins (the small GTPase rab18 and cav3DGV) were co-expressed they labeled different subpopulations of similar size within the same cell [23]. While this could maybe attributed to stochastic polymerization of caveolin (thereby excluding rab18 from the LD surface), it was also reported that endogenous rab18 and ADP showed reciprocal labeling in HepG2 cells [22]. Moreover, cav3DGV appeared within minutes on nascent LDs induced by fatty acid addition whereas rab18 arrived only after 3 h. Together with the observation that there was enhanced LD association of rab18 after stimulation of lipolysis of 3T3-L1 adipocytes, this suggests that the protein coat of LDs is dynamically changed to adapt to different functional requirements. This is further strengthened by the observation that perilipin family proteins exchange during maturation and growth of LDs in 3T3-L1 adipocytes: Early LDs in the periphery contain preferentially TIP47 whereas later on adipophilin and perilipin predominate [62,77]. Subsets of LDs were also observed using invertebrate model systems (Drosophila; [78]). Uneven staining of a particular LD-associated protein at the surface is often caused by fixation artifacts (e.g. [79]; see also Supplement #2). It seems therefore essential to us to analyze potential heterogeneities by live microscopy. Indeed, rab18 was shown to localize to distinct domains at the LD surface also in living cells [23]. A tempting speculation is that rab18 is involved in establishing LD-ER associations that could be involved in lipid exchange between these organelles [2,22]. 5.2. Uneven distribution of lipids Staining of neutral lipids usually does not indicate differences in hydrophobic content among LDs. A more specific approach employed the electron dense chemical osmium tetroxide which preferentially binds to unsaturated fatty acid chains; especially lipids containing polyunsaturated moieties are therefore amenable to high-resolution EM microscopy. Incubation of 3T3-L1 adipocytes with docosahexaenoic acid (DHA) surprisingly showed that middle-sized LDs (2–5 lm) gain DHA containing lipids much faster than smaller or larger LDs [66]. We consider this as direct evidence for subpopulations of LDs differing in their metabolic activity; fibroblasts however showed uniform incorporation into lipid droplets of different size. At the level of light microscopy, application of autofluorescent polyene lipids [80] also indicated that LDs acquired lipids at different rates but this was not quantified [59]. 5.3. Heterogeneity within single lipid droplets When LDs were first filled with cholesteryl esters and only later with DHA lipids (stained again by osmium tetroxide), EM analysis indicated that individual LDs were not homogeneously stained. The images looked suspiciously like phase separation had occurred but if that happened before or after fixation and preparation for EM

imaging is not clear [66]. More direct evidence for heterogeneity within single droplets was gained by live CARS microscopy [81]. 3T3-L1 adipocytes incubated with saturated (palmitate) and unsaturated (linolenate) in a ratio of 3:1 showed a phase separation for some of their LDs. It should be mentioned here that only few studies quantitatively investigated the lipid composition of LDs so far. Remarkably, the LD surface is enriched in phosphatidylcholine but decreased in sphingomyelin and phosphatidylethanolamine when compared to total membrane (CHO cells; [82]). Lyso-phospholipids which allow higher LD curvature due to their shape were also more abundant. It will be very interesting to directly compare LD phospholipids with the putative parental ER organelle, or to learn about phospholipid changes after stimulation of lipolysis. Freeze-fracturing indicated that perilipin family proteins and caveolin are localized within the hydrophobic core of LDs [83]. However, to date this has not been confirmed by other techniques. At the level of light microscopy, neither antibody labeled nor fluorescently tagged perilipin family proteins have been observed internally. Internal membrane structures and even ribosomes within leukocyte lipid droplets have also been reported by electron microscopy [84]. This implies unexpected functions but also awaits substantiation by other methods.

6. Perspective – future challenges Rather than presenting satisfying conclusions on lipid droplet biology (which seems premature anyway), we chose to outline a personal view of the most pressing unsolved issues (see Fig. 1). 6.1. Zooming Elucidation of the molecular mechanism of LD biogenesis has been prevented so far by technical limitations. One challenge is to label nascent LDs unambiguously so that 25 nm or smaller lipid aggregates can be identified and followed over time. Another difficulty is obtaining sufficient resolution. Live STED (stimulated emission depletion) fluorescence microscopy has been demonstrated at a resolution of 50 nm (five times higher as the resolution achievable by standard light microscopy; [85]), but strong photobleaching and long acquisition times might still limit the application to LD biogenesis. Better imaging technology should also resolve some of the contentious issues regarding LD morphology: Are there internal hydrophilic proteins or even membranous structures with attached ribosomes? Are there local areas of phospholipid bilayer? Many powerful technologies which are unsuitable for living cells might still yield insight with artificial systems. Maybe in vitro reconstitution of TG synthesis using recombinant DGAT2 in bilayer liposomes will allow the observation and characterization of lipid aggregates accumulating within the membrane. 6.2. Surfing Assuming that after continued synthesis by ER-localized enzymes neutral lipids have accumulated, the next critical step would be the acquisition of specific proteins. It is a complete enigma how peripherally associated proteins are able to sense neutral lipids beyond their reach. Maybe there simply is a cytosolic phospholipase sensing the increased curvature at the ER membrane surrounding the phase separated lens of neutral lipids; generation of cone shaped lyso-phospholipids would then increase the local curvature, helping to shape a bud. Another idea would be that a locally confined change in lipid composition (by the continuous intercalation of neutral lipids) would offer a more accommodating environment to proteins destined for LDs. Continuing diffusion (‘‘surfing”) of initially ER-localized proteins would be sufficient to

M. Digel et al. / FEBS Letters 584 (2010) 2168–2175

2173

Fig. 1. Future challenges of lipid droplet biology. ZOOMING relates to the strong demand for higher sensitivity and resolution of live imaging in elucidating the biogenesis of lipid droplets. This is likely again a case where technical advances will precede scientific progress. The SURFING hypothesis suggests that the recruitment of proteins to LDs is lipid mediated and essentially a self-organizing process. Only COUNTING lipid droplet molecules will enable a satisfying understanding of the lipid droplet as an organelle. It is overdue to start quantitative approaches in earnest. The enlarged area visualizes triglycerides intercalating into the phospholipid monolayer, especially where cone shaped lysophospholipids are present. Lipid homeostasis both on a cellular and whole organism level is crucially dependent on the extent of lipid oxidation (BREATHING), and one way of regulation may simply be spatial proximity.

allow trapping at lipid microdomains. Elucidation of this mechanism will be supported by the structural analysis of the amphipathic helices of LD proteins which move from the ER to nascent LDs after stimulation of TG synthesis. Likewise, cytosolic proteins could probe the ER membrane all the time but remain attached only when they hit a TG modified area. To us, lipid mediated sorting triggered by the triglycerides themselves is appealing because of the inherent simplicity of a self-organizing principle. Why look for convoluted models when there are easier ways? 6.3. Counting Proteomics have helped define the LD as an organelle with specific protein composition. The next step partially under way is the confirmation of LD association for individual proteins by independent methods. We strongly believe that now is the time to go from qualitative assessment to quantification on a bigger scale. For sure it is functionally important to know how abundant a protein on the LD surface really is. For instance, if a lipid enzyme is the second most abundant protein, might that not be a clue to another function independent of the enzyme activity (e.g. scaffolding)? Another clue regarding functional relevance will come from quantifying the distribution of proteins shared between LDs and a second compartment, especially the ER. If only 2% of a protein are present at the LD surface, will it still perform an important function there? Ultimately it should be feasible to quantitatively describe a standardized 2 lm lipid droplet on a molecular level, attaching numbers to individual lipid species and proteins. As an example, about 45 000 perilipin molecules on a 2 lm LD are to be expected (for an adipocyte LD; assuming a globular shape for perilipin and 10% coverage of the LD surface). Achieving this will be far more difficult than the molecular anatomy of synaptic vesicles [86], especially because of the size heterogeneity of LDs. Another important aspect will be the comparative and quantitative analysis of different cell types by the same technology. Proteins present in different amounts or missing at all will provide important functional insights.

6.4. Breathing There has been much recent speculation on the interactions of LDs with other organelles (e.g. [87]). Admittedly there are many tantalizing hints and lots of circumstantial evidence, but convincing studies directly addressing this are scarce. There is little question that connections to the ER have to be there, and associations to peroxisomes or endosomes might give unexpected functional insights. However more interesting to us would be investigating an alliance between mitochondria and lipid droplets. After all, oxidizing lipids is the only sure way to get rid of them! Surplus storage of lipids is an enormous health problem worldwide, and it is important to investigate all kinds of strategies ranging from molecular changes over neural circuits to behavioral and society aspects. Neutral lipids stored in LDs appear to undergo futile cycles of hydrolysis and resynthesis to a substantial degree. Spatial proximity to mitochondria could already be sufficient to loose some of the liberated fatty acids to b-oxidation. In a very general sense, we would like to point out here that investigating the localization of organelles, proteins and lipids is not about descriptive mapping but hugely important when it comes to functional relevance and molecular mechanisms. There is no doubt that the knowledge about lipid droplet biology will continue to increase at a formidable speed. It will be fascinating to follow how such a fermenting scientific area will eventually converge on generally accepted principles. We hope to have contributed a stimulating if wayward jigsaw piece on the road to this. Acknowledgements Space constraints excluded the citation of many important papers from our colleagues, and in several cases we were able to quote only exemplary studies. In addition, on several issues we chose to present more provocative views rather than being finely balanced and comprehensive. Special thanks to Regina Großmann for providing the fluorescent image and to Berenice Rudolph for

2174

M. Digel et al. / FEBS Letters 584 (2010) 2168–2175

the supplemental movie. Financial support by the German research foundation (FU 340/5-1 to J.F.; EH 196/5-1 to R.E.) is gratefully acknowledged. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.febslet.2010.03.035. References [1] Beckman, M. (2006) Cell biology: great balls of fat. Science 311, 1232–1234. [2] Martin, S. and Parton, R.G. (2006) Lipid droplets: a unified view of a dynamic organelle. Nat. Rev. Mol. Cell Biol. 7, 373–378. [3] Ohsaki, Y., Cheng, J., Suzuki, M., Shinohara, Y., Fujita, A. and Fujimoto, T. (2009) Biogenesis of cytoplasmic lipid droplets: from the lipid ester globule in the membrane to the visible structure. Biochim. Biophys. Acta (BBA) – Mol. Cell Biol. L 1791, 399–407. [4] Thiele, C. and Spandl, J. (2008) Cell biology of lipid droplets. Curr. Opin. Cell Biol. 20, 378–385. [5] Walther, T.C. and Farese Jr., R.V. (2009) The life of lipid droplets. Biochim. Biophys. Acta (BBA) – Mol. Cell Biol. L 1791, 459–466. [6] Murphy, D.J. and Vance, J. (1999) Mechanisms of lipid-body formation. Trends Biochem. Sci. 24, 109–115. [7] Londos, C., Brasaemle, D.L., Schultz, C.J., Segrest, J.P. and Kimmel, A.R. (1999) Perilipins, ADRP, and other proteins that associate with intracellular neutral lipid droplets in animal cells. Semin. Cell Dev. Biol. 10, 51–58. [8] Brasaemle, D.L., Dolios, G., Shapiro, L. and Wang, R. (2004) Proteomic analysis of proteins associated with lipid droplets of basal and lipolytically-stimulated 3T3-L1 adipocytes. J. Biol. Chem. 279, 46835–46842. [9] Fujimoto, Y. et al. (2004) Identification of major proteins in the lipid dropletenriched fraction isolated from the human hepatocyte cell line HuH7. Biochim. Biophys. Acta 1644, 47–59. [10] Martinez-Botas, J. et al. (2000) Absence of perilipin results in leanness and reverses obesity in Lepr(db/db) mice. Nat. Genet. 26, 474–479. [11] Nishino, N. et al. (2008) FSP27 contributes to efficient energy storage in murine white adipocytes by promoting the formation of unilocular lipid droplets. J. Clin. Invest. 118, 2808–2821. [12] Welte, M.A. (2007) Proteins under new management: lipid droplets deliver. Trends Cell Biol. 17, 363–369. [13] Hamilton, J., Miller, K. and Small, D. (1983) Solubilization of triolein and cholesteryl oleate in egg phosphatidylcholine vesicles. J. Biol. Chem. 258, 12821–12826. [14] Robenek, H., Hofnagel, O., Buers, I., Robenek, M.J., Troyer, D. and Severs, N.J. (2006) Adipophilin-enriched domains in the ER membrane are sites of lipid droplet biogenesis. J. Cell Sci. 119, 4215–4224. [15] Ploegh, H.L. (2007) A lipid-based model for the creation of an escape hatch from the endoplasmic reticulum. Nature 448, 435–438. [16] Murphy, D.J. (2001) The biogenesis and functions of lipid bodies in animals, plants and microorganisms. Prog. Lipid Res. 40, 325–438. [17] Guo, Y. et al. (2008) Functional genomic screen reveals genes involved in lipiddroplet formation and utilization. Nature 453, 657–661. [18] Beller, M., Sztalryd, C., Southall, N., Bell, M., Jackle, H., Auld, D.S. and Oliver, B. (2008) COPI complex is a regulator of lipid homeostasis. PLoS Biol. 6, e292. [19] Kimmel, A.R., Brasaemle, D.L., McAndrews-Hill, M., Sztalryd, C. and Londos, C. (2010) Adoption of PERILIPIN as a unifying nomenclature for the mammalian PAT-family of intracellular lipid storage droplet proteins. J. Lipid Res. 51, 468– 471. [20] Liu, P., Ying, Y., Zhao, Y., Mundy, D.I., Zhu, M. and Anderson, R.G. (2004) Chinese hamster ovary K2 cell lipid droplets appear to be metabolic organelles involved in membrane traffic. J. Biol. Chem. 279, 3787–3792. [21] Hodges, B.D.M. and Wu, C.C. (2009) Proteomic insights into an expanded cellular role for cytoplasmic lipid droplets. J. Lipid Res. 51, 262–273. [22] Ozeki, S., Cheng, J., Tauchi-Sato, K., Hatano, N., Taniguchi, H. and Fujimoto, T. (2005) Rab18 localizes to lipid droplets and induces their close apposition to the endoplasmic reticulum-derived membrane. J. Cell Sci. 118, 2601–2611. [23] Martin, S., Driessen, K., Nixon, S.J., Zerial, M. and Parton, R.G. (2005) Regulated localization of rab18 to lipid droplets: effects of lipolytic stimulation and inhibition of lipid droplet catabolism. J. Biol. Chem. 280, 42325–42335. [24] Liu, P., Bartz, R., Zehmer, J.K., Ying, Y.-s., Zhu, M., Serrero, G. and Anderson, R.G.W. (2007) Rab-regulated interaction of early endosomes with lipid droplets. Biochim. Biophys. Acta (BBA) – Mol. Cell Res. 1773, 784–793. [25] Fujino, T., Kang, M.-J., Suzuki, H., Iijima, H. and Yamamoto, T. (1996) Molecular characterization and expression of rat Acyl-CoA synthetase 3. J. Biol. Chem. 271, 16748–16752. [26] Becker, M., Wunsch, J., Rudolph, B., Stremmel, W. and Fuellekrug, J. (2007) Subcellular targeting of acyl-CoA synthetases. FEBS J. 274, 111. [27] Mountford, C.E. and Wright, L.C. (1988) Organization of lipids in the plasma membranes of malignant and stimulated cells: a new model. Trends Biochem. Sci. 13, 172–177. [28] Ohsaki, Y., Cheng, J., Suzuki, M., Fujita, A. and Fujimoto, T. (2008) Lipid droplets are arrested in the ER membrane by tight binding of lipidated apolipoprotein B-100. J. Cell Sci. 121, 2415–2422.

[29] Bulankina, A.V., Deggerich, A., Wenzel, D., Mutenda, K., Wittmann, J.G., Rudolph, M.G., Burger, K.N.J. and Honing, S. (2009) TIP47 functions in the biogenesis of lipid droplets. J. Cell Biol. 185, 641–655. [30] Hickenbottom, S.J., Kimmel, A.R., Londos, C. and Hurley, J.H. (2004) Structure of a lipid droplet protein: the PAT family member TIP47. Structure 12, 1199– 1207. [31] Ohsaki, Y., Maeda, T., Maeda, M., Tauchi-Sato, K. and Fujimoto, T. (2006) Recruitment of TIP47 to lipid droplets is controlled by the putative hydrophobic cleft. Biochem. Biophys. Res. Commun. 347, 279–287. [32] Garcia, A., Sekowski, A., Subramanian, V. and Brasaemle, D.L. (2003) The central domain is required to target and anchor perilipin a to lipid droplets. J. Biol. Chem. 278, 625–635. [33] Subramanian, V., Garcia, A., Sekowski, A. and Brasaemle, D.L. (2004) Hydrophobic sequences target and anchor perilipin A to lipid droplets. J. Lipid Res. 45, 1983–1991. [34] Lippincott-Schwartz, J., Snapp, E. and Kenworthy, A. (2001) Studying protein dynamics in living cells. Nat. Rev. Mol. Cell Biol. 2, 444–456. [35] Wang, H. et al. (2009) Activation of hormone-sensitive lipase requires two steps: protein phosphorylation and binding to the PAT-1 domain of lipid droplet coat proteins. J. Biol. Chem. 284, 32116–32125. [36] Targett-Adams, P., Chambers, D., Gledhill, S., Hope, R.G., Coy, J.F., Girod, A. and McLauchlan, J. (2003) Live cell analysis and targeting of the lipid dropletbinding adipocyte differentiation-related protein. J. Biol. Chem. 278, 15998– 16007. [37] Bickel, P.E., Tansey, J.T. and Welte, M.A. (2009) PAT proteins, an ancient family of lipid droplet proteins that regulate cellular lipid stores. Biochim. Biophys. Acta (BBA) – Mol. Cell Biol. L 1791, 419–440. [38] Brasaemle, D.L. (2007) Thematic review series: adipocyte Biology. The perilipin family of structural lipid droplet proteins: stabilization of lipid droplets and control of lipolysis. J. Lipid Res. 48, 2547–2559. [39] Schweiger, M. et al. (2008) The C-terminal region of human adipose triglyceride lipase affects enzyme activity and lipid droplet binding. J. Biol. Chem. 283, 17211–17220. [40] Soni, K.G., Mardones, G.A., Sougrat, R., Smirnova, E., Jackson, C.L. and Bonifacino, J.S. (2009) Coatomer-dependent protein delivery to lipid droplets. J. Cell Sci. 122, 1834–1841. [41] Barba, G. et al. (1997) Hepatitis C virus core protein shows a cytoplasmic localization and associates to cellular lipid storage droplets. Proc. Natl. Acad. Sci. USA 94, 1200–1205. [42] Miyanari, Y. et al. (2007) The lipid droplet is an important organelle for hepatitis C virus production. Nat. Cell Biol. 9, 1089–1097. [43] Boulant, S., Montserret, R., Hope, R.G., Ratinier, M., Targett-Adams, P., Lavergne, J.P., Penin, F. and McLauchlan, J. (2006) Structural determinants that target the hepatitis C virus core protein to lipid droplets. J. Biol. Chem. 281, 22236–22247. [44] Shavinskaya, A., Boulant, S., Penin, F., McLauchlan, J. and Bartenschlager, R. (2007) The lipid droplet binding domain of hepatitis C virus core protein is a major determinant for efficient virus assembly. J. Biol. Chem. 282, 37158– 37169. [45] Horiguchi, Y., Araki, M. and Motojima, K. (2008) 17[beta]-Hydroxysteroid dehydrogenase type 13 is a liver-specific lipid droplet-associated protein. Biochem. Biophys. Res. Commun. 370, 235–238. [46] Zehmer, J.K., Bartz, R., Liu, P. and Anderson, R.G.W. (2008) Identification of a novel N-terminal hydrophobic sequence that targets proteins to lipid droplets. J. Cell Sci. 121, 1852–1860. [47] Ingelmo-Torres, M. et al. (2009) Hydrophobic and basic domains target proteins to lipid droplets. Traffic 10, 1785–1801. [48] Pol, A., Luetterforst, R., Lindsay, M., Heino, S., Ikonen, E. and Parton, R.G. (2001) A caveolin dominant negative mutant associates with lipid bodies and induces intracellular cholesterol imbalance. J. Cell Biol. 152, 1057–1070. [49] Ostermeyer, A.G., Paci, J.M., Zeng, Y., Lublin, D.M., Munro, S. and Brown, D.A. (2001) Accumulation of caveolin in the endoplasmic reticulum redirects the protein to lipid storage droplets. J. Cell Biol. 152, 1071–1078. [50] Fujimoto, T., Kogo, H., Ishiguro, K., Tauchi, K. and Nomura, R. (2001) Caveolin-2 is targeted to lipid droplets, a new ‘‘membrane domain” in the cell. J. Cell Biol. 152, 1079–1086. [51] Pol, A., Martin, S., Fernandez, M.A., Ferguson, C., Carozzi, A., Luetterforst, R., Enrich, C. and Parton, R.G. (2004) Dynamic and regulated association of caveolin with lipid bodies: modulation of lipid body motility and function by a dominant negative mutant. Mol. Biol. Cell 15, 99–110. [52] Turro, S. et al. (2006) Identification and characterization of associated with lipid droplet protein 1: a novel membrane-associated protein that resides on hepatic lipid droplets. Traffic 7, 1254–1269. [53] Zehmer, J.K., Bartz, R., Bisel, B., Liu, P., Seemann, J. and Anderson, R.G.W. (2009) Targeting sequences of UBXD8 and AAM-B reveal that the ER has a direct role in the emergence and regression of lipid droplets. J. Cell Sci. 122, 3694– 3702. [54] Pol, A., Martin, S., Fernandez, M.A., Ingelmo-Torres, M., Ferguson, C., Enrich, C. and Parton, R.G. (2005) Cholesterol and fatty acids regulate dynamic caveolin trafficking through the Golgi complex and between the cell surface and lipid bodies. Mol. Biol. Cell 16, 2091–2105. [55] Cole, N.B., Smith, C.L., Sciaky, N., Terasaki, M., Edidin, M. and LippincottSchwartz, J. (1996) Diffusional mobility of Golgi proteins in membranes of living cells. Science 273, 797–801. [56] Hirschberg, K., Miller, C.M., Ellenberg, J., Presley, J.F., Siggia, E.D., Phair, R.D. and Lippincott-Schwartz, J. (1998) Kinetic analysis of secretory protein traffic

M. Digel et al. / FEBS Letters 584 (2010) 2168–2175

[57]

[58]

[59]

[60] [61]

[62]

[63]

[64]

[65] [66]

[67]

[68]

[69] [70] [71]

and characterization of Golgi to plasma membrane transport intermediates in living cells. J. Cell Biol. 143, 1485–1503. Dahm, T., White, J., Grill, S., Fullekrug, J. and Stelzer, E.H.K. (2001) Quantitative ER (left-right-arrow) Golgi transport kinetics and protein separation upon Golgi exit revealed by vesicular integral membrane protein 36 dynamics in live cells. Mol. Biol. Cell 12, 1481–1498. Skinner, J.R., Shew, T.M., Schwartz, D.M., Tzekov, A., Lepus, C.M., Abumrad, N.A. and Wolins, N.E. (2009) Diacylglycerol enrichment of endoplasmic reticulum or lipid droplets recruits perilipin 3/TIP47 during lipid storage and mobilization. J. Biol. Chem. 284, 30941–30948. Kuerschner, L., Moessinger, C. and Thiele, C. (2007) Imaging of lipid biosynthesis: how a neutral lipid enters lipid droplets. Traffic 9, 338– 352. Lingwood, D. and Simons, K. (2010) Lipid rafts as a membrane-organizing principle. Science 327, 46–50. Bostrom, P., Rutberg, M., Ericsson, J., Holmdahl, P., Andersson, L., Frohman, M.A., Boren, J. and Olofsson, S.-O. (2005) Cytosolic lipid droplets increase in size by microtubule-dependent complex formation. Arterioscler. Thromb. Vasc. Biol. 25, 1945–1951. Wolins, N.E., Quaynor, B.K., Skinner, J.R., Schoenfish, M.J., Tzekov, A. and Bickel, P.E. (2005) S3-12, Adipophilin, and TIP47 package lipid in adipocytes. J. Biol. Chem. 280, 19146–19155. Bostrom, P. et al. (2007) SNARE proteins mediate fusion between cytosolic lipid droplets and are implicated in insulin sensitivity. Nat. Cell Biol. 9, 1286– 1293. Szymanski, K.M. et al. (2007) The lipodystrophy protein seipin is found at endoplasmic reticulum lipid droplet junctions and is important for droplet morphology. Proc. Natl. Acad. Sci. USA 104, 20890–20895. Fei, W. et al. (2008) Fld1p, a functional homologue of human seipin, regulates the size of lipid droplets in yeast. J. Cell Biol. 180, 473–482. Cheng, J., Fujita, A., Ohsaki, Y., Suzuki, M., Shinohara, Y. and Fujimoto, T. (2009) Quantitative electron microscopy shows uniform incorporation of triglycerides into existing lipid droplets. Histochem. Cell Biol. 132, 281– 291. Valetti, C., Wetzel, D.M., Schrader, M., Hasbani, M.J., Gill, S.R., Kreis, T.E. and Schroer, T.A. (1999) Role of dynactin in endocytic traffic: effects of dynamitin overexpression and colocalization with CLIP-170. Mol. Biol. Cell 10, 4107– 4120. Nan, X., Potma, E.O. and Xie, X.S. (2006) Nonperturbative chemical imaging of organelle transport in living cells with coherent anti-stokes Raman scattering microscopy. Biophys. J. 91, 728–735. Xie, X.S., Yu, J. and Yang, W.Y. (2006) Living cells as test tubes. Science 312, 228–230. Huff, T.B. and Cheng, J.-X. (2007) In vivo coherent anti-Stokes Raman scattering imaging of sciatic nerve tissue. J. Microsc. 225, 175–182. Welte, M.A. (2009) Fat on the move: intracellular motion of lipid droplets. Biochem. Soc. Trans. 37, 991–996.

2175

[72] Fujimoto, Y. et al. (2007) Involvement of long chain acyl-CoA synthetase in local synthesis of neutral lipids in cytoplasmic lipid droplets in human hepatocyte HuH7. J. Lipid Res. 48, 1280–1292. [73] Stone, S.J., Levin, M.C., Zhou, P., Han, J., Walther, T.C. and Farese, R.V. (2009) The endoplasmic reticulum enzyme DGAT2 is found in mitochondriaassociated membranes and has a mitochondrial targeting signal that promotes its association with mitochondria. J. Biol. Chem. 284, 5352–5361. [74] Mashek, D.G., Li, L.O. and Coleman, R.A. (2007) Long-chain acyl-CoA synthetases and fatty acid channeling. Future Lipidol. 2, 465–476. [75] Digel, M., Ehehalt, R., Stremmel, W. and Fullekrug, J. (2009) Acyl-CoA synthetases: fatty acid uptake and metabolic channeling. Mol. Cell Biochem. 326, 23–28. [76] Milger, K. et al. (2006) Cellular uptake of fatty acids driven by the ER-localized acyl-CoA synthetase FATP4. J. Cell Sci. 119, 4678–4688. [77] Wolins, N.E., Brasaemle, D.L. and Bickel, P.E. (2006) A proposed model of fat packaging by exchangeable lipid droplet proteins. FEBS Lett. 580, 5484–5491. [78] Beller, M., Riedel, D., Jansch, L., Dieterich, G., Wehland, J., Jackle, H. and Kuhnlein, R.P. (2006) Characterization of the Drosophila lipid droplet subproteome. Mol. Cell Proteomics 5, 1082–1094. [79] Fukumoto, S. and Fujimoto, T. (2002) Deformation of lipid droplets in fixed samples. Histochem. Cell Biol. 118, 423–428. [80] Kuerschner, L., Ejsing, C.S., Ekroos, K., Shevchenko, A., Anderson, K.I. and Thiele, C. (2004) Polyene-lipids: a new tool to image lipids. Nat. Methods 2, 39–45. [81] Rinia, H.A., Burger, K.N.J., Bonn, M. and Müller, M. (2008) Quantitative labelfree imaging of lipid composition and packing of individual cellular lipid droplets using multiplex CARS microscopy. Biophys. J. 95, 4908–4914. [82] Bartz, R. et al. (2007) Lipidomics reveals that adiposomes store ether lipids and mediate phospholipid traffic. J. Lipid Res. 48, 837–847. [83] Robenek, H., Robenek, M.J. and Troyer, D. (2005) PAT family proteins pervade lipid droplet cores. J. Lipid Res. 46, 1331–1338. [84] Wan, H.C., Melo, R.C., Jin, Z., Dvorak, A.M. and Weller, P.F. (2007) Roles and origins of leukocyte lipid bodies: proteomic and ultrastructural studies. FASEB J. 21, 167–178. [85] Hein, B., Willig, K.I. and Hell, S.W. (2008) Stimulated emission depletion (STED) nanoscopy of a fluorescent protein-labeled organelle inside a living cell. Proc. Natl. Acad. Sci. USA 105, 14271–14276. [86] Takamori, S. et al. (2006) Molecular anatomy of a trafficking organelle. Cell 127, 831–846. [87] Murphy, S., Martin, S. and Parton, R.G. (2009) Lipid droplet-organelle interactions; sharing the fats. Biochim. Biophys. Acta (BBA) – Mol. Cell. Biol. L 1791, 441–447. [88] Nakamura, N. and Fujimoto, T. (2003) Adipose differentiation-related protein has two independent domains for targeting to lipid droplets. Biochem. Biophys. Res. Commun. 306, 333–338. [89] Wolins, N.E. et al. (2006) OXPAT/PAT-1 is a PPAR-induced lipid droplet protein that promotes fatty acid utilization. Diabetes 55, 3418–3428.