FEBS Letters 587 (2013) 3696–3702

journal homepage: www.FEBSLetters .org

Inhibition of ADP-ribosylation suppresses aberrant accumulationof lipidated apolipoprotein B in the endoplasmic reticulum

0014-5793/$36.00 � 2013 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.febslet.2013.09.036

Abbreviations: ADP, adenosine diphosphate; ALLN, N-acetyl-L-leucyl-L-leucyl-L-norleucinal; ApoB, apolipoprotein B-100; Arf1, ADP-ribosylation factor 1; BFA,brefeldin A; DAPI, 4,6-diamidino-2-phenylindole; DHA, docosahexaenoic acid; ER,endoplasmic reticulum; GDP, guanine diphosphate; GEF, guanine nucleotideexchange factor; GTP, guanine triphosphate; LD, lipid droplet; LPDS, lipoproteindeficient serum; MIBG, m-iodobenzylguanidine hemisulfate salt; MTP, microsomaltriglyceride transfer protein; NAM, nicotinamide; NDGA, nordihydroguaiaretic acid;OA, oleic acid; VLDL, very low-density lipoprotein⇑ Corresponding author. Address: Department of Anatomy and Molecular Cell

Biology, Nagoya University Graduate School of Medicine, 65 Tsurumai, Showa,Nagoya 466-8550, Japan. Fax: +81 52 744 2011.

E-mail address: tfujimot@med.nagoya-u.ac.jp (T. Fujimoto).

Yuki Ohsaki a, Jinglei Cheng a, Kazushi Yamairi a, Xiaoyue Pan b, M. Mahmood Hussain b, Toyoshi Fujimoto a,⇑a Department of Anatomy and Molecular Cell Biology, Nagoya University Graduate School of Medicine, Nagoya 466-8550, Japanb Departments of Cell Biology and Pediatrics, State University of New York Medical Center, NY 11203, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 21 August 2013Revised 17 September 2013Accepted 17 September 2013Available online 4 October 2013

Edited by Felix Wieland

Keywords:Brefeldin AADP-ribosylationApoB-crescentLipid dropletER

ApoB-crescent, an endoplasmic reticulum (ER)-lipid droplet amalgamation structure, is a usefulmarker to indicate aberrant lipidated apolipoprotein B accumulation in the hepatocyte ER. Blockadeof the ER-to-Golgi transport by either vesicle transport inhibitors or dominant-negative Arf1 causeda significant increase in ApoB-crescents. However, a low concentration of Brefeldin A induced thesame result without affecting protein secretion, suggesting ADP-ribosylation as an additionalmechanism. ADP-ribosylation inhibitors not only suppressed the increase of ApoB-crescents, butalso rapidly dissolved existing ApoB-crescents. These results implicate the involvement of ADP-ribo-sylation in the ApoB-crescent formation and maintenance process at the ER.� 2013 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

1. Introduction

Apolipoprotein B-100 (ApoB) is a protein that constitutes verylow-density lipoproteins (VLDL). In hepatocytes, ApoB is lipidatedcotranslationally to form pre-VLDL, which thereafter acquiresadditional lipids to become mature VLDL. Because the VLDLconcentration in the blood needs to be strictly controlled tomaintain systemic lipid metabolism, it is important to understandthe regulatory mechanism of VLDL secretion.

The VLDL secretion from hepatocytes is regulated mainly byApoB protein degradation [1]. If the co-translational lipidation doesnot occur properly, nascent ApoB is dislocated from the Sec61translocon to the cytoplasm, ubiquitinated, and degraded byproteasomes [2]. In addition to this well-characterized mechanism

to degrade poorly-lipidated ApoB, ApoB after lipidation is alsosubjected to intracellular degradation [2–6]. In previous studies,we defined a degradation pathway of lipidated ApoB at thejuncture of ER and lipid droplets (LD) and showed that Derlin-1and UBXD8 play critical roles [7,8]. When this ApoB degradationpathway is suppressed by proteasome inhibition, docosahexaenoicacid (DHA), UBXD8 knockdown, or other methods, a unique ER-LDamalgamation structure called an ‘ApoB-crescent’ increasedsignificantly [7,8]. These findings along with other resultssuggested that the ApoB-crescent formation is induced by anabnormal accumulation of lipidated ApoB in the ER lumen.Conversely, ApoB-crescents are regarded as a convenient way tomonitor the lipidated ApoB degradation process.

In the present study, we explored the mechanism of ApoB-cres-cent formation in detail. We hypothesized that if an increase inlipidated ApoB induces ApoB-crescents, then blockade of the ER-to-Golgi transport should also increase their numbers. To test thishypothesis, we applied several different methods to suppress thesecretory pathway and found that all of them indeed increasedApoB-crescents. Surprisingly, brefeldin A (BFA) at a low concentra-tion increased ApoB-crescents without significantly affectingsecretion. Because this BFA effect was repressed by ADP-ribosyla-tion inhibitors, we concluded that an unknown target(s) of ADP-ribosylation is involved in the degradation process of lipidatedApoB. The results are discussed with regards to the mechanismof ApoB-crescent formation.

Y. Ohsaki et al. / FEBS Letters 587 (2013) 3696–3702 3697

2. Materials and methods

Please see Supplementary Data for details.

2.1. Cells and transfection

Huh7, HepG2, and 3Y1 cells were used. When appropriate, cellswere incubated in medium containing 10 lM N-acetyl-L-leucyl-L-leucyl-L-norleucinal (ALLN) (Sigma–Aldrich) to inhibit proteasomalfunctions. Oleic acid (OA) and DHA (Sigma–Aldrich) werecomplexed with fatty acid-free BSA (Wako) at a molar ratio of6:1 and applied at a final fatty acid concentration of 0.4 mM.

2.2. Plasmids

pcDNA3.1(+)/HA tagged-human Arf1(WT) and Arf1(T31N) werekindly donated by Dr. Naoko Morinaga (Chiba University). HumanCtBP1-S/BARS cDNA was obtained by reverse transcription andcloned to the pEGFP-C2 vector (Clontech).

2.3. Subcellular fractionation

Cells were disrupted by nitrogen cavitation and subjected to asucrose density gradient ultracentrifugation [9]. Fractions wereprecipitated with 10% trichloroacetic acid, dissolved in the samplebuffer, and analyzed by Western blotting.

2.4. Immunofluorescence microscopy and data analysis

Cells were fixed with 3% formaldehyde and permeabilizedeither with 0.01% digitonin or with 0.1% Triton X-100. LDs andnuclei were stained with BODIPY493/503 (Invitrogen) and 4,6-diamidino-2-phenylindole (DAPI) (Sigma–Aldrich), respectively.

To analyze effects on ApoB-crescents, at least 10 pictures wererandomly taken, and the ratio of ApoB-crescent-positive cells wasquantified. The results from three independent experiments wereaveraged, and the statistical difference from the control wasexamined using the Student’s t-test.

2.5. Electron microscopy

Cells cultured on coverslips were fixed with 2.5% glutaralde-hyde in 0.1 M sodium cacodylate buffer and post-fixed in a mixtureof 1% osmium tetroxide and 0.1% potassium ferrocyanide in thesame buffer.

2.6. Microsomal triglyceride transfer protein (MTP) activity assay

The MTP activity of mouse liver tissue homogenate (50 lgprotein), HepG2 cells (50 lg protein), and 1 lg purified MTP wasmeasured using the fluorescence assay as described [10].

3. Results

3.1. ApoB-crescents increased significantly by blockade of the ER-to-Golgi transport

The ApoB-crescent is an ER-LD amalgamation structure, whichis recognized as crescent-shaped ApoB labeling around LDs byimmunofluorescence microscopy [9]. To examine whether block-ade of the ER-to-Golgi transport increases ApoB-crescents, wetested three vesicle transport inhibitors that work by differentmechanisms. BFA down-regulates COPI vesicle formation by sup-pressing the GDP/GTP exchange activity of Arf1-GEF [11,12],whereas Exo1 accelerates GTP hydrolysis through the activationof Arf1-GAP [13]. NDGA has been suggested to modulate the

dynein-dynactin-related process [14] and to disrupt cisternalorganization of the Golgi complex [15]. When Huh7 cells weretreated with these reagents, secretion of ApoB was virtuallyabolished, but the intracellular amount of ApoB did not decreasesignificantly (Fig. 1A), suggesting an increase of ApoB in the ER.Under this condition, a significant increase of ApoB-crescents wasobserved (Figs. 1B, C).

We also tested the effect of transient expression of Arf1(T31N),the dominant-negative form of Arf1 that down-regulates the ER-to-Golgi transport [16]. The ApoB-crescent formation was in-creased in cells expressing Arf1(T31N), but not in cells expressingwild-type Arf1 or EGFP alone (Fig. 1D).

The increase of ApoB-crescents by the above treatments wascompletely inhibited when MTP inhibitor (MTPi; BAY13-9952 at100 nM) was given simultaneously (data not shown), confirmingthat lipidated ApoB is responsible for the ApoB-crescent formation.The result corroborated that blockade of the ER-to-Golgi transportcaused accumulation of lipidated ApoB in the ER and therebyincreased ApoB-crescents.

3.2. A low concentration of BFA increased ApoB-crescents withoutinhibiting secretion

In the above experiment, cells were treated with BFA for 1 h at aconcentration of 5 lg/ml. This condition is generally used to blockthe ER-to-Golgi transport. We found, however, that ApoB-crescentsincreased even when BFA was applied to Huh7 cells at much lowerconcentrations, i.e., 5–10 ng/ml (Fig. 2A). In cells treated with10 ng/ml BFA for 12 h, the Golgi delineated by immunofluores-cence labeling of GM130 showed disintegration (Fig. 2B) and ApoBand transferrin secretion decreased (Fig. 2C), but neither of thesechanges were observed in cells treated with 5 ng/ml BFA for 12 h(Figs. 2B, C). It was further confirmed that ApoB-crescents inducedto form by 5 ng/ml BFA take a similar ultrastructure to thosecaused by proteasomal inhibition or DHA [7] (Fig. 2D). This resultsuggested that BFA does not increase ApoB-crescents simply byblocking the ER-to-Golgi transport, but that other mechanismsmay be involved.

3.3. ADP-ribosylation inhibitors suppressed ApoB-crescent formation

Besides inhibition of Arf1-GEF, BFA is known to increase ADP-ribosylation of several proteins [17,18]. To test whether ADP-ribo-sylation is related to the ApoB-crescent formation, we examinedthe effect of three ADP-ribosylation inhibitors: MIBG, which is spe-cific to mono-ADP-ribosylation [19,20], and coumarimycin A1 andNAM, which are effective for both mono- and poly-ADP-ribosyla-tion [21]. When Huh7 cells were treated for 6 h with 5 lg/mlBFA in the presence of one of the ADP-ribosylation inhibitors,ApoB-crescent formation was suppressed significantly (Figs. 3A,B). This result indicated that the effect of BFA on the ApoB-crescentformation was exerted mainly through facilitating ADP-ribosyla-tion. Because MIBG is thought to be most specific to mono-ADP-ribosylation among the three reagents [21], the following experi-ments were mainly performed using MIBG, but the other twoinhibitors were also tested in most cases.

MIBG also inhibited the increase of ApoB-crescents that wasinduced by other reagents. In the presence of MIBG, either prote-asomal inhibition with 10 lM ALLN or treatment with 0.4 mMDHA failed to increase ApoB-crescents (Fig. 3C). The inhibitoryeffect of MIBG on ApoB-crescent formation was similar even whencells were treated with a mixture of ALLN and 3-methyladenine tosuppress both autophagy and proteasomes (Fig. 3D).

The effect of ADP-ribosylation inhibition on ApoB was furtherexamined by western blotting of subcellular fractions (Fig. 3E). InHuh7 cells treated with 10 lM ALLN alone for 12 h, a significant

Fig. 1. Blockade of the ER-to-Golgi transport increased the number of ApoB-crescents. (A) Effect of vesicular transport inhibitors on intracellular and secreted ApoB. Huh7cells were treated with 5 lg/ml BFA, 100 lM Exo1, or 30 lM NDGA for 4 h. Cell lysates and the culture media were subjected to Western blotting. ApoB secreted into theculture medium was reduced significantly by any of the reagents, whereas the decrease in intracellular ApoB was relatively small. (B) BFA caused a significant increase inApoB-crescents. Huh7 cells treated with 5 lg/ml BFA for 1 h were labeled for ApoB (red), LD (green), and DNA (blue). Arrowheads indicate ApoB-crescents that are observedas a crescent-shaped or circular ApoB labeling around LDs. Bar, 10 lm. (C) Vesicle transport inhibitors increased ApoB-crescents. In Huh7 cells treated with 5 lg/ml BFA,100 lM Exo1, or 30 lM NDGA for 6 h, the ratio of ApoB-crescent-positive cells increased significantly. (D) Overexpression of a dominant-negative Arf1, Arf1(T31N), increasedApoB-crescents. Huh7 cells were transfected with cDNA of either EGFP alone, HA-tagged wild-type Arf1 [HA-Arf1(WT)], or HA-tagged Arf1(T31N), and examined one daylater. The introduced proteins (red) along with ApoB (green) and DNA (blue) are shown. Arrowheads indicate ApoB-crescents. Please note that only ApoB in ApoB-crescents islabeled when cells are permeabilized with digitonin [9]. Bars, 10 lm.

3698 Y. Ohsaki et al. / FEBS Letters 587 (2013) 3696–3702

amount of ApoB was recovered in the LD fraction, reflecting accu-mulation of lipidated ApoB in ApoB-crescents [7]. In contrast, whenHuh7 cells were treated for 12 h with both 10 lM ALLN and 25 lMMIBG, ApoB in the LD fraction was negligible, consistent with thedecrease of ApoB-crescents in the MIBG-treated cell (Fig. 3C). Sim-ilar results were obtained with the other ADP-ribosylation inhibi-tors (data not shown).

In addition, MIBG caused a rapid decrease in the number ofexisting ApoB-crescents. ApoB-crescents that had formed by treat-ment with 10 lM ALLN for 6 h decreased significantly within 1 hafter MIBG was applied (Fig. 3F). Consistent with this result, theamount of ApoB in the LD fraction of the ALLN-treated cells was re-duced to a negligible level by treatment with 25 lM MIBG for 1 h(Fig. S1).

Several results indicated that MIBG did not affect the amountof LDs. First, MIBG did not affect the amount of ADRP (perilipin2),an authentic LD marker protein, in the LD fraction (Fig. 3E).Second, MIBG did not influence de novo LD formation that wasinduced by fatty acids (Fig. S2). These results suggested thatADP-ribosylation inhibitors suppressed the process of ApoB-crescent formation per se without affecting the LD-related lipidmetabolism.

It was also confirmed that treatment with MIBG did not de-crease either the mRNA or the protein level of ApoB in the cell(Figs. S3A, B). On the other hand, even when transcription wasinhibited with actinomycin D, secretion of ApoB was not affectedafter incubation for up to 6 h (Fig. S3C). The result indicated thatthe effect of MIBG on ApoB-crescents was exerted by a mechanismunrelated to the ApoB expression.

3.4. MIBG suppressed ApoB-crescent formation without affecting theMTP activity

MTP is critical for the assembly and secretion of apoB-contain-ing lipoproteins and its inhibition reduces ApoB secretion [22]. Wehave previously shown that MTP inhibition suppresses the ApoB-crescent formation by blocking ApoB lipidation [7]. The drasticreduction of ApoB-crescents by ADP-ribosylation inhibitors mayhave been caused because the inhibitors suppressed the MTPactivity by an unknown mechanism. To address this possibility,the effect of ADP-ribosylation inhibitors on the MTP activity wasexamined by an in vitro assay. MIBG (25 or 100 lM) or NAM (10or 30 mM) did not decrease the lipid transfer activity when mouseliver tissue homogenate was used to assess MTP activity (Fig. 4A).Similarly, MIBG and NAM neither reduced MTP activity in eitherHepG2 cell lysates or the purified protein (Fig 4A). Rather, MIBG in-creased the triglyceride transfer activity of MTP in vitro, butwhether it induced the same effect in cells needs to be examinedin a separate study. Further, we studied the effects of MIBG andNAM on MTP activity using different amounts of liver tissuehomogenates. MIBG increased, but NAM had no effect, on MTPactivity (Fig. 4A). It was confirmed that MIBG and NAM had noeffect on blank and total fluorescent measurements (Fig. S4). Thus,MIBG and NAM do not reduce MTP activity.

Consistent with the above result, the effect of MIBG on ApoB-crescents was clearly different from that of an MTPi, BAY13–9952. MIBG inhibited the increase of ApoB-crescents induced byeither ALLN or DHA (Fig. 3C and D), whereas BAY13–9952suppressed the increase of ApoB-crescents induced by ALLN, but

Fig. 2. BFA at low concentration increased ApoB-crescents without affecting ApoB secretion. (A) Treatment with BFA at 5–10 ng/ml for 12 h caused a significant increase inApoB-crescents. ApoB (red) and LD (green) were labeled. ApoB-crescents are indicated by arrowheads. Bar, 10 lm. (B) The Golgi showed structural disintegration in cellstreated with 10 ng/ml BFA for 12 h (left), but not in cells treated with 5 ng/ml BFA for 12 h (right). GM130, a Golgi matrix protein, was labeled. Bars, 10 lm. (C) Secretion ofboth ApoB and transferrin was decreased with 10 ng/ml BFA but not with 5 ng/ml BFA. The culture media were collected 24 h or 48 h after the BFA addition. (D) Electronmicroscopy of ApoB-crescents in cells treated with 5 ng/ml BFA for 12 h. The LD-ER amalgamation structures were frequently observed (arrow). The ER cistern constitutingthe ApoB-crescent was relatively wide in many cases. Bar, 0.5 lm.

Y. Ohsaki et al. / FEBS Letters 587 (2013) 3696–3702 3699

not the increase caused by DHA (Fig. 4B). In addition, MIBGreduced existing ApoB-crescents quickly (Fig. 3F), whereasBAY13-9952 did not (Fig. 4C). Electron microscopy confirmed thatthe LD-associated ER cistern of existing ApoB-crescents shrank andeventually disappeared with the MIBG treatment (data not shown).The results indicated that the effect of MIBG on the ApoB-crescentwas not due to suppression of the MTP activity.

4. Discussion

VLDL secretion from hepatocytes is regulated mainly bymodulating intracellular ApoB degradation [1]. The degradationof poorly-lipidated nascent ApoB has been studied extensively,

but the mechanism whereby ApoB is processed after lipidationis less well understood [2–6]. In a series of studies, we charac-terized an LD-resident mechanism whereby lipidated ApoB isdegraded and showed that UBXD8 and Derlin-1 play criticalroles in this process [7,8,23]. Importantly, abnormal accumula-tion of lipidated ApoB in the ER lumen was found to form aunique ApoB-crescent structure that can be easily observedusing immunofluorescence microscopy [9]. This structure in-creased significantly when degradation of lipidated ApoB wasabrogated, and it was totally abolished by MTP inhibition [7].These results indicate that ApoB-crescents can be used as aspecific indicator to show abnormalities in the post-lipidationprocess in the ER.

Fig. 3. ADP-ribosylation inhibitors decreased the number of ApoB-crescents. (A) The increase of ApoB-crescents induced by BFA was suppressed by MIBG. Huh7 cells werecultured for 6 h in the presence of 5 lg/ml BFA with or without 100 lM MIBG. The cells were labeled for ApoB (red) and LD (green). ApoB-crescents are indicated byarrowheads. Bars, 10 lm. (B) ADP-ribosylation inhibitors suppressed the increase of ApoB-crescents induced by BFA. Huh7 cells were treated with 5 lg/ml BFA for 6 h and theeffect of 100 lM MIBG, 100 lM coumermycin A1, and 30 mM NAM on the ratio of ApoB-crescent-positive cells was quantified. (C) MIBG suppressed the increase of ApoB-crescents induced by either ALLN or DHA. Huh7 cells were treated for 6 h with 10 lM ALLN (a) or 0.4 mM DHA (b) in the presence of 0, 25, or 100 lM MIBG. (D) MIBGsuppressed the increase of ApoB-crescents induced by simultaneous inhibition of proteasomes and autophagy. Huh7 cells were treated for 6 h with both 10 lM ALLN and10 mM 3-MA with or without 25 lM MIBG. (E) The amount of ApoB in the LD fraction decreased significantly by MIBG. Huh7 cells were cultured for 12 h in the presence of10 lM ALLN with or without 100 lM MIBG. Fractions obtained by sucrose density-gradient centrifugation were subjected to Western blotting. ApoB in the LD fraction(fraction #1) decreased significantly in cells treated with MIBG (left panel). The ADRP reactivity confirmed that equivalent amounts of LDs were loaded for the two samples(right panel). (F) MIBG decreased existing ApoB-crescents within 1 h. Huh7 cells were first treated with 10 lM ALLN for 6 h to induce ApoB-crescents, and then incubatedwith 0, 25, or 100 lM MIBG in the continual presence with ALLN.

3700 Y. Ohsaki et al. / FEBS Letters 587 (2013) 3696–3702

In agreement with this finding, the number of ApoB-crescentswas found to increase significantly in the following conditions[7–9,24]: (1) inhibition of protein degradation by proteasomes orautophagy, (2) suppression of the functionality of UBXD8 orDerlin-1, (3) inhibition of the ER chaperone related to ApoB folding,and (4) acceleration of the production of lipidated ApoB by fattyacids. In the present study, we showed that blockade of the ER-to-Golgi vesicular transport by any means increases the numberof ApoB-crescents significantly. These results support our

contention that ApoB-crescents indicate aberrant accumulation oflipidated ApoB in the ER.

However, it was unexpected that BFA at low concentrationsincreased ApoB-crescents without significantly suppressing pro-tein secretion. The ER-to-Golgi transport may be slightly af-fected even by low concentrations of BFA, but the prominentinhibitory effect of ADP-ribosylation inhibitors indicate thatBFA causes the increase of ApoB-crescents by activating ADP-ribosylation.

Fig. 4. MIBG did not affect the MTP activity. (A) Neither MIBG nor NAM inhibited the MTP activity. The in vitro fluorescence assay for the lipid transfer activity using mouseliver tissue homogenate (a, d, e), HepG2 cell lysate (b), or purified MTP (c). The effect of MIBG (25 lM or 100 lM) and NAM (10 mM or 30 mM) was examined. The validity ofthe assay was confirmed by the significant suppression of the activity in the presence of an MTPi (0.06 lM BMS200150). The assay was done using triplicate samples. (B) MTPisuppressed increases in ApoB-crescents induced by proteasome inhibition but not by DHA. Huh7 cells were treated for 6 h with 10 lM ALLN or 0.4 mM DHA with or without100 nM BAY13–9952. (C) MTPi did not affect existing ApoB-crescents. Huh7 cells were treated with 10 lM ALLN for 6 h, and then treated for an additional 1 h with or withoutBAY13-9952 in the continuous presence of ALLN.

Y. Ohsaki et al. / FEBS Letters 587 (2013) 3696–3702 3701

How do ADP-ribosylation inhibitors suppress ApoB-crescentformation? First, we suspected that ADP-ribosylation inhibitorsmay have suppressed MTP activity nonspecifically, but this pos-sibility was refuted by the in vitro MTP assay (Fig. 4A). Second,down-regulation of ApoB protein expression was not likelybecause the cellular ApoB was not changed significantly by

ADP-ribosylation inhibitors (Figs. 3E and S3B). Third, accelera-tion of protein degradation by proteasomes or autophagymay not be involved because ADP-ribosylation inhibitors sup-pressed ApoB-crescent formation even when both degradationpathways were inhibited (Fig. 3D). Fourth, acceleration ofsecretion is also unlikely, because ApoB-crescent formation

3702 Y. Ohsaki et al. / FEBS Letters 587 (2013) 3696–3702

was suppressed even in the presence of a high concentration ofBFA (Fig. 3A). Thus, ADP-ribosylation inhibition appears to sup-press ApoB-crescent formation even though lipidated ApoB waspresent in the cell.

Currently, the mechanism by which ADP-ribosylation inhibitorssuppress ApoB-crescent formation is unknown. An important clueto elucidate the mechanism is the fact that ADP-ribosylationinhibitors rapidly reduced the number of existing ApoB-crescents(Fig. 3E), an effect that was not observed with MTPi or otherreagents (Fig. 4B). This result suggests that inhibition of ADP-ribo-sylation facilitates the degradation or export of lipidated ApoB outof the ER via a non-canonical pathway. Another non-exclusivepossibility is that ADP-ribosylation inhibitors down-regulateApoB-crescents by repressing the structural and functional liaisonbetween the ER and LDs.

BFA is known to induce mono-ADP-ribosylation of two intracel-lular proteins, C-terminal-binding protein-1 (CtBP1) short-form/BFA-ADP-ribosylation substrate (CtBP1-S/BARS) and the glycolyticenzyme GAPDH [17,18,25,26]. In a previous study, LDs were shownto decrease either by BFA or CtBP1 knockdown and this decrease wassuppressed by ADP-ribosylation inhibitors [27]. This result led theauthors to speculate that transcription of the genes involved in lipidsynthesis was suppressed by the BFA-induced co-repressor activityof the long isoform of CtBP1 (CtBP1-L) [27]. The effect of ADP-ribo-sylation inhibitors that we observed in the present study is not likelyrelated to the above phenomenon, because it was evident within 1 h,which is too short of a time to affect gene transcription. Consistently,the ApoB mRNA level was not changed by MIBG, and transcriptioninhibition by actinomycin D did not suppress ApoB secretion at leastfor 6 h (Figs. S3A, C). On the other hand, the possibility remains thatCtBP1-S/BARS is involved in regulating ApoB-crescent formationwithout affecting transcription.

The present results reinforce our contention that the ApoB-cres-cent is a reliable marker for monitoring the process of lipidated ApoBaccumulation in hepatocytes. The ApoB-crescent is not limited tohepatocellular carcinoma cell lines, but is also found in human andmouse primary hepatocytes in culture ([9]; Imai et al., unpublishedobservation). The molecular mechanism of ApoB-crescent forma-tion warrants further studies because it is likely to be related tothe mechanism of intracellular ApoB degradation and the ER�LDrelationship. We expect that more molecular insights will be ob-tained by identification of novel ADP-ribosylation targets.

Acknowledgments

We thank Dr. Harris Jamil, Dr. Naoko Morinaga, and Dr. NobuhiroNakamura for reagents. This study was supported by Grants-in-Aidfor Scientific Research of the Ministry of Education, Culture, Sports,Science and Technology of the Government of Japan (to Y.O. andT.F.), a grant from Takeda Science Foundation (to Y.O.), and a Na-tional Institutes of Health Grant DK081879 (to M.M.H.).

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.febslet.2013.09.036.

References

[1] Fisher, E.A. (2012) The degradation of apolipoprotein B100: multipleopportunities to regulate VLDL triglyceride production by differentproteolytic pathways. Biochim. Biophys. Acta 1821, 778–781.

[2] Fisher, E.A., Pan, M., Chen, X., Wu, X., Wang, H., Jamil, H., Sparks, J.D. andWilliams, K.J. (2001) The triple threat to nascent apolipoprotein B. Evidence formultiple, distinct degradative pathways. J. Biol. Chem. 276, 27855–27863.

[3] Liao, W., Yeung, S.C. and Chan, L. (1998) Proteasome-mediated degradation ofapolipoprotein B targets both nascent peptides cotranslationally beforetranslocation and full-length apolipoprotein B after translocation into theendoplasmic reticulum. J. Biol. Chem. 273, 27225–27230.

[4] Olofsson, S.O. and Boren, J. (2005) Apolipoprotein B: a clinically importantapolipoprotein which assembles atherogenic lipoproteins and promotes thedevelopment of atherosclerosis. J. Intern. Med. 258, 395–410.

[5] Pan, M., Cederbaum, A.I., Zhang, Y.L., Ginsberg, H.N., Williams, K.J. and Fisher,E.A. (2004) Lipid peroxidation and oxidant stress regulate hepaticapolipoprotein B degradation and VLDL production. J. Clin. Invest. 113,1277–1287.

[6] Pan, M. et al. (2008) Presecretory oxidation, aggregation, and autophagicdestruction of apoprotein-B: a pathway for late-stage quality control. Proc.Natl. Acad. Sci. USA 105, 5862–5867.

[7] Ohsaki, Y., Cheng, J., Suzuki, M., Fujita, A. and Fujimoto, T. (2008) Lipid dropletsare arrested in the ER membrane by tight binding of lipidated apolipoproteinB-100. J. Cell Sci. 121, 2415–2422.

[8] Suzuki, M., Otsuka, T., Ohsaki, Y., Cheng, J., Taniguchi, T., Hashimoto, H.,Taniguchi, H. and Fujimoto, T. (2012) Derlin-1 and UBXD8 are engaged indislocation and degradation of lipidated ApoB-100 at lipid droplets. Mol. Biol.Cell 23, 800–810.

[9] Ohsaki, Y., Cheng, J., Fujita, A., Tokumoto, T. and Fujimoto, T. (2006)Cytoplasmic lipid droplets are sites of convergence of proteasomal andautophagic degradation of apolipoprotein B. Mol. Biol. Cell 17, 2674–2683.

[10] Athar, H., Iqbal, J., Jiang, X.C. and Hussain, M.M. (2004) A simple, rapid, andsensitive fluorescence assay for microsomal triglyceride transfer protein. J.Lipid Res. 45, 764–772.

[11] Mossessova, E., Corpina, R.A. and Goldberg, J. (2003) Crystal structure ofARF1⁄Sec7 complexed with Brefeldin A and its implications for the guaninenucleotide exchange mechanism. Mol. Cell 12, 1403–1411.

[12] Rouhana, J. et al. (2013) Kinetics of interaction between ADP-ribosylationfactor-1 (Arf1) and the Sec7 domain of Arno guanine nucleotide exchangefactor, modulation by allosteric factors, and the uncompetitive inhibitorbrefeldin A. J. Biol. Chem. 288, 4659–4672.

[13] Feng, Y. et al. (2003) Exo1: a new chemical inhibitor of the exocytic pathway.Proc. Natl. Acad. Sci. USA 100, 6469–6474.

[14] Arasaki, K., Tani, K., Yoshimori, T., Stephens, D.J. and Tagaya, M. (2007)Nordihydroguaiaretic acid affects multiple dynein-dynactin functions ininterphase and mitotic cells. Mol. Pharmacol. 71, 454–460.

[15] Tagaya, M., Henomatsu, N., Yoshimori, T., Yamamoto, A., Tashiro, Y. andMizushima, S. (1996) Inhibition of vesicle-mediated protein transport bynordihydroguaiaretic acid. J. Biochem. 119, 863–869.

[16] Dascher, C. and Balch, W.E. (1994) Dominant inhibitory mutants of ARF1 blockendoplasmic reticulum to Golgi transport and trigger disassembly of the Golgiapparatus. J. Biol. Chem. 269, 1437–1448.

[17] De Matteis, M.A. et al. (1994) Stimulation of endogenous ADP-ribosylation bybrefeldin A. Proc. Natl. Acad. Sci. USA 91, 1114–1118.

[18] Di Girolamo, M. et al. (1995) Evidence that the 50-kDa substrate ofbrefeldin A-dependent ADP-ribosylation binds GTP and is modulated by theG-protein beta gamma subunit complex. Proc. Natl. Acad. Sci. USA 92,7065–7069.

[19] Loesberg, C., van Rooij, H. and Smets, L.A. (1990) Meta-iodobenzylguanidine (MIBG), a novel high-affinity substrate for choleratoxin that interferes with cellular mono(ADP-ribosylation). Biochim.Biophys. Acta 1037, 92–99.

[20] Smets, L.A., Janssen, M., Metwally, E. and Loesberg, C. (1990) Extragranularstorage of the neuron blocking agent meta-iodobenzylguanidine (MIBG) inhuman neuroblastoma cells. Biochem. Pharmacol. 39, 1959–1964.

[21] Weigert, R. et al. (1997) Characterization of chemical inhibitors of brefeldin A-activated mono-ADP-ribosylation. J. Biol. Chem. 272, 14200–14207.

[22] Hussain, M.M., Rava, P., Walsh, M., Rana, M. and Iqbal, J. (2012) Multiplefunctions of microsomal triglyceride transfer protein. Nutr. Metab. (Lond.) 9,14.

[23] Suzuki, M., Shinohara, Y., Ohsaki, Y. and Fujimoto, T. (2011) Lipid droplets: sizematters. J. Electron Microsc. 60, S101–S116.

[24] Fujimoto, T. and Ohsaki, Y. (2006) Proteasomal and autophagic pathwaysconverge on lipid droplets. Autophagy 2, 299–301.

[25] Hottiger, M.O., Hassa, P.O., Luscher, B., Schuler, H. and Koch-Nolte, F. (2010)Toward a unified nomenclature for mammalian ADP-ribosyltransferases.Trends Biochem. Sci. 35, 208–219.

[26] Colanzi, A. et al. (2013) Molecular mechanism and functional role of brefeldinA-mediated ADP-ribosylation of CtBP1/BARS. Proc. Natl. Acad. Sci. USA 110,9794–9799.

[27] Bartz, R., Seemann, J., Zehmer, J.K., Serrero, G., Chapman, K.D., Anderson, R.G.and Liu, P. (2007) Evidence that mono-ADP-ribosylation of CtBP1/BARSregulates lipid storage. Mol. Biol. Cell 18, 3015–3025.