CHAPTER

7
Analysis of Lipid Droplets
in Hepatocytes

Huajin Wang*, Ariel D. Quiroga{, and Richard Lehner{

*
Department of Cell Biology, Yale University School of Medicine, New Haven,
Connecticut, USA
{
Instituto de Fisiologı´a Experimental, Consejo Nacional de Investigaciones Cientı
´ficas y Te´cnicas (CONICET), Facultad de Ciencias Bioquı´micas y Farmace´uticas, Universidad
Nacional
de Rosario, Rosario, Argentina
{
Group on Molecular and Cell Biology of Lipids, Departments of Pediatrics and Cell Biology,
University of Alberta, Edmonton, Alberta, Canada

CHAPTER OUTLINE
Introduction and Rationale..........................................................................................................108
7.1 Materials.......................................................................................................................111
7.1.1 Isolation of CLDs.............................................................................................111
7.1.2 Isolation of LLDs.............................................................................................111
7.1.3 Use of BODIPY Fatty Acids to Visualize CLD Dynamics..................................112
7.2 Methods........................................................................................................................113
7.2.1 Isolation of CLDs from Mouse Liver..............................................................113
7.2.1.1 Mice, Diets, and Feeding States........................................................113
7.2.1.2 CLD Fractionation................................................................................113
7.2.1.3 Solubilization of CLD-Associated Proteins for Immunoblotting.....115
7.2.2 Isolation of LLDs from Mouse Liver...............................................................115
7.2.2.1 Preparation of the Mouse...................................................................115
7.2.2.2 Isolation of the Liver and Tissue Homogenization..........................115
7.2.2.3 Isolation of Microsomes......................................................................116
7.2.2.4 Release of Microsomal Lumenal Content........................................117
7.2.2.5 Separation of LLDs from VLDL Precursors......................................117
7.2.2.6 Isolation of Microsomal LLDs by Density Gradient
Ultracentrifugation.........................................................................................117
7.2.2.7 Sample Analysis...................................................................................118

Methods in Cell Biology, Volume 116

Copyright © 2013 Elsevier Inc. All rights reserved.
ISSN 0091-679X
http://dx.doi.org/10.1016/B978-0-12-408051-5.00007-3
107
108 CHAPTER 7 Analysis of Lipid Droplets in Hepatocytes

7.2.3 Use of BODIPY Fatty Acids for Visualization of CLD Dynamics..................119

7.2.3.1 Preparation of Mouse Hepatocytes Grown on the Cover Slip or
Glass Bottom Dish.........................................................................................119
7.2.3.2 Oleic Acid and BODIPY Fatty Acid Incubations of Hepatocytes. . .120
7.2.3.3 Microscope Setup and Imaging......................................................121
7.2.3.4 Image Processing and Analysis.....................................................122
7.3 Discussion.....................................................................................................................122
Conclusion............................................................................................................................125
Acknowledgments................................................................................................................125
References.............................................................................................................................126

Abstract
The liver plays an important role in triacylglycerol (TG) metabolism. It can store
large amounts of TG in cytosolic lipid droplets (CLDs), or it can package TG into
very-low density lipoproteins (VLDL) that are secreted from the cell. TG packaged
into VLDL is derived from TG stored within the endoplasmic reticulum in lumenal
lipid droplets (LLDs). Therefore, the liver contains at least three kinds of LDs that
differ in their protein composition, subcellular localization, and function. Hepatic
LDs undergo tremendous changes in their size and protein composition depending
on the energetic (fasting/feeding) and pathological (viral infection, nonalcoholic
fatty liver disease, etc.) states. It is crucial to develop methodologies that allow
the isolation and analyses of the various hepatic LDs in order to gain insight into
the differential metabolism of these important lipid storage/transport particles in
health and disease. Here, we present detailed protocols for the isolation and
analysis of CLDs and LLDs and for monitoring CLD dynamics.

INTRODUCTION AND RATIONALE

All organisms and cell types investigated so far can store triacylglycerols (TGs),
ste- rol esters in lipid droplets (LDs). Until recently, LDs were largely considered
to be inert organelles, functioning only as lipid-storing bodies. However, it is now
clear that LDs are active and dynamic organelles that play multiple roles in lipid
metabolism, signal transduction, protein storage, and lipid trafficking (Olofsson
et al., 2009; Walther & Farese, 2012).
The liver plays a central role in lipid metabolism and storage. In addition to the
critical role of this organ in TG synthesis, storage, and provision of substrates for
b-oxidation and ketogenesis, the liver is a specialized organ that secretes TG into
the blood in apolipoprotein B (apoB)-containing very-low density lipoprotein
(VLDL) particles. Therefore, hepatocytes contain several types of LDs, each with
a specific subcellular localization and distinct protein composition. In addition to
cytosolic LDs (CLDs) that are present in most other cell types, hepatocytes contain
at least two more types of LDs in the lumen of the endoplasmic reticulum (ER) where
 Introduction and Rationale 109

VLDL is assembled: the lumenal apoB-free lipid droplets (LLDs) and the apoB-
containing particles (VLDL and its precursors). LLDs have been proposed to
provide TG source for the bulk lipidation of VLDL precursors. Because the hepatic
CLDs, LLDs, and VLDL are important contributors to whole body energy
homeosta- sis, it has become important to characterize these intracellular TG
storage entities. The most challenging issue in characterization of CLDs, LLDs, and
the nascent VLDL is a reliable method to separate these very distinct LDs. Added
complication is the rap- idly changing composition of LD-associated proteins with
various metabolic states of the hepatocyte (feeding/fasting, stress induced by viral
infection, etc.). It is therefore important to take into consideration these factors
when studying hepatic LDs.
CLDs are composed of a lipid core, mainly TG, cholesterol esters (CE), and
retinyl esters, surrounded by an amphipathic lipid monolayer (phospholipids and free
cholesterol) decorated with LD-associated proteins including the PAT (Perilipin1,
ADRP/Perilipin2, TIP47/Perilipin3) protein family (Bickel, Tansey, & Welte,
2009; Farese & Walther, 2009; Kuhnlein, 2012; Yang et al., 2012). Hepatic CLDs
also contain numerous proteins that are also found on CLDs in adipocytes and
other cell types, such as ER-resident proteins, Rab GTPases, and cytoskeleton
components, indicating that LDs might share similar regulation in all cell types.
To date, only a few studies have been performed investigating protein composition
of hepatic CLDs. One of these studies was performed in rat liver following partial
hepatectomy, where 50 proteins were identified, including perilipin 2, ER-resident
proteins, lipid and vitamin metabolism enzymes, cytoskeletal components, cell
signaling and cell activation regulation proteins, and a number of proteins that par-
ticipate in diverse intracellular trafficking pathways and exocytosis (Turro et al.,
2006). Another work performed in the HuH7 hepatoma cell line identified 17 pro-
teins, including perilipin 2 and lipid and steroid metabolism enzymes as the most
abundant proteins (Fujimoto et al., 2004). Except for the presence of PAT protein
family, hydroxysteroid dehydrogenases, and Rab5 GTPase, no similarities in
protein contents between the two LD compositions were observed, possibly reflect-
ing the differences between human hepatoma cell line (HuH7) and the liver, and/or
differences in the metabolic state of the cells. Lack of overlap between the
proteomes published by different groups also to a certain degree reflects the
various degree of contamination present in all subcellular organelles isolated by
density centrifugation. Even though a large amount of contaminants are
removed while LDs partition across layers of a density gradient, it is impossible to
remove all, especially since most common contaminations arise from
hydrophobic proteins nonspecifically interacting with LDs, as well as from
abundantly expressed proteins. It is also likely that some LDs are in the
continuum with the ER (the site of LD origin), thus a small amount of the ER
proteome may copurify with LDs.
Similarly, studies investigating hepatic LLDs are limited, mainly because they
have proved very challenging. Light microscopy is not suitable for observing LLDs
since their size is within or below the range of lipoproteins (ranging from 7 to
200 nm) and is under the detection limit of conventional light microscopy. Even
with the state-of-the-art super resolution microscopy techniques, it is difficult to
110 CHAPTER 7 Analysis of Lipid Droplets in Hepatocytes

distinguish LLDs from presecretory VLDL or VLDL precursor particles. The only
reported success in observing LLDs was by immunogold electron microscopy, where
they were identified as VLDL-sized particles in the smooth ER lacking immunode-
tectable apoB (Alexander, Hamilton, & Havel, 1976). Subsequently, chylomicron-
sized LLDs were detected in the ER of enterocytes lacking apoB expression,
further supporting this observation (Hamilton, Wong, Cham, Nielsen, & Young,
1998). No further progress has been made to provide additional evidence for the
presence of these LDs. We have developed a method to purify LLDs by subcellular
fractionation and to biochemically characterize protein and lipid properties of the
isolated LLDs, thus providing an approach to biochemically study the formation
and metabolism of LLDs. The purification of LLDs will become a powerful tool
to study events involved in the lipidation step of VLDL assembly. LLDs were
found to be heterogeneous in size, possibly reflecting the various states of synthe-
sis/turnover (Wang, Gilham, & Lehner, 2007). Whereas in mouse CLDs, TG
accounts for up to 80% of total lipids, phospholipid for about 15%, and cholesterol
with CE for the remaining 5%, LLDs contained on average a lower percentage of
TG (60%) and a higher percentage of phospholipid (25%) (Wang et al., 2007); lipid
ratios similar to those found in mature VLDL. Proteomic analysis of LLDs
revealed the presence of microsomal TG transfer protein (MTP), protein disulfide
isomerase (PDI), apolipoprotein E (apoE), and several other ER-resident proteins
including two members of the carboxylesterase family, carboxylesterase
3/triacylglycerol hydrolase (Ces3/TGH, Ces1d) and carboxylesterase 1/esterase-x
(Ces1/Es-x, Ces1g) (Wang et al., 2007).
Despite the important functions LDs serve in many cellular processes, our
knowledge of the cell biology of LDs lacks behind that of other intracellular organ-
elles. It is generally believed that LD biogenesis in eukaryotes initiates from the ER
where TG biosynthesis takes place. However, little is known about the mechanism by
which nascent LDs accrue additional TG and grow in size after nascent formation.
This gap is in part due to the lack of good tools to visualize the flux of lipids.
Traditional lipophilic dyes such as BODIPY 493/503, Nile Red, and LD540 are ex-
cellent tools to visualize the morphology of already formed CLDs by fluorescence
microscopy; however, they do not allow tracking the dynamics of initial LD
formation and cannot distinguish different pools of LDs at the different stages of
biogenesis. Thus, a method is needed that distinguishes between preformed and
newly synthesized CLDs. Additionally, CLD formation is a rapid process that
occurs almost immediately (within 15 min) after OA addition (Wang et al., 2010).
Thus, real-time cell imaging is necessary to capture this short window during
CLD formation.
Many lipid analogues have been developed in the past decade, including a
variety of fluorescent fatty acid analogues, thus enabling the tracking of initial lipid
incorporation into CLDs with real-time microscopy. These analogues include, but
are not limited to, NBD-conjugated fatty acids (Chattopadhyay, 1990), BODIPY-
conjugated fatty acids (Pagano, Martin, Kang, & Haugland, 1991),
 7.1 Materials 111

polyene-fatty acids (Kuerschner et al., 2005), etc. With these tools in hands, we are
able to start answering important questions regarding the mechanism of CLDs bio-
genesis, such as the formation of nascent CLDs and addition of lipids to preformed
CLDs.
This protocol covers the isolation of CLDs and LLDs from the liver of fasted
C57BL/6J mouse maintained on chow diet. Analysis protocols related to the char-
acterization of CLDs and LLDs and visualization of CLD dynamics will also be
briefly discussed.

7.1 MATERIALS
7.1.1 Isolation of CLDs
Reagents:
1. Phosphate-buffered saline (PBS): NaCl 137 mM, KCl 2.7 mM,
Na2HPO4 2H · 2O 10 mM, KH2PO4 2.0 mM, pH 7.4
2. Hypotonic medium (HM): 20 mM Tris–HCl, pH 7.4, 1 mM EDTA. Prepare
fresh, keep refrigerated
3. 60% sucrose-HM: HM supplemented with 60 g/100 ml sucrose. Prepare fresh,
keep refrigerated for no more than 1 day
4. 5% sucrose-HM: HM supplemented with 5 g/100 ml sucrose. Prepare fresh,
keep refrigerated for no more than 1 day
All hypotonic media contain protease and phosphatase inhibitors at their optimal
concentrations.
5. Sodium dodecyl sulfate (SDS) 10% (w/v)
Other materials and equipment:
6. Plastic tubes
7. Eppendorf tubes (1.5 ml)
8. Low-speed refrigerated centrifuge with appropriate centrifuge tubes
9. Ultracentrifuge with swinging-bucket rotor
10. Thin-walled polycarbonate ultracentrifuge tubes
11. Reagents and equipment for SDS–PAGE and immunoblotting

7.1.2 Isolation of LLDs

Reagents:
1. Tris-buffered saline (TBS): 20 mM Tris–HCl, 150 mM NaCl
2. Homogenizationbuffer: 250 mMsucrose, 20 mMTris–HCl, pH 7.4, 1 mMEDTA
3. Washing buffer: 20 mM Tris–HCl, 0.5 M NaCl (pH 7.4)
4. 1 mM Tris–HCl, pH 8.8
5. Glycerol
6. Protease inhibitor cocktail (Roche)
112 CHAPTER 7 Analysis of Lipid Droplets in Hepatocytes

7. Goat anti-apoB antibody (Chemicon)

8. 4 ×SDS–PAGE sample buffer: 200 mM Tris–HCl, pH 6.8, 8% SDS, 40%
glycerol (v/v), 40% b-mercaptoethanol (v/v), 0.4% Bromophenol Blue

Other materials and equipment:

9. Surgical tools for removing the mouse liver

10. Motor-driven Potter–Elvehjem tissue homogenizer with loose-fitting Teflon
pestle (Wheaton), 15 ml capacity
11. Table-top refrigerated centrifuge with swinging-bucket rotor
12. Beckman ultracentrifuge with SW41Ti swinging-bucket rotor
13. 15 ml Falcon conical bottom tubes
14. Beckman 13.2 ml Ultra-ClearTM or thin-wall polyallomer ultracentrifuge tubes

Additional reagents and equipment for:

15. SDS–PAGE and immunoblotting, lipid extraction, thin-layer chromatography

(TLC) plates (lipid analysis), and fast protein liquid chromatography (FPLC)
and Superose 6 (gel filtration column for determining particle size).

7.1.3 Use of BODIPY fatty acids to visualize CLD dynamics

Reagents:

1. Collagen solution from calf skin (type I): 0.1% solution in 0.1 M acetic acid
2. PBS: 137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 2 mM KH2PO4, pH 7.4
3. Bovine serum albumin (BSA): essentially fatty acid free (Sigma)
4. Hepatocyte culture media: Dulbecco’s modified Eagle’s medium (DMEM),
10% FBS (heat inactivated at 56 ◦C for 30 min), 100 U/ml penicillin, and
100 mg/ml streptomycin
5. 20 ×oleic acid complexed to BSA (OA/BSA): 7.5 mM OA, 10% BSA,
dissolved in DMEM (see Section 2.3.2. for protocol)
6. Labeling medium A: DMEM, 1 × OA/BSA, 6 mM BODIPY FL C12 or BODIPY
558/568 C12 (Molecular Probes)
7. Labeling medium B: QTB fatty acid uptake reagent (Molecular Devices)
reconstituted in 10 ml DMEM containing 1 ×OA/BSA
8. BODIPY 493/503 dye: 1 mg/ml in DMSO (Molecular Probes)

Other materials and equipment:

9. Spinning-disk confocal microscope with 488 and 543 nm laser or similar

10. Environment chamber for live-cell imaging with temperature control and CO2
supply
11. Stage adaptor and culture chamber for live-cell imaging
12. Volocity software for image capture and processing
 7.2 Methods 113

7.2 METHODS
7.2.1 Isolation of CLDs from mouse liver
Before we start: all samples and solutions throughout the preparation of LDs
should be kept on ice, and make sure that protease inhibitor cocktail is freshly
added to all buffers used during the preparation.

7.2.1.1 Mice, diets, and feeding states

In order to have a successful (purer) CLD fractionation, the amount of lipid present
in the liver should be considered before starting any LD isolation procedures. As a
general rule, the larger amount of lipid in the tissue, the higher the chance of con-
tamination of LDs with other organelles. Therefore, three key interrelated issues
must be taken into account before isolating CLDs from mouse livers: (i) the diet,
(ii) the fasting time, and (iii) the amount of TG present in the tissue. First, high-fat
diets increase the amount of TG in the liver but also make the tissue much more
brittle thereby increasing the likelihood of obtaining contaminated LD fractions.
On the other hand, low-fat diets may result in livers with less TG, but CLDs
become easier to isolate since the chances of contamination with other intracellular
mem- branes are minimized. Second, and equally important, is the feeding state of
the animals: it is known that fasting induces the accumulation of hepatic TG and
there- fore CLD number, size, and also protein composition vary depending on the
feed- ing states of the animals. Fasting times longer than 24 h are not
recommended for the mouse, since the animal enters into a starvation mode at such
a prolong fasting period, and the proteome does not represent a normal
physiological state. On the other hand, we determined that refeeding times (after a
24 h fast) should be no lon- ger than 6 h if one is to study CLD composition in
animals with defined nutritional states. This is as a mean of synchronicity: after a
24 h fast, animals are eager to eat, thus all mice feed at a fairly constant pace for at
least 5 h. From 6 h on, mice start to eat less and at a different pace from each other,
which has an impact on CLD size, number, and protein/lipid composition.
Refeeding times shorter than 6 h might re- sult in a significant overlap in CLD
protein composition between fasting and refeeding states. Third, the amount of
lipid present in the liver may be first eval- uated taking into account the genetic
background of the animal, since defined liver lipid phenotypes have been described
for different strains of mice. For instance, Ces1—/— mice present with obesity and
mild steatosis (Quiroga et al., 2012). Sim- ilarly, virus-induced knockdown of Atgl
also results in hepatic steatosis (Ong, Mashek, Bu, Greenberg, & Mashek, 2011).

7.2.1.2 CLD fractionation

CLDs are isolated by methods developed by Brasaemle and Wolins (2006) with
some modifications specific to the liver tissue (Table 7.1). Mice should be anesthe-
tized by inhalation with isoflurane and exsanguinated by cardiac puncture. Quickly
but carefully, livers are harvested, perfused with ice-cold PBS, and cleaned from
any
114 CHAPTER 7 Analysis of Lipid Droplets in Hepatocytes

Table 7.1 Lipid Droplet Isolation from Mouse Livers

Harvest livers quickly and carefully
Clean livers from any other tissues
Rinse livers thoroughly in ice-cold PBS
Homogenize livers by soft homogenization in HM on ice
Transfer homogenates to 15 ml tubes

6. Centrifuge at 500 × g for 10 min at 4 ◦C

7. Spin the supernatant at 15,000 × g for 10 min at 4 ◦C
8. Recover fat cakes into new clean tubes
9. Wash twice at 15,000 × g for 10 min at 4 ◦C
Dilute fat cakes with 1/3 volume of ice-cold HM containing 60% sucrose (final
concentration 20% sucrose)
Disperse aggregates of LDs by gently pipetting
Layer the diluted aggregates carefully at the bottom of ultracentrifuge tubes
Overlay these tubes with double the volume HM containing 5% sucrose
Carefully overlay the tubes with same volume of HM
15. Ultracentrifuge tubes at 28,000 × g for 30 min
Recover LDs carefully (use a pipet or a slicer) into Eppendorf tubes
Analysis
(modified from )

leftover adipose tissue. Individual livers are transferred to a beaker containing ice-
cold PBS and are thoroughly rinsed. Then, livers are weighed and immediately
sub- jected to soft homogenization on ice in HM (20% homogenate, w/v) by 10
gentle strokes with a motor-driven Potter–Elvehjem homogenizer set at a speed of
3. This step is crucial to preserve LD integrity and to prevent damage to
intracellular organ- elles (such as the ER, which would otherwise leak lumenal
contents including VLDL and LLDs that would contaminate CLD preparations).
◦
Homogenates are transferred to 15 ml tubes and × centrifuged 10 min at 500 g, 4 C.
The supernatant containing the floating fat layer (fat cake)
× is then spun at 15,000 g
for 10 min to remove mito- chondria and to further allow fat cake separation. Fat
cakes are then recovered into new tubes and washed×twice with ice-cold HM at
15,000 g for 10 min. Recovered fat cakes are then diluted with 1:3 (v/v) of ice-cold
HM containing 60% sucrose to yield a final 20% sucrose-adjusted homogenate.
Aggregates of LDs should be finely and thoroughly dispersed by gentle pipetting.
It is highly recommended to use a pipet tip with a wide opening. Diluted CLD
aggregates are now carefully layered at the bot- tom of ultracentrifuge tubes and
overlayed with double the volume HM containing 5% sucrose. These tubes are now
carefully overlayed with same volume of HM. Tubes should be ultracentrifuged at
×
28,000 g for 30 min. After ultracentrifugation, fat cakes are carefully recovered
(using a pipet or a slicer) and kept in Eppendorf tubes for fur- ther analysis. This
step could be repeated after further dilution of recovered fat cake with 60% sucrose
in order to get a highly enriched (purified) CLD fraction.
After isolation, recovery and purity of LDs should be evaluated by SDS–PAGE
followed by immunoblotting (see below how to solubilize CLD-associated proteins).
 7.2 Methods 115

Known CLD proteins should be immunoblotted as controls using specific antibodies.

For this purpose, perilipin 2 is an excellent marker for CLD recovery in the liver.
In order to determine the purity of the preparation, cytosolic proteins and resident
poly- topic membrane proteins from other organelles such as the ER and
mitochondria should be analyzed by immunoblotting with specific antibodies.
Polytopic ER mem- brane proteins (such as phosphatidylethanolamine-N-
methyltransferase) are good markers for the absence of ER contaminations from
CLDs. Likewise, fumarase, with mitochondrial and cytosolic localizations, is a
reliable marker for contamination of CLDs with mitochondrial components.

7.2.1.3 Solubilization of CLD-associated proteins for immunoblotting

In order to solubilize CLD-associated proteins, fresh CLD fractions should be mixed
with 10% SDS (1:1, v/v) and incubated for 1–2 h at 37 ◦C in a sonicating water bath
with con- stant agitation. Then, samples should be centrifuged in a microcentrifuge
for 10 min at maximum speed at room temperature and the infranatants containing
the solubilized pro- teins should be collected from beneath the floating lipid layer. For
this purpose, it is very helpful to use a 200 ml tip inserted very carefully through the
floating fat cake all the way down to the bottom of the tube. It is important not to
disrupt the lipid cake; should this happen, recentrifugation should be performed.
Transfer the infranatant to a 1.5 ml× tube and add equivalent volumes of 2 SDS
electrophoresis sample buffer. Then, boil sam- ples for 10 min and finally load them
onto a discontinuous SDS–PAGE gel.

7.2.2 Isolation of LLDs from mouse liver

7.2.2.1 Preparation of the mouse
To prepare mice for LLD isolation, we fast mice overnight (12–16 h) before exper-
iments because under fasting conditions, large amounts of LDs are accumulated in
the liver. Instead of simply removing food from the mice, fasting should be done
by transferring mice into a clean cage with free access to water. Usually up to four
livers are used for an experiment. If the study requires comparing different feeding
condi- tions, then much more material is required. The following method is
performed using fasted mice.

7.2.2.2 Isolation of the liver and tissue homogenization

Anesthetize the mice to a state suitable for surgery by inhaling of metophane or
iso- flurane. Vital signs should be monitored during this procedure. Exsanguinate
the mice via cardiac puncture and carefully remove the liver, which is immediately
transferred to ice-cold TBS (Table 7.2). After rinsing with TBS, livers are weighed
and homogenization buffer is added to make a final 20% (w/v) homogenate. For
homogenization, we use a motor-driven Potter–Elvehjem homogenizer (Wheaton
Science Products, Millville, NJ) at medium speed (levels 3–4) for 10 strokes in ho-
mogenization buffer. Proper homogenization is one of the most important steps for
the preparation in order to efficiently break cells while maintaining the integrity of
subcellular content. Alternative devices such as Polytron or loose-fitting Dounce
116 CHAPTER 7 Analysis of Lipid Droplets in Hepatocytes

Table 7.2 Isolation of Lumenal Lipid Droplets from Mouse Liver

Harvest livers quickly and carefully
Clean livers from any other tissues
Rinse livers thoroughly in ice-cold TBS
Gently homogenize livers with Potter homogenizer in homogenization buffer on ice to make 20% homogenate
Transfer homogenates to 15 ml tubes

6. Centrifuge at 500 × g for 10 min at 4 ◦C

7. Spin the supernatant at 15,000 × g for 10 min at 4 ◦C
8. Spin the supernatant at 106,000 × g for 1 h at 4 ◦C
9. Resuspend the pellet in ice-cold washing buffer
10. Spin at 106,000 × g for 1 h at 4 ◦C and collect pellet (microsomes)
Resuspend microsomes in 1 mM Tris–HCl (pH 8.8)
Incubate on ice for 30 min
Gently homogenize to disrupt microsomes and release lumenal content
14. Spin at 106,000 × g for 1 h at 4 ◦C and collect supernatant
Adjust supernatant to 50 mM Tris–HCl, 150 mM NaCl (pH 7.4)
Add anti-apoB polyclonal antibodies
Rotate end-over-end overnight at 4 ◦C
Add protein A Sepharose beads
Rotate end-over-end for at least 2 h at 4 ◦C
20. Pellet beads by centrifugation for 30 s at 6000 × g and collect supernatant (post-IP)
Combine 2 ml of post-IP with an equal volume of glycerol
Transfer mixture to a Beckman 13.2 ml Ultra-ClearTM tube
Sequentially overlay with 4 ml of homogenization buffer and 4 ml of TBS
24. Centrifuge in Beckman SW40 rotor at 35,000 rpm (160,000 × g) for 2 h at 8 ◦C
Carefully puncture the bottom of the tube to collect fractions
Collect 2 ml per fraction for a total of six fractions for analysis
Analyses

homogenizer can also be used but conditions should be optimized carefully. In

order to monitor the proper separation, recovery, and integrity of subcellular
fractions, an aliquot should be saved for each of the following purification steps.
These aliquots will include: cytosol, microsomes, microsomal membrane,
immunoprecipitate (IP), post-IP supernatant, microsomal lumen, and fractions #1–
6.

7.2.2.3 Isolation of microsomes

Microsomes from C57BL/6J mouse liver homogenates are prepared by sequential
cen- trifugation essentially as previously described (Lehner & Kuksis, 1993).
Remove cel- lular debris by centrifugation
× at 500 g for 10 min in a table-top
centrifuge and the resulting supernatant is centrifuged subsequently
× at 15,000 g
for 10 min to pellet crude mitochondria.
× The 15,000 g supernatant contains
microsomes and cytosol (plus CLDs), which can be separated by ×
ultracentrifugation at 106,000 g for 1 h. The supernatant (cytosol plus CLDs)
from this step is transferred into a new tube for further analysis and pellet
(microsomes) is resuspended in washing buffer. To resuspend the pellet, we use
the Potter homogenizer to gently disrupt the pellet before pipetting up and down
since the pellet tends to stick to pipette tips. Washing is an
 7.2 Methods 117

important step to remove proteins and CLDs peripherally associated with the
micro- somal membrane. Microsomal membranes are pelleted by
ultracentrifugation
× at 106,000 g for 1 h, while peripheral proteins (cytosolic side)
remain in the supernatant.

7.2.2.4 Release of microsomal lumenal content

To release microsomal lumenal contents, pelleted microsomes are resuspended in
1 mM Tris–HCl (pH 8.8) in a volume equivalent to 1/3 of the original homogeniza-
tion buffer and incubate on ice for 30 min. This step swells microsomes by osmotic
pressure and preserves the integrity of LLDs and the associated proteins. At the
end of the incubation, the swelled microsomes are disrupted by homogenization
with Potter by three strokes at low setting to release lumenal contents. The mixture
is then centrifuged at 106,000 g for 1 h to obtain lumenal contents (supernatant)
×
and microsomal membranes (pellet).

7.2.2.5 Separation of LLDs from VLDL precursors

At least two types of LDs are present in the microsomal lumen: the apoB-
containing particles (including mature VLDL and VLDL precursors) and the apoB-
free LLDs. To remove apoB-containing particles from LLDs, the microsomal
lumenal contents obtained from the previous step are adjusted to 50 mM Tris–HCl,
150 mM NaCl (pH 7.4), and apoB-containing particles are immunoprecipitated
using anti-apoB polyclonal antibodies (goat anti-apoB from Chemicon). It is
important that the im- munoprecipitation is performed in the absence of detergents
in order to preserve the integrity of LLDs. Two forms of apoB are present in
the rodent liver: apoB48 ( 250 kDa) and apoB100 ( 500 kDa). The antibodies we
~ recognize both forms of ~
use apoB. For each milliliter of lumenal content, usually 5
ml of antibodies are added and the mixture is incubated while rotating the tubes
end-over-end at 4 ◦C overnight. Then 20 ml protein A Sepharose beads prewashed
with PBS are added and incubated with the mixture end-over-end at 4 ◦C for 2 h to
form protein–bead complex, which is subsequently pelleted by centrifugation for
×
30 s at 6000 g. The IP is then washed
3 with
× PBS and prepared for SDS–PAGE and immunoblotting. The supernatant
(post-IP) is collected for further purification, which is described below.

7.2.2.6 Isolation of microsomal LLDs by density gradient

ultracentrifugation
Two milliliters of post-IP supernatant are combined with an equal volume of
glycerol and transferred to a transparent ultracentrifuge tube suitable for SW40
swing bucket rotor. We prefer the Beckman Ultra-ClearTM centrifuge tube, since it
allows for good inspection of the gradient set up. The glycerol-adjusted samples
are then overlayed with 4 ml of homogenization buffer and an additional layer of 4
ml of TBS. The gra- dient should be set with great care. We find using a long
needle of 18G or narrower is helpful to gently deliver overlay solutions without
disturbing the lower layer. Once the gradient is set, the samples are centrifuged in a
Beckman SW40 rotor at 35,000 rpm (160,000× g) for 2 h at 8 ◦C. Be sure to use a
swinging bucket and avoid fixed angle
118 CHAPTER 7 Analysis of Lipid Droplets in Hepatocytes

rotors for this step since LLDs tend to stick to the wall of the centrifuge tube and
thus cause cross-contamination in the various gradient fractions. Immediately after
centri- fugation, the bottom of the tube is carefully punctured with a fine needle and
fractions are collected into collection tubes. Alternatively, a tube slicer can be used
to collect fractions and minimize carry over from other fractions. Two milliliters
per fraction are collected for a total of six fractions, labeled fractions #1–6, with #6
being the top (most buoyant) fraction. The fractions, together with aliquots from
the various steps of preparation, are then used for analysis for their biochemical
compositions and physical prosperities. The following are examples of analyses we
performed.

7.2.2.7 Sample analysis

7.2.2.7.1 Analysis of protein composition by SDS–PAGE and
immunoblotting
After purification, samples should first be analyzed for recovery and contamination
by immunoblotting. We use ER-resident proteins such as PDI as markers for recov-
ery of ER lumenal content, and immunoblot for apoB to monitor the degree of con-
tamination of LLDs with apoB-containing VLDL precursors. SDS–PAGE and
immunoblotting procedures are similar to standard procedure used for most
proteins; however, it is important that 2–5% SDS (final concentration) should be
included to dissolve lipids. The sample buffer we use for immunoblotting of LLD-
associated proteins is listed in Section 1. Due to the low abundance of LLDs and
large volume collected from the density gradient, fractions #4–6 are concentrated
15-fold using centrifugation-based protein concentrators (Pierce) with molecular
weight cutoff of 10 kDa. By comparing proteins enriched in each fraction obtained
from the gra- dient with aliquots saved during the preparation, fractions #4–6 are
LLD fractions with distinct associated proteins. An ER lumenal lipase,
Ces3/TGH, can be used as a marker for fraction #4 (corresponding to smaller
LLDs), and apoE for fraction #6 (corresponding to larger LLDs).
Fractions can be analyzed for protein and lipid compositions using established
methodologies described elsewhere.

7.2.2.7.2 Analysis of LD particle size by FPLC and native PAGE

These analyses are performed in combination with protein and lipid composition
analysis. Information obtained from these analyses is useful not only to understand
the physical property of LLDs, but more importantly, as a read out for protein and
lipid compositions associated with the various subsets of LLDs. This information is
especially beneficial when comparing between different experimental conditions,
such as genetic backgrounds, nutritional states, or drug treatments.
Gel filtration chromatography. Combined LLD fractions are concentrated and
ap- plied to a Superose 6 size exclusion FPLC column (Pharmacia, Uppsala,
Sweden). Elu- tion of various TG-rich LLD subfractions is determined by on-line
detections of TG content. Specifically, eluted fractions are mixed in-line with the
InfinityTM Triglycer- ide Reagent (Thermo Fisher Scientific, Inc., Waltham, MA)
using a post-column T-connector/Solvent Delivery Module (model 110B, Beckman
Coulter, Mississauga,
 7.2 Methods 119

Ontario, Canada) and passed through a CH-30 Column Heater (Eppendorf, Missis-
sauga, Ontario, Canada) set at 37 ◦C. Reaction products are monitored at 500 nm in
real time using a Programmable Detector Module (model 166, Beckman Coulter).
Un- der this setup, fractions that elute from the 22nd to 25th minute contain
VLDL-sized particles and fractions that elute from the 38th to 53th minute contain
HDL-sized par- ticles. To analyze protein contained in different fractions, we collect
one fraction every 4 min (2 ml) from the 22nd to 58th minute, precipitate proteins
with 2 volumes of ice- cold—acetone for 30 min at 20 ◦C, resuspend the protein pellet in
50 ml of SDS–PAGE sample buffer, and analyze the protein content by SDS–PAGE
and immunoblotting. Several reference proteins, including Ces3/TGH, MTP, and
apoE, can be used to eval- uate successful separation of LLDs of varies size. We
found that Ces3/TGH is present in eluent from 26 to 58 min and peaks around 42 to
58 min, apoE is eluted at around 26–38 min, while MTP elutes later than apoE, at
around 42–58 min.
Gradient native PAGE. Twenty five microliter aliquots of isolated LLDs are mixed
with an equal volume of 2 native
× PAGE loading buffer and applied to a 2–10% gra-
dient nondenaturing polyacrylamide gel, casted with a gradient maker (Hoefer,
Inc.). Proteins are resolved in native PAGE running buffer without SDS at 80–120
V using a BioRad Mini-PROTEAN system. Particle size is determined by
comparison with the migration of purified lipoprotein standards listed in the table
below:

Size Category Molecular Mass (kDa) Diameter (nm)

HDL1 440–669 12.2–17.0
HDL2 232–440 10.5–12.2
HDL3/preb1 66–232 7.1–10.5

7.2.3 Use of BODIPY fatty acids for visualization of CLD

dynamics (Table 7.3)
7.2.3.1 Preparation of mouse hepatocytes grown on the cover slip
or glass bottom dish
This protocol uses primary mouse hepatocytes isolated by collagenase perfusion of
the mouse liver. Details of the preparation can be found in previous publications
(Yao & Vance, 1988). We use fasted mice. The isolation should be performed im-
mediately before each experiment since cultured primary hepatocytes gradually
lose their metabolic properties, such as expression of some key enzymes involved
in lipid metabolism. Hepatocytes should not be older than 72 h and preferably used
within 48 h after seeding (Table 7.3). We seed 0.2 106 cells for each well in a six-
×
well plate. For proper imaging of cells by confocal microscopy, cells should be
plated on glass cover slips with thickness compatible with the microscope
objectives to be used. If a microscope has a culture chamber or adaptor for live-cell
imaging, cells can be seeded onto a regular round cover slip placed at the bottom of
a six-well plate;
120 CHAPTER 7 Analysis of Lipid Droplets in Hepatocytes

Table 7.3 Visualization of Lipid Droplet Dynamics using BODIPY Fatty Acids
1. Coat cover slips with collagen solution and place in a six-well plate
6
2. Seed 0.2 × 10 freshly prepared hepatocytes per well
Incubate cells in serum-free DMEM overnight
Incubate cells with labeling medium A containing 558/568 C12 for at least 4 h
Setup microscope for imaging with 491 and 543 nm laser lines
Transfer cover slips onto a culture chamber placed on the microscope in an environment chamber thermostated at 37 ◦C a
Find fields of interest and record stage location
Carefully rinse once with PBS without disturbing stage locations
Carefully add labeling media B (time 0)
Immediately start image acquisition every 1 min over a period of 30 min with Z-slices of
0.5 mm steps

the diameter of the cover slip should be suitable for the adaptor. Otherwise, a glass
bottom culture dish (e.g., MatTek dishes) can be used instead. To coat cover slips
or glass bottom dishes with collagen, pipette collagen solution (Sigma) into wells
con- taining the cover slip, let the solution sit briefly, then remove the collagen
solution, and rinse wells twice with PBS. Collagen solutions can be reused if kept
sterile. Cells are then maintained in hepatocyte culture media at 37 ◦C in humidified
air containing 5% CO2 for at least 4 h to allow attachment of the cells to the
collagen matrices.

7.2.3.2 Oleic acid and BODIPY fatty acid incubations of hepatocytes

It is recommended that endogenously preformed CLDs should be reduced as much as
possible by incubating cells grown on collagen-coated cover slips in serum-free
DMEM overnight. Even though preformed CLDs always exist in hepatocytes iso-
lated from wild-type mice, this step helps reduce the stored TG pool and reduce
the background.
To visualize preformed LDs using exogenous fatty acid source, the intracellular
TG storage is first augmented by incubations with OA for at least 4 h. OA is com-
plexed to BSA to avoid lipotoxicity induced by free fatty acids. To make a 20×OA/
BSA complex, dissolve 5 g of FA-free BSA in 50 ml DMEM and warm up the
mix- ture to 56 ◦C in a water bath. Weigh out 106 mg OA in a glass beaker (100
ml) using analytical balance and warm up at 56 ◦C water bath for 2 min. This step
should be done immediately before adding BSA in order to minimize oxidation.
Add BSA so- lution to the warmed up OA and stir vigorously using a stirring bar
for 2 min. The solution should clear, an indication that OA has been complexed
with BSA. Filter- sterilize the solution while it is still hand-warm and store at 4
◦
C.
Labeling of preformed LDs is performed by mixing the red fluorescent fatty
acid analogue, BODIPY 558/568 C12, during OA loading (labeling medium A).
After 4 h incubation, cells are washed with PBS and incubated with labeling media
containing the fluorescent green fatty acid analogue, BODIPY FL C 12. For imaging
and quan- tification of nascent LD formation under real-time conditions, we use
labeling media B for this step since a quencher is included in the labeling
medium B to absorb
 7.2 Methods 121

background fluorescent signals in the media so that only signals from the fluoro-
phores incorporated into LDs are collected. For this purposes, labeling medium B
should be added into the culture chamber on the microscope, immediately before
image capture (see Section 2.3.3).
In principle, the choice of fatty acid analogues can be reversed, that is,
BODIPY FL C12 for preformed and BODIPY 558/568 C12 for nascent CLDs. We
have demonstrated by both microscopy and TLC that these analogues are
similar in their capabilities to incorporate into TG/LDs. However, we prefer using
BODIPY FL C12 for nascent LD formation in live-cell imaging for the following
reasons: (1) compared to BODIPY 558/568 C 12, BODIPY FL C12 gives brighter
signal and sharper image; (2) as a consequence, lower laser power can be used
for live-cell imaging, which is crucial to keep cells healthy during imaging and
for avoiding artifacts in quantification due to bleaching of the fluorophore; (3)
chemical quencher for BODIPY FL C12 is readily available in the commercial kit
mentioned above.

7.2.3.3 Microscope setup and imaging

The spinning-disk confocal microscopes suitable for this experiment should be
equipped with green and red laser lines for excitation of the fatty acid analogues.
It is essential to have a temperature-controlled environment chamber. It is
preferable to have CO2 supply; however, cells can be maintained in medium
containing 50 mM HEPES for at least 30 min without CO2. Samples should be
secured to the stage with an adaptor that completely restricts movement of the
samples when reagents are added during imaging. For this reason, a culture
chamber is preferable to a glass bot- tom dish, as the latter usually does not fit
tightly to the stage adaptor. It is also ex- tremely helpful to have multistage control
and autofocusing to image multiple cells during the same imaging session. Taking
Z stacks of cells is helpful to bypass focus fluctuations during long-term imaging.
For our setup, cells grown on cover slips are mounted onto a culture chamber
(Chamlide, Seoul, Korea) and placed in an environment chamber thermostated at
37 ◦C and supplied with 5% CO2. Labeling media B is carefully added to cells in
the culture chamber immediately before image capture (time 0). Confocal micros-
copy is performed on a spinning-disk microscope (WaveFx from Quorum Technol-
ogies, Guelph, Canada) setup on an Olympus IX-81 inverted stand (Olympus,
Markham, Canada). Images are acquired through a 60 ×objective (N.A. 1.42) with
an EMCCD camera (Hamamatsu, Japan). Fluorescent fatty acid analogues
BODIPY FL C12 and BODIPY 558/568 C12 are successively excited by a 491 nm
(GFP chan- nel) and a 543 nm (Cy3 channel) laser line (Spectral Applied Research,
Richmond Hill, Canada), respectively. Z-slices of 0.5 mm steps are acquired using
Volocity soft- ware (Improvision) through the cells using a piezo z-stage (Applied
Scientific Instru- mentation, Eugene, USA) with image capture every 1 min over a
period of 30 min. Quantification of fluorescent intensity is done using Volocity
software (Ver. 5.0.0) (see below).
122 CHAPTER 7 Analysis of Lipid Droplets in Hepatocytes

7.2.3.4 Image processing and analysis

Image analysis can be performed by a variety of software, including open architec-
ture software, such as Image J and CellProfiler, commercial imaging software,
such as MetaMorph and Volocity, or a high-level language for data analysis and
visual- ization, such as MATLAB. Each of the abovementioned software has its
own strength and weakness, but in all cases, object segregation is the most
challenging task in order to identify objects as well as possible. This is especially
valid when CLDs are vastly different in their sizes, because it might be difficult,
if possible at all, to precisely identify all objects. Parameters should be optimized
so that when different samples are compared, it is reasonable to assume the same
error for the ob- ject of interest in all samples. Therefore, relatively large data sets
should be obtained for quantification.
For our processing and analysis, images captured with time-lapse microscopy
are processed with Volocity software (PerkinElmer). Brightness, contrast, density,
and blackness are adjusted to obtain sharp images, and photobleaching is corrected
for quantification. For our data set, one can calculate the transfer of newly
synthesized lipids (BODIPY FL C12) to preformed LDs (BODIPY 558/568 C 12).
Performed CLDs are defined as areas containing red fluorescent signal, selected by
choosing objects within a defined density threshold in the Cy3 channel. A single
cell is defined as the region of interest (ROI), and objects within the ROI are
refined by setting ob- ject size and shape references. Touching objects are
separated and objects selected under these criteria are defined as the preformed
CLD area. Data from all time points are corrected for photobleaching.
Fluorescence intensity within preformed CLD area at each point is quantified for
both GFP and Cy3 channels and presented as the per- centage of initial fluorescent
intensity (at time “0”). The results are exported to Microsoft Excel for plotting as
line graphs, in which the rate of incorporation can be reflected clearly: red signals
in the preformed LD are relatively stable, suggesting of low turnover, while signals
from the green channel increase gradually with time, suggesting continuous
incorporation of the newly synthesized TG, either by lipolysis and reesterification
of new LDs, or by fusion of newly synthesized and preformed LDs. When
hepatocytes from different genetic background were compared, differences in fatty
acid transfer rate from nascent to preformed CLDs can be clearly observed (Wang
et al., 2010).

7.3 DISCUSSION
The importance of CLD metabolism in tissues other than adipose resulted in the
need of methodologies that would lead to the recovery of highly purified CLD
prepara- tions. Due to the high proportion of neutral lipids and the relative low
concentration of proteins, CLDs would appear to be easy to isolate compared with
other organelles. However, a care has to be undertaken in homogenization and
centrifugation proto- cols that allow preparations of highly purified CLD
fractions.
 7.3 Discussion 123

Technically, the study of CLD metabolism in the liver should not represent a
problem, provided mild homogenization protocols are followed that preserve the
in- tegrity of intracellular organelles and avoid contamination of CLDs with
lumenal TG-rich lipoproteins and LLDs. Therefore, the study of hepatic CLDs
components has added complexity compared to CLDs from other cells and tissues.
A few studies have been performed to describe the proteome of hepatic CLDs
(Fujimoto et al., 2004; Turro et al., 2006). An early method for isolating hepatic
CLDs involved dis- continuous density gradient centrifugation and yielded six discrete
bands of lipid par- ticles, rich in TG and cholesterol (Ontko, Perrin, & Horne,
1986). Unfortunately, due to the lack of knowledge of CLD-associated proteins
(PAT family of proteins was discovered nearly a decade later), the purity and
protein composition of the various fractions method has not been adequately
validated.
Here, we describe a method for isolating CLDs from the liver, based on a
method by Brasaemle and Wolins (2006), and a method to isolate LLDs. The use
of a soft tissue homogenization is crucial to preservation of CLD integrity of LDs.
Simple two-step low-speed centrifugation and a single ultracentrifugation step
using a discontinuous density gradient yield highly purified CLD preparations. We
have used this method to analyze a CLD proteome in fasted and re-fed conditions.
The CLD proteome changes dramatically depending on the feeding state of the
mouse and therefore it is highly advisable that CLDs are prepared from animals in
a controlled metabolic state.
It is important to note that this method of CLD isolation can be applied to
evaluate a wide range of metabolic processes such as lipid metabolism in different
feeding states, biochemical determinations such as enzyme activities, particle
size, etc. The analysis of the dynamic nature of these organelles not only provides
the tools for the understanding of molecular mechanisms involved in CLD
formation and mo- bilization, but also paves the road to development of novel
therapies for treatment of pathological conditions.
The research on LLDs has been challenging for multiple reasons: (1) it is dif-
ficult to resolve LLDs from apoB-containing VLDL and its precursors; (2) it is
difficult to compromise the integrity of the microsomal membranes without af-
fecting the integrity of LLDs; (3) LLDs are present in low abundance; (4) con-
tamination from ER-resident proteins needs to be avoided; (5) LLDs are too
small to be visualized by light microscopy, and there had been only limited suc-
cess with electron microscopy in studying LLDs. The protocol we developed
overcomes most of these difficulties and represents a practical and effective
way to purify LLDs.
To separate LLDs from apoB-containing VLDL and its precursors, we used im-
munoprecipitation to remove apoB-containing particles, which was proved
success- ful since apoB was absent from the post-IP fraction, while the LLD-
associated carboxylesterase 3/TGH was recovered in this fraction (Wang et al.,
2007). How- ever, care should be taken to maintain the integrity of particles during
immunopre- cipitation procedure; immunoprecipitation should be performed in the
absence of detergents, as they would destroy lipid particles, and/or change protein
and lipid composition of isolated LLDs. Similarly, the method using high pH
carbonate
124 CHAPTER 7 Analysis of Lipid Droplets in Hepatocytes

(0.2 M Na2CO3, pH 11) for extraction used for studying topologies of polytopic
membrane proteins is unsuitable as this method might denature proteins and/or
strip proteins from the LLD surface. We used hypotonic solution to swell and
compromise microsomal membrane integrity to release LLDs. Because LDs do not
contain aque- ous phase within the core, they are not susceptible to hypotonic
osmotic pressure. One of the challenges we have encountered in the preparation
of LLDs is their low abundance. Our studies have suggested that less than 3%
of the intrahepatic TG is contained in LLDs. This problem is compounded by the
relatively low recov- ery. For obtaining sufficient amount of LLDs for subsequent
studies, usually two to four mouse livers are necessary. However, it is challenging
to adapt this method to cultured hepatocytes unless one uses radioactive tracers to
monitor LLD recovery. As mentioned above, an alternative approach to release
microsomal lumenal con- tents is through the use of high pH Na2CO3 (Sundaram
et al., 2010). This method is more efficient at breaking microsomes than the
hypotonic method we used, and therefore results in a higher recovery of lumenal
contents. However, because Na2CO3 will also strip proteins and CLDs peripherally
associated with membranes in addition to LLDs, this might result in contamination
of LLDs with CLDs that remained associated with microsomes. Readers should
choose carefully which procedure is preferable depending on the experimental
needs. The following publications maybe referred to for comparison (Wang et al.,
2007; Yao, Zhou, Figeys, Wang, & Sundaram, 2013).
To assess the purity of the isolated LLDs and to estimate the degree of contam-
ination, protein composition needs to be determined. The LLD fractions should
be free of apoB, transmembrane proteins, and CLD markers. However, many ER
lumenal proteins are found in LLDs. This may be due to the ER–LD connection,
as by high-confidence LD proteomics studies, many proteins were found to have dual
localization in the ER and on LDs (Krahmer et al., 2013). Using mass
spectrometry, we also found some cytosolic and mitochondria associated proteins in
LLDs fraction. These are inevitable contaminations that exist in essentially all
subcellular purifica- tions. Thus, further combining LLD purification with a high-
confidence proteomics approach such as SILAC would be beneficial to identify
bona fide LLD proteins. This approach would further assist in identifying specific
protein markers for LLDs. So far, proteins confirmed to be present on mouse liver
LLDs are carboxylesterases 1 and 3, apoE, MTP (Wang et al., 2007), and apoCIII
(Sundaram et al., 2010). How- ever, these proteins are not exclusively associated
with LLDs but are also present in their “lipid-free” form in the ER lumen.
The isolation and analysis of LLDs provided direct biochemical evidence that
LLDs are true entities that possess different biochemical properties from that of
CLDs and VLDL. This protocol can be adapted to a wide variety of applications
to study the mechanism of LLD formation, the role of LLD-associated proteins in
VLDL secretion, as well as genetic, nutritional, and pharmaceutical influence on
these processes.
This protocol also provides an example of using BODIPY fatty acid analogues
to study dynamics of LD formation and metabolism. BODIPY fatty acids are
available
 Acknowledgments 125

in different chain length, and both BODIPY-C16 and BODIPY-C12 are reported to
incorporate well into cells (Thumser & Storch, 2007). We have only used BOD-
IPY-C12 since with the inclusion of the BODIPY head group the analogue mimics
the natural chain length of OA. BODIPY-C 12 is esterified into both phospholipids
and neutral lipids and labels the same LDs as those stained with conventional LD
dyes (BODIPY 493/503 and Nile Red), thus showing properties similar to natural
long-chain fatty acids (Wang et al., 2010).
We used primary hepatocytes in this protocol due to our specific need to
compare cells of different genetic backgrounds; however, this method can be easily
adapted to other cell types. When a new cell type is used, several control
experiments should be performed to ensure that the analogues are metabolized as
would be expected from native fatty acids. Controls should include: (1)
confirmation of esterification into phospholipids and neutral lipids by TLC; (2)
monitoring dynamics of analogue up- take and incorporation, preferably compare
these with radiolabeled OA; (3) coloca- lization with conventional neutral lipid
dyes.
This method can also be used to study the localization of nascent CLD
formation, interactions of proteins, and nascent LDs immediately after fatty acid
loading of cells. However, these applications require either transient expression of
fluorescently tagged proteins, or immunostaining of a protein of interest.
Immunofluorescence staining of LD-associated proteins is challenging since lipids
cannot be fixed by general fixatives. Some recently published protocols addressed
this problem (DiDonato & Brasaemle, 2003; Ohsaki, Maeda, & Fujimoto, 2005).
Primary hepatocytes are known for being difficult to transfect. Liposome-based
transfection reagents, such as Lipofectamine 2000, are able to deliver plasmid
DNAs into hepatocytes when the ratio of transfection reagent DNA is optimized;
however, the transfection efficiency is generally very low, less than 15%. We found
that a cell type-specific reagent, Targefect-Hepatocytes, together with its enhancer
Virofect (Targeting Systems, CA), provides a much higher transfection efficiency
(Wang et al., 2010).

CONCLUSION
We have provided detailed protocols and expected outcomes for the purification
and analysis of hepatic LDs, including CLDs and LLDs. We have also presented a
cell biological method for monitoring LD dynamics in hepatocytes. These
protocols can be extended to study liver LD metabolism under different metabolic
states.

Acknowledgments
Supported by grants from the Canadian Institutes of Health Research and Natural Sciences and
Engineering Research Council of Canada. Huajin Wang is supported postdoctoral fellowship
from the Canadian Institutes of Health Research. Richard Lehner is an Alberta Innovates-
Health Solutions Scientist.
126 CHAPTER 7 Analysis of Lipid Droplets in Hepatocytes

References
Alexander, C. A., Hamilton, R. L., & Havel, R. J. (1976). Subcellular localization of B apo-
protein of plasma lipoproteins in rat liver. The Journal of Cell Biology, 69, 241–263.
Bickel, P. E., Tansey, J. T., & Welte, M. A. (2009). PAT proteins, an ancient family of lipid
droplet proteins that regulate cellular lipid stores. Biochimica et Biophysica Acta, 1791,
419–440.
Brasaemle, D. L., & Wolins, N. E. (2006). Isolation of lipid droplets from cells by density
gradient centrifugation. Current Protocols in Cell Biology, Chapter 3, Unit 3.15.
Chattopadhyay, A. (1990). Chemistry and biology of N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)-
labeled lipids: Fluorescent probes of biological and model membranes. Chemistry and
Physics of Lipids, 53, 1–15.
DiDonato, D., & Brasaemle, D. L. (2003). Fixation methods for the study of lipid droplets
by immunofluorescence microscopy. Journal of Histochemistry and Cytochemistry, 51,
773–780.
Farese, R. V., Jr., & Walther, T. C. (2009). Lipid droplets finally get a little R-E-S-P-E-C-T.
Cell, 139, 855–860.
Fujimoto, Y., Itabe, H., Sakai, J., Makita, M., Noda, J., Mori, M., et al. (2004). Identification of
major proteins in the lipid droplet-enriched fraction isolated from the human hepatocyte
cell line HuH7. Biochimica et Biophysica Acta, 1644, 47–59.
Hamilton, R. L., Wong, J. S., Cham, C. M., Nielsen, L. B., & Young, S. G. (1998).
Chylomi- cron-sized lipid particles are formed in the setting of apolipoprotein B
deficiency. Journal of Lipid Research, 39, 1543–1557.
Krahmer, N., Hilger, M., Kory, N., Wilfling, F., Stoehr, G., Mann, M., et al. (2013). Protein
correlation profiles identify lipid droplet proteins with high confidence. Molecular &
Cel- lular Proteomics, 12, 1115–1126.
Kuerschner, L., Ejsing, C. S., Ekroos, K., Shevchenko, A., Anderson, K. I., & Thiele, C.
(2005). Polyene-lipids: A new tool to image lipids. Nature Methods, 2, 39–45.
Kuhnlein, R. P. (2012). Thematic review series: Lipid droplet synthesis and metabolism:
from yeast to man. Lipid droplet-based storage fat metabolism in Drosophila. Journal of
Lipid Research, 53, 1430–1436.
Lehner, R., & Kuksis, A. (1993). Triacylglycerol synthesis by an sn-1,2(2,3)-diacylglycerol
transacylase from rat intestinal microsomes. Journal of Biological Chemistry, 268,
8781–8786.
Ohsaki, Y., Maeda, T., & Fujimoto, T. (2005). Fixation and permeabilization protocol is
crit- ical for the immunolabeling of lipid droplet proteins. Histochemistry and Cell
Biology, 124, 445–452.
Olofsson, S. O., Bostrom, P., Andersson, L., Rutberg, M., Perman, J., & Boren, J. (2009). Lipid
droplets as dynamic organelles connecting storage and efflux of lipids. Biochimica et
Bio- physica Acta, 1791, 448–458.
Ong, K. T., Mashek, M. T., Bu, S. Y., Greenberg, A. S., & Mashek, D. G. (2011). Adipose
triglyceride lipase is a major hepatic lipase that regulates triacylglycerol turnover and
fatty acid signaling and partitioning. Hepatology, 53, 116–126.
Ontko, J. A., Perrin, L. W., & Horne, L. S. (1986). Isolation of hepatocellular lipid droplets:
The separation of distinct subpopulations. Journal of Lipid Research, 27, 1097–1103.
Pagano, R. E., Martin, O. C., Kang, H. C., & Haugland, R. P. (1991). A novel fluorescent
cer- amide analogue for studying membrane traffic in animal cells: Accumulation at the
Golgi
 References 127

apparatus results in altered spectral properties of the sphingolipid precursor. The Journal of
Cell Biology, 113, 1267–1279.
Quiroga, A. D., Li, L., Trotzmuller, M., Nelson, R., Proctor, S. D., Kofeler, H., et al. (2012).
Deficiency of carboxylesterase 1/esterase-x results in obesity, hepatic steatosis, and hyper-
lipidemia. Hepatology, 56, 2188–2198.
Sundaram, M., Zhong, S., Bou Khalil, M., Zhou, H., Jiang, Z. G., Zhao, Y., et al. (2010).
Functional analysis of the missense APOC3 mutation Ala23Thr associated with human
hypotriglyceridemia. Journal of Lipid Research, 51, 1524–1534.
Thumser, A. E., & Storch, J. (2007). Characterization of a BODIPY-labeled fluorescent fatty
acid analogue. Binding to fatty acid-binding proteins, intracellular localization, and me-
tabolism. Molecular and Cellular Biochemistry, 299, 67–73.
Turro, S., Ingelmo-Torres, M., Estanyol, J. M., Tebar, F., Fernandez, M. A., Albor, C. V., et
al. (2006). Identification and characterization of associated with lipid droplet protein
1: A novel membrane-associated protein that resides on hepatic lipid droplets. Traffic, 7,
1254–1269.
Walther, T. C., & Farese, R. V., Jr. (2012). Lipid droplets and cellular lipid metabolism. An-
nual Review of Biochemistry, 81, 687–714.
Wang, H., Gilham, D., & Lehner, R. (2007). Proteomic and lipid characterization of
apolipo- protein B-free luminal lipid droplets from mouse liver microsomes:
Implications for very low density lipoprotein assembly. Journal of Biological
Chemistry, 282, 33218–33226.
Wang, H., Wei, E., Quiroga, A. D., Sun, X., Touret, N., & Lehner, R. (2010). Altered lipid
droplet dynamics in hepatocytes lacking triacylglycerol hydrolase expression. Molecular
Biology of the Cell, 21, 1991–2000.
Yang, L., Ding, Y., Chen, Y., Zhang, S., Huo, C., Wang, Y., et al. (2012). The proteomics of
lipid droplets: Structure, dynamics, and functions of the organelle conserved from
bacteria to humans. Journal of Lipid Research, 53, 1245–1253.
Yao, Z. M., & Vance, D. E. (1988). The active synthesis of phosphatidylcholine is required
for very low density lipoprotein secretion from rat hepatocytes. Journal of Biological
Chem- istry, 263, 2998–3004.
Yao, Z., Zhou, H., Figeys, D., Wang, Y., & Sundaram, M. (2013). Microsome-associated lu-
menal lipid droplets in the regulation of lipoprotein secretion. Current Opinion in
Lipidol- ogy, 24, 160–170.