Mechanisms of lipid-body formation

Mechanisms of lipid-body formation

REVIEWS TIBS 24 – MARCH 1999 gemeinschaft (SFB 323 project B1, Schwerpunktsprogramm ‘Molekulare Analyse von Regulationsnetzwerken in Bakterien’ and G...

658KB Sizes 0 Downloads 5 Views

REVIEWS

TIBS 24 – MARCH 1999 gemeinschaft (SFB 323 project B1, Schwerpunktsprogramm ‘Molekulare Analyse von Regulationsnetzwerken in Bakterien’ and Graduiertenkolleg ‘Mikrobiologie’) and the Fonds der Chemischen Industrie for support, Wolfgang Köster, Athanasios Mademidis and Stefan Plantör for help in designing figures, and Uwe Stroeher and Karen A. Brune for critical reading of the manuscript.

References 1 Drechsel, H. and Winkelmann, G. (1997) in Transition Metals in Microbial Metabolism (Winkelmann, G. and Carrano, C. J., eds), pp. 1–49, Harwood Academic 2 Raymond, K. N., Müller, G. and Matzanke, B. (1984) Top. Curr. Chem. 123, 249–302 3 Braun, V., Hantke, K. and Köster, W. (1998) in Metal Ions in Biological Systems. Iron Transport and Storage in Microorganisms, Plants and Animals (Vol. 35) (Sigel, A. and Sigel, D., eds), pp. 67–145, Marcel Dekker

4 Gray-Owen, S. D. and Schryvers, A. B. (1996) Trends Microbiol. 4, 185–191 5 Genco, C. and Desai, P. J. (1996) Trends Microbiol. 4, 179–184 6 Letoffe, S., Redeker, V. and Wandersman, C. (1998) Mol. Microbiol. 179, 3572–3579 7 Schneider, R. and Hantke, K. (1993) Mol. Microbiol. 8, 111–121 8 Rosenberg, M. F., Callaghan, R., Ford, R. C. and Higgins, C. F. (1997) J. Biol. Chem. 272, 10685–10694 9 Killmann, H., Benz, R. and Braun, V. (1993) EMBO J. 12, 3007–3016 10 Bonhivers, M., Ghazi, A., Boulanger, P. and Letellier, L. (1996) EMBO J. 15, 1850–1856 11 Plancon, L., Chami, M. and Letellier, L. (1997) J. Biol. Chem. 272, 16868–16872 12 Letellier, L., Locher, K., Plancon, L. and Rosenbusch, J. P. (1997) J. Biol. Chem. 272, 1448–1445 13 Ferguson, A. et al. (1998) Science 282, 2215–2220 14 Locher, K. P., Rees, B., Koebnik, R., Mitschler, A., Moulinier, L., Rosenbusch, P. and Moras, D. (1998) Cell 95, 771–778 15 Rutz, J. M. et al. (1992) Science 258, 471–475

Mechanisms of lipid-body formation Denis J. Murphy and Jean Vance Most organisms transport or store neutral lipids as lipid bodies – lipid droplets that usually are bounded by specific proteins and (phospho)lipid. Neutral-lipid bodies vary considerably in their morphology and are associated with an extremely diverse range of proteins. However, the mechanisms by which they are generated in plants, animals and microorganisms appear to share many common features: lipid bodies probably arise from microdomains of the endoplasmic reticulum (or the plasma membrane in prokaryotes) that contain lipid-biosynthesis enzymes, and their synthesis and size appear to be controlled by specific protein components. LIPID-BODY FORMATION occurs at some point in the life cycle of nearly all organisms and is an integral part of energy storage and/or transport in most eukaryotes. Malfunctions in neutrallipid storage are implicated in several serious human diseases, such as fatty liver, obesity, atherosclerosis and type 2 D. J. Murphy is at the Dept of Brassica and Oilseeds Research, John Innes Centre, Norwich Research Park, Norwich, UK NR4 7UH; and J. Vance is at the Lipid and Lipoprotein Research Group and Dept of Medicine, 315 HMRC, University of Alberta, Edmonton, Alberta, Canada T6G 2S2. Email: [email protected]

diabetes. There is also considerable biotechnological interest in manipulation of lipid storage for both medical and agricultural purposes. Storage and transport lipids are contained in spheroidal droplets, which can range in diameter from 0.1 mm to 50 mm. The neutral lipids contained in these droplets are primarily triacylglycerol (TAG), diacylglycerol (in some insect tissues) and cholesteryl esters. Despite its evident importance, however, there have, until recently, been surprisingly few studies of the fundamental mechanism(s) of lipid-body biogenesis. During the past few years, this situation has begun to

0968 – 0004/98/$ – See front matter © 1998, Elsevier Science. All rights reserved.

16 Buchanan, S. et al. (1999) Nat. Struct. Biol. 6, 56–63 17 Braun, V. (1995) FEMS Microbiol. Rev. 16, 295–307 18 Postle, K. (1993) J. Bioenerg. Biomembr. 25, 591–601 19 Moeck, G. S. and Coulton, J. W. (1998) Mol. Microbiol 28, 675–681 20 Kadner, R. J. (1990) Mol. Microbiol. 4, 2027–2033 21 Jiang, X. et al. (1997) Science 276, 1261–1264 22 Cornelissen, C. A., Anderson, J. E. and Sparling, P. F. (1997) Mol. Microbiol. 26, 25–35 23 Köster, W. (1991) Biol. Metals 4, 23–32 24 Mietzner, T. A. et al. (1998) Curr. Top. Microbiol. Immun. 225, 113–135 25 Boos, W. and Lucht, J. M. (1996) in Escherichia coli and Salmonella typhimurium (Neidhardt, F.C., ed.), pp. 1175–1209, ASM Press 26 Mademidis, A. et al. (1997) Mol. Microbiol. 26, 1109–1123 27 Groeger, W. and Köster, W. (1998) Microbiology 144, 2759–2769 28 Bruns, C. M. et al. (1998) Nat. Struct. Biol. 4, 919–924 29 Hantke, K. (1997) J. Bacteriol. 179,

change, and several common themes are now emerging. Here, we compare recent findings from a range of cell types from plants, animals and microorganisms (see Table 1).

Plants Seeds and fruits. Intracellular storagelipid bodies in plants are particularly abundant in oil-rich fruit and seed tissues, which can contain as much as 50–75% (w/w) lipid. As in animals, such cytosolic lipid bodies are believed to arise from specific microdomains of the endoplasmic reticulum (ER) membrane that contain the full complement of TAGbiosynthesis enzymes1 (Fig. 1). Data from several labeling studies suggest that these enzymes channel intermediates towards lipid-body formation and, hence, segregate them from the bulk lipid-bilayer components2. Lipid bodies from all desiccationtolerant seeds analysed to date are bounded by a continuous surface layer of unique amphipathic proteins, which are termed oleosins2. Recent evidence suggests that oleosins are cotranslationally inserted into the ER membrane and have an unusual topology: both termini are directed towards the cytoplasm. Site-directed mutagenesis studies have shown that targeting of oleosin to lipid bodies is regulated by the protein’s characteristic hydrophobic central domain and, in particular, by a tripleproline knot motif3. Recent studies on the ectopic expression of oleosin in

PII: S0968-0004(98)01349-8

109

REVIEWS

TIBS 24 – MARCH 1999

Table 1. Lipid-body formation in different cell types Tissue/cell type

Site of formation

Targeting

Major associated Refs protein(s)

Plants Seeds and fruits Anther – pollen Anther – tapetum Plastids

ER ER ER Envelope membrane

Cytosol Cytosol Release via cell lysis Stroma

Oleosina None Oleosin-like Fibrillin/PLP

2, 3 5 5, 6 10, 11

ER

Secretion via ER–Golgi Cytosol Cytosol

Apolipoprotein

12

Perilipin/ADRP Perilipin/ADRP

21, 22 22

Secretion via exocytosis Cytosol

(i) ADRP, (ii) butyrophilinb ADRP?

28 29 27

Cytosol Cytosol

? Phasin

9, 34, 35 36, 38

Cytosol

?

39

Animals Liver – hepatocytes, intestine – enterocytes Adipose – adipocytes Adrenal, testis, ovary (steroidogenic) Mammary – epithelium Others – e.g. leukocytes

ER ER ER? ER?

Microorganisms Yeast ER PolyhydroxyalkanoatePlasma membrane accumulating prokaryotes Streptomyces spp Plasma membrane

aOnly in desiccating seeds. bAssociated with the enfolding bilayer membrane rather than the lipid body itself. Abbreviations used: ADRP, adipocyte-differentiation-related protein; ER, endoplasmic reticulum; PLP, plastid lipid-associated protein.

transgenic plants indicate that, although the protein is synthesized ubiquitously, it is rapidly degraded in all non-seed tissues except anthers (D. J. Murphy and M. M. Moloney, unpublished). Because anthers contain lipid-body-synthesizing tapetal and pollen cells, oleosin accumulation in plants, like that of the lipid-body proteins in animals [adipose-differentiation-related protein (ADRP) and perilipin] appears to depend on concurrent formation of lipid bodies. Oleosins are unlikely to play a major role in lipid-body biogenesis per se, because they are absent from lipid-rich tissues of fruits and many tropical oilseeds that do not undergo desiccation as a normal part of development. However, such tissues contain much larger (5–50 mm) lipid bodies than do seeds of desiccation-tolerant species (which have a diameter of 0.5–3 mm); this suggests that oleosins regulate the size of lipid bodies and that they also play an important role in enabling seeds to withstand dehydration2. In fact, lipid bodies in seeds of oleosin-deficient plants are stable to drying but undergo a catastrophic and fatal coalescence and phase inversion upon rehydration, which does not occur in oleosin-rich species4. These data indicate that oleosins protect lipid bodies from coalescence during the extreme water stress caused by sudden hydration from an almost dry state. Other tissues. Pollen grains can contain 20–30% (w/w) storage lipid5. The lipid bodies of pollen grains contain no

110

detectable oleosin, although pollen grains, like oleosin-rich seeds, undergo dehydration and rehydration during maturation and germination. Instead, pollen lipid bodies are often encircled by ER cisternae, which act as a physical barrier to coalescence5. A similar scenario is evident in numerous animal tissues (ranging from heart to ovary) that possess ER-bound cytosolic lipid bodies, and might therefore represent a common mechanism for packaging lipid bodies in cells. Another major lipid-producing cell type in many plants is the tapetum, which surrounds and nourishes developing microspores and pollen grains within the anther. Tapetal cytosolic lipid bodies appear to be unique: they are made up of a complex network of fibres, vesicles and TAG-rich droplets. Most of their associated proteins contain 7–8-kDa Nterminal hydrophobic domains, which are very similar to a domain present in seed oleosins, and hypervariable 7–40kDa C-terminal domains5,6. Following tapetal apoptosis, the lipid bodies are released into the anther lumen and relocate to form an extracellular lipidic coating around the maturing pollen grains. This process is associated with an endoproteolytic cleavage that removes the oleosin-like hydrophobic domain; the variable C-terminal domain remains and is the major protein component of the pollen coat6. One can draw an analogy between this plant lipid-associated protein and the vitellogenin class of serum

lipoproteins that is specific to laying hens: following its uptake into oocytes, vitellogenin II is cleaved specifically by the protease cathepsin D to yield four polypeptides, which have different locations and functions in egg development7. Recently, Thompson and co-workers8 postulated that lipid bodies play a key role in remobilization of membrane lipids during senescence in some, and possibly all, plant tissues. Small lipid bodies might also be involved in various aspects of intracellular lipid metabolism and trafficking in animals – as Leber et al.9 have proposed for yeast. Plastids. Plastids are organelles that are unique to plants and almost certainly are derived from prokaryotic endosymbionts. Several plastid types, such as chromoplasts and elaioplasts, accumulate neutral-lipid bodies that contain TAG, carotenoids and/or sterol esters. In chloroplasts, formation of TAG-rich lipid bodies can be induced by oxidative stress or during senescence. Recently, Pozueto-Romero and co-workers10 identified a class of ~32-kDa plastid proteins, termed fibrillin or plastoglobule lipidassociated protein (PLP), that are present in the lipid bodies of a wide range of plant species. Plastid neutral lipids probably arise from the inner-envelope membrane, which is equivalent to the plasma membrane of prokaryotes. The ratio of PLP to neutral lipid appears to determine the shape of the lipid bodies, which can range from large spherical globules to rod-shaped fibrils11.

Animals Hepatocytes and enterocytes. In mammalian liver and intestine, TAG and cholesteryl esters can be stored in the cytosol as lipid droplets or, instead, secreted into the circulation as very low-density lipoproteins (VLDLs) from the liver or as chylomicrons from the intestine (Fig. 2). These plasma lipoproteins deliver waterinsoluble lipids from the liver and intestine to other tissues in the body for use as an energy source. Plasma lipoproteins all have a common spherical structure in which a surface monolayer that consists of specific apolipoproteins, unesterified cholesterol and phospholipids surrounds a neutral-lipid core. The hepatic assembly of the lipid components of VLDLs and apolipoprotein B, the principal apolipoproteins component of nascent VLDLs, occurs in the ER (Ref. 12). Subsequently, VLDLs are secreted into the circulation via the Golgi. Secretion of apolipoprotein B and lipids is obligatorily linked. Under most

REVIEWS

TIBS 24 – MARCH 1999 physiological conditions, cells synthesize apolipoprotein B constitutively at a rate in excess of that at which VLDLs are secreted; this suggests that VLDL secretion is not regulated by the rate of apolipoprotein B synthesis. Indeed, most data available indicate that the supply of lipids such as TAG, phospholipids and cholesteryl esters probably determines how much of the apolipoprotein B made moves across the ER membrane and into the ER lumen and, therefore, how much VLDL is secreted. Excess apolipoprotein B that is not translocated/secreted is degraded by proteasomes in the cytosol and possibly by other proteases13,14. The phospholipids and cholesterol used for assembly of the VLDL surface monolayer are synthesized on ER membranes. However, recent data suggest that .70% of the TAG in VLDL is not synthesized de novo but produced from pre-existing cytosolic TAG droplets through a lipolysis–re-esterification cycle15. The process by which the neutral lipids of VLDLs (i.e. TAG and, to a lesser extent, cholesteryl esters) are concentrated into a core structure is not well understood. The gene that encodes the ER-luminal protein microsomal TAGtransfer protein is defective in individuals suffering from abetalipoproteinemia (from whose plasma apolipoprotein-B-containing lipoproteins are absent). This discovery led to the demonstration that this transfer protein is required for secretion of all apolipoprotein-B-containing lipoproteins (i.e. VLDLs and chylomicrons)16. The microsomal TAG-transfer protein can transfer lipids, particularly TAG, between membranes in an in vitro assay and, consequently, might mediate transfer of TAG to apolipoprotein B during VLDL assembly. However, the question of whether this protein is involved in cotranslational/cotranslocational assembly of apolipoprotein B and TAG (see Fig. 2), or in the fusion of a luminal TAG droplet with a small, lipid-poor apoprotein-B particle in the ER lumen, is still being debated. Other suggestions are that the microsomal TAG-transfer protein acts as a chaperone and/or participates directly in translocation of apolipoprotein B across the ER membrane16–18. An important, but unanswered, question remains: what determines whether TAG is directed to the ER lumen for assembly into VLDL and secretion into the circulation, or into lipid-storage droplets in the cytosol? Cytosolic neutral-lipid droplets also consist of a core of TAG and/or cholesteryl esters that is surrounded by a surface monolayer of

TAG Oil-body oleosins

ER-bound oleosins

Figure 1 Lipid-body formation in seeds. In most seeds, storage-lipid bodies are produced by budding from the ER in association with oleosins. Nascent lipid bodies will fuse until surrounded by a continuous layer of oleosins, which prevents further fusion. Some tropical and subtropical oilseeds that do not normally undergo dehydration do not produce oleosins around their lipid bodies, which has led to suggestions that oleosins stabilize lipid bodies during the drastic changes in water potential that occur in imbibing seeds. TAG, triacylglycerol.

phospholipids and proteins (but not apolipoprotein B). Formation of cytosolic lipid droplets might take place in a specific domain of the ER; another domain could be used for lipoprotein assembly and targeting to the lumen. A region of the ER termed the mitochondria-associated membranes contains severalfold-elevated levels (in comparison with the bulk of ER) of enzyme activities that participate in neutral-lipid biosynthesis (i.e. diacylglycerol acyltransferase and acyl-CoA: cholesterol acyltransferase)19, and might be involved in one of these processes. Adipocytes. Adipose tissue is the major long-term energy store in mammals. A typical adipocyte from white adipose tissue contains one or a few large (10–100mm diameter) cytoplasmic triacylglycerolrich lipid bodies. Some hibernating animals and the foetuses/neonates of many mammals also accumulate brown adipose tissue, which consists of adipocytes that contain many smaller (2–10mm diameter) storage-lipid bodies and numerous mitochondria. Brown adipose tissue is responsible for thermogenesis through storage-body lipolysis and uncoupled respiration of the products. The early phases of lipid-body biogenesis probably proceed by a similar mechanism in white- and brownadipose-tissue adipocytes. Early in adipocyte differentiation, nascent lipid bodies appear to arise from, and sometimes appear to be enfolded by, ER membranes (Fig. 3). Such lipid bodies contain an ~50-kDa surface-bound acylated protein – adipose differentiationrelated protein (ADRP) – that originally was isolated as a strongly induced

marker of adipocyte differentiation20. However, Brasaemle and co-workers21 have now shown that ADRP is present on the surface of the tiny cytoplasmic lipid droplets found in undifferentiated adipocytes and that it is also widely distributed, probably ubiquitously, in all animal lipid-accumulating cell types. During the further differentiation of adipocytes, perilipins (a class of 42–56kDa polypeptides) replace ADRP as the predominant protein on the surfaces of lipid bodies. The N-terminal 105 residues of the perilipins are similar to ADRP, although perilipins are found only in adipocytes and steroidogenic cells22. As in the case of ADRP, perilipin expression is concurrent with lipid-body formation. If perilipins adopt a globular conformation, they should cover at least 15–20% of the lipid-body surfaces (C. Londos, pers. commun.), whereas if they assume a more extended linear conformation similar to that of the apolipoproteins, they should cover most of the droplet surface. Perilipins appear to have several functions. First, they play a role in lipid-body biogenesis: undifferentiated wild-type 3T3-L1 fibroblasts accumulate minimal TAG and no perilipins, but those that constitutively express a transgene that encodes perilipin A accumulate numerous small TAG droplets bounded by perilipin22. Second, they stabilize lipid bodies: the perilipin surface layer around mature lipid bodies appears to shield TAG from premature lipolysis. Third, they mobilize TAG after hormonal activation: adrenalin-activated phosphorylation of perilipin leads to the latter’s

111

REVIEWS

TIBS 24 – MARCH 1999

Ribosome mRNA

TAG lipolysis and re-esterification

Cytosol

TAG ? N

TAG N TAG

?

MTP

Amphipathic lipids Apolipoprotein B

(a)

(b)

TAG N MTP

TAG

?

MTP TAG

Lumen

(c)

Apolipoprotein B

(d)

(e)

ER

Figure 2 A cotranslational model for hepatic assembly of very low-density lipoproteins (VLDLs). (a) Apolipoprotein B translation initiates on ribosomes attached to the ER membrane. The protein associates with some amphipathic lipids (phospholipids and cholesterol) as it is cotranslationally translocated across the membrane. (b) Cytosolic triacylglycerol (TAG) droplets are hydrolysed and TAG is re-synthesized on the ER for assembly with apolipoprotein B. (c) Translation/translocation continues. (d) As more TAG, phospholipids and cholesterol are added to the particle, its size increases. (e) A fully formed VLDL particle is released into the ER lumen for secretion into the circulatory system via the Golgi. The microsomal TAG-transfer protein (MTP) might be involved in translocation of apolipoprotein B, assembly of phospholipids with apolipoprotein B, and/or transfer of TAG to the core of the nascent VLDL particle.

redistribution on the lipid-body surface. This disrupts the perilipin shell and allows hormone-sensitive lipase, which is also phosphorylated in response to adrenalin, access to the lipid to initiate TAG breakdown22. Perilipins are unlikely to play a role in lipid-body fusion, given that their expression in lipogenic cells that normally express only ADRP does not alter lipid-body size (C. Londos, pers. commun.). Adipocyte differentiation in white adipose tissue involves the coalescence of small nascent lipid bodies to form the one or more large TAG droplets found in mature cells. By contrast, most other cell types contain numerous smaller lipid bodies. Franke and co-workers23 have proposed that, in adipocytes, cytoskeletal intermediate-filament proteins, particularly vimentin, form cages around small lipid bodies and thereby prevent the latter from fusing. Indeed, a vimentin network is seen in the vicinity of lipid bodies in newly differentiating adipocytes24. As the cells mature, a decrease in the number of vimentin fibres coincides with, and might play a role in, lipid-body coalescence. This hypothesis is appealing; however, transgenic mice that lack vimentin exhibit completely normal adipocyte lipid-body biogenesis, even though no compensatory formation of other intermediate-filament proteins could be demonstrated25. Nevertheless, coalescence of adipocyte lipid

112

bodies must be highly regulated – witness the thousandfold difference in the volume of lipid bodies in white and brown adipose tissue – but the mechanism of adipocyte lipid-body maturation remains to be elucidated. Other animal cell types. Most, if not all, animal tissues contain cells that accumulate lipids for short-term storage. The distribution of ADRP reflects this: ADRP is expressed in all tissues examined in mammals and is expressed at particularly high levels in lung, brain, testis and mammary glands21. By contrast with ADRP, perilipin is found only in adipocytes and in steroidogenic cells. The latter, which are found in the adrenal cortex, ovarian follicles and testis, are the primary sites of steroid-hormone biosynthesis. They accumulate numerous small lipid bodies in which cholesteryl esters are surrouned by several isoforms of perilipin, including one that might be specific for this cell type22. As in adipocytes, lipid-body formation and lipid mobilization in steroidogenic cells are probably regulated, at least in part, by perilipin. The reason for the occurrence of perilipin both in adipocytes and in steroidogenic cells probably lies in a common mechanism of lipid-body mobilization. In both cases, external hormonal stimulation elevates intracellular cyclic AMP levels and, thereby, activates protein kinase A. In steroidogenic cells, a cholesteryl

esterase that is probably identical to the hormone-sensitive lipase of adipocytes26 binds to the lipid-body surface after disruption of the perilipin shell by phosphorylation – as Londos et al.22 have proposed for adipocytes – and hydrolyses the cholesteryl esters stored within. The released cholesterol is then transported to mitochondria for further metabolism to steroid hormones. In many cases, stress can induce lipid storage; this is particularly evident in leukocytes from joints of patients who suffer from inflammatory arthritis, the airways of patients who suffer from acute respiratory-distress syndrome, and caesin-elicited guinea-pig peritoneal exudates27. Leukocyte lipid-storage bodies are inducible within one hour and accumulate eicosanoids that are part of the generalized inflammatory response. It would be interesting to determine whether ADRP or another lipid-binding protein is co-induced by such stresses. Milk lipid globules. Lactating mammals produce milk that contains variable amounts of lipid: the lipid content ranges from only 0.2% (w/w), for the black rhinoceros, to as much as 61% (w/w), for the hooded seal28. Milk lipids normally consist of TAG droplets enfolded by a protein-rich lipid bilayer, which is termed the milk-lipid-globule membrane (Fig. 4). Specialized epithelial cells that line the alveoli of mammary

TIBS 24 – MARCH 1999

REVIEWS

glands secrete the soluble components (e.g. sugars and Catecholamines some proteins) and lipidic Plasma membrane components of milk. Recent studies indicate that milk lipid HSL globules arise from cytosolic cAMP PKA lipid bodies that undergo protein-mediated exocytosis29. Milk TAGs are produced in specific regions of the ER P-HSL membrane, which become ADRP P-Perilipin distended and bud off into Perilipin small (0.15–0.50-mm diameter) Perilipin 30 Perilipin nascent lipid bodies . Enzymes involved in TAG biosynthesis, such as acyltransferases, are concentrated in a particular membrane subfraction, which ER Nascent might be an ER domain that lipid specializes in lipid-body forbodies mation. This domain might be Mature lipid body similar to the TAG-enriched ER domains found in hepatocytes19, Leydig cells21 and oilFigure 3 seed embryos1. A ,10-kDa Lipid-body assembly in adipocytes. The figure shows a possible mechanism for the formation and mocytosolic factor stimulates rebilization of the large storage-lipid bodies of white adipose tissue. Similar mechanisms probably exist lease of nascent microlipid in brown adipose tissue and steroidogenic cells. Small lipid bodies produced from the ER in undifferentiated cells and at early stages of adipocyte differentiation tend to be associated with adiposedroplets from the ER memdifferentiation-related protein (ADRP). At later stages, perilipin is the major lipid-body protein. The lipid brane in vitro31. Fusion of these bodies undergo a regulated series of fusions until they achieve a mature size that is determined by the nascent microlipid droplets cell type. Perilipin-coated lipid bodies break down only when triggered by hormonally induced phosproduces the 1–4-mm mature phorylation, by protein kinase A (PKA), of perilipin and cytosolic hormone-sensitive lipase (HSL). cytosolic lipid bodies that acPhosphorylated perilipin probably rearranges on the lipid-body surface to allow HSL to access, and count for .90% (v/v) of bovine consequently to hydrolyse, the lipid core. milk lipid. Such fusions can be reconstituted in vitro after the addition of calcium and cytosolic proteins membrane. The third major milk-lipid- Microorganisms or exogenous gangliosides32. Although ma- globule protein is xanthine oxidase, which Yeast. Most yeasts produce small numture cytosolic lipid bodies of many sizes is a 155-kDa soluble protein that lacks se- bers of cytosolic lipid bodies, but the can be found, each mammalian species cretory signals. Xanthine oxidase is pro- oleagenous yeasts can accumulate up to tends to accumulate milk lipids in drop- posed to bind to the cytosolic domain of 25% (w/w) storage lipid in response to a lets of a narrow size range; this suggests butyrophilin on the plasma membrane. high carbon : nitrogen ratio34. Lipid bodies that lipid-body fusion is tightly regulated. Butyrophilin and xanthine oxidase to- in Saccharomyces cerevisiae contain Milk lipid globules from a range of mam- gether act as a receptor for the binding of almost equal amounts of TAGs and sterol malian species contain three particularly cytosolic lipid bodies, probably through esters35. As in other eukaryotes, yeast abundant proteins: ADRP, butyrophilin interactions with ADRP (Ref. 29). The lipid bodies probably arise from the ER. and xanthine oxidase29 (Fig. 4). Of these, binding of ADRP on the lipid-body surface However, several enzymes involved in only ADRP is found on cytosolic lipid bod- to the butyrophilin–xanthine-oxidase re- yeast lipid metabolism, including glycies in secretory epithelial cells and, as in ceptors on the plasma membrane might erophosphate acyltransferases, sterol other lipogenic cell types, it is probably in- act like a zip: the lipid body could be sur- D24-methyltransferase and squalene epoxivolved in lipid-body formation28. By con- rounded by an evaginated plasma mem- dase, localize (at least partially) to lipid trast with that in adipocytes and steroido- brane bilayer, which pinches off and bodies9,35. Changes in the ratio of the genic cells, ADRP in the secretory releases a milk lipid globule into the alve- binding of these enzymes to the ER and epithelium persists as the lipid bodies un- olar lumen. The globule thus contains a their binding to lipid bodies during yeast dergo fusion in the cytosol until they TAG droplet that is surrounded by a sur- development suggest that there is a dyreach their mature size; it is probably the face layer of ADRP, which is itself encir- namic interaction, and possibly physical major lipid-body protein component im- cled by the plasma membrane remnant contact, between the two compartments9. mediately prior to secretion28. Butyro- and its associated butyrophilin–xanthine- These studies imply that, in yeast, lipid philin is a 66-kDa glycoprotein that con- oxidase complex. Heid and co-workers28 bodies do not serve simply as inert lipid tains two approximately equally sized have isolated the three proteins from de- stores but play an important role in the globular domains, which are separated by tergent-solubilized milk lipid globules as biosynthesis, mobilization and trafficka single membrane-spanning region33. The a stable complex that has a molar ratio of ing of intracellular neutral lipids9. protein is cotranslationally inserted into 5 butyrophilin : 3 ADRP : 1 xanthine oxiProkaryotes. Very few prokaryotes acthe ER membrane; it has a type I topology dase; the proteins must therefore form a cumulate neutral lipids as energy stores. and is probably targeted to the plasma very strong association. They tend instead to sequester glycogen,

113

REVIEWS

TIBS 24 – MARCH 1999

polyphosphates and polyhydroxyalkanoates. Polyhydroxyalkanoates are present in over 90 genera of prokaryotes and, as in the case of neutral lipids, are stored as insoluble cytosolic inclusions36. The polyhydroxyalkanoate globules are surrounded by amphipathic surface proteins termed phasins, which exhibit molecular weights of 14–25 kDa37,38. These proteins, as in the cases of ADRP and perilipin in animals and oleosin in plants, might play a role in globule formation, maturation and/or mobilization. The few prokaryotes that accumulate TAG belong to Gram-positive genera such as Streptomyces or Rhodococcus. Indeed, in S. coelicolor, TAG can constitute .80% of the total cell volume during the stationary phase39. These lipid bodies probably arise from the plasma membrane but it is not known whether they contain specific surface proteins.

Conclusions and future directions The major, and perhaps unique, site of lipid-body assembly (except in prokaryotes and plastids) is in specialized regions of the ER, where biosynthetic enzymes might be grouped functionally

(e.g. as metabolons). Most lipid bodies accumulate with specific surface-bound proteins that are also found on the ER. The tiny nascent lipid bodies normally undergo a series of highly regulated fusions and can alter their surface proteins before reaching their mature size and composition. Targeting of lipid bodies to the cytosol might be the default pathway, whereas targeting to the ER lumen requires cotranslation of specific proteins, such as apolipoprotein B, and possibly the participation of chaperones. Several unanswered questions remain. (1) How do nascent lipid bodies bud off from membranes, and are proteins involved? (2) Can lipid bodies re-fuse with membranes? (3) How do lipid bodies fuse with each other? (4) How do some lipid bodies enlarge without fusing, and is a lipid-transfer protein involved? (5) How is the mature size of lipid bodies determined? (6) How are lipid bodies mobilized? Are specific lipases involved, such as the recently characterized hepatic TAG

hydrolase (R. Lehner and D. E. Vance, pers. commun.)? (7) What determines whether hepatic TAG is directed to the ER lumen for lipoprotein secretion or to the cytosol for storage as lipid droplets? (8) Is neutral-lipid biosynthesis coordinated with the synthesis of specific lipid-body-associated proteins? Although these are formidable questions, the increasing sophistication of molecular and biochemical techniques, and a greater awareness of comparative studies in different tissues and organisms, makes us optimistic that many of these questions will soon be answered. For example, the recent isolation both of mammalian diacylglycerol acyltransferase40 and a higher-plant diacylglycerol acyltransferase (M. J. Hills, pers. commun.) and various acyl-CoA:cholesterol acyltransferases41 provides important new tools for the characterization of the mechanisms of lipid-body biogenesis in living cells.

Acknowledgements We thank C. Londos, J. E. Thompson and T. W. Keenan for their advice, and

Butyrophilin

Ca2+

Butyrophilin

Microlipid body

Xanthine oxidase

Milk lipid globule

ADRP Cytoplasmic lipid body

Secretory epithelial cell

Microvilli

Alveolar lumen

Figure 4 One possible mechanism for the formation and secretion of milk lipid globules. See elsewhere29 for alternative models. Microlipid bodies arise from the ER in association with adipose-differentiation-related protein (ADRP). Although some microlipid bodies might be secreted directly, the vast majority undergo a series of fusions, which might be regulated by Ca21 and other cytosolic factors, until they reach their mature size range. Mature cytosolic lipid bodies then dock with regions of the plasma membrane (or possibly with cytosolic secretory vesicles) through interactions between ADRP and a butyrophilin–xanthine-oxidase receptor complex28. This results in evagination of the lipid body and its associated membrane bilayer into the alveolar lumen, which is followed by their release to form a milk lipid globule.

114

REVIEWS

TIBS 24 – MARCH 1999 C. Londos and M. J. Hills for the provision of unpublished data. We thank D. E. Vance, M. J. Hills and Z. Poghosyan for critically reading the manuscript, T. Bown for the graphics, and the Heart and Stroke Foundation of Alberta and BBSRC for funding (to J. E. V. and D. J. M., respectively).

References 1 Lacey, D. J. and Hills, M. J. (1996) Planta 199, 545–551 2 Murphy, D. J. (1993) Prog. Lipid. Res. 32, 247–280 3 Abell, B. M. et al. (1997) Plant Cell 9, 1481–1493 4 Leprince, O., van Aelst, A. C., Pritchard, H. W. and Murphy, D. J. (1998) Planta 204, 109–119 5 Piffanelli, P., Ross, J. H. E. and Murphy, D. J. (1998) Sex. Plant Reprod. 11, 65–80 6 Murphy, D. J. and Ross, J. H. E. (1998) Plant J. 13, 1–16 7 Yamamura, J. et al. (1995) Biochim. Biophys. Acta 1244, 384–394 8 Thompson, J. E., Froese, C. D. and Madey, E. (1998) Prog. Lipid Res. 37, 119–141 9 Leber, R. et al. (1998) Mol. Biol. Cell 9, 375–386 10 Pozueta-Romero, J. et al. (1997) Plant Physiol. 115, 1185–1194 11 Chen, H-C. et al. (1998) Plant J. 14, 317–326 12 Davis, R. A. and Vance, J. E. (1996) in Biochemistry of Lipids, Lipoproteins and Membranes (Vance, D. E. and Vance, J. E., eds), pp. 473–493, New Comprehensive Biochemistry, Elsevier 13 Yeung, S. J., Chen, S. H. and Chan, L. (1996) Biochemistry 35, 13843–13848 14 Yao, Z., Tran, K. and McLeod, R. S. (1997) J. Lipid Res. 38, 1937–1953 15 Wiggins, D. and Gibbons, G. F. (1996) Biochem. J. 320, 673–679 16 Gordon, D. A., Wetterau, J. R. and Gregg, R. E. (1995) Trends Cell Biol. 5, 317–321 17 Du, E. Z. et al. (1996) J. Lipid Res. 27, 1309–1315 18 Pease, R. J. and Leiper J. M. (1996) Curr. Opin. Lipidology 7, 132–138 19 Vance, J. E. (1990) J. Biol. Chem. 265, 7248–7256 20 Jiang, H-P. and Serrero, G. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 7856–7860 21 Brasaemle, D. L. et al. (1997) J. Lipid Res. 38, 2249–2263 22 Londos, C. et al. (1995) Biochem. Soc. Trans. 23, 611–615 23 Franke, W. W., Hergt, M. and Grund, C. (1987) Cell 49, 131–141 24 Blanchette-Mackie, E. J. et al. (1995) J. Lipid Res. 36, 1211–1226 25 Guyon-Colucci, E. et al. (1994) Cell 79, 679–694 26 Yeaman, S. J. et al. (1994) Adv. Enzyme Regul. 34, 355–370 27 Bozza, P. T. et al. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 11091–11096

28 Heid, H. W., Schnolzer, M. and Keenan, T. W. (1996) Biochem. J. 320, 1025–1030 29 Mather, I. H. and Keenan, T. W. (1998) J. Mammary Gland Biol. Neoplasia 3, 259–273 30 Zaczek, M. and Keenan, T. W. (1990) Protoplasma 159, 179–182 31 Keenan, T. W., Dylewski, D. P., Ghosal, D. and Keon, B. H. (1992) Eur. J. Cell Biol. 57, 21–29 32 Valivullah, H. M., Bevan, D. R., Peat, A. and Keenan, T. W. (1988) Proc. Natl. Acad. Sci. U. S. A. 85, 8775–8779 33 Banghart, L. R. et al. (1998) J. Biol. Chem. 273, 4171–4179 34 Holdsworth, J. E. and Ratledge, C. (1991) Lipids 26, 111–118 35 Leber, R. et al. (1994) Yeast 10, 1421–1428 36 Poirier, Y., Nawrath, C. and Somerville, C. (1995) Biotechnology 13, 142–150 37 Pieper-Furst, U., Madkour, M. H., Mayer, F. and Steinbuchel, A. (1994) J. Bacteriol. 176, 4328–4337 38 Steinbuchel, A. et al. (1995) Can. J. Microbiol. 41, 94–105 39 Packter, N. M. and Olukoshi, E. R. (1995) Arch. Microbiol. 164, 420–427 40 Cases, S. et al. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 13018–13024 41 Chang, C. Y. et al. (1995) J. Biol. Chem. 270, 29532–29540

Students Did you know that you are entitled to a 50% discount on a subscription to TiBS? See the bound-in subscription order card for details Pete Jeffs is a freelancer working in Paris, France.

115