Plant Physiology and Biochemistry 42 (2004) 501509

www.elsevier.com/locate/plaphy

Original article

Protein composition of oil bodies in Arabidopsis thaliana ecotype WS

Pascale Jolivet a, Emeline Roux a, Sabine dAndrea a, Marlne Davanture b,
Luc Negroni b, Michel Zivy b, Thierry Chardot a,*
a

INRA INA-PG, (UMR 206 Chimie Biologique), Laboratoire de Chimie Biologique, Centre de Biotechnologie Agro-Industrielle,
Thiverval-Grignon 78850, France
b
Plate-forme de Protomique, INRA CNRS INA-PG, (UMR 320 Gntique vgtale), Ferme du Moulon 91190 Gif-sur-Yvette, France
Received 11 December 2003; accepted 8 April 2004
Available online 13 May 2004

Abstract
Till now, only scattered data are available in the literature, which describes the protein content of plant oil bodies. Especially, the proteins
closely associated with the model plant Arabidopsis thaliana oil bodies have never been previously purified and characterized. Oil bodies have
been purified using flotation techniques, combined with incubations under high salt concentration, in the presence of detergents and urea in
order to remove non-specifically trapped proteins. The identity and integrity of the oil bodies have been characterized. Oil bodies exhibited
hydrodynamic diameters close to 2.6 m, and a ratio fatty acid-protein content near 20. The proteins composing these organelles were
extracted, separated by SDS-PAGE, digested by trypsin, and their peptides were subsequently analyzed by nano-chromatographymass
spectrometry (nano-LCMS/MS). This led to the identification of a limited number of proteins: four different oleosins, ATS1, a protein
homologous to calcium binding protein, a 11-b-hydroxysteroid dehydrogenase-like protein, a probable aquaporin and a
glycosylphosphatidylinositol-anchored protein with no known function. The two last proteins were till now never identified in plant oil bodies.
Structural proteins (oleosins) represented up to 79% of oil body proteins and the 18.5 kDa oleosin was the most abundant among them.
2004 Elsevier SAS. All rights reserved.
Keywords: Arabidopsis thaliana; Caleosin; Oil bodies; Oleosin; Protein composition; Steroleosin

1. Introduction
All types of eukaryote cells contain intracellular particles
to store lipids as food reserves, which will be mobilized
during a forthcoming period of active metabolism. Of all
these lipid particles, those from seeds have been studied most
extensively. Plant seeds store triacylglycerols (TAG), esters
of glycerol and fatty acids, in oil bodies also called oleosomes. The TAG, which will be later broken down for germi-

Abbreviations: ACN, acetonitrile; BSA, bovine serum albumin; EMBL,

European Molecular Biology Laboratory; FAME, fatty acid methyl ester;
GPA, glycosylphosphatidylinositol-anchored protein; GPI, glycosylphosphatidylinositol; LCMS/MS, liquid chromatography coupled with tandem
mass spectrometry; MES, morpholinoethanesulfonic acid; NCBI, National
Center for Biotechnology Information; PAGE, polyacrylamide gel electrophoresis; SDS, sodium dodecyl sulfate; TAG, triacylglycerols; TBST, Trisbuffered saline with Tween-20; TIP, tonoplast intrinsic protein.
* Corresponding author.
E-mail address: thierry.chardot@grignon.inra.fr (T. Chardot).

2004 Elsevier SAS. All rights reserved.

doi:10.1016/j.plaphy.2004.04.006

nation and growth of the young seedling, serve both for

carbon skeleton and energy resources. The TAG matrix of an
oil body is surrounded by a monolayer of phospholipids
embedded with abundant oleosins [14] and some minor proteins of higher molecular mass [37]. It is still not clear how
the components are assembled into oil bodies during seed
formation [28]. Oleosins have received considerable attention in recent years, particularly in relation to oil body biogenesis and their structural role in stabilizing the
TAG/cytosol oil body interface. Several cDNA and genomic
sequences encoding for these proteins have been described
from various species. An oleosin molecule is thought to
contain three distinct structural domains: a central hydrophobic anchoring domain, highly conserved and containing a
typical sequence called proline knot, and two terminal amphipathic domains. The steric hindrance and electronegative
repulsion provided by oleosins seem to be involved in the
stability of the oil bodies. It has been suggested that the entire
surface of an oil body is covered by oleosins such that the
compressed oil bodies never coalesce or aggregate in the
cells of a mature seed. Moreover, the maintenance of the oil
bodies as small entities provides a large surface area per unit

502

P. Jolivet et al. / Plant Physiology and Biochemistry 42 (2004) 501509

mass of TAG and would facilitate lipase binding and lipolysis

during germination [14]. Structural proteins, which may
function as barriers to control the lipolysis of core lipids are
also found in animal lipid droplets. Proteins like adipophilin
and perilipin are expressed highly in adipocytes [21] and in
milk lipid globule membrane [35]. On the contrary, these
structural proteins are not found in yeast lipid droplets [1].
Besides abundant oleosins, three minor proteins of 27,
37 and 39 kDa have been identified in sesame oil bodies [7].
The gene encoding 27 kDa protein has been cloned [8].
Sequence analysis reveals the presence of a putative calciumbinding motif and the protein has been named caleosin.
Similar to oleosin structure, caleosin contains three structural
domains: a N-terminal hydrophilic domain including the
calcium-binding motif, a central hydrophobic anchoring domain with a proline knot and a C-terminal hydrophilic domain containing a potential phosphorylation site [8]. In the
same way, the gene encoding 37 kDa protein has been cloned
[20] and the sequence analysis reveals that the corresponding
protein, named steroleosin, possesses a N-terminal hydrophobic anchoring segment preceding a soluble domain homologous to sterol-binding dehydrogenases/reductases that
are involved in signal transduction in diverse organisms.
The drawback to the structural model of oil bodies described above is that it is supported by a relatively poor
characterization of the proteins involved. This is because
these proteins have proved difficult to purify to a state in
which they can be solubilized for detailed structural analysis.
Moreover, oil body preparations can be contaminated extensively by storage proteins and therefore the knowledge of the
specific protein composition of the oil bodies has not been
obtained. It is, however, essential that genomic studies be
validated by the analysis of the intact oil body proteins. To
date, no exhaustive description of the protein complement of
oil bodies from Arabidopsis thaliana has been given, although the complete sequencing of the A. thaliana genome
has made possible an extensive identification of oleosin,
caleosin and steroleosin genes indicating that A. thaliana has
the potential to express these proteins. Sixteen oleosin genes
have been characterized in the A. thaliana genome [17]. They
have been divided into three groups especially according to
their tissue specific expression: five genes are specifically
expressed in maturing seeds, three genes are expressed in
both maturing seeds and floral microspores and eight genes
are specifically expressed in the floral tapetum. Caleosin- and
steroleosin-like genes are also present in the A. thaliana
genome. The caleosin-like proteins are encoded by a multigene family. Four of these genes are weakly expressed in
various tissues and only one is highly expressed in developing embryos and mature seeds [24]. An alignment of eight
putative sterol-binding dehydrogenase/reductase sequences
in A. thaliana has been reported by Lin et al. [20] but it is not
known whether the corresponding genes are transcriptionally
active and the cellular localization of the potential proteins
has not been deduced.
Furthermore, it is of interest to understand which geneproducts are expressed under various states of growth and

environmental stresses. One approach is to isolate and identify the entire protein complement of oil bodies. In this work,
we describe the preparation and purification of oil bodies
from A. thaliana in order to carry out the most detailed
analysis of their proteins as possible.

2. Results
2.1. Oil body purification
Oil bodies were purified from mature seeds of A. thaliana.
The oil body fraction was observed under a light microscope
and proved to be constituted of spheres of similar size able to
emit fluorescence light upon incubation in the presence of
Nile red. No coalescence was observed, even 3 d after oil
body purification. The size of the globules was estimated to
be 2.81 0.72 m by light microscopy and 2.63 0.29 m
using dynamic light scattering.
A. thaliana seeds were characterized by weight measurements and protein or fatty acid accumulation. The mature
desiccated seed weighed about 16 g with a mean protein
content of 6.2 g and a fatty acid mass in TAG of 6.3 g, i.e.
the ratio fatty acidtotal protein is near 1 in the mature seed.
In a typical oil body preparation with 200 mg of seeds as
starting material, we obtained 1 ml of an oil body fraction
containing 1.1 mg proteins and 22.3 mg of fatty acid in TAG.
Oil bodies were enriched 20 times according to the fatty
acidprotein ratio.
2.2. SDS-PAGE analysis
The proteins contained in the oil body fractions CF3, CF4
and CF6 obtained through the last steps of preparation (see
Section 4) were analyzed by SDS-PAGE and compared with
the proteins from the crude extract. Results are presented in
Fig. 1. Obviously, the protein patterns of crude extract (lane
4) and of oil body fractions (lanes 13) are different. Oil body
fractions were especially enriched with proteins migrating
1a

1b

2a 2b 3a 3b

kDa
200
116.3
97.4
66.3
55.4

8
7
6
5
4
3
2
1

36.5
31
21
14.4

Fig. 1. SDS-PAGE of proteins from seed crude extract and purified oil
bodies. Proteins from seeds (20 g, lane 4) and from purified oil body
fractions CF3 (10 g, lanes 1), CF4 (10 g, lanes 2) and CF6 (10 g, lanes 3)
have been separated on 12% Nu PAGE gels. Gels have been stained with
G250 Coomassie blue (lanes 1a, 2a, 3a and 4) or AgNO3 (lanes 1b3b).
Molecular mass marker (M) was Mark 12 from Novex.

P. Jolivet et al. / Plant Physiology and Biochemistry 42 (2004) 501509

503

Table 1
Identified proteins in oil bodies purified from Arabidopsis thaliana seeds
Band

Identified protein a

1
2
3
4
5
6
7
8

Oleosin 18.5 kDa

Putative oleosin
Oleosin 20.3 kDa
Oleosin 21.2 kDa
Probable aquaporin TIP3.2
Putative embryo-specific protein ATS1
Predicted GPI-anchored protein
11-b-Hydroxysteroid dehydrogenase-like

Molecular massb
(kDa)
18.569
19.755
20.313
21.279
28.182
28.012
21.569
39.087

Sequence
coverage (%)
24.9
23.5
7.9
34.2
10.5
8.6
22.5
22.1

Accession number a

Gene b

PI b

Rename c

CAA44225
AAF01542
CAA63011
BAB11599
O22588
AAL36241
AAO50480
BAB09145

At4g25140
At3g01570
At3g27670
At5g40420
At1g17810
At4g26740
At1g54860
At5g50600

9.43
7.11
6.91
9.36
6.54
5.92
6.09
5.92

S3
S1
S2
S4
Clo-1
Arab-1

Identified protein and accession number were obtained with Sequest research using A. thaliana protein sequences downloaded from NCBI site
(htpp://www.ncbi.nlm.nih.gov/).
b
Theoretical average molecular mass, pI and gene localization were predicted from SWISS-PROT database (htpp://expasy.ch/sprot/).
c
Nomenclature used by Kim et al. [17] for oleosins, by Naested et al. [24] for caleosins, by Lin et al. [20] for steroleosins is reported.

within the 1520 kDa region. It is observed that some bands

disappeared more or less completely upon urea and hexane
treatment, especially bands migrating at almost 36 and
21 kDa.
At least six different protein bands were clearly visible
within the 655 kDa range in the CF6 fraction after Coomassie blue staining (lane 3a). Silver staining revealed some
other bands with a very low intensity (lane 3b). The protein
patterns were obtained in a highly reproducible way (data not
shown).

dylinositol and the protein recovered from A. thaliana oil

bodies possesses four potential glycosylation sites. These
structural modifications could explain the high mass observed or the electrophoretical behaviour. The other proteins
visible upon silver staining were contaminating storage proteins. Upon scanning of SDS-PAGE gels and image analysis,
we obtained an approximate quantification on the basis of
band intensity. We found that oleosins represented almost
10% of proteins in the crude extract and 79% in the oil body
fraction and that 18.5 kDa oleosin was by far the most
abundant protein.

2.3. Mass spectrometry analysis

2.4. Some properties of oil body proteins
Protein bands were identified through the analysis of their
trypsic peptides with LCMS/MS (see Section 4). The major
protein bands present in the crude extract (Fig. 1, lane 4) and
disappearing with purification were identified as polypeptides from storage proteins of A. thaliana: 2S albumin
(18.8 kDa), b sub-units of 12S cruciferins CRA1, CRB and
CRU3 (20.821.2 kDa) and a sub-units of 12S cruciferins
(29.837 kDa). Table 1 shows the identity of proteins present
in oil body CF6 fraction (Fig. 1, lane 3). Four oleosins were
recovered in bands 14, with molecular mass ranging from
18 569 to 21 279 Da. Their characteristics (accession number, gene localization and theoretical pI) are described in
Table 1. As one might expect, all analyzed trypsic peptides
belonged to the two terminal amphipathic domains of the
proteins. Besides oleosins, two other proteins were visible
upon Coomassie blue staining and identified. The band
6 contained a putative embryo-specific protein (28 012 Da),
the gene of which was designated ATS1. Band 8 was identified as an 11-b-hydroxysteroid dehydrogenase (EC
1.1.1.146)-like protein (39 087 Da). A probable aquaporin
TIP3.2 (band 5) and a predicted glycosylphosphatidylinositol (GPI)-anchored protein (band 7) were also detected in a
very low abundance since these proteins were only visible
after silver staining. The predicted molecular mass of GPIanchored protein from the sequence (21.6 kDa) was much
lower than the apparent molecular mass observed upon SDSPAGE (31 kDa). In fact, the C-terminal hydrophobic region
of GPI-anchored proteins is replaced by a glycosylphosphati-

In order to identify oil body minor proteins, we have

performed extraction of the proteins from the oil bodies by
organic solvents with mixtures containing various
chloroform/methanol ratios (3/6, 6/3, v/v). Proteins from the
oil body fraction CF6 (Fig. 2, lane 1) were absent from the
aqueous or organic phases upon chloroform/methanol (6/3)
extraction (Fig. 2, lanes 2 and 4), and were exclusively
detected at the interface between aqueous and organic phases
1

kDa

66.3
55.4
8
7
6
5
4
3
2
1

36.5
31
21
14.4

Fig. 2. Extraction of oil bodies proteins by chloroform/methanol mixtures.

Proteins from oil bodies (fraction CF6, 20 g, lane 1) have been extracted
using chloroform/methanol (6/3, v/v) and proteins from the aqueous phase
(lane 2), from the interface (lane 3) and from the organic phase (lane 4) have
been solubilized in the same volume of SDS-PAGE buffer and equal volumes of each fraction have been submitted to SDS-PAGE. Gel has been
stained with AgNO3. Molecular mass marker (M) was Mark 12 from Novex.

504

P. Jolivet et al. / Plant Physiology and Biochemistry 42 (2004) 501509

of A. thaliana purified oil bodies was 2.6 0.3 m and was

slightly higher than the diameters observed by electron microscopy by Tzen et al. [36] for oil bodies from different
species which ranged from 0.65 m for rape to 2 m for
sesame. However, Leprince et al. [19] showed from microscopic observation that the diameter of oil bodies could reach
up to 7 m in mature seeds from Theobroma cacao. Our
value was closer to the values determined by dynamic laser
light scattering for almond oil bodies [3].
Fig. 3. Western Blot. Twenty micrograms of oil body proteins (lane 2) and
0.3 g of purified recombinant His tagged 21.2 kDa oleosin (lane 4) were
separated on a 12% Nu PAGE gel. The gel was transferred onto Nylon
membrane, which was subjected to Western blot using anti-oleosin antibodies. Lanes 1 and 3 correspond to See Blue Novex pre-stained marker.

(Fig. 2, lane 3). Band 5 (probable aquaporin) was weak but

present and band 7 (probable GPI-anchored protein) was
much more visible (Fig. 2, lane 3).
In order to confirm the presence of 21.2 kDa oleosin in the
oil body fraction, we have raised antibodies in a rabbit
against this bacterially expressed oleosin. Results are presented in Fig. 3. The recombinant His tagged protein (lane 4)
was recognized by the anti-oleosin serum. The control with
pre-immune serum was negative (data not shown). The serum was able to detect a single band in the oil body fraction
(lane 2), showing that it did not cross react with other oleosins present in the oil body. Molecular mass of the recombinant protein, harboring a His tag, was slightly higher than
the molecular mass of authentic 21.2 kDa oleosin.

3. Discussion
3.1. Oil bodies purity and size
A. thaliana seeds used in this study contained 39% of
protein and fatty acids. These values were close to the values
determined by Baud et al. [2], ONeill et al. [27] and Ruuska
et al. [30]. The oil body preparation used several flotation
steps. Proteins non-specifically associated with or trapped
within oil bodies were subsequently removed by detergent
washing, ionic elution and urea treatment. In particular,
washing the membrane preparation with salts has been reported to strip the preparation of proteins, which do not show
a strong association in membranes [30]. TAG in defective or
broken oil bodies were removed by hexane extraction. Oil
bodies were then enriched 20 times according to the fatty
acidprotein ratio. The value of this ratio is close to the
27 value determined for rapeseed oil bodies [36], but comparison should, however, be taken with caution, as we did not
use the same methods for lipid and protein determination.
Especially, the protein quantification method utilizing the
protein-Coomassie dye binding failed, in our hands, to give
reproducible results.
Oil bodies were stable with time in the absence of any
added protease inhibitor as no lipid coalescence was observed after 3 d storage at 4 C. The hydrodynamic diameter

3.2. Proteins from A. thaliana oil bodies

3.2.1. Major proteins from oil bodies
Oleosins, the major proteins of oil bodies were especially
poor candidates for bidimensional electrophoresis, due to
their alkaline pIs, and their low solubility in aqueous media,
some of them being described as being soluble only in chloroform [5]. The under-representation of membrane proteins
on 2D gels has already been pointed out by Santoni et al.
[32]. On the other hand, the A. thaliana pollen coat proteome
was solved upon separation of proteins using SDS-PAGE
[23]. Because high concentrations of SDS can be used, a 1D
SDS-PAGE procedure is therefore more efficient in analyzing hydrophobic proteins than 2D-PAGE using isoelectric
focusing in the first dimension [33]. For these reasons, we
choose to separate oil body proteins using SDS-PAGE instead of 2D electophoresis.
It was noticed that the protein patterns were highly reproducible. This suggests that the signature proteins of this CF6
fraction were present and that unspecific co-isolation of proteins with oil bodies is unlikely. Upon use of the method
developed by Seigneurin-Berny et al. [33], we found all the
identified oil body proteins at a water organic solvent interface confirming their amphipathic property previously deduced from their primary sequence structure. Oil bodies
purified from seeds of A. thaliana contain only a limited
population of proteins detected upon Coomassie blue staining: (a) four oleosins, the gene of which corresponded to
S1S4 according to the gene nomenclature proposed by Kim
et al. [17], (b) a putative embryo-specific protein ATS1;
similar proteins, termed caleosins, were encoded by homologous genes in storage oil bodies of sesame [8] and rice [11],
(c) an 11-b-hydroxysteroid dehydrogenase-like protein;
cDNA sequence alignment of Arabidopsis and sesame
showed that this protein was equivalent to a minor protein
present in sesame oil bodies and named steroleosin for its
homology with a sterol-binding dehydrogenase/reductase
[20]. Among the proteins of A. thaliana oil bodies, oleosins
are the most abundant proteins as we estimated them to
represent 79% of the protein content. These data confirm the
high abundance of oleosins in seeds and their enrichment in
purified oil bodies. A. thaliana S1 and S2 oleosins exhibit a
neutral to sligthly acidic pI. This confirms that alkaline pI is
not a restricted characteristic of oleosins as already pointed
out by Kim et al. [17]. The genes encoding for seed oil body
oleosins are scattered on three different chromosomes con-

P. Jolivet et al. / Plant Physiology and Biochemistry 42 (2004) 501509

trary to the genes coding for tapetum oleosins, which are all
found on chromosome V [17]. The most abundant oleosin
observed in seeds was the 18.5 kDa oleosin (S3). This agrees
with Kim et al. [17] who found that RNA corresponding to
this protein was the most abundant at all stages of silique
maturation. We could detect in mature seeds S1S4 oleosins,
but we failed to detect any peptide corresponding either to S5
or to any other oleosins. This may be due to the fact that
among mRNA detected by Kim et al. [17], some may not be
translated into proteins at the stage we observed them (mature seeds), or that their abundance in mature seeds is far
lower than abundance of either S4, S1 or S2 oleosin. Finally,
variability of protein expression among ecotypes cannot be
excluded.
3.2.2. Presence of caleosin and steroleosin-like proteins
One caleosin- and one steroleosin-like protein were associated with the oil body fraction but were much less abundant
than oleosins. A. thaliana genome analysis reveals that there
are seven genes related to caleosin and eight to steroleosin
but it is not known whether the genes are all transcriptionally
active. More, few data are available on tissue distribution and
subcellular localization of the proteins potentially encoded
by these genes. In other plants, caleosin-like proteins could
be expressed not only in oil bodies but also in various non-oil
storage tissues. In Brassica napus, at least one isoform is
tightly bound to oil bodies during seed development and
other isoforms accumulate in vegetative tissues and are probably integral components of the endoplasmic reticulum
membrane [12]. Out of the seven caleosin-like genes in
A. thaliana, four have corresponding mRNA sequences, i.e.
the genes are definitely expressed (Clo-1 to Clo-4). Only one
of these four caleosins (Clo-1) was highly expressed in developing embryos and mature seeds [24]. As the caleosin-like
protein identified in A. thaliana oil bodies was identical to
Clo-1, our results confirmed the location of Clo-1 in mature
seeds and more precisely in oil bodies. In addition to hydrophobic and proline-rich domains, caleosin also possesses an
immediately adjacent potential amphipathic a-helical domain which may play a role in its binding to oil bodies [8,24].
Caleosin function is presumably modulated by calciumbinding and its phosphorylation state. It has been suggested
that caleosin could be involved in calcium-mediated fusion
of oil bodies [11]. Oil body proteins could spontaneously
target to artificial oil emulsions. The proline knot motif in
oleosin and caleosin has been proposed to play an essential
role for this targeting [9].
Steroleosin seems to exist in seed oil bodies of diverse
species and comprises an oil body-anchoring segment,
NADPH-binding sub-domain, active site and sterol-binding
sub-domain [20]. Eight steroleosin homologous genes are
present in the Arabidopsis genome (Arab-1 to Arab-8). From
the sequence analysis of the N-terminal appendices, Lin et al.
[20] have suggested that Arab-1 and Arab-2 are oil body
associated, Arab-3 and Arab-4 are membrane associated,
Arab-5, Arab-6 and Arab-7 may or may not be membrane

505

bound and Arab-8 is water soluble and presumably present in

cytosol. From our results, only one protein from these eight
proteins, Arab-1, seemed to be really oil body associated.
This protein could be involved in diverse signal transduction.
3.2.3. Presence of two other minor proteins
Two other proteins were detected upon silver staining,
which proves that these proteins existed in a very low abundance, an aquaporin and a GPI-anchored protein. Aquaporins
belong to a group of abundant membrane proteins. They have
six trans-membrane a helices and two additional hydrophobic membrane-embedded domains including a highly conserved Asn-Pro-Ala (NPA) motif. Aquaporins form a hydrophilic channel allowing non-specific diffusion of water and in
addition of glycerol, urea, ammonia or other uncharged small
metabolites across the membrane barrier. The two asparagine
residues in NPA motifs control water passage through the
formation of hydrogen bonds. The putative aquaporin identified here seems to belong to the tonoplast intrinsic protein
(TIP) family more similar to the animal aquaporins than the
bacterial ones but the subcellular distribution of aquaporins
in plant cells appears to have a broad variety [31]. TIP
aquaporins are classically water facilitators but can present
glycerol transport activity [39]. Quigley et al. [29] have
analyzed the Arabidopsis genome for the presence of
aquaporin-related sequences and they have observed that the
aquaporin TIP3.2 transcripts were predominantly isolated
from developing seeds. The presence and function of aquaporin in our preparations remain to be precise: to isolate oil
bodies from water or on the contrary to facilitate the intake of
water and uptake of glycerol during lipolysis and germination.
The second minor protein we have identified with a high
sequence coverage yield was a putative glycosylphosphatidylinositol-anchored protein (GPA): At1g54860.
In a recent in silico analysis, Borner et al. [6] have predicted
the presence of numerous of these cell surface proteins detected often with difficulty in classical proteomic analyses.
Some are not recovered from 2D gels and must be analyzed
by one-dimensional SDS-PAGE. Trypsin digestion of others
yields few or no peptides for LCMS/MS analysis due to
their heavy glycosylation. At1g54860 belongs to the subfamily of GPA with unknown function. Some others GPA could
be involved in lipid transfer or associated with membrane
lipid microdomains [15]. At1g54860 presents N and
C-terminal hydrophobic domains presumably embedded into
membranes. The association of GPA with lipid microdomains has been investigated in Saccharomyces cerevisiae
[15] but few informations exist with plants. Strong association of proteins with lipid membranes achieved by glycolipid
anchors such as GPI has been noticed in A. thaliana [31,34].
Phosphatidylinositol is not the major constituent of phospholipids in plant seeds but can represent 720% of phospholipids [36].
Proteins involved in membrane traffic or signalization are
also reported in mammalian lipid droplets [21,35,40] and the

506

P. Jolivet et al. / Plant Physiology and Biochemistry 42 (2004) 501509

presence of a porin in oil bodies from Chinese hamster ovary

K2 cells has been described recently [21]. Besides, mammalian lipid droplets appear to contain the molecular machinery
to synthesize, store, utilize and degrade various lipids
[21,26]. In the same way, the predominant proteins in the
lipid droplet fraction of yeast are enzymes involved in sterol
and triglyceride metabolism [1] and diacylglycerol acyl
transferase activity was detected in the lipid body fraction of
an oleaginous fungus [16]. In our case, we could not detect
any peptide belonging to enzymes from the lipid metabolic
pathways such as diacylglycerol O-acyltransferase (EC
2.3.1.20), or triacylglycerol acylhydrolase (EC 3.1.1.3) in
A. thaliana oil bodies. Until now, these enzymes were only
detected by their activity [4,5,13,38], and could be present in
the oil bodies only in minute amounts, or at different phases
of seed genesis or germination. More, in oil bodies from
A. thaliana, oleosins are very predominant. Due to the high
abundance of these oleosins, other integral proteins are likely
to constitute very minor proteins that were not analyzed here
due to limitations in the amount available.
In conclusion, our results and literature data prove that the
protein composition of oil bodies has a high variability
among considered organisms (plant, fungi and animal cells).
It appears that seed oil bodies present the simplest protein
composition with a great abundance in structural proteins
and on the contrary a relative poor abundance in enzymes.
This variability might reflect the various role, biologic function and dynamic of oil bodies. Further work will be needed
to better understand time dependent composition, assembly
and function of seed oil body and targeting of its integral
proteins. It is also of interest to study global patterns of
protein content and activity and how these change during
development to boost our understanding of systems-level
cellular behaviour.

4. Methods
4.1. Seeds
Mature seeds of A. thaliana ecotype WS (Wassilewskija)
were kindly donated by M. Miquel (UMR Biologie des
Semences, INRA, Versailles, France).
4.2. Oil body purification and characterization
4.2.1. Purification
Oil bodies were purified using the method described by
Tzen et al. [37]. In a typical oil body purification, 100
200 mg of seeds were soaked in Milli Q grade water for 1 h at
room temperature, and were subsequently ground 20 times
for 15 s in 4 ml of 10 mM sodium phosphate buffer (pH 7.5)
containing 0.6 M sucrose (buffer 1) with a Potter grinder
driven by a Heidolph motor (rate 7). The sample was cooled
on ice between each grinding cycle, and the potter was rinsed
by 8 ml of buffer 1. The suspension was overlaid by one

volume of 10 mM sodium phosphate buffer (pH 7.5) containing 0.4 M sucrose (buffer 2) and spun at 10,000 g and 4 C
for 30 min in a Kontron Ultracentrifuge equipped with a
swinging-bucket rotor. The floating oil body fraction was
collected, resuspended in 12 ml of 5 mM sodium phosphate
buffer (pH 7.5) containing 0.2 M sucrose and 0.1% (v/v)
Tween 20. The suspension was overlaid by one volume of
10 mM sodium phosphate buffer (pH 7.5) and spun as before.
The oil body fraction was resuspended in 12 ml of buffer
1 additionally containing 2 M NaCl, overlaid by one volume
of 10 mM sodium phosphate buffer containing 0.25 M sucrose and 2 M NaCl and spun. The oil body fraction was
collected (CF3), resuspended in 8 ml of 9 M urea and left on
a shaker (60 rpm) at room temperature for 10 min. Then the
suspension was placed in centrifuge tubes, overlaid by one
volume of 10 mM sodium phosphate buffer (pH 7.5) and
spun. The oil body fraction was collected (CF4), resuspended
in 4 ml of buffer 1 and then mixed with 4 ml of hexane. After
centrifugation and removal of the upper hexane layer, the oil
body fraction was collected and resuspended in 4 ml of buffer
1. In a last step, the resuspension was overlaid by one volume
of buffer 2 and centrifuged. The floating oil body fraction
was collected (CF6), resuspended in a minimal volume of
buffer 1 and stored at 4 C till further use.
4.2.2. Microscopy
Nile red (Molecular Bioprobe, Montluon, France, solution at 1 mg ml1 in acetone) was added to an aliquot of oil
body suspension (1/10, v/v). After 1 h incubation at room
temperature, oil bodies were observed at 1000 focus
through WIG or WB filters (for fluorescence) with an Olympus BX 51 light microscope equipped with 100 oil immersion objective. Images were recorded using the Photometrics
Cool SNAP software. The diameter of the oil bodies was
estimated by comparison with a scale.
4.2.3. Light scattering
Oil bodies were diluted in water (generally 10 l of purified fraction in 500 l), and their hydrodynamic ratio was
determined using a Malvern HPPS particle sizer. We used a
lipid particle refraction index equal to 1.46. Experiments
were performed at 20 C in a quartz cuvette. Measurements
were repeated 310 times. Standard latex particles (200 nm,
Nanospheres, Duke Scientific, Palo Alto, USA) were suspended in Milli Q grade ultrapure water (Millipore Corp,
Molsheim, France) and used to check measurement accuracy.
4.3. Protein determination
Proteins were precipitated with two volumes of cold acetone, resuspended in 0.1% (w/v) SDS in 0.1 M NaOH
solution by sonication and quantified with the Folin Cioccalteu reagent [22] using BSA as standard.

P. Jolivet et al. / Plant Physiology and Biochemistry 42 (2004) 501509

4.4. Extraction of proteins by chloroform/methanol

mixtures
The oil body fraction CF6 containing 20 g of proteins has
been mixed with chloroform/methanol mixtures (3/6 or 6/3,
v/v), vigorously vortexed and spun at 15000 g during
5 min. Aqueous phase, organic phase and interface were
collected, and entirely resuspended in electrophoresis
sample buffer and submitted to SDS-PAGE (see below).
4.5. TAG extraction and quantification
Lipids were extracted according to Folch et al. [10], after
addition of triheptadecanoylglycerol as internal standard.
Lipid extract resuspended in chloroform was separated by
thin layer chromatography using silica gel 60 F254 plates
(Merck, VWR, Fontenay-sous-bois, France) and a mixture of
hexane/diethylether/acetic acid (40/10/1, v/v/v) for development. The spot corresponding to TAG was visualized under
UV after spraying with rhodamine (0.006% in 5 M ammonium hydroxide in ethanol/water, 1/1, v/v), scraped off the
plate and eluted with diethylether. Fatty acid methyl esters
(FAMEs) were prepared from extracted TAG using BF3methanol. After solvent evaporation, TAG were resuspended
in 1 ml of 0.5 M NaOH in methanol and incubated 10 min at
70 C. After addition of 1 ml of 14% BF3 in methanol,
saponified TAG were incubated for 10 min at 70 C. FAMEs
were extracted with pentane and analyzed by GC using a
Varian chromatograph with a moving needle-type injector at
220 C and a bonded silica capillary column with a stationary
phase of 90% biscyanopropyl/10% cyanopropyl siloxane
(30 m 0.25 mm I.D., 0.25 m film thickness, SP-2380 from
Supelco, Saint-Quentin Fallavier, France). Helium was used
as gas vector (1 bar). The column temperature program
started at 120 C, ramping to 210 C at 2 C min1. The flame
ionization detector was at 250 C. Identification of FAME
peaks was based upon retention times obtained for FAME
standards (Supelco). The total amount of fatty acids in TAG
was calculated from the ratio between the sum of FAME peak
areas and the heptadecanoic acid methyl ester peak area.
4.6. Electrophoresis
SDS-PAGE of proteins was carried out using 12% ready
to use Nu PAGE polyacrylamide gels (Novex, San Diego).
Sample buffer was containing 2% (w/v) SDS and 5% (v/v)
2-mercaptoethanol. Gels were run under 100 V for 180 min,
using 50 mM MES buffer (pH 7.3), and were stained with
Coomassie blue (G-250) according to Neuhoff and Arold
[25] or with AgNO3 using the Xpress silver staining kit
(Invitrogen, Cergy Pontoise, France). The molecular mass
marker was Mark 12 (Novex, San Diego). Proteins from the
crude extract and from the oil body fractions were precipitated with acetone for 1 h at 20 C, and the pellets were
dried and resuspended in electrophoresis sample buffer. Gels
were scanned (300 dpi) using an EPSON Perfection

507

1200 PHOTO scanner, and the TIFF resulting gels were

analyzed using the Image Quant (version 4.2a) software
(Molecular Dynamics).
4.7. Identification of proteins of the oil bodies
Protein bands stained with Coomassie blue were excised
from the polyacrylamide gel and stored at 20 C in 1%
acetic acid (v/v). As silver staining is not compatible with
proteomic analysis by LCMS/MS, protein bands were located upon silver staining and closely related bands were
excised from the blue stained gel. Before trypsin digestion,
gel slices were washed for 5 min with water, dehydrated for
15 min with ACN and dried by vacuum centrifugation. Proteins were reduced for 30 min with 150 l of 10 mM dithiothreitol0.1 M NH4HCO3 at 56 C and alkylated in the dark
with 100 l of 55 mM iodoacetamide0.1 M NH4HCO3 for
20 min at room temperature. The gel pieces were washed in
0.1 M NH4HCO3, dehydrated with ACN and vacuum dried.
Then proteins were digested overnight at 37 C with sequencing grade trypsin (EC 3.4.21.4, Roche Diagnostics,
Meylan, France) at the concentration of 12.5 mg l1 in the
presence of 25 mM NH4HCO3 and 5 mM CaCl2. The resulting peptides were extracted with 5% formic acid (v/v),
ACN/H2O (50/50, v/v) and ACN. Combined extracts were
dried and samples were dissolved in 0.15% (v/v) trifluoroacetic acid containing 3% (v/v) ACN before nano-liquid
chromatographymass spectrometry analysis.
HPLC was performed with Ultimate LC system combined
with Famos autosample and Switchos II microcolumn
switching for pre-concentration (LC Packings, Amsterdam,
The Netherlands). The samples were loaded on the column
(PEPMAP C18, 5 m, 75 m i.d., 15 cm, LC Packing) using
a pre-concentration step on a micro pre-column cartridge
(300 m i.d., 5 mm). Sample (4 l) was loaded on pre-column
at 5 l min1. After 3 min, the pre-column was connected
with the separating column and the gradient was started at
200 nl min1. Buffers were 0.1% HCOOH in water (A) and
0.1% HCOOH in ACN (B). A linear gradient from 5% to
45% B for 25 min was applied. Including the regeneration
step, one run was 60 min length. The LCQ Deca xp (Thermofinnigan, Les Ulis, France) was used with a nanoelectrospray interface. Ionisation (1.21.4 kV ionisation potential)
was performed with liquid junction and non-coated capillary
probe (New Objective, Cambridge, USA). Peptides ions
were analyzed by the Nth-dependent method as follows: (a)
full MS scan (m/z 5001500), (b) ZoomScan (scan of the two
major ions with higher resolution), (c) MS/MS of these two
ions. Identification was performed with Sequest using
Arabidopsis protein sequences downloaded from the National Center for Biotechnology Information FTP site. The
data base-searching algorithm Sequest uses a crosscorrelation (Xcorr) function to assess the quality of the match
between a tandem mass spectrum and amino acid sequence
information in a database. To minimize false positives, only
proteins with peptides sequences with Xcorr scores >2.2 (+2
charged peptide) or >1.8 (+1 charged peptide) were considered.

508

P. Jolivet et al. / Plant Physiology and Biochemistry 42 (2004) 501509

4.8. Serum against recombinant 21.2 kDa oleosin

The 21.2 kDa oleosin coding sequence (EMBL Accession
No. X91956) was amplified from a kZAPII vector (generous
gift of L. Lepiniec and V. Guyon, INRA Versailles, France).
DNA was amplified in a mini-cycler apparatus (MJ Research), using a ready to go PCR bead (Amersham Pharmacia Biotech, Les Ulis, France) and the following program:
25 cycles of 1 min at 94 C, 30 s at 52 C, 40 s at 72 C. The
forward primer (5-g gaa ttc cat atg gcg gat aca cac cgt-3)
included the translation start codon (in italics) and a NdeI
restriction site (underlined). The reverse primer (5-ggg ggg
acg acg gct gca ctc gag cgg-3) included a XhoI restriction
site (underlined). Oligonucleotides were synthesized by
MWGBiotech AG. The resulting amplification product was
treated with NdeIXhoI before cloning into pET20b vector
(Novagen,VWR), and the integrity of the construct was confirmed by DNA sequencing. Expression of the protein comprising a N-terminal His Tag was induced by isopropylthiob-D galactoside in transformed BL21 E. coli bacteria. Protein
was purified using NiNTA gel in the presence of 8 M urea,
and used to immunize.

References
[1]

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

4.9. Anti-21.2 kDa oleosin western blotting

[10]

Oleosins were detected with the anti-recombinant oleosin

serum after electro transfer onto Nylon Membranes (Millipore, Guyancourt, France) according to Laurire [18]. Membrane was then saturated with 1% BSA (w/v) in 20 mM Tris
buffer (pH 7.4) containing 150 mM NaCl (TBS) and 0.1%
Tween (TBST-BSA), and then rinsed for 30 min with TBSTBSA, incubated for 30 min with the anti-oleosin serum
(d = 1/1000) in TBST-BSA, rinsed for 30 min with TBSTBSA and then incubated with goat alkaline phosphatase
conjugated anti-rabbit antibodies (d = 1/1000) (Sigma, LIsle
dAbeau, France). The membrane was rinsed three times
with TBST, once with TBS, and revelation with alkaline
phosphatase substrate (nitro blue tetrazolium/5-bromo-4chloro-3-indolyl phosphate) was performed.

[11]
[12]

[13]

[14]
[15]
[16]

[17]

Acknowledgements
Authors wish to thank M. Canonge for expert oil bodies
preparation and protein separation and S. Denery (INRA
Nantes) for raising anti bodies against recombinant
A. thaliana 21.2 kDa oleosin. Thanks to D. Dalgleish (University of Guelph) for critical reading of the manuscript. The
mass spectrometry and proteomic equipments were purchased with funds from IFR87 la Plante et son Environnement, Rgion Ile de France, INRA, CNRS, Universit Paris
Sud and Gnoplante.

[18]

[19]

[20]

[21]

K. Athenstaedt, D. Zweytick, A. Jandrositz, S.D. Kohlwein, G. Daum,

Identification and characterization of major lipid particle proteins of
the yeast Saccharomyces cerevisiae, J. Bacteriol. 181 (1999) 6441
6448.
S. Baud, J.-P. Boutin, M. Miquel, L. Lepiniec, C. Rochat, An integrated overview of seed development in Arabidopsis thaliana ecotype
WS, Plant Physiol. Biochem. 40 (2002) 151160.
F. Beisson, N. Fert, S. Bruley, R. Voultoury, R. Verger, V. Arondel,
Oil-bodies as substrates for lipolytic enzymes, Biochim. Biophys.
Acta 30 (2001) 4758.
F. Beisson, N. Fert, J. Nari, G. Noat, V. Arondel, R. Verger, Use of
naturally fluorescent triacylglycerols from Parinari glaberrimum to
detect low lipase activities from Arabidopsis thaliana seedlings, J.
Lipid Res. 140 (1999) 23132321.
F. Beisson, N. Fert, R. Voultoury, V. Arondel, Large scale purification
of an almond oleosin using an organic solvent procedure, Plant
Physiol. Biochem. 39 (2001) 18.
G.H.H. Borner, K.S. Lilley, T.J. Stevens, P. Dupree, Identification of
glycosylphosphatidylinositol-anchored proteins in Arabidopsis. A
proteomic and genomic analysis, Plant Physiol. 132 (2003) 568577.
E.C.F. Chen, S.S.K. Tai, C.C. Peng, J.T.C. Tzen, Identification of three
novel unique proteins in seed oil bodies of sesame, Plant Cell Physiol.
39 (1998) 935941.
J.C.F. Chen, C.C.Y. Tsai, J.T.C. Tzen, Cloning and secondary structure analysis of caleosin, a unique calcium-binding protein in oil
bodies of plant seeds, Plant Cell Physiol. 40 (1999) 10791086.
J.C.F. Chen, J.T.C. Tzen, An in vitro system to examine the effective
phospholipids and structural domain for protein targeting to seed oil
bodies, Plant Cell Physiol. 42 (2001) 12451252.
J. Folch, M. Lees, G.H. Sloane-Stanly, A simple method for the
isolation and purification of total lipids from animal tissues, J. Biol.
Chem. 226 (1957) 497509.
G.I. Frandsen, J. Mundy, J.T.C. Tzen, Oil bodies and their associated
proteins, oleosin and caleosin, Physiol. Plant. 112 (2001) 301307.
I. Hernandez-Pinzon, K. Patel, D.J. Murphy, The Brassica napus
calcium-binding protein, caleosin, has distinct endoplasmic
reticulum- and lipid body-associated isoforms, Plant Physiol. Biochem. 39 (2001) 615622.
M.J. Hills, D.J. Murphy, Characterization of lipases from the lipid
bodies and microsomal membranes of erucic acid-free oilseed-rape
(Brassica napus) cotyledons, Biochem. J. 249 (1988) 687693.
A.H. Huang, Oleosins and oil bodies in seeds and other organs, Plant
Physiol. 110 (1996) 10551061.
E. Ikonen, Roles of lipid rafts in membrane transport, Curr. Opin. Cell
Biol. 13 (2001) 470477.
Y. Kamisaka, S. Mishra, T. Nakahara, Purification and characterization of diacylglycerol acyltransferase from the lipid body fraction of
an oleaginous fungus, J. Biochem. 121 (1997) 11071114.
H.U. Kim, K. Hsieh, C. Ratnayake, A.H. Huang, A novel group of
oleosins is present inside the pollen of Arabidopsis, J. Biol. Chem.
277 (2002) 2267722684.
M. Laurire, A semidry electroblotting system efficiently transfers
both high- and low-molecular-weight proteins separated by SDSPAGE, Anal. Biochem. 231 (1993) 206211.
O. Leprince, A.C. van Aelst, H.W. Pritchard, D.J. Murphy, Oleosins
prevent oil-body coalescence during seed inhibition as suggested by a
low-temperature scanning electron microscope study of desiccationtolerant and -sensitive oilseeds, Planta 204 (1997) 109119.
L.-J. Lin, S.S.K. Tai, C.-C. Peng, J.T.C. Tzen, Steroleosin, a sterolbinding dehydrogenase in seed oil bodies, Plant Physiol. 128 (2002)
12001211.
P. Liu, Y. Ying, Y. Zhao, D.I. Mundy, M. Zhu, R.G.W. Anderson,
Chinese hamster ovary K2 cell lipid droplets appear to be metabolic
organelles involved in membrane traffic, J. Biol. Chem. 279 (2004)
37873792.

P. Jolivet et al. / Plant Physiology and Biochemistry 42 (2004) 501509

[22] O.H. Lowry, N.J. Rosebrough, A.L. Farr, R.J. Randall, Protein measurement with the Folin-Phenol reagents, J. Biol. Chem. 193 (1951)
265275.
[23] J.A. Mayfield, A. Fiebieg, S.E. Johnstone, D. Preuss, Gene families
from the Arabidopsis thaliana pollen coat proteome, Science 29
(2001) 24822485.
[24] H. Naested, G.I. Frandsen, G.-Y. Jauh, I. Hernandez-Pinzon,
H.B. Nielsen, D.J. Murphy, J.C. Rogers, J. Mundy, Caleosins: Ca2+binding proteins associated with lipid bodies, Plant Mol. Biol. 44
(2000) 463476.
[25] V. Neuhoff, N. Arold, Improved staining of proteins in polyacrylamide gels including isoelectric focusing gels with clear background at
nanogram sensitivity using Coomassie Brilliant Blue G-250 and
R-250, Electrophoresis 9 (1988) 255262.
[26] M. Ohashi, N. Mizushima, Y. Kabeya, T. Yoshimori, Localization of
mammalian NAD(P)H steroid dehydrogenase-like protein on lipid
droplets, J. Biol. Chem. 278 (2003) 3681936829.
[27] C.M. ONeill, S. Gill, D. Hobbs, C. Morgan, I. Bancroft, Natural
variation for seed composition in Arabidopsis thaliana, Phytochemistry 64 (2003) 10771090.
[28] C.-C. Peng, J.T.C. Tzen, Analysis of the three essential constituents of
oil bodies in developing sesame seeds, Plant Cell Physiol. 39 (1998)
3542.
[29] F. Quigley, J. Rosenberg, Y. Shachar-Hill, H.J. Bohnert, From genome
to function: the Arabidopsis aquaporins, Genome Biol. 3 (2001) 117.
[30] S.A. Ruuska, T. Girke, C. Benning, J.B. Ohlrogge, Contrapuntal
networks of gene expression during Arabidopsis seed filling, Plant
Cell 6 (2002) 11911206.
[31] V. Santoni, P. Gerbeau, H. Javot, C. Maurel, The high diversity of
aquaporins reveals novel facets of plant membrane functions, Curr.
Opin. Plant Biol. 3 (2000) 476481.

509

[32] V. Santoni, M. Molloy, T. Rabilloud, Membrane proteins and

proteomics: un amour impossible? Electrophoresis 21 (2000) 1054
1070.
[33] D. Seigneurin-Berny, N. Rolland, J. Garin, J. Joyard, Differential
extraction of hydrophilic proteins from chloroplast envelope
membranes: a subcellular-specific proteomic approach to identify rare
intrinsic membrane proteins, Plant J. 51 (1999) 217228.
[34] D.J. Sherrier, T.A. Prime, P. Dupree, Glycosylphosphatidylinositolanchored cell-surface proteins from Arabidopsis, Electrophoresis 20
(1999) 20272035.
[35] N.G. Than, B. Sumegi, S. Bellyei, T. Berki, G. Szekeres, T. Janaky,
A. Szigeti, H. Bohn, G.N. Than, Lipid droplet and milk lipid globule
membrane associated placental protein 17b (PP17b) is involved in
apoptotic and differentiation processes of human epithelial cervical
carcinoma cells, Eur. J. Biochem. 270 (2003) 11761188.
[36] J. Tzen, Y. Cao, P. Laurent, C. Ratnayake, A. Huang, Lipids, proteins,
and structure of seed oil bodies from diverse species, Plant Physiol.
101 (1993) 267276.
[37] J.T.C. Tzen, C.-C. Peng, D.-J. Cheng, E.C.F. Chen, J.M.H. Chiu, A
new method for seed oil body purification and examination of oil body
integrity following germination, J. Biochem. 121 (1997) 762768.
[38] L. Valencia-Turcotte, R. Rodriguez-Sotres, The treatment of purified
maize oil bodies with organic solvents and exogenous diacylglycerol
allows the detection and solubilization of diacylglycerol acyltransferase, Biochim. Biophys. Acta 1534 (2001) 1426.
[39] A.R. Weig, C. Jakob, Functional identification of the glycerol permease activity of Arabidopsis thaliana NLM1 and NLM2 proteins by
heterologous expression in Saccharomyces cerevisiae, FEBS Lett.
481 (2000) 293298.
[40] C.C. Wu, K.E. Howell, M.C. Neville, J.R.Yates III, J.L. McManaman,
Proteomics reveal a link between the endoplasmic reticulum and lipid
secretory mechanisms in mammary epithelial cells, Electrophoresis
21 (2000) 34703482.