Báo cáo khoa học: Glutathione transferases kappa 1 and kappa 2 localize in peroxisomes and mitochondria, respectively, and are involved in lipid metabolism and respiration in Caenorhabditis elegans pot

Glutathione transferases kappa 1 and kappa 2 localize in
peroxisomes and mitochondria, respectively, and are
involved in lipid metabolism and respiration in
Caenorhabditis elegans
´
Elise Petit1,2,3, Xavier Michelet4,5, Claudine Rauch1,2,3, Justine Bertrand-Michel6, Francois Terce6,7,8,
¸
Renaud Legouis4,5 and Fabrice Morel1,2,3
1
2
3
4
5
6
7
8

´
Inserm U620, Universite de Rennes 1, France
´
EA-MDC, Universite de Rennes 1, France
IFR 140, Rennes, France
´ ´
´
CNRS Centre de Genetique Moleculaire – UPR2167, Gif-sur-Yvette, France
´
´
Universite Paris-Sud Orsay, Universite Paris-6, France
´ ´
´
IFR150, Institut Federatif de Recherche Bio-Medicale de Toulouse, Plateau technique de Lipidomique, France

INSERM, U563, Toulouse, France
´
´
´
´
Universite Toulouse III Paul Sabatier, Departement Lipoproteines et Mediateurs Lipidiques, IFR150, France

Keywords
Caenorhabditis elegans; fatty acids;
glutathione transferase kappa; mitochondria;
peroxisomes
Correspondence
ˆ
F. Morel, INSERM U522 ⁄ EA MDC, Hopital
Pontchaillou, 35033 Rennes, France
Fax: +33 299540137
Tel: +33 299543737
E-mail:
´ ´
R. Legouis, Centre de Genetique
ˆ
´
Moleculaire – UPR2167 CNRS Batiment 26,
Avenue de la terrasse, 91198 Gif-sur-Yvette
Cedex, France
Tel: +33 169824374
Fax: +33 169824386
E-mail:
(Received 28 May 2009, revised 3 July
2009, accepted 7 July 2009)

doi:10.1111/j.1742-4658.2009.07200.x

To elucidate the function of kappa class glutathione transferases (GSTs) in
multicellular organisms, their expression and silencing were investigated in
Caenorhabditis elegans. In contrast with most vertebrates, which possess
only one GST kappa gene, two distinct genes encoding GSTK-1 and
GSTK-2 are present in the C. elegans genome. The amino acid sequences
of GSTK-1 and GSTK-2 share around 30% similarity with the human
hGSTK1 sequence and, like the human transferase, GSTK-1 contains a
C-terminal peroxisomal targeting sequence. gstk-1 and gstk-2 genes show
distinct developmental and tissue expression patterns. We show that
GSTK-2 is localized in the mitochondria and expressed mainly in the pharynx, muscles and epidermis, whereas GSTK-1 is restricted to peroxisomes
and expressed in the intestine, body wall muscles and epidermis. In order
to determine the potential role(s) of GST kappa genes in C. elegans, specific
silencing of the gstk-1 and gstk-2 genes was performed by an RNA interference approach. Knockdown of gstk-1 or gstk-2 had no apparent effect
on C. elegans reproduction, development, locomotion or lifespan. By contrast, when biological functions (oxygen consumption and lipid metabolism) related to peroxisomes and ⁄ or mitochondria were investigated, we
observed a significant decrease in respiration rate and a lower concentration of the monounsaturated fatty acid cis-vaccenic acid (18:1x7) when
worms were fed on bacteria expressing RNA interference targeting both
gstk-1 and gstk-2. These results demonstrate that GST kappa, although not
essential for the worm’s life, may be involved in energetic and lipid metabolism, two functions related to mitochondria and peroxisomes.

Abbreviations
Dsba, protein disulfide isomerase A; FAME, fatty acid methyl ester; GFP, green fluorescent protein; GST, glutathione transferase; PTS1,
peroxisomal targeting signal 1; RNAi, RNA interference.

5030

FEBS Journal 276 (2009) 5030–5040 ª 2009 The Authors Journal compilation ª 2009 FEBS

E. Petit et al.

Glutathione transferase kappa in C. elegans

Introduction
Glutathione transferase (GST) kappa is a 26.5 kDa
protein that was initially isolated from the rat liver
mitochondrial matrix and classified as a class theta
GST [1]. The determination of the three-dimensional
structure of the class kappa enzyme from rat
(rGSTK1-1), complexed with glutathione (GSH),
showed a different folding topology from that of the
other GST classes, and revealed that the enzyme shows
similarity with the protein disulfide bond isomerase,
DsbA, from Escherichia coli and a bacterial 2-hydroxychromene-2-carboxylate isomerase, an enzyme involved
in the naphthalene degradation pathway [2,3].
Although class kappa GST showed an activity towards
aryl halides, such as 1-chloro-2,4-dinitrobenzene, and
can reduce cumene hydroperoxide and (S)-15-hydroperoxy-5,8,11,13-eicosatetraenoic acid [4], this activity
remained quite low when compared with that of other
soluble GSTs. Interestingly, a recent study has shown
that GST kappa might also possess a function independent of its glutathione conjugation activity in adipose tissue [5]. Indeed, Liu et al. [5] have identified
GSTK1 as a key regulator for the multimerization of
adiponectin, which is an adipocyte-derived hormone,
in both human and rodent.
Tissue distribution, analysed by RT-qPCR, showed
that the hGSTK1 gene is expressed in the 24 different
human tissues examined [4]. In the mouse, the
mGSTK1 protein is present in large amounts in the
liver, kidney, stomach and heart, and its association

with liver and kidney mitochondria has been demonstrated by electron microscopy [6]. GSTK1 transcript
tissue expression is similar in the rat and in the mouse
[7]. With regard to subcellular localization, in contrast
with soluble GSTs, which are mainly present in the
cytosol, GST kappa is localized in peroxisomes and
mitochondria [4]. Although the process of GST kappa
targeting to mitochondria is unclear, it has been
reported to associate with the Hsp60 chaperone [3],
and a possible cleavage site for a mitochondrial presequence exists at the N-terminus. A peroxisomal targeting sequence (tripeptide ARL) has been identified in
the C-terminus of the hGSTK1 subunit [4].
The recent demonstration of GST kappa as a regulator of protein multimerization and its particular subcellular location have led to questions about its further
role(s) and substrate(s) [5]. Common peroxisomal and
mitochondrial functions are related to lipid metabolism, including a- and b-oxidation of fatty acids that
generate acetyl-CoA and different acyl-CoA intermediates [8,9]. Thus, the presence of GSTK1 in both organelles suggests that it may be specifically involved in the

b-oxidation of fatty acids, either through its catalytic
activity, a certain transport function or interaction
with other proteins. Interestingly, its role in adiponectin regulation is also related to lipid and glucose
metabolism.
The nematode Caenorhabditis elegans is a genetically
well-characterized model organism [10] which presents
several advantages: (a) small size; (b) rapid reproduction as a self-fertile hermaphrodite; (c) large number of
offspring (250–300 progeny); (d) growth on a solid surface medium; and (e) transparent body allowing the
observation of cells in mature and developing animals.
Furthermore, as about 60% of C. elegans genes show
similarity to human genes, and transient RNA interference (RNAi) allows specific gene silencing, this model
organism represents a powerful tool for gene function
analysis.
The aim of our study was to characterize GST
kappa gene(s) and proteins in C. elegans and, by

means of RNAi, to investigate the effects of gene
silencing on the nematode phenotype. Our results
showed that the C. elegans genome contains two
GST kappa genes encoding GSTK-1 and GSTK-2,
which localize in peroxisomes and mitochondria,
respectively. Double inactivation by RNAi affects
the worm’s metabolism through a reduction in its
rate of respiration and modification of its lipid
content.

Results
The C. elegans genome contains two GST kappa
genes
We have previously described a C. elegans protein
showing 33% homology with the human GST kappa,
hGSTK1, amino acid sequence [4]. Database analyses
revealed the presence of two genes previously named
ZK1320.1 and D2024.7 in the C. elegans genome.
These two genes are located on chromosomes II
(ZK1320.1) and IV (D2024.7) and have probably
arisen by gene duplication. Both genes are composed
of three exons and two introns (Fig. 1A), the nucleotide sequence at the splice junctions is consistent with
the canonical GT–AG rule and the corresponding
encoded amino acid sequences comprise 226 and 225
residues, respectively, and share 32% identity. Orthologues of these genes are observed in other nematode
species, including C. briggsae, C. remanei, C. japonica,
Ancylostoma ceylanicum, Heterorhabditis bacteriophora
and Meloidogyne. Three arguments strongly suggest

FEBS Journal 276 (2009) 5030–5040 ª 2009 The Authors Journal compilation ª 2009 FEBS

5031


Glutathione transferase kappa in C. elegans

A

69

gstk-1 (zk1320.1)

B
AA position
rGSTK1
GSTK-1
GSTK-2

462

51 150

2

182

3

1
151

gstk-2 (d2024.7)

E. Petit et al.

380

599

1
50

469

100

150

200

C

that C. elegans ZK1320.1 and D2024.7 genes belong to
the kappa class of GSTs and share a common ancestral gene with rat GSTK1. Firstly, there are conserved
amino acids in the protein sequences of rat GSTK1
and C. elegans ZK1320.1 and D2024.7. Secondly, we
showed common exon–exon junctions in the translated
amino acid sequences of the two C. elegans genes
and the rat GST kappa gene (Fig. 1B). Finally, using
the psortii program (),

the presence of a C-terminal peroxisomal targeting
signal 1 (PTS1), composed of the three amino acids
serine-lysine-leucine (SKL), was demonstrated in some,
but not all, nematode species and rat GST kappa
(tripeptide ARL) amino acid sequences (Fig. 1C).
Furthermore, psortii also predicted a mitochondrial
presequence with a putative cleavage site after residue
5 (MPNRK ⁄ VV) at the N-terminus of the GSTK-2
sequence. For these reasons, ZK1320.1 and D2024.7
genes were renamed gstk-1 and gstk-2.
5032

147

3

2

Fig. 1. Genomic structure and intron positions in C. elegans gst kappa genes. (A)
Genomic structure of hGSTK1. Exons are
represented as black boxes and introns
are represented by lines; the numbers
indicate the size in nucleotides. The gene
structure is drawn to scale. (B) Intron
positions in rat and C. elegans amino acid
sequences. Filled and open triangles mark
common and unique intron positions,
respectively. (C) Amino acid sequence
alignment of rat GSTK1 with GST kappa of
nematodes. The aligned sequences are

listed below, followed by the species’
names and accession numbers in parentheses. rGSTK1 (Rattus norvegicus, UniProt:
P24473), CelGSTK-1 (Caenorhabditis elegans, UniProt: Q09652), CbrGSTK-1 (Caenorhabditis briggsae, UniProt: A8XB52),
CelGSTK-2 (Caenorhabditis elegans, UniProt:
Q18973), CbrGSTK-2 (Caenorhabditis briggsae, UniProt: A8X1K2), CreGSTK-2 (Caenorhabditis remanei, WormBase: RP16274),
CjaGSTK-1 (Caenorhabditis japonica,
WormBase: JA07681), AcGSTK (Ancylostoma ceylanicum, GenBank: CB175111.1),
MhGSTK (Meloidogyne hapla, GenBank:
EX007447.1), HbGSTK (Heterorhabditis
bacteriophora, GenBank: BM883827.1).
*Residues involved in glutathione binding
site. #Residues involved in dimer interface.

GSTK-1 and GSTK-2 are localized in peroxisomes
and mitochondria, respectively
The analysis of GST kappa transcript levels has shown
a ubiquitous expression in human [4] and mouse [6]
tissues. In order to determine the spatial and temporal
expression patterns of gstk-1 and gstk-2 genes in
C. elegans, we constructed gfp::gstk-1 and gstk-2::gfp
fusions under the control of approximately 1 kb of the
5¢ regulatory gstk-1 and gstk-2 sequences, respectively.
In these reporter fusion proteins, green fluorescent
protein (GFP) was inserted in frame immediately
upstream or downstream of gstk-1 and gstk-2
sequences, respectively. Transgenic strains were
obtained by microinjection, and the localization of
fusion proteins in animals was revealed by the examination of GFP fluorescence (Fig. 2).
The expression of the gfp::gstk-1 transgene was first
detected in 100 cell embryos, as shown in Fig. 2A.

FEBS Journal 276 (2009) 5030–5040 ª 2009 The Authors Journal compilation ª 2009 FEBS


E. Petit et al.

Glutathione transferase kappa in C. elegans

A

A'

E

B

F

C

G

D

E'

H

Fig. 2. Analysis of expression pattern of gstk-1 and gstk-2 in embryo and adult C. elegans. Projection of confocal images of GFP::GSTK-1
(A–D) and GSTK-2::GFP (E–H) at different developmental stages. (A) GFP::GSTK-1 is first detected at mid-embryogenesis in the primordium

of the intestine (white arrow) and in the epidermis (arrowheads). A¢ is the corresponding Nomarski picture. (B) During larval development,
GFP::GSTK-1 is very strongly expressed with a vesicular localization in the intestine (arrows). (C) A weaker expression of GFP::GSTK-1 is
present in the muscles (arrowheads) and the epidermis (arrows). (D) Faint diffuse expression is detected in the rectal gland cells (compare
with intestinal signal indicated with an arrow). (E) GSTK-2::GFP is first detected in muscle quadrants (arrowheads) during morphogenesis of
the embryo. E¢ is the corresponding Nomarski picture. (F) In larvae, a strong punctate staining is present in the pharynx (arrows) and the
body wall muscles (arrowheads), and a weaker signal is observed in the intestine. (G) In the pharynx, a very regular expression of GSTK2::GFP in myo-epithelial cells (arrow) is characteristic of a mitochondrial localization. (H) In body wall muscle cells, GSTK-2::GFP is detected
as a tubular network with stronger punctata (arrows) typical of the mitochondrial system. Scale bar, 10 lm.

Expression was increased during morphogenesis of the
embryo. Fluorescence was observed as strong punctate
staining in intestinal cells and in the epidermis. The
number and intensity of fluorescent structures
increased strongly in the intestine during larval development (Fig. 2D), whereas the epidermal fluorescent
punctata weakened (Fig. 2B). In addition, a diffuse
localization of GFP::GSTK-1 was observed in the
body wall muscles (Fig. 2C) and in the rectal gland
cells in larvae and adults (Fig. 2D). The presence of a
peroxisomal targeting signal in its C-terminus (Fig. 1)
and the punctate localization pattern suggest that
GSTK-1 is a peroxisomal protein. To confirm the peroxisomal localization, we masked PTS1 by fusing the
GFP at the C-terminus of GSTK-1. Transgenic worms
for GSTK-1::GFP only presented diffuse staining (data
not shown), further supporting a peroxisomal localization of GSTK-1.
In the GSTK-2::GFP transgenic strain, fluorescence
was first detected during the second half of embryogenesis (Fig. 2E). A strong signal was detected as
punctate staining in the pharynx (Fig. 2G) and the
body wall muscles (Fig. 2F), and a weaker signal was

also observed in the intestine. In muscle cells, the
strong GSTK-2::GFP punctata were part of a tubular

network which was weakly fluorescent (Fig. 2H). Interestingly, similar staining has been observed previously
by the expression of GFP fused to specific subcellular
targeting sequences [11], strongly suggesting the
presence of GSTK-2 in the mitochondria. To confirm
this mitochondrial localization, we stained both
GFP::GSTK-1- and GSTK-2::GFP-expressing worms
with MitoTracker Red (Fig. 3). Although GFP::
GSTK-1 did not show any colocalization with MitoTracker Red staining (Fig. 3A¢¢), GSTK-2::GFP fully
colocalized with the mitochondrial dye (Fig. 3B¢¢).
Together, these data indicate that GSTK-1 and
GSTK-2 have a peroxisomal and mitochondrial localization, respectively.
Impairment of oxygen consumption and lipid
content in gstk-1 and gstk-2 double-knockdown
worms
Post-transcriptional gene silencing of specific genes by
RNAi is a well-established method in C. elegans [12].

FEBS Journal 276 (2009) 5030–5040 ª 2009 The Authors Journal compilation ª 2009 FEBS

5033


Glutathione transferase kappa in C. elegans

E. Petit et al.

A

A'

A''

B

B'

B''

Fig. 3. GSTK-2, but not GSTK-1, localizes in
mitochondria. Single confocal images of
GFP::GSTK-1 (A) and GSTK-2::GFP (B) in
adult animals. GSTK-2::GFP fully colocalizes
with the mitochondrial-specific marker
MitoTracker (A¢), whereas GFP::GSTK-1 is
not localized in the mitochondrial network
(B¢). Scale bar, 10 lm.

5034

100

Control

90

gstk-1 (RNAi)

80
gstk-2 (RNAi)

70
Viability (%)

In our study, C. elegans (strain N2) was fed with bacteria producing dsRNA of the gstk-1 and ⁄ or gstk-2
coding regions. In each experiment, four feeding conditions were defined: one group of worms was fed with
control bacteria containing the empty plasmid pL4440,
one with bacteria expressing gstk-1(RNAi), another
with bacteria expressing gstk-2(RNAi) and one with a
mix (1 : 1) of both RNAi-expressing bacteria. To test
the efficiency of RNAi, fluorescence levels in RNAitreated GSTK::GFP transgenic worms were compared
with the levels observed in control worms. Figure S1
(see Supporting information) shows that RNAi
directed against either gstk-1 (Fig. S1A) or gstk-2
(Fig. S1E) was efficient, with the exception of the
pharynx in gstk-2::gfp transgenic worms, where silencing was incomplete (Fig. S1E,F). It is noteworthy that
gstk-2(RNAi) had no effect on the transgenic strain
expressing GFP::GSTK-1 (Fig. S1B), and GSTK2::GFP expression remained unchanged in transgenic
animals fed with bacteria producing gstk-1 dsRNA
(Fig. S1D), indicating the specificity of these RNAi
forms and the absence of compensatory adaptation
[i.e. upregulation of gstk-1 in gstk-2(RNAi) worms].
RNAi silencing of gstk-1 or gstk-2 had no apparent
effect on C. elegans reproduction or development (data
not shown). This absence of obvious phenotype has
been reported previously in wide RNAi screens (http://

gstk-1/k-2 (RNAi)

60
50

40
30
20
10
0
0

5

10

15
Days

20

25

30

Fig. 4. gstk-1(RNAi) and gstk-2(RNAi) do not affect the C. elegans
lifespan. Effects of RNAi-mediated knockdown of gstk-1 and ⁄ or
gstk-2 on the lifespan in wild-type worms. Worms were fed either
with control bacteria not expressing any dsRNA or bacteria
expressing dsRNA that targets gstk-1, gstk-2 or both gstk-1 and
gstk-2. Nematode survival was analysed by the Kaplan–Meier
method using Graphpad Prism 5. The same software was used to
test the equality of survival with the log-rank (Wilcoxon) test. Each
experimental condition was tested in triplicate.

www.wormbase.org). We also found that these RNAi
forms did not affect the C. elegans lifespan (Fig. 4).
The first animal died after 5 days and all animals were

FEBS Journal 276 (2009) 5030–5040 ª 2009 The Authors Journal compilation ª 2009 FEBS


O2 absorbed
(nmol·min–1·1000–1 worms)

E. Petit et al.

Glutathione transferase kappa in C. elegans

*

20
15
10
5
0

Control

gstk-1
(RNAi)

gstk-1/k-2
(RNAi)

gstk-2
(RNAi)

RNAi Strain
Fig. 5. gstk-1 ⁄ gstk-2(RNAi) animals present an altered respiratory
rate. Oxygen consumption was assessed in the fourth larval stage
of animals fed either with control bacteria not expressing any
dsRNA or bacteria expressing dsRNA that targets gstk-1, gstk-2
or both gstk-1 and gstk-2. Results are the mean of six values ± standard deviation, and are expressed as nmoles of O2 per
minute per 1000 worms. Student’s t-test was applied for statistical
studies between RNAi-fed worms and control worms (*P £ 0.05).

dead after 32 days. The mean lifespan for wild-type,
gstk-1(RNAi), gstk-2(RNAi) and gstk-1+gstk-2
(RNAi) were 14, 14, 13 and 13 days, respectively.
These results indicate that gst kappa genes are not
essential for the survival of C. elegans.
As our expression data for GSTK-1 and GSTK-2
supported peroxisomal and mitochondrial localizations, we further investigated cellular functions, such
as lipid metabolism and oxygen consumption, which
are closely related to these two organelles. The control
worms showed an oxygen consumption of 12.2 nmolỈ
min)1 per 1000 worms. Interestingly, a significant

decrease in oxygen consumption of about 40%
(7.5 nmolỈmin)1 per 1000 worms) was observed when
worms were fed on bacteria expressing dsRNAs targeting both gstk-1 and gstk-2 (Fig. 5). However, no
significant decrease in oxygen consumption was
observed in worms fed with gstk-1(RNAi) or gstk2(RNAi) alone compared with the control condition.
Thereafter, in order to investigate the effect of gstk-1

and ⁄ or gstk-2 silencing on lipid metabolism, we
measured the following lipid fractions: phospholipids,
diglycerides, triglycerides, free or esterified cholesterol,
free fatty acids and total fatty acids. Although most
lipid concentrations were unchanged (Tables S1–S4,
see Supporting information) between wild-type and
gstk-1(RNAi), gstk-2(RNAi) and gstk-1 ⁄ gstk-2(RNAi)
worms, a difference was observed for cis-vaccenic acid
(18:1x7) for the fatty acid methyl ester (FAME) fraction, which was decreased significantly in worms
depleted for both gstk-1 and gstk-2 (Fig. 6). These
data strongly suggest that GSTK-1 and GSTK-2 have
overlapping functions, as oxygen consumption and
cis-vaccenic acid levels were unchanged in gstk1(RNAi) and gstk-2(RNAi) worms and decreased only
in double-knockdown animals.

Discussion
In this study, we investigated the localization and
potential role(s) of GST kappa in C. elegans. In
contrast with most vertebrates, the C. elegans genome
contains two GST kappa genes, gstk-1 and gstk-2. The
two genes are located on different chromosomes and
contain three exons. Interestingly, orthologues of
C. elegans gstk-1 and gstk-2 genes are also found in

Control

gstk-1(RNAi)

gstk-2(RNAi)

gstk-1/k-2(RNAi)

*

18
16

Fig. 6. gstk-1 ⁄ gstk-2(RNAi) animals display
an abnormal FAME composition. Simplified
FAME composition of control and gstk-1,
gstk-2 and double gstk-1(RNAi) and
gstk-2(RNAi)-treated animals (see also
Tables S1–S4). Depletion of both gstk-1 and
gstk-2 leads to a decrease in the 18:1w7
fatty acid, but does not affect other lipids.
The results are the mean of three experiments ± standard deviation. Student’s t-test
was applied for statistical studies between
RNAi-fed worms and control worms
(*P £ 0.05).

% of total fatty acids

14
12
10
8
6
4
2
0

16:0

18:0

16:1w7

FEBS Journal 276 (2009) 5030–5040 ª 2009 The Authors Journal compilation ª 2009 FEBS

18:1w9

18:1w7

18:2w6

20:5w3

17 Cyclo

Fatty acid mono esters

5035


Glutathione transferase kappa in C. elegans

E. Petit et al.

the nematode species C. briggsae and C. remanei, indicating that a gene duplication took place before the
speciation of these three Caenorhabditis species.
Sequence conservation between the GSTs of C. elegans

gstk-1 ⁄ 2 and those of other nematode species
(C. briggsae, C. remanei, C. japonica, Ancylostoma
ceylanicum, Heterorhabditis bacteriophora and Meloidogyne hapla), as well as with rat GSTK1, is also
observed, the most highly conserved residues being
those that contribute to the glutathione-binding site
and dimerization of the protein. Interestingly, structure
prediction and molecular modelling studies have
shown that, despite the low sequence similarity
(30%), rGSTK1 and hGSTK1 structures are recognized as the closest structural homologues of C. elegans GSTK-1 and GSTK-2 (data not shown).
Together, these observations suggest that C. elegans
GSTK-1 and GSTK-2 might have similar activities to
their mammalian orthologues [3,4,7], and may contribute, at least in part, to detoxification processes.
Moreover, a tripeptide sequence (SKL), known as
PTS1, is present at the C-terminal end of the C. elegans
GSTK-1 protein. PTS1 is involved in protein import
into peroxisomes, and the importance of this signal for
peroxisome targeting in C. elegans has been shown previously by Motley et al. [13]. Interestingly, GFP fused
to the N-terminus of GSTK-1 was found to localize in
punctate bodies of C. elegans cells in several tissues,
strongly suggesting a peroxisomal localization. By contrast, colabelling with MitoTracker Red showed that
GSTK-2::GFP was present mainly in the mitochondria.
It is noteworthy that, in human cells, hGSTK1 is both
peroxisomal and mitochondrial and contains a C-terminal PTS1 [4]. This preserved intracellular localization
of GST kappa of both nematodes and vertebrates,
together with sequence conservation and intron positions in amino acid sequences, indicate that kappa class
genes probably originated from a common ancestral
gene which was present before the protostome ⁄ deuterostome split. Interestingly, the presence of two duplicated paralogous genes, gstk-1 and gstk-2, in the
C. elegans genome allowed specialization of the subcellular localization for each gene.
The use of reporter fusion proteins allowed the
study of tissue expression patterns of gst kappa genes.

Although both GFP::GSTK-1 and GSTK-2::GFP
fusion proteins are observed in common tissues, such
as the intestine, they also have a specific localization in
other tissues, such as the epidermis or pharynx for
GFP::GSTK-1 and GSTK-2::GFP, respectively. In
C. elegans, the intestine and epidermis are at the
interface between the organism and its environment.
Therefore, these tissues represent defence barriers
5036

against toxic agents, such as gut-derived oxidants or
endogenously generated reactive oxygen species. The
intestine is also a highly metabolically active organ
and represents the tissue in which most C. elegans
peroxisomes are found, as shown by immunostaining
for catalase [14] and by electron microscopy [15]. Interestingly, GSTK2::GFP is predominantly expressed in
muscle cells (pharynx and body wall muscle). The
expression of GST kappa genes in body wall muscle
and pharynx might be related to the large number of
mitochondria in these two tissues, which are associated
with high energy consumption.
In order to gain further insight into the potential
function(s) of GSTK-1 and GSTK-2, RNAi was used
to knock down the expression of the two corresponding genes, either separately or simultaneously. Knockdown of gstk-1 and ⁄ or gstk-2 had no effect on worm
lifespan, locomotion or development, suggesting that
GST kappa genes are not essential for the worm’s life.
Because, as in mammals, peroxisomes and mitochondria in C. elegans play a key role in the production of
reactive oxygen species and in lipid metabolism,
including fatty acid b-oxidation [16], we investigated
the potential role of gstk-1 and ⁄ or gstk-2 on the lipid

composition of worms. For this purpose, phospholipids, diglycerides, triglycerides, free and esterified
cholesterol, and free and total fatty acid levels were
measured in worms fed on gstk-1(RNAi) and ⁄ or gstk2(RNAi). With the exception of cis-vaccenic acid
(18:1x7) from the FAME fraction, there was no modification of lipid composition between worms fed on
the empty vector control RNAi and those fed on
gstk-1(RNAi) and ⁄ or gstk-2(RNAi). It is noteworthy
that the concentration of cis-vaccenic acid methyl ester
was decreased only in double-knockdown (gstk-1 and
gstk-2) worms. Vaccenic acid is the most abundant
fatty acid in phospholipids and triglycerides [17], and
is elongated from palmitoleic acid (16:1x7).
Another phenotypic feature of double-knockdown
(gstk-1 and gstk-2) worms was the impairment of oxygen consumption. It is also noteworthy that vaccenic
acid synthesis and worm respiration are closely linked
to peroxisomal and ⁄ or mitochondrial activities. Interestingly, vaccenic acid is an important component of
cardiolipin in different animal species [18], and this
phospholipid plays a key role in mitochondrial function, particularly at the respiratory chain level [19].
Although the link between decreased vaccenic acid
levels and impairment in oxygen consumption merits
further investigation, the presence of altered phenotypes
only in double-knockdown worms indicates compensatory roles for GSTK-1 and GSTK-2 and suggests overlapping functions. As peroxisomes and mitochondria

FEBS Journal 276 (2009) 5030–5040 ª 2009 The Authors Journal compilation ª 2009 FEBS


E. Petit et al.

are metabolically linked, cooperate and cross-talk,
especially in the b-oxidation of various fatty acids and
in the maintenance of homeostasis in cellular reactive

oxygen species [20,21], our results further strengthen
the close relationship between these two organelles. It
is noteworthy that specific knockdown of either gstk-1
or gstk-2 was not accompanied by an upregulation of
the paralogous gene. Such compensatory responses
have been demonstrated in Gst alpha 4 (Gsta4) or Gst
zeta 1 (Gstz1) knockout mice, where the expression of
other Gst classes and antioxidant enzymes was induced
[22,23]. By contrast, knockout of Gst pi1 ⁄ 2 (Gstp1 ⁄ 2)
did not lead to the upregulation of at least class alpha
and mu transferases, as reported by Henderson et al.
[24]. As the C. elegans genome contains more than 50
genes encoding zeta, sigma, pi and omega class GSTs
[25], further studies are needed to determine whether
GSTs belonging to these classes or other genes are
upregulated in gstk-1 and ⁄ or gstk-2 knockout worms.
With regard to the potential role of GST kappa,
either direct or indirect, in lipid metabolism and respiration, it stills remain unclear. One hypothesis might
be that GST kappa plays a role in the folding of proteins involved in lipid synthesis or respiration. Indeed,
it has been demonstrated that GST kappa shares
sequence and secondary structure homology with
E. coli protein disulfide bond isomerase (DsbA) and
has the same general folding as DsbA. The DsbA family is a subfamily of the thioredioxin family and catalyses disulfide bond formation during the folding of
secreted proteins in bacterials [26]. Recently, Liu et al.
[5] have shown that mouse and human GST kappa,
renamed DsbA-L by these authors, are highly
expressed in adipose tissue and interact with adiponectin. Adiponectin is an adipokine specifically secreted
from adipose tissue, which plays a key role in glucose
and lipid metabolism in insulin-sensitive tissues [27].
Overexpression of GST kappa promotes adiponectin

multimerization by the formation of disulfide bonds
between trimers [5]. Although there is no adiponectin
gene in the C. elegans genome, GST kappa might have
conserved such a role in the regulation of protein multimerization or interaction. Certain proteins involved in
lipid metabolism can exist as both monomers and
dimers, for example the fatty acid synthase complex,
and it has been demonstrated that some desaturases
also form dimers [28,29]. Similarly, several mitochondrial proteins form disulfide-linked multimeric complexes [30]. Thus, a possible role of GST kappa might
be to favour specific protein–protein interactions, and
the absence of such interactions in double-knockdown
(gstk-1 and gstk-2) worms may lead to lipid metabolism
and respiration impairment. Further investigation will

Glutathione transferase kappa in C. elegans

be needed to confirm this hypothesis and to determine
which proteins might be regulated by GST kappa.
In conclusion, this work has allowed the characterization of two GST kappa genes, gstk-1 and gstk-2, in
the C. elegans genome. The products of these genes are
differentially expressed in worm tissues and show distinct subcellular localizations, namely peroxisomal for
GSTK-1 and mitochondrial for GSTK-2. Specific
repression of each gene has no consequences on the
worm phenotype. By contrast, double-knockdown
(gstk-1 and gstk-2) worms show decreased vaccenic
acid levels and lower oxygen consumption when
compared with wild-type worms.

Materials and methods
Caenorhabditis elegans strains
Caenorhabditis elegans cultures were grown and maintained

at 20 °C using NGM agar plates supplemented with
5 lgỈmL)1 of cholesterol [10]. The wild-type reference strain
Bristol N2 was used. All experiments were performed at
20 °C.

Fluorescent-tagged protein constructs and the
production of transgenic animals
Reporter gene constructs were obtained by a PCR fusionbased approach [31]. Genomic gstk-2 (D2024.7), with
1.8 kb immediately upstream of the start codon, was PCR
amplified from wild-type genomic DNA using a TripleMasterPCR System (Eppendorf, Hamburg, Germany). This
product was then coamplified with a 1.8 kb PCR fragment
containing the GFP coding sequence and the 3¢ untransformed region (UTR) of unc-54 (from plasmid pPD95.75
kindly provided by A. Fire). For gstk-1 (ZK1320.1), a 1.1 kb
promoter fragment was amplified and fused with the GFP
coding sequence and then coamplified with the gstk-1 genomic and 3¢ UTR. Sequences were checked and the resulting
gfp::gstk-1 and gstk-2::gfp fragments were microinjected [32]
at 50 ngỈlL)1 into the syncytial gonad of young wild-type
adult hermaphrodites, together with 200 ngỈlL)1 of the plasmid pRF4 containing the dominant marker rol-6(su1006)
[33]. For each construct, at least three independent lines were
analysed for expression.

Immunofluorescence microscopy
Routinely, fluorescence expression patterns and phenotypic
analyses were carried out on a Zeiss axioskop 2 plus
equipped with Nomarski optics (Zeiss, Le Pecq, France).
Confocal stacks of images every 0.3–0.5 lm were captured
on an inverted Leica SP2 confocal microscope (Leica,
Rueil-Malmaison, France). Z projections were analysed

FEBS Journal 276 (2009) 5030–5040 ª 2009 The Authors Journal compilation ª 2009 FEBS

5037


Glutathione transferase kappa in C. elegans

E. Petit et al.

using Image J software and then processed using Adobe
PhotoshopÒ. To stain mitochondria, animals were incubated for 10 min with 10 lm of MitoTracker Red (Invitrogen Molecular Probes, Cergy Pontoise, France), and then
moved to a fresh plate for 2 h. Worms were anaesthetized
in either 1 mgỈmL)1 levamisole or 10 lm azide.

RNAi experiments
RNAi by feeding bacteria was performed using the N2
strain, as described previously [34,35], with the following
modifications. The bacterial HT115(DE3) E. coli strains
used for feeding experiments were obtained from J. Ahringer (University of Cambridge, UK) (gstk-1, ref: WBRNAi00021881) and OpenBiosystems (Fisher Scientific-Open
Biosystems, Illkirch, France) (gstk-2, ref : RCE1182). In
brief, control and RNAi strain cultures were grown for 6 h
in LB medium containing 100 mgỈmL)1 ampicillin, and
then spread onto NGM agar containing isopropyl thio-b-dgalactoside (1 mm) and carbenicillin (25 lgỈmL)1). For double-RNAi treatment, equal concentrations of both strains
were mixed before seeding. The next day, animals at the
fourth larval stage were placed onto RNAi plates, grown
for 2 days and then harvested by rinsing with M9 buffer
(0.1 m NaCl, 0.05 m potassium phosphate, pH 6.0). Adults
were allowed to settle, and eggs were recovered from
hermaphrodites by alkaline hypochlorite lysis (5 min at
room temperature in 0.5 m NaOH, 5% hypochlorite) [36].
The eggs were rinsed with M9 buffer and the resulting L1

larvae were transferred the next day to fresh agar plates
containing the different dsRNA conditions.

Lifespan assays
First-generation progeny from RNAi and control conditions were picked at the fourth larval stage and transferred
onto fresh RNAi plates. The day of the shift was counted
as day 0 in the adult lifespan assay. To prevent mixing test
worms with their progeny during the reproduction period,
adult nematodes were transferred daily to fresh plates.
Monitoring of lethality was performed every day and
worms were considered to be dead when they failed to
move, either spontaneously or in response to touch, and
showed no pharyngeal pumping. Worms that crawled off
the plate were excluded (considered to have escaped). One
hundred worms per condition were used in each lifespan
experiment, conducted in triplicate. Nematode survival was
analysed by the Kaplan–Meier method using Graphpad
Prism 5. The same software was used to test the equality of
survival with the log-rank (Wilcoxon) test.

Oxygen consumption assays
Oxygen consumption rates were measured using a DW1 ⁄ AD
Clark-type oxygen electrode (Hansatech, Norfolk, UK).

5038

Young adult worms that were maintained on NGM agar
plates covered with the corresponding RNAi bacteria were
washed twice and resuspended in 50 lL of M9, and then
transferred into the chamber already containing 450 lL of

M9 buffer, and respiration was measured at 20 °C for at
least 10 min. All washes and measurements were performed
in oxygenated M9 buffer. Samples were carefully recovered
from the chamber and the number of worms was counted.
For each condition, the mean rate was calculated from
triplicate experiments.

Western blotting
To prepare total extracts, worm pellets were resuspended in
Laemmli sample buffer, vortexed three times for 15 s after
the addition of broken glass beads, and then denatured for
5 min at 100 °C and separated by 10% SDS–PAGE.
Proteins were transferred to nitrocellulose membranes
(Schleicher & Schuell BioScience, Dassel, Germany) and
probed with the mouse anti-GFP IgG1k (Roche Diagnostics, Meylan, France). Immunoreactive proteins were
revealed with a chemiluminescent detection system (SuperSignal Pico Chemiluminescent Substrate; Pierce Inc.,
Rockford, IL, USA).

Lipid analyses
Aliquots of C. elegans were crushed in 2 mL of methanol–
5 mm EGTA (2 : 1, v ⁄ v) with an Ultra Turax; 100 lL of
homogenate were evaporated and the pellet was dissolved
in 0.25 mL of NaOH (0.1 m) overnight for protein measurement using the Bio-Rad assay. For each analysis, lipids
from the homogenate were extracted according to Bligh
and Dyer [37] in chloroform–methanol–water (2.5 : 2.5 :
2.1, v ⁄ v ⁄ v) in the presence of the internal standards. For
total fatty acid analysis, lipids from a 200 lL homogenate
were extracted and transmethylated with 1 mL
BF3 ⁄ CH3OH (SUPELCO 10% w ⁄ w) for 1 h at 150 °C.
FAMEs were extracted with 2 mL of hexane–1 mL of

water. The organic phase was evaporated to dryness and
dissolved in 20 lL of ethyl acetate. One microlitre of
FAME was analysed by gas–liquid chromatography [38] on
a 5890 Hewlett Packard system using a Famewax RESTEK
fused silica capillary column (30 m · 0.32 mm inside diameter, 0.25 mm film thickness). The oven temperature was
programmed from 110 to 220 °C at a rate of 2 °CỈmin)1
and the carrier gas was hydrogen (0.5 bar). The injector
and detector were maintained at 225 and 245 °C, respectively. Finally, 2 lg of glyceryl triheptadecanoate were used
as internal standard.
For free fatty acid analysis, 400 lL of homogenate were
extracted and dissolved in 1 mL of hexane. Free fatty acids
were transmethylated in 1 mL of BF3 ⁄ CH3OH (10% w ⁄ w)
for 5 min at room temperature and free FAMEs were
extracted with 2 mL of hexane–1 mL of water. The organic

FEBS Journal 276 (2009) 5030–5040 ª 2009 The Authors Journal compilation ª 2009 FEBS


E. Petit et al.

phase was evaporated to dryness and dissolved in 10 lL of
ethyl acetate. Analysis was performed as above with 1 lg
of nonadecanoic acid as internal standard. All chemicals
were obtained from Sigma-Aldrich, Lyon, France.

Acknowledgements
This work was supported in part by the Institut
´
´
National de la Sante et de la Recherche Medicale,

Centre National de la Recherche Scientifique and
Association pour la Recherche sur le Cancer. Elise
Petit was founded by the Ligue National Contre le
Cancer and Xavier Michelet by the Association pour
la Recherche contre le Cancer. We are grateful to Professor A. Guillouzo and Drs E. Culetto and B. Fromenty for critical reading of the manuscript. The
Imaging and Cell Biology Facility of the IFR87
(FR-W2251) ‘La plante et son environnement’ is sup`
ported by the Action de Soutien a la Technologie et la
Recherche en Essonne, Conseil de l’Essonne.

References
1 Harris JM, Meyer DJ, Coles B & Ketterer B (1991)
A novel glutathione transferase (13-13) isolated from
the matrix of rat liver mitochondria having structural
similarity to class theta enzymes. Biochem J 278,
137–141.
2 Ladner JE, Parsons JF, Rife CL, Gilliland GL &
Armstrong RN (2004) Parallel evolutionary pathways
for glutathione transferases: structure and mechanism
of the mitochondrial class kappa enzyme rGSTK1.
Biochemistry 43, 352–361.
3 Robinson A, Huttley GA, Booth HS & Board PG
(2004) Modelling and bioinformatics studies of the
human Kappa-class glutathione transferase predict a
novel third glutathione transferase family with similarity
to prokaryotic 2-hydroxychromene-2-carboxylate
isomerases. Biochem J 379, 541–552.
4 Morel F, Rauch C, Petit E, Piton A, Theret N, Coles B
& Guillouzo A (2004) Gene and protein characterization of the human glutathione S-transferase kappa and
evidence for a peroxisomal localization. J Biol Chem

279, 16246–16253.
5 Liu M, Zhou L, Xu A, Lam KS, Wetzel MD, Xiang R,
Zhang J, Xin X, Dong LQ & Liu F (2008) A disulfidebond A oxidoreductase-like protein (DsbA-L) regulates
adiponectin multimerization. Proc Natl Acad Sci USA
105, 18302–18307.
6 Thomson RE, Bigley AL, Foster JR, Jowsey IR,
Elcombe CR, Orton TC & Hayes JD (2004) Tissuespecific expression and subcellular distribution of
murine glutathione S-transferase class kappa.
J Histochem Cytochem 52, 653–662.

Glutathione transferase kappa in C. elegans

7 Jowsey IR, Thomson RE, Orton TC, Elcombe CR &
Hayes JD (2003) Biochemical and genetic characterization of a murine class Kappa glutathione S-transferase.
Biochem J 373, 559–569.
8 Bartlett K & Eaton S (2004) Mitochondrial betaoxidation. Eur J Biochem 271, 462–469.
9 Poirier Y, Antonenkov VD, Glumoff T & Hiltunen JK
(2006) Peroxisomal beta-oxidation – a metabolic pathway with multiple functions. Biochim Biophys Acta
1763, 1413–1426.
10 Brenner S (1974) The genetics of Caenorhabditis elegans.
Genetics 77, 71–94.
11 Labrousse AM, Zappaterra MD, Rube DA & van der
Bliek AM (1999) C. elegans dynamin-related protein
DRP-1 controls severing of the mitochondrial outer
membrane. Mol Cell 4, 815–826.
12 Timmons L & Fire A (1998) Specific interference by
ingested dsRNA. Nature 395, 854.
13 Motley AM, Hettema EH, Ketting R, Plasterk R &
Tabak HF (2000) Caenorhabditis elegans has a single
pathway to target matrix proteins to peroxisomes.

EMBO Rep 1, 40–46.
14 Togo SH, Maebuchi M, Yokota S, Bun-Ya M,
Kawahara A & Kamiryo T (2000) Immunological
detection of alkaline-diaminobenzidine-negative
peroxisomes of the nematode Caenorhabditis elegans:
purification and unique pH optima of peroxisomal
catalase. Eur J Biochem 267, 1307–1312.
15 Yokota S, Togo SH, Maebuchi M, Bun-Ya M,
Haraguchi CM & Kamiryo T (2002) Peroxisomes of the
nematode Caenorhabditis elegans: distribution and
morphological characteristics. Histochem Cell Biol 118,
329–336.
16 Gurvitz A, Langer S, Piskacek M, Hamilton B, Ruis H
& Hartig A (2000) Predicting the function and subcellular location of Caenorhabditis elegans proteins similar to
Saccharomyces cerevisiae beta-oxidation enzymes. Yeast
17, 188–200.
17 Tanaka T, Ikita K, Ashida T, Motoyama Y, Yamaguchi
Y & Satouchi K (1996) Effects of growth temperature on
the fatty acid composition of the free-living nematode
Caenorhabditis elegans. Lipids 31, 1173–1178.
18 Wolff RL, Combe NA & Entressangles B (1985)
Positional distribution of fatty acids in cardiolipin of
mitochondria from 21-day-old rats. Lipids 20, 908–914.
19 Mileykovskaya E, Zhang M & Dowhan W (2005)
Cardiolipin in energy transducing membranes.
Biochemistry (Mosc) 70, 154–158.
20 Camoes F, Bonekamp NA, Delille HK & Schrader M
(2009) Organelle dynamics and dysfunction: a closer
link between peroxisomes and mitochondria. J Inherit
Metab Dis 32, 163–180.

21 Schrader M & Yoon Y (2007) Mitochondria and
peroxisomes: are the ‘big brother’ and the ‘little sister’
closer than assumed? BioEssays 29, 1105–1114.

FEBS Journal 276 (2009) 5030–5040 ª 2009 The Authors Journal compilation ª 2009 FEBS

5039


Glutathione transferase kappa in C. elegans

E. Petit et al.

22 Engle MR, Singh SP, Czernik PJ, Gaddy D, Montague
DC, Ceci JD, Yang Y, Awasthi S, Awasthi YC &
Zimniak P (2004) Physiological role of mGSTA4-4, a
glutathione S-transferase metabolizing 4-hydroxynonenal: generation and analysis of mGsta4 null mouse.
Toxicol Appl Pharmacol 194, 296–308.
23 Fernandez-Canon JM, Baetscher MW, Finegold M,
Burlingame T, Gibson KM & Grompe M (2002)
Maleylacetoacetate isomerase (MAAI ⁄ GSTZ)-deficient
mice reveal a glutathione-dependent nonenzymatic
bypass in tyrosine catabolism. Mol Cell Biol 22,
4943–4951.
24 Henderson CJ, Smith AG, Ure J, Brown K, Bacon EJ
& Wolf CR (1998) Increased skin tumorigenesis in mice
lacking pi class glutathione S-transferases. Proc Natl
Acad Sci USA 95, 5275–5280.
25 McElwee JJ, Schuster E, Blanc E, Piper MD, Thomas
JH, Patel DS, Selman C, Withers DJ, Thornton JM,

Partridge L et al. (2007) Evolutionary conservation of
regulated longevity assurance mechanisms. Genome Biol
8, R132.
26 Ito K & Inaba K (2008) The disulfide bond formation
(Dsb) system. Curr Opin Struct Biol 18, 450–458.
27 Oh DK, Ciaraldi T & Henry RR (2007) Adiponectin in
health and disease. Diabetes Obes Metab 9, 282–289.
28 Behrouzian B & Buist PH (2002) Fatty acid desaturation: variations on an oxidative theme. Curr Opin Chem
Biol 6, 577–582.
29 Zhang S, Yang Y & Shi Y (2005) Characterization of
human SCD2, an oligomeric desaturase with improved
stability and enzyme activity by cross-linking in intact
cells. Biochem J 388, 135–142.
30 Bihlmaier K, Mesecke N, Terziyska N, Bien M, Hell K
& Herrmann JM (2007) The disulfide relay system of
mitochondria is connected to the respiratory chain.
J Cell Biol 179, 389–395.
31 Hobert O (2002) PCR fusion-based approach to create
reporter gene constructs for expression analysis in
transgenic C. elegans. BioTechniques 32, 728–730.
32 Mello C & Fire A (1995) DNA transformation.
Methods Cell Biol 48, 451–482.
33 Kramer JM, French RP, Park EC & Johnson JJ (1990)
The Caenorhabditis elegans rol-6 gene, which interacts
with the sqt-1 collagen gene to determine organismal
morphology, encodes a collagen. Mol Cell Biol 10,
2081–2089.

5040

34 Kamath RS, Martinez-Campos M, Zipperlen P,
Fraser AG & Ahringer J (2001) Effectiveness of
specific RNA-mediated interference through ingested
double-stranded RNA in Caenorhabditis elegans.
Genome Biol 2, 1–10.
35 Timmons L, Court DL & Fire A (2001) Ingestion of
bacterially expressed dsRNAs can produce specific and
potent genetic interference in Caenorhabditis elegans.
Gene 263, 103–112.
36 Emmons SW, Klass MR & Hirsh D (1979) Analysis of
the constancy of DNA sequences during development
and evolution of the nematode Caenorhabditis elegans.
Proc Natl Acad Sci USA 76, 1333–1337.
37 Bligh EG & Dyer WJ (1959) A rapid method of total
lipid extraction and purification. Can J Biochem Physiol
37, 911–917.
38 Lillington JM, Trafford DJ & Makin HL (1981) A
rapid and simple method for the esterification of fatty
acids and steroid carboxylic acids prior to gas–liquid
chromatography. Clin Chim Acta 111, 91–98.

Supporting information
The following supplementary material is available:
Fig. S1. gstk-1 and gstk-2 are specifically silenced by
RNAi feeding.
Table S1. Total fatty acid composition of wild-type,
gstk-1, gstk-2 and gstk-1 ⁄ gstk-2 double-knockdown
worms.
Table S2. Free fatty acid composition of wild-type,
gstk-1, gstk-2 and gstk-1 ⁄ gstk-2 double-knockdown

worms.
Table S3. Neutral lipid composition of wild-type, gstk1, gstk-2 and gstk-1 ⁄ gstk-2 double-knockdown worms.
Table S4. Phospholipid composition of wild-type, gstk1, gstk-2 and gstk-1 ⁄ gstk-2 double-knockdown worms.
This supplementary material can be found in the
online article.
Please note: As a service to our authors and readers,
this journal provides supporting information supplied
by the authors. Such materials are peer-reviewed and
may be re-organized for online delivery, but are not
copy-edited or typeset. Technical support issues arising
from supporting information (other than missing files)
should be addressed to the authors.

FEBS Journal 276 (2009) 5030–5040 ª 2009 The Authors Journal compilation ª 2009 FEBS