This article was published in the above mentioned Springer issue.The material, including all portions thereof, is protected by copyright;all rights are held exclusively by Springer Science + Business Media.

The material is for personal use only;commercial use is not permitted.

Unauthorized reproduction, transfer and/or usemay be a violation of criminal as well as civil law.

ISSN 1138-7548, Volume 66, Number 1

ORIGINAL PAPER

Association of SND1 protein to low density lipid dropletsin liver steatosis

I. Garcia-Arcos & Y. Rueda & P. González-Kother & L. Palacios & B. Ochoa & O. Fresnedo

Received: 2 July 2009 /Accepted: 3 November 2009 /Published online: 23 April 2010# University of Navarra 2010

Abstract Although the human homologue of SNDp102, p100 coactivator, was initially described as anuclear protein, the p100 coactivator protein familymembers have non-nuclear localization in mammaliancells with active lipid handling, storage, and secretion.However, their role in lipid homeostasis remainsunresolved. Here, we investigate the distribution ofthe rat homologue SND p102 (also called SND1) andits association with newly formed lipid droplets in theliver parenchyma and cultured hepatocytes. Sucrosegradient fractionation showed that SND p102 cofrac-tionated with endoplasmic reticulum and Golgimarkers. Such cofractionation was not altered inregenerating steatotic rat liver. However, SND p102was also detected in lipid droplets from regeneratingliver, showing a specific directionalization to the leastdense ones. Confocal microscopy of cultured hepato-cytes confirmed the findings of gradient fractionation.

In addition, p100 coactivator was consistently en-countered in microsomes and lipid droplets in controland oleate-treated HepG2 cells. The total amount ofSND p102 in hepatocytes was similar in bothconditions, suggesting a specific translocation of theprotein. Our findings indicate that SND p102 and thehuman p100 coactivator have a ubiquitous cytoplas-mic distribution in hepatocytes and that steatogenicconditions promote the targeting of SND p102 fromother cell compartments to specific low density lipiddroplets.

Keywords SND p102 . SND1 . Lipid droplets .

Hepatocytes . Hepatic steatosis

Introduction

Lipid droplets (LD) are lipid storage structures foundin most eukaryotic cells. LD play dynamic roles in theturnover of neutral lipids, including triacylglycerols,cholesteryl esters, and retinyl esters [18], and also invesicular transport and lipid trafficking events [36]. Inaddition to enzymes of lipid metabolism and proteinslike small GTPases or caveolin involved in cellularsignaling and traffic, mammalian LD usually contain atleast one member of the perilipin–adipophilin–TIP47proteins [19]. These proteins seem to be incorporatedinto the LD coat during its biogenesis or in responseto specific signals from either the cytosol or cellmembrane systems as the endoplasmic reticulum (ER).

J Physiol Biochem (2010) 66:73–83DOI 10.1007/s13105-010-0011-0

I. Garcia-Arcos :Y. Rueda : P. González-Kother :L. Palacios :B. Ochoa (*) :O. FresnedoDepartment of Physiology, Faculty ofMedicine andDentistry,University of the Basque Country, Leioa, Spaine-mail: begona.ochoa@ehu.es

P. González-KotherScience Faculty, Sma. Concepción Catholic University,Caupolicán, Chile

L. PalaciosProgenika Biopharma S.A., Bizkaia Technologic Park,Derio, Spain

Author's personal copy

SND1 gene product was first described by Tonget al. [33] as p100 coactivator (p100; gi. 77404397).In rat, the formerly named p105 coactivator wascloned and renamed SND p102 (gi. 51512261)reflecting its structural properties [24]. Both thefunction and the subcellular location of SND1 proteinfamily members have been studied from very differ-ent points of view. Initially, p100 was described as atranscriptional coactivator [33], but further studiesdetected p100 in a wide variety of tissues and relatedit to a variety of physiologic events, such as arteryinjuries connected to oxidative stress [29] and milkprecursory lipid droplets in the mammary gland [15].

The functions attributed to SND proteins include,among others, transcriptional coactivation [25, 37] andmRNA maturation [32, 43], both of them connected toa nuclear location in the cell [33, 37]. The SND1 geneproduct has also been related to extranuclear mRNAprocessing, since it has been described as a componentof RNA-induced silencing complex [8, 31]. Nonethe-less, other studies have focused on functions notconnected to nucleic acids. The organisms and celltypes used as models for these experiments areheterogeneous and resulting from these studies; SNDp102 has been related to multiple and diverse cell type-and species-specific cellular processes. In zebrafish,two different-sized products of SND1 gene have beendescribed in cytoplasmatic localization [44]. In plants,the p100 homologue has been identified as acytoskeleton-associated protein [30], and in melano-cytes, it has been detected in melanosomes [2, 9]. TheSND1 protein has been identified in the ER and milkprecursory LD in rat and cow mammary gland cells[15]. In fact, the amount of p100 has been found to riseduring lactation in response to prolactin in culturedsecretory cells from bovine, murine, and ovine models[7]. Other studies have contributed to this tumult ofpossibilities: in rat, SND1 gene expression increases incertain cancers [5, 14, 35], as well as in balloon-injuredrat carotid arteries and in endothelial cells exposed toreactive oxygen species, related to lipid deposition inartery walls during atherogenic processes [29]. In arecent study [24], we have established a directrelationship between SND p102 and the amount ofphospholipids in apoB-containing lipoproteins secretedby primary cultures of hepatocytes differentiallyexpressing the protein.

The ER is particularly important in vesicle traf-ficking and secretion, and both LD and melanosomes

are organelles closely related to the ER [21], suggest-ing a connection of SND1 with secretory processes.The cell types in which this protein family has beenidentified in non-nuclear compartments share withliver parenchymal cells a notable secretion capacityand a very high lipid handling capacity. In fact, theliver is able to accommodate changes in lipid fluxesby modification of specific metabolic, secretory, andstoring functions, playing a central role in theregulation of whole-body lipid homeostasis. To ourknowledge, none of the studies published so far haveused the hepatocyte as a working model.

The aim of this work was to study in detail thecytoplasmatic localization of SND p102 in hepato-cytes. To further evaluate SND p102 location on lipiddroplets, we used two well-established pro-steatoticconditions: the physiological regeneration of the liverafter a partial hepatectomy (70%) and the treatment ofhepatocyte cultures with exogenous oleate. The twolines of evidence point to a translocation of SNDp102 to low density LD under stimulation of hepato-cellular esteatosis.

Materials and methods

Animals and surgery Experiments were conducted incompliance with institutional guidelines. FemaleSprague-Dawley rats weighing 175–200 g were usedthroughout the study. Partial hepatectomy (70%) wasperformed according to Waynford et al. [39], withresection of the upper median and left lobes. Theremnant liver mass was removed 24 h later.

Subcellular fractionation of liver tissue The proce-dure was carried out at 4°C, and the homogenate wasobtained following the method described in Palacioset al. [24]. Briefly, the freshly extracted liver washomogenized using a Potter homogenizer in 4volumes of Tris–HCl 20 mM, pH 7.4, EDTA 1 mM,and protease inhibitors as recommended (Roche).Next, the nuclei were sedimented by centrifugationat 1,000×g for 10 min, and the postnuclear superna-tant (PNSN) was adjusted to a sucrose concentrationof 20%. Then, 1.5 ml PNSN was loaded onto adiscontinuous sucrose gradient composed of fourphases (1.5 ml each) ranging from 40 to 25% (w/v)sucrose, and overlaid with three more 1.5 ml phasesof 15%, 10%, and 5% (w/v) sucrose. The tubes were

74 I. Garcia-Arcos et al. Author's personal copy

centrifuged for 4 h at 150,000×g in a TST 41.14 rotor(Kontron). Finally, 18 fractions were obtained fromthe top of the gradient. The upper fraction, asemisolid layer, was collected by adsorption on ahydrophobic spatula and dissolved in 60 mM Tris–HCl buffer pH 6.8, containing 10% sodium dodecylsulfate (SDS) and 50% sucrose. The followingfractions were pipetted.

For LD isolation from rat hepatocyte primarycultures, the gradient was scaled down to a totalvolume of 4 ml and was composed by 35% sucrose(0.5 ml), 25% sucrose (0.5 ml) PNSN (1.5 ml), 15%,10%, and 5% sucrose layers (0.5 ml each).

Culture of HepG2 cells and rat hepatocytes Humanhepatoblastoma HepG2 cells were cultured as de-scribed in Fresnedo et al. [12] and used at approxi-mately 80% confluence. Cells were washed twice andharvested with cold phosphate-buffered saline (PBS).Next, they were homogenized by gentle sonication onice followed by being passed through a 25-gageneedle ten times. The PNSN was obtained andprocessed as above, with the exception that it wasloaded onto a gradient scaled down to a total volumeof 4 ml (0.5 ml/phase) and centrifuged in a TST 55.5rotor (Kontron).

Hepatocyte isolation and culture was performedexactly as described elsewhere [24]. Briefly, hepato-cytes were isolated from Sprague-Dawley female ratsand cultured in dishes coated with a matrix ofcollagen and fibronectin. In order to induce steatosis,the culture medium was supplemented with 0.6 mMoleic acid bound to 100 mg/ml fatty acid-free bovineserum albumin (BSA), while the controls weresupplemented with BSA alone. The hepatocytes werekept overnight in these culture conditions.

Immunofluorescence and immunoblotting studies Im-munocytochemistry was performed as in Palacioset al. [24]. Hepatocytes were cultured on coverslips.After the treatments, the coverslips were washed withPBS and the cells fixed with 3.7% formaldehyde.They were permeabilized with 0.1% Triton X-100 andblocked with 10% fetal bovine serum in PBS. Next,the antibodies shown in Table 1 were used at thedefined dilution in blocking solution. The primaryand secondary antibodies were incubated for 1.5 and1 h, respectively. Optical sections were obtained usingan Olympus Fluoview FV500 confocal microscope

(SGIKER Facility of the University of the BasqueCountry). The fluorofores were excited sequentially,ensuring no channel bleed-through.

For immunoblotting, hepatocytes were scraped offin PBS, pelleted, and incubated for 30 min in lysisbuffer with 50 mM Tris–HCl, pH 8 with 1% Igepal,0.5 mM DTT, 1 mM EDTA, and protease inhibitors asrecommended (Roche). They were centrifuged at1,000×g, and the supernatant was collected. Westernblotting was performed exactly as described previ-ously [24], using Superblock Blocking BufferTM

(Pierce) with 0.1% Tween-20 as blocking solutionand the antibodies listed in Table 1.

Protein measurement The amount of protein wasmeasured by the bicinconinic acid method followingthe manufacturer (Pierce) instructions using BSA asa standard. Two percent SDS was included in allsamples to be measured to avoid overestimation in thehighly lipidic samples [17].

Table 1 Antibodies used in this work

Antibody

Rabbit anti-SND p102 IC (1/75)WB (0.3 mg/ml)

Mouse anti-ADRP IC (1/75) Progen(Denmark)WB (1/10000)

Mouse anti-GAPDH IC (1/2000) AB CamCambridge

Chicken anti-calregulin IC (1/75) AB CamCambridge

Goat anti-calregulin WB (1/1000) Santa CruzBiotechnology

Mouse anti-GM130 WB (1/1000) AB CamCambridge

Alexa FluorTM

594-Goat anti-rabbit IgGIC (1/200) Molecular

Probes

Alexa FluorTM

488-Goat anti-mouse IgGIC (1/200) Molecular

Probes

Alexa FluorTM

488-Donkeyanti-chicken IgG

IC (1/2000) MolecularProbes

HRP-goat anti-rabbit IgG WB (1/2000) Sigma,St. Louis, MO

HRP-horse anti-mouse IgG WB (1/2000) Cell signalingtechnology

HRP-bovine anti-goat IgG WB (1/7000) Santa CruzBiotechnology

For each antibody, the application and working dilutions aredetailed (IC immunocytochemistry, WB Western blotting)

SND1 targeting to liver lipid droplets 75 Author's personal copy

Lipid extraction and quantification Lipids from thedifferent density fractions were extracted followingthe method of Folch et al. [11] and dried. The lipidextracts were dissolved in toluene, and an internalstandard of cholesteryl formate was added. Lipidswere quantified as described [28].

Results

Subcellular distribution of SND p102 in liver paren-chyma In previous work, indirect immunofluores-cence revealed extranuclear localization of SNDp102 in rat hepatocytes. In addition, cell fractionationusing differential centrifugation and immunoblottingresulted in us recovering the most SND p102 proteinin microsomal pellets, which suggests that this proteinresides mainly in the endoplasmic reticulum [24].

To analyze the cytoplasmic localization of SNDp102 in more detail, we applied sucrose densitygradient centrifugation to isolate 18 fractions usingPNSN of liver homogenate as the starting material.Density fractions were collected and characterized forthe distribution profiles of total protein and organellemarker proteins. To ascertain whether SND p102 celldistribution is dependent on steatosis induction, thestudy was performed on quiescent, control, andregenerating liver at 24 h after partial hepatectomy.

The mass of protein recovered in the fractions, andthe resulting profile along the gradient are plotted inFig. 1a. Most of the cellular protein was accumulatedin the fractions below the loading point, that is to say,below fractions 9 and 10 (F9 and F10), whichcorrespond to 20% sucrose. The fractions below theloading point did not show differences in proteinbetween quiescent and regenerating tissue. By con-trast, above the loading point, all the fractions derivedfrom regenerating tissue contained higher proteinlevels than those in quiescent liver. This was par-ticularly evident in fraction 1, which contained5.8 mg protein/g of tissue in the regenerating liverand only 0.2 mg/g of tissue in the quiescent liver.

The fractions were blotted against GM130 andcalregulin to identify the microsomal fractions con-taining Golgi (Fig. 1b) and ER (Fig. 1c) membranes,respectively. The distribution of these two proteinsalong the gradient was practically identical in quies-cent and regenerative tissue. GM130 was detected infractions 10 to 15 and calregulin in fractions 10 to 14.

Therefore, Golgi cofractionated with a minor part ofthe cytosol, which presumably corresponds with theloading point and extensively with the ER, not beingpossible to separate the intracellular membranesystems that form the microsomes. As shown in theblots in Fig. 1b, c (left), the fractionated PNSNcontained similar amounts of calregulin and GM130before and after the hepatectomy.

To identify the fractions corresponding to LD, weused antibodies against adipophilin (ADRP). Thisprotein was found only in the light fractions and, asexpected for being a LD characteristic protein, ADRPexpression was much higher in regenerating tissue.ADRP was more intensely detected and in a greaternumber of fractions of different densities in regener-ating tissue than in quiescent liver (Fig. 1d). Thehydrophobic lipid content (triglyceride+cholesterylesters) measured in the fractions above the loadingpoint corresponding to LD is shown in the upper partof Fig. 1d. As expected, the regenerating livercontained higher amounts of neutral lipid and mostof it was detected in the lightest fraction, whichalso contained the highest amount of the LD markeradipophilin.

SND p102 was mainly detected in the microsomalfractions both in quiescent and regenerating liver(Fig. 1e). It showed cofractionation with the endo-plasmic reticulum and Golgi marker proteins, asexpected from our own previous results [24]. SNDp102 was also detected in the fractions identified asLD by the presence of adipophilin and hydrophobiclipid enrichment. The amount of SND p102 in LDfractions was greater in regenerating than in quiescenttissue, where, even extending the exposure timeduring the Western blot membrane development, theSND p102 band was practically undetectable. How-ever, the protein was unequivocally detected infraction 1 of steatotic liver. The proportion of SNDp102 in LD fractions increased from 0.37% inquiescent tissue to 3.11% 24 h after the hepatectomy.However, the total amount of SND p102 in PNSN didnot differ between quiescent and regenerating tissue(Fig. 1e, left blot).

Selective targeting of SND p102 to the lowest densitylipid droplets Although all the fractions above theloading point contained LD in regenerating liver,fractions 1 to 3 contained the most anti-ADRPimmunoreactive material (Fig. 1e). Figure 2 shows

76 I. Garcia-Arcos et al. Author's personal copy

the distribution profiles of SND p102 and ADRP inLD fractions in three independent experiments. Theblots of quiescent and regenerating tissue of onerepresentative experiment are presented. The findingsare shown in sufficient detail for it to be seen thatADRP distribution was specific for each individual.In regenerating liver, and regardless of both itsdistribution pattern in quiescent liver and ADRPdistribution in any condition, SND p102 showed acharacteristic adscription to fraction 1. It was detectedto a lesser extent in the denser LD fractions. However,SND p102 distribution in quiescent tissue wasdifferent for each animal, and in all cases, detectionwas extremely weak in LD fractions. Hence, in thesteatotic tissue, SND p102 acquires a characteristic

profile, showing maximal abundance in the fractionscontaining the lightest LD.

SND p102 visualization in primary hepatocytecultures We performed immunocytochemistry experi-ments in primary hepatocyte cultures to visualizeSND p102 in its cellular context. The primaryantibodies we used were the same as in the gradientcharacterization experiments of liver tissue. Asexpected from our own results [24], SND p102 wasfound to be distributed heterogeneously in thecytoplasm of primary hepatocytes, with a pattern thatresembles reticular localization (Fig. 3). Consistentwith its detection in the sucrose density gradient forliver, SND p102 clearly overlapped with GM130 and

Fig. 1 Subcellular distribution of SND p102 in quiescent andregenerating rat liver. PNSN from 0.32 g of regenerating ratliver was fractionated on a sucrose gradient as described inMaterials and methods. The collected volume was 500 µl forfractions 1 to 4, and 750 for fractions 5 to 18. a The totalamount of protein in each fraction is plotted. The PNSN and thefractions were blotted with antibodies against the Golgi marker

GM-130 (b), the endoplasmic reticulum marker calregulin (c),the lipid droplet marker adipophilin (ADRP) (d), and SNDp102 (e). For the detection, 10 µl of fractions 1 to 4 and 7.5 µlof the others were loaded in the gel. In d, the absolute amountof hydrophobic lipids, given as the sum of triglycerides andcholesterol esters in each fraction, is shown above the graph

SND1 targeting to liver lipid droplets 77 Author's personal copy

calregulin. SND p102 also overlapped with glyceral-dehyde 3-phosphate dehydrogenase (GAPDH), indi-cating a cytosolic localization for the protein as well.

Growth and proliferation of LD in hepatocyteswere promoted by treatment of the cultures witholeate, as described in the Methods section. LD werevisualized as an ADRP ring structure (Fig. 3a, lowpanel). Due to the ubiquitous distribution of SNDp102 and to the close proximity of the LD to the ER,it is difficult to perceive any overlap of SND p102and ADRP in these micrographs.

We investigated the presence of SND p102 in LDpromoted by oleate treatment of primary hepatocytesafter density gradient subcellular fractionation. The

experiment followed the same design as for livertissue, but the gradient was scaled down to a totalvolume of 4 ml. Figure 3b shows the expression ofSND p102 in some representative fractions of the

Fig. 2 Distribution profile of SND p102 in density gradientfractions in quiescent and regenerating liver. The ADRP-positivefractions were blotted against SND p102 and the integratedoptical density (IOD) was calculated for the total volume of thefraction. The IOD of SND p102 (a) and ADRP (b) bands for the

total volume of LD fractions of three independent experiments isshown. Note that the distribution of SND p102 in the regener-ating tissue is independent from that in the quiescent liver andfrom those of ADRP in quiescence and regeneration

Fig. 3 Confocal microscopy analysis of SND p102 distribution incultured hepatocytes. a Rat hepatocyte cultures were fixed andimmunostained for SND p102 (red) and the organelle markers:GM130 for Golgi, calregulin for endoplasmic reticulum, GAPDHfor cytosol, and ADRP for lipid droplets. Before ADRP staining,the cultures were treated overnight with 0.6 mM oleic acid in BSA(oleate-treated) to promote lipid droplet growth and proliferation.The images were acquired with a 60× objective and 2× opticalzoom. b The PNSN from control and oleate-treated hepatocyteswas centrifuged in a density gradient, and nine fractions werecollected and blotted with antibodies against ADRP and SND p102

b

78 I. Garcia-Arcos et al. Author's personal copy

SND1 targeting to liver lipid droplets 79 Author's personal copy

gradient in oleate-treated and non-treated hepatocytes.There is a visible cofractionation of ADRP and SNDp102 in the lightest fractions, identified as LD (F1 toF4), while not in the denser fractions. Therefore, thereis an increased presence of SND p102 in the lowdense LD isolated from the steatotic condition in bothliver tissue and primary hepatocytes.

SND p102 in human hepatocyte cultures The humanhepatoblastoma cell line HepG2 was utilized to deter-mine whether the anti-SND p102 antibodies wouldreveal a p100 distribution similar to that observed for itshomologue in rat liver after a partial hepatectomy. Thecellular distribution revealed for p100 in the confocalmicroscopy micrographs (Fig. 4a) was similar to that

Fig. 4 Association of p100coactivator to microsomesand lipid droplets in HepG2cells. a Control and oleate-treated HepG2 cells werestained for p100 and ADRPand observed under a con-focal microscope. b ThePNSN from control andoleate-treated HepG2 cellswas centrifuged in a densitygradient, and 17 fractionswere collected and blottedwith antibodies againstADRP, calregulin, and SNDp102. We show fractions 2and 3 as paradigms of lipiddroplet containing fractions,and 9 and 10 as microsome-containing fractions

80 I. Garcia-Arcos et al. Author's personal copy

seen in primary hepatocyte cultures [24]. SND1 wasmainly localized in the cytoplasm, showing an irregularand reticular staining.

The PNSN was prepared from control and 0.6 mMoleate-supplemented HepG2 cultures and fractionatedin a density gradient identical to that designed for ratliver. The human homologue of SND p102 had aheterogeneous distribution, it being detected both inlight and dense fractions. In Fig. 4, we show againfractions 9 and 10 as a paradigm for the densefractions, and fractions 2 and 3 as representative ofthe light fractions. Distribution of p100 in HepG2PNSN was similar to that of SND p102 in rat liverPNSN. Most of the protein was recovered in the highdensity fractions of the gradient, corresponding tomicrosomes. Once more, this shows that oleateexposure causes p100 overexpression in lipid drop-lets, which indicates that p100 has a behavior similarto that of SND p102 in rat liver.

Discussion

The human homologue of SND p102, p100 coactiva-tor, was described for the first time by Tong et al. [33]in the nuclei of several cell types, and some studiesabout the possible functions of this protein werecarried out in proliferative cell lines on the basis of anuclear localization [37, 42]. However, p100 hassince been detected in a wide variety of organisms,tissues, and cell types. The functions attributed tothis family of proteins involve multiple cellular events[31, 32, 37, 43].

Using rat liver tissue, primary cultures of rathepatocytes and human-derived HepG2 cells, wehave identified the main microsomal location forSND p102 and its human homologue p100. Theconfocal micrographs were consistent with the densitygradient subfractionation data and showed that themajority of SND p102 had microsomal localization,as there is visible overlapping with the markers forER and Golgi. Clearly visible overlap with thecytosolic marker GAPDH was also found in theconfocal micrographs. Overall, these observationsindicate that SND p102 has ubiquitous cytoplasmicdistribution.

The location and function of SND p102 have beenstudied using diverse approaches and cell types ortissues [15, 29, 30, 35, 38, 44]. It seems that the

subcellular localization of SND p102 is organism-and tissue/cell-specific [6], pointing to the possibilitythat this protein fulfills diverse specific and differen-tiated functions.

In the hepatocyte, the ER is a highly activeorganelle. It plays crucial roles in lipid and proteinsynthesis, detoxification, vesicle trafficking, andsecretory functions. In cells with intense secretoryactivity, SND p102 localization in the ER could berelated with vesicle trafficking. Indeed, the correspon-dent homologue in bovine was detected in the ERfrom mammary gland cells of lactating cows, wherean increase in SND1 expression was found inresponse to a rise in prolactin secretion [15].

Both mammary gland cells and hepatocytes areable to handle large amounts of lipid. In line with this,our own previous results [24] connect SND p102 withlipid metabolism and the VLDL secretory pathway inprimary hepatocytes, and, in a study by Sakamoto etal. [29], SND p102 was found to be overexpressed inendothelial cells exposed to reactive oxygen speciesand subsequently related with atherogenic processes.Additional evidence linking SND1 protein with lipidmetabolism is provided by the work of Keenan et al.[15], where the bovine homologue of SND p102 wasdetected in ER and LD from cow mammary gland.

To analyze the association of SND p102 tohepatocyte LD, we used two well-known models ofhepatocellular steatosis: liver regeneration after partialhepatectomy and treatment of cultured hepatocyteswith oleate, a highly potent inducer of LD formation[13]. Following a partial hepatectomy, the remainingquiescent hepatocytes quickly divide to restore theliver mass. During this process, the liver suffers atransient accumulation of large amounts of fat. Suchfat is stored in cytosolic LD, which proliferate andgrow after the partial hepatectomy [34].

LD are composed of a hydrophobic core, mainlytriglycerides and cholesteryl esters, and a surroundingmonolayer of phospholipid and cholesterol, withassociated proteins. Among them, ADRP is one ofthe best studied and is used as an organelle markerdue its exclusive location in LD. It was then expectedthat an increase in ADRP expression in the regenerat-ing liver would be found, it was detected insubcellular fractions of different densities. The highamount of neutral lipid found in these fractions,especially in the regenerating tissue, confirmed thepresence of LD. In the steatotic condition, SND p102

SND1 targeting to liver lipid droplets 81 Author's personal copy

clearly cofractionated with ADRP, while it was onlyweakly detected in the quiescent liver. The amount ofSND p102 was similar before and 24 h after thehepatectomy, so that these data seem to point to asubcellular translocation in response to such steato-genic condition. However, since the proportion ofSND p102 in LD is minor, it cannot be ruled out thatsuch a slight overexpression could be beyond theresolution limit of the Western blotting.

Oleic acid was efficient at producing the prolifer-ation of LD in primary cultures of hepatocytes, as theconfocal micrographs demonstrate. LD are organellesthought to derive from the ER [21, 27]. It is not yetclear, even when the LD are mature organelles, ifthere is continuity between the LD and the ERmembranes or if it is a simple apposition [23]. Giventhis and the fact that the main location of SND p102is the ER, it is perhaps not surprising that it is difficultto identify any overlap of SND p102 and ADRP inconfocal micrographs. However, the density gradientfractionation of HepG2 cells and rat primary hepato-cytes did unambiguously show the presence of SNDp102 in the LD containing fractions also for the oleatetreatment model of steatosis.

Notably, SND p102 cofractionated with ADRPmainly in the lightest fraction (Fig. 2). This targetingof SND p102 to such a specific subpopulation sug-gests the existence of different kinds of LD. Ontkoet al. [22] also separated and analyzed LD subpopu-lations in a density gradient, obtaining a lipidcomposition specific for each density fraction. Theheterogeneity of LD is nowadays starting to bestudied. It is recognized that LD-associated proteinsdiffer and change depending on the cell type andorganism characteristics [4, 10, 40], as well as on themetabolic status of the cell [1, 3, 16, 20, 23, 26, 41].However, studies focused on examining the differ-ences between LD subpopulations are scarce [1, 3,16, 26].

In summary, SND p102 in hepatocytes is mainlylocalized in microsomal membranes, experiencing anoverexpression in LD under the studied steatogenicconditions. SND p102 targeting is specifically directedtowards the lightest subpopulation of lipid droplets,suggesting some kind of involvement in the metabolicavailability of their components. Proteome analysis ofLD populations of different sizes and densities wouldhelp to clarify their metabolic differences and tounderstand their physiological significance.

Acknowledgments The authors wish to thank C. Enrich forhis contribution in the initial stage of this work and M. Busto forher excellent technical assistance. This research was supported bythe Spanish Ministry of Education and Science (SAF2007-60211), Basque Government (PE06UN24 and IT-325-07) and theUniversity of the Basque Country (15942/2004). Y.R. is recipientof a FPI fellowship from the Basque Government.

References

1. Bartz R, Li B, Venables WH, Zehmer JK, Roth MR, Welti R,Anderson RG, Liu P, Chapman KD (2007) Lipidomicsreveals that adiposomes store ether lipids and mediatephospholipid traffic. J Lipid Res 48:837–847

2. Basrur V, Yang F, Kushimoto T, Higashimoto Y, Yasumoto K,Valencia J, Muller J, Vieira WD, Watabe H, Shabanowitz J,Hearing VJ, Hunt DF, Appella E (2003) Proteomic analysis ofearly melanosomes: identification of novel melanosomalproteins. J Proteome Res 2:69–79

3. Blaner WS, O'Byrne SM, Wongsiriroj N, Kluwe J,D'Ambrosio DM, Jiang H, Schwabe RF, Hillman EM,Piantedosi R, Libien J (2008) Hepatic stellate cell lipiddroplets: a specialized lipid droplet for retinoid storage.Biochim Biophys Acta. doi:10.106/j.bbalip. 2008.11.001

4. Brasaemle DL, Dolios G, Shapiro L, Wang R (2004)Proteomic analysis of proteins associated with lipid dropletsof basal and lipolytically stimulated 3T3-L1 adipocytes.J Biol Chem 279:46835–46842

5. Brito GC, Fachel AA, Vettore AL, Vignal GM, Gimba ERP,Campos FS, Barcinski MA, Verjovski-Almeida S, Reis EM(2008) Identification of protein-coding and intronic non-coding RNAs down-regulated in clear cell renal carcinoma.Mol Carcinog 47:757–767

6. Broadhurst MK, Lee RS, Hawkins S, Wheeler TT (2005)The p100 EBNA-2 coactivator: a highly conserved proteinfound in a range of exocrine and endocrine cells and tissuesin cattle. Biochim Biophys Acta 1681:126–133

7. Broadhurst MK, Wheeler TT (2001) The p100 coactivator ispresent in the nuclei of mammary epithelial cells and its abun-dance is increased in response to prolactin in culture and inmammary tissue during lactation. J Endocrinol 171:329–337

8. Caudy AA, Ketting RF, Hammond SM, Denli AM, BathoornAM, Tops BB, Silva JM, Myers MM, Hannon GJ, PlasterkRH (2003) A micrococcal nuclease homologue in RNAieffector complexes. Nature 425:411–414

9. Chi A, Valencia JC, Hu ZZ, Watabe H, Yamaguchi H,Mangini NJ, Huang H, Canfield VA, Cheng KC, Yang F,Abe R, Yamagishi S, Shabanowitz J, Hearing VJ, Wu C,Appella E, Hunt DF (2006) Proteomic and bioinformaticcharacterization of the biogenesis and function of melano-somes. J Proteome Res 5:3135–3144

10. Dugail I, Hajduch E (2007) A new look at adipocyte lipiddroplets: towards a role in the sensing of triacylglycerolstores? Cell Mol Life Sci 64:2452–2458

11. Folch J, Lees M, Sloane Stanley GH (1957) A simplemethod for the isolation and purification of total lipids fromanimal tissues. J Biol Chem 226:497–509

12. Fresnedo O, De Heredia ML, Martinez MJ, Cristobal S, RejasMT, Cuezva JM, Ochoa B (2001) Immunolocalization of a

82 I. Garcia-Arcos et al. Author's personal copy

novel cholesteryl ester hydrolase in the endoplasmic reticulumof murine and human hepatocytes. Hepatology 33:662–667

13. Fujimoto Y, Onoduka J, Homma KJ, Yamaguchi S, Mori M,Higashi Y,MakitaM, Kinoshita T, Noda J, Itabe H, Takanoa T(2006) Long-chain fatty acids induce lipid droplet formationin a cultured human hepatocyte in a manner dependent ofAcyl-CoA synthetase. Biol Pharm Bull 29:2174–2180

14. Ho J, Kong JWF, Choong LY, Loh MCS, Toy W, Chong PK,Wong CH, Wong CY, Shah N, Lim YP (2009) Novelbreast cancer metastasis-associated proteins. J Proteome Res8:583–594

15. Keenan TW, Winter S, Rackwitz HR, Heid HW (2000)Nuclear coactivator protein p100 is present in endoplasmicreticulum and lipid droplets of milk secreting cells.Biochim Biophys Acta 1523:84–90

16. Kellner-Weibel G, Hendry-Rinde B, Haynes MP, Adelman S(2001) Evidence that newly synthesized esterified cholesterolis deposited in existing cytoplasmic lipid inclusions. J LipidRes 42:768–777

17. Kessler RJ, Fanestil DD (1986) Interference by lipids in thedetermination of protein using bicinchoninic acid. AnalBiochem 159:138–142

18. Martin S, Parton RG (2006) Lipid droplets: a unified viewof a dynamic organelle. Nat Rev Mol Cell Biol 7:373–378

19. Miura S, Gan JW, Brzostowski J, Parisi MJ, Schultz CJ,Londos C, Oliver B, Kimmel AR (2002) Functionalconservation for lipid storage droplet association amongperilipin, ADRP, and TIP47 (PAT)-related proteins inmammals, Drosophila, and Dictyostelium. J Biol Chem277:32253–32257

20. Moore HP, Silver RB, Mottillo EP, Bernlohr DA, GrannemanJG (2005) Perilipin targets a novel pool of lipid droplets forlipolytic attack by hormone-sensitive lipase. J Biol Chem280:43109–43120

21. Murphy DJ, Vance J (1999) Mechanisms of lipid-bodyformation. Trends Biochem Sci 24:109–115

22. Ontko JA, Perrin LW, Horne LS (1986) Isolation ofhepatocellular lipid droplets: the separation of distinctsubpopulations. J Lipid Res 27:1097–1103

23. Ozeki S, Cheng J, Tauchi-Sato K, Hatano N, Taniguchi H,Fujimoto T (2005) Rab18 localizes to lipid droplets andinduces their close apposition to the endoplasmic reticulum-derived membrane. J Cell Sci 118:2601–2611

24. Palacios L, Ochoa B, Jose Gomez-Lechon M, Vicente CJ,Fresnedo O (2006) Overexpression of SND p102, a rathomologue of p100 coactivator, promotes the secretion oflipoprotein phospholipids in primary hepatocytes. BiochimBiophys Acta 1761:698–708

25. Paukku K, Yang J, Silvennoinen O (2003) Tudor andnuclease-like domains containing protein p100 function ascoactivators for signal transducer and activator of transcription5. Mol Endocrinol 17:1805–1814

26. Rinia HA, Burger KN, Bonn M, Muller M (2008)Quantitative label-free imaging of lipid composition andpacking of individual cellular lipid droplets using multiplexCARS microscopy. Biophys J 95:4908–4914

27. Robenek H, Hofnagel O, Buers I, Robenek MJ, Troyer D,Severs NJ (2006) Adipophilin-enriched domains in the ERmembrane are sites of lipid droplet biogenesis. J Cell Sci119:4215–4224

28. Ruiz JI, Ochoa B (1997) Quantification in the subnanomolarrange of phospholipids and neutral lipids by monodimen-

sional thin-layer chromatography and image analysis. J LipidRes 38:1482–1489

29. Sakamoto K, Yamasaki Y, Kaneto H, Fujitani Y, Matsuoka T,Yoshioka R, Tagawa T, Matsuhisa M, Kajimoto Y, Hori M(1999) Identification of oxidative stress-regulated genes inrat aortic smooth muscle cells by suppression subtractivehybridization. FEBS Lett 461:47–51

30. Sami-Subbu R, Choi SB, Wu Y, Wang C, Okita TW (2001)Identification of a cytoskeleton-associated 120 kDa RNA-binding protein in developing rice seeds. Plant Mol Biol46:79–88

31. Scadden AD (2005) The RISC subunit Tudor-SN binds tohyper-edited double-stranded RNA and promotes its cleavage.Nat Struct Mol Biol 12:489–496

32. Shaw N, Zhao M, Cheng C, Xu H, Saarikettu J, Li Y, Da Y,Yao Z, Silvennoinen O, Yang J, Liu ZJ, Wang BC, Rao Z(2007) The multifunctional human p100 protein “hooks”methylated ligands. Nat Struct Mol Biol 14:779–784

33. Tong X, Drapkin R, Yalamanchili R, Mosialos G, Kieff E(1995) The Epstein-Barr virus nuclear protein 2 acidicdomain forms a complex with a novel cellular coactivatorthat can interact with TFIIE. Mol Cell Biol 15:4735–4744

34. Trotter NL (1964) A fine structure study of lipid in mouseliver regenerating after partial hepatectomy. J Cell Biol21:233–244

35. Tsuchiya N, Ochiai M, Nakashima K, Ubagai T, Sugimura T,Nakagama H (2007) SND1, a component of RNA-inducedsilencing complex, is up-regulated in human colon cancersand implicated in early stage colon carcinogenesis. CancerRes 67:9568–9576

36. Umlauf E, Csaszar E,MoertelmaierM, Schuetz GJ, Parton RG,Prohaska R (2004) Association of stomatin with lipid bodies.J Biol Chem 279:23699–23709

37. Valineva T, Yang J, Palovuori R, Silvennoinen O (2005)The transcriptional co-activator protein p100 recruitshistone acetyltransferase activity to STAT6 and mediatesinteraction between the CREB-binding protein and STAT6.J Biol Chem 280:14989–14996

38. Valineva T, Yang J, Silvennoinen O (2006) Characteriza-tion of RNA helicase A as component of STAT6-dependentenhanceosome. Nucl Acids Res 34:3938–3946

39. Waynforth HB, Fleccknell PA (1992) Experimental andsurgical technique in the rat. Academic Press

40. Welte MA (2007) Proteins under new management: lipiddroplets deliver. Trends Cell Biol 17:363–369

41. Wolins NE, Brasaemle DL, Bickel PE (2006) A proposedmodel of fat packaging by exchangeable lipid dropletproteins. FEBS Lett 580:5484–5491

42. Yang J, Aittomaki S, Pesu M, Carter K, Saarinen J,Kalkkinen N, Kieff E, Silvennoinen O (2002) Identificationof p100 as a coactivator for STAT6 that bridges STAT6with RNA polymerase II. EMBO J 21:4950–4958

43. Yang J, Valineva T, Hong J, Bu T, Yao Z, Jensen ON,Frilander MJ, Silvennoinen O (2007) Transcriptional co-activator protein p100 interacts with snRNP proteins andfacilitates the assembly of the spliceosome. Nucl Acids Res35:4485–4494

44. Zhao CT, Shi KH, Su Y, Liang LY, Yan Y, Postlethwait J,Meng AM (2003) Two variants of zebrafish p100 areexpressed during embryogenesis and regulated by nodalsignaling. FEBS Lett 543:190–195

SND1 targeting to liver lipid droplets 83 Author's personal copy