Báo cáo khoa học: The secretory omega-class glutathione transferase OvGST3 from the human pathogenic parasite Onchocerca volvulus pot

The secretory omega-class glutathione transferase
OvGST3 from the human pathogenic parasite
Onchocerca volvulus
Eva Liebau
1
, Jana Ho
¨
ppner
1
, Mareike Mu
¨
hlmeister
1
, Cora Burmeister
2
, Kai Lu
¨
ersen
1
,
Markus Perbandt
3
, Christel Schmetz
4
, Dietrich Bu
¨
ttner
4
and Norbert Brattig
4
1 Institute of Animal Physiology, University of Mu

¨
nster, Germany
2 Institute of Parasitology, Justus-Liebig-University, Giessen, Germany
3 Institute of Biochemistry, Center for Structural and Cell Biology, University of Lu
¨
beck, Germany
4 Bernhard Nocht Institute, Hamburg, Germany
The glutathione S-transferases (GSTs) constitute a
highly versatile superfamily that is thought to have
evolved from a thioredoxin-like ancestor in response to
the development of oxidative stress, sharing sequence
and structural similarities with several stress-related
proteins in a widespread range of organisms. Addition-
ally, several GST-related proteins have been described,
demonstrating that this ancient protein fold has been
‘recycled’ by nature for new functions, such as plant
stress-induced proteins, bacterial stringent starvation
Keywords
glutathione S-transferase; nematode;
Onchocerca; parasite
Correspondence
E. Liebau, Institute of Animal Physiology,
University of Mu
¨
nster, Hindenburgplatz 55,
Mu
¨
nster 48143, Germany
Fax: +49 251 8321766
Tel: +49 251 8321710

E-mail:
Database
Additional sequence data obtained in this
study have been reported to GenBank. The
original sequence data available under
accession number AF203814 have been
changed accordingly
(Received 9 February 2008, revised 22 April
2008, accepted 1 May 2008)
doi:10.1111/j.1742-4658.2008.06494.x
Onchocerciasis or river blindness, caused by the filarial nematode Oncho-
cerca volvulus, is the second leading cause of blindness due to infectious
diseases. The protective role of the omega-class glutathione transferase 3
from O. volvulus (OvGST3) against intracellular and environmental reactive
oxygen species has been described previously. In the present study, we con-
tinue our investigation of the highly stress-responsive OvGST3. Alternative
splicing of two exons and one intron retention generates five different tran-
script isoforms that possess a spliced leader at their 5¢-end, indicating that
the mechanism of mature mRNA production involves alternative-, cis- and
trans-splicing processes. Interestingly, the first two exons of the ovgst3 gene
encode a signal peptide before sequence identity to other omega-class gluta-
thione transferases begins. Only the recombinant expression of the isoform
that encodes the longest deduced amino acid sequence (OvGST3 ⁄ 5) was
successful, with the purified enzyme displaying modest thiol oxidoreductase
activity. Significant IgG1 and IgG4 responses against recombinantly
expressed OvGST3 ⁄ 5 were detected in sera from patients with the general-
ized as well as the chronic hyperreactive form of onchocerciasis, indicating
exposure of the secreted protein to the human host’s immune system and
its immunogenicity. Immunohistological localization studies performed at
light and electron microscopy levels support the extracellular localization

of the protein. Intensive labeling of the OvGST3 was observed in the egg
shell at the morula stage of the embryo, indicating extremely defined,
stage-specific expression for a short transient period only.
Abbreviations
CDNB, 1-chloro-2,4-dinitrobenzene; GSH, reduced glutathione; GST, glutathione S-transferase; mf, microfilaria; Ni-NTA, nickel–nitrilotriacetic
acid; SL, spliced leader.
3438 FEBS Journal 275 (2008) 3438–3453 ª 2008 The Authors Journal compilation ª 2008 FEBS
proteins, yeast nitrogen metabolism regulator URE2,
the c-subunit of the elongation factor 1 or even ion
channels [1].
A prominent catalytic activity of the GSTs is the
conjugation of reduced glutathione (GSH) to numer-
ous electrophilic substrates, usually promoting their
inactivation, degradation and excretion. The family is
characterized by a broad range of substrate specificity
with low affinity K
m
values. This lower catalytic effi-
ciency has probably been an integral part of the evolu-
tion of GSTs as detoxifiers of a wide range of
endogenous and environmental chemicals. Moreover, a
growing number of nondetoxification functions have
now been attributed to GSTs, essentially making them
multifunctional enzymes, devoted to various aspects of
cell defense. They participate in the catabolism of aro-
matic amino acids, the synthesis of eicosanoids, the
binding and transport of potentially toxic nonsubstrate
molecules (ligands), the clearance of oxidative stress
products and they can physically interact with kinases
involved in signal transduction [2–5]. Evidence of func-

tional flexibility can be found within one particular
GST-class, as exemplified by the sigma-class, where
members fulfill structural functions (e.g. S-crystallins)
[6] or participate in prostaglandin synthesis [7]. Inter-
estingly, comparative studies of free-living and para-
sitic nematodes demonstrated prostaglandin synthesis
activity only in sigma-class GSTs of parasites [8].
Whereas the model nematode Caenorhabditis elegans
only has cytosolic sigma-class GSTs, the filarial para-
site Onchocerca volvulus has the secreted form OvGST1
that acts as a prostaglandin D
2
synthase directly at the
parasite–host interface, making an interception of the
local immune response appear to be feasible [9,10].
Divergent preferences of ligands, such as hemin, have
also been observed within the same GST-class of free-
living and parasitic nematodes, with this function
appearing to be an adaptation to parasitism or, specifi-
cally, to blood feeding [11]. Similarly, kinetic and
structural data obtained from the sole GST from the
malarial parasite Plasmodium falciparum indicate that
the enzyme optimized its binding property with the
parasitotoxic hemin rather than its catalytic efficiency
towards electrophilic compounds, possibly responding
to specific evolutionary pressures [12,13].
Distinct from the prototypical tyrosine or serine resi-
dues characteristic of other GST-classes, the omega-
class has a cysteine residue in the active site that can
form a mixed disulfide bond with GSH. It is therefore

not surprising that the omega-class GSTs have a dis-
tinct substrate profile, most notably GSH-dependent
thiol-transferase and dehydroascorbate reductase activ-
ity, reflecting their structural similarity to glutaredoxins
[14]. Recently, their participation in the multistep bio-
transformation of inorganic arsenic has been demon-
strated and variations in the human omega-1 genes
have been found that modify the age-at-onset of Alz-
heimer and Parkinson diseases [15]. Other described
functions of the omega-class GSTs include a modula-
tion of ryanodine receptor calcium release channels
[16], a participation in the post-translational processing
of interleukin-1b in monocytes [17] and synthesis of an
important intermediate in drosopterin biosynthesis [18].
A role of omega-class GSTs in the oxidative stress
response has been shown [19,20], including studies of
the omega-class GST from the human pathogenic fil-
arial worm O. volvulus (OvGST3) [21,22]. In the pres-
ent study, we continue our investigation of the
OvGST3. Gene analysis identified an additional exon
at the 5¢-end that encodes the first part of a signal
peptide. Alternative splicing of two exons and one
intron retention results in five different transcripts that
have the spliced-leader (SL1) trans-spliced to their
5¢-end. To analyze the capacity of the secretory protein
to stimulate host immune responses, the antibody
responses of onchocerciasis patients against the recom-
binant OvGST3 were studied. Immunohistological
localization by light and electron microscopy
demonstrates an intensive staining of the egg shell at

the morula stage of the embryo, indicating defined
expression for a short transient period only.
Results and Discussion
Genomic structure and alternative splicing of
the OvGST3
The gene of the ovgst3 was isolated by screening an
O. volvulus lambda Fix II genomic library. In addition
to the previously described ovgst3 gene structure [22],
438 bp of 5¢-upstream region and one new exon,
encoding the first part of a signal peptide, were identi-
fied. The gene now consists of 2117 bp composed of
eight exons and seven introns (Fig. 1). The nucleotide
sequence at the splice junctions is consistent with the
canonical GT-AG rule. The cDNA sequence confirms
the intron–exon boundaries predicted from the geno-
mic sequence. Interestingly, we discovered one cDNA
where exon 5 was absent. Using 5¢ RACE, additional
cDNA clones were obtained and a total of five differ-
ent types of mRNA variants were detected. These were
generated by exon skipping (‘alternative splice region’)
(Fig. 1) and one intron retention (intron 5), with the
potential to produce five different proteins (OvGST3 ⁄
1–OvGST3 ⁄ 5). The three isoforms OvGST3 ⁄ 1to
OvGST3 ⁄ 3 identified by Kampko
¨
tter et al. [22] have a
E. Liebau et al. Omega-class glutathione transferase from O. volvulus
FEBS Journal 275 (2008) 3438–3453 ª 2008 The Authors Journal compilation ª 2008 FEBS 3439
novel 5¢-exon previously not described. The isoform
OvGST3 ⁄ 5 encodes the longest deduced amino acid

sequence and most likely is the ancestral form because
it is thought that, in multi-intron genes, constitutive
splicing predates exon skipping [23]. Retention of
intron 5 results in the inclusion of a premature termi-
nation codon exon, potentially producing the truncated
OvGST3 ⁄ 1. The isoforms OvGST3 ⁄ 2, OvGST3 ⁄ 3 and
OvGST3 ⁄ 4 either do not include cassette exons 4 or 5
or both, respectively (Fig. 1). Up to now, scarce infor-
mation about alternative splicing in GST genes is
available; however, this mechanism is probably more
prevalent than previously assumed. Unlike the
situation observed for the ovgst3, spliced transcripts of
other described GSTs appear to share the same
N-terminus involved in glutathione binding, whereas
splicing occurs in the C-terminal domain, conferring
variation in substrate specificity and expanding the
substrate range of the enzyme [24].
The most comprehensive structural and functional
information about alternative spliced variants comes
from the insect delta-class GSTs in the Anopheline
mosquitoes. All alternative transcripts share a common
NH
2
-terminal domain (exon 2), which is spliced to one
of several alternative exons encoding variable COOH-
termini to yield mature transcripts. The resulting alter-
natively spliced products share high amino
acid sequence identity but possess different catalytic
efficiencies and substrate specificities. Splicing is an
efficient means of expanding substrate diversity recog-

nized by GSTs with a minimal increase in gene dupli-
cation [25,26]. Recently, it has been shown that
individual GST-isoforms from insects can differentially
interact with components of the c-Jun N-terminal
kinase pathway and their role as positive or negative
regulators of signalling through this pathway is
suggested [27].
Alternative splicing is a powerful mechanism generat-
ing multiple forms of mRNA from a single gene and
thereby expanding the diversity of expressed transcripts.
The system either produces nonfunctional truncated
proteins or proteins with altered regulation, distribution
or physiological function. In an alternative mode, alter-
native splicing can also function as an on ⁄ off switch by
producing mRNA in which translation is suppressed
due to the presence of a premature termination codon,
such as the one observed in the OvGST3 ⁄ 1-mRNA.
Blotting of O. volvulus homogenate followed by immun-
odetection with affinity-purified anti-OvGST3 serum
revealed a faint band of around 18 kDa only after pro-
longed staining and it is not clear whether this protein is
the OvGST3 ⁄ 1 (predicted molecular mass without
signal peptide = 17.4 kDa), the OvGST3⁄ 3(19.9kDa),
a proteolytic product or even a nonspecific cross-
reacting antigen (data not shown). Therefore, it
remains uncertain whether the OvGST3 ⁄ 1-transcript is
AAAA
AAAA
AAAA
AAAA

AAAA
OvGST3/5
OvGST3/1
OvGST3/2
OvGST3/4
OvGST3/3
E1
E1
E1
E1
E1
E2
E2
E2
E2
E2
E3
E3
E3
E3
E3
E4
E4
E4
E4
E5
E5
E5
E5
E6

E6
E6
E6
E7
E7
E7
E7
E8
E8
E8
E8
E1 E2 E3 E6 E7 E8
SLA
SLB
SLA
SLA
SLA
SLA
SLB
* stop
23 bp 77 bp 133 bp 97 bp 123 bp 115 bp 158 bp 72 bp
372 bp 217 bp 97 bp 216 bp 63 bp 133 bp 221 bp 122 bp
438 bp
Alternative splice region
Fig. 1. Schematic diagram of alternative splicing in the ovgst3 gene. Schematics illustrate the exon and intron organization of the ovgst3
gene, the location of the ‘alternative splice region’ (exon 4 and 5) and five different isoforms. Exons are shown as numbered boxes. The five
different transcripts (OvGST3 ⁄ 1–OvGST3 ⁄ 5) obtained, possess the spliced leader sequence at their 5¢ end, indicating that the mechanism of
mature mRNA production involves both cis- and trans-splicing processes. SLA indicates the acceptor site located 119 nucleotides from the
start codon ATG, SLB is an alternative acceptor site located 30 nucleotides downstream of the acceptor site SLA. Retention of intron 5 intro-
duces a stop codon (marked by the asterisk) in the ORF, leading to a premature termination of translation.

Omega-class glutathione transferase from O. volvulus E. Liebau et al.
3440 FEBS Journal 275 (2008) 3438–3453 ª 2008 The Authors Journal compilation ª 2008 FEBS
translated into a truncated protein or is a candidate
for nonsense-mediated mRNA decay. In addition to its
role in eliminating faulty transcripts and thereby pre-
venting the accumulation of truncated and potentially
toxic protein fragments, nonsense-mediated mRNA
decay also has a role in controlling gene expression
and is implicated in several essential physiological pro-
cesses [28–30].
Protein sequence analysis
An alignment with representative sequences from dif-
ferent GST classes (data not shown) was used to
generate a phylogenetic tree, clearly grouping the
OvGST3 into the omega-class (Fig. 2A). The
OvGST3 ⁄ 5 was aligned with the omega-class GST1
from human (GSTO1) (AF212303) demonstrating
approximately 30% identity. Residues that contribute
to the binding of GSH, as described for the human
omega-class GST, are either conserved or conserva-
tively replaced. Another distinguishing feature is the
active site cysteine (Cys33). The first two exons of the
ovgst3 gene encode a signal peptide before sequence
identity to other omega-class GSTs begins. The pre-
dicted cleavage site and the start of the mature
protein lies between amino acid residues Ala20 and
Ile21 (Fig. 2B).
Even though the human GSTO1 does not have a
signal peptide, the enzyme was recently found in spu-
tum supernatant, whereas intracellular markers were

negative. This demonstrates that GSTO1 is excreted
into airway secretions, where its role in the mainte-
nance of GSH homeostasis in the extracellular space is
postulated [31].
To obtain a clearer picture of the potential effects
of splicing on protein structure, a model of the
OvGST3 ⁄ 5 was generated based on the structure of
the human omega-class GSTO1 (Fig. 3). In general,
the principal isoform OvGST3 ⁄ 5 consists of two
domains that are linked by a loop between helices a3
and a4. The N-terminal thioredoxin-like domain har-
bours the glutathione-binding (G)-site. The G-site is
formed by helix a2, by residues connecting helix a2
and strand b3 and by a segment connecting strand b4
to helix a3. The C-terminal domain is largely a-helical
and consists of five a-helices that are connected by a
variety of loops.
Based on the full-length model, we have deduced
models of the observed splice variants (Fig. 3A–D).
Translation of the OvGST3 ⁄ 1 transcript leads to a
truncated protein with questionable conformation of
a4 (Fig. 3A). Whereas the N-terminal thioredoxin-
like domain is maintained, the complete C-terminal
domain with the exception of a4 is lost. In the iso-
form OvGST3 ⁄ 2, helices a3, a4 and b4 are missing
(Fig. 3B) and loss of exon 4 leads to the alternative
isoform OvGST3 ⁄ 4, lacking b2, a2 and b3 (Fig. 3C).
The lack of the fragment encoded by exons 4 and 5
forces the most drastic changes in structure and
results in the protein product OvGST3 ⁄ 3 missing a

2,
b2, b3, b4, a3 and a4 (Fig. 3D). Removing these
important secondary structures would certainly affect
folding and especially function because the G-site is
destroyed.
In general, GSTs are biologically active as homodi-
mers. The interactions occurring at the intersubunit
interface of the homodimers are dominated by hydro-
phobic interactions between residues from domain 1 of
one subunit and domain 2 of the other. Because many
subunit interface residues are located in a4, a5 and b3,
exon 4 and ⁄ or exon 5 deletions will break the con-
served subunit interactions at the dimer interface area.
Accordingly, none of the truncated splicing forms will
have the ability to form intact dimers.
Although it has been possible to confirm the
ovgst3-splice variants at transcript level, it is impor-
tant to analyze whether these splice variants are
actually translated into proteins or whether isoforms
with extreme deletions are misfolded and quickly
degraded. To obtain evidence at the protein level,
western blot analysis of homogenate of adult
O. volvulus was carried out using affinity-purified
anti-OvGST3 ⁄ 5 (Fig. 4B,C). Surprisingly, only one
dominant isoform of approximately 30 kDa was
observed, corresponding to the long isoform
OvGST3 ⁄ 5. Whereas western blotting revealed signifi-
cant expression of the principle splice isoform
OvGST3 ⁄ 5 in adult female worms, only minor levels
were detected in adult males (Fig. 4C). This result is

in good agreement with immunolocalization of the
OvGST3, where intensive staining is observed in the
egg shell (Figs 6 and 7).
The ‘alternative splice region’ of the ovgst3, com-
prising both exon 4 and 5 (Figs 1 and 2b), is almost
identical to exon 4 of the human omega-class GST2
(gsto2). Pronounced skipping of exon 4 is the only
observed alternative splicing difference, affecting
GSTO2-transcripts. Calarco et al. [32] demonstrated
that GSTO2 transcripts that include or skip exon 4
have similar stabilities. However, transient expression
in HeLa cells resulted in minor protein levels of the
exon 4-skipped splice variant, indicating that
skipping leads to expression of an unstable protein.
Levels of active GSTO2 are thus determined by
expression of the exon 4-containing splice variant.
Interestingly, in chimpanzees, skipping of exon 4
E. Liebau et al. Omega-class glutathione transferase from O. volvulus
FEBS Journal 275 (2008) 3438–3453 ª 2008 The Authors Journal compilation ª 2008 FEBS 3441
is not so pronounced, resulting in species-specific
differences in the expression of the active splice vari-
ant of GSTO2 [32].
Because no animal host has been identified that can
be used to provide the various stages of O. volvulus in
quantity (e.g. in particular, the infectious larvae can
A
B
Fig. 2. Phylogenetic analysis of the OvGST3. (A) After CLUSTALW multiple alignment of the indicated GST proteins (data not shown),
sequences were adjusted manually using
BIOEDIT, version 5.0.9 [55] and phylogenetic relationships were estimated using MEGA, version 3.1

[56]. The accession numbers of the compared proteins are: Arabidopsis thaliana (A.t.ph, CAA72413), Caenorhabditis elegans (C.e.o,
NP_498728; C.e.z, CAA91449), Drosophila melanogaster (D.m.d, NP_524326), Homo sapiens (H.s.a, AAB24012; H.s.m, AAA60963; H.s.o,
AAF73376; H.s.pi, NP_000843; H.s.t, NM_000854; H.s.z, AAC33591), Musca domestica (M.d.d, CAA43599; M.d.s, AAA03434), Mus muscu-
lus (M.m.a, AAI32577; M.m.m, P10649; M.m.o, NP_034492; M.m.t, CAA66666), Nostoc punctiforme (N.p.l, ZP_00105965), Ommastrephes
sloani (O.s.s, M36938); Onchocerca volvulus (O.v.o, AAF99575; O.v.pi, P46427; O.v.s, AAG44696), Ostreococcus tauri (O.t.l, CAL49924),
Petunia · hybrida (P.h.ph, CAA68993), Rattus norvegicus (R.n.pi, AAB59718). (B) Sequence alignment of OvGST3 ⁄ 5 from O. volvulus and
the human omega-class hsGSTO1. Residues of the assumed active site are shown underlined and in bold. Residues that are identical are
contained in black boxes and are indicated by an asterisk (*), whereas sequence similarity is indicated by a colon (:). Gaps indicated by a
dash were introduced to optimize the alignment. The secondary structural elements a-helices are colored red and b-strands are in blue.
Arrows indicate positions of exons and intron–exon boundaries. The putative signal peptide (italics) is based on prediction made by
SIGNALP
software, with the proposed cleavage site between amino acid residues Ala20 and Ile21.
Omega-class glutathione transferase from O. volvulus E. Liebau et al.
3442 FEBS Journal 275 (2008) 3438–3453 ª 2008 The Authors Journal compilation ª 2008 FEBS
only be obtained by dissecting infected blackflies), it
is not possible to perform western blots of different
developmental stages of O. volvulus. Furthermore, this
inaccessibility of O. volvulus means that in vitro inves-
tigations of stress-responsive genes at the protein level
cannot be performed, comprehensive studies of the
oxidative stress-response are unfeasible and partial
purification of possibly existing low-abundant
isoforms is impossible. Therefore, questions regarding
possible stage- or stress-regulated OvGST3-isoform
expression cannot be settled conclusively.
AB
CD
Fig. 3. Ribbon presentation of a three dimensional model of the
OvGST3 ⁄ 5. The model is based on the structure of the human
omega-class GSTO1 (protein databank code 1EEM) with a-helices

colored red and b-strands in blue. The N to C direction of the struc-
tural elements can be deduced by the labeling of the secondary
structures. The four splice isoforms (A, OvGST3 ⁄ 1; B, OvGST3 ⁄ 2;
C, OvGST3 ⁄ 4; D, OvGST3 ⁄ 3) are mapped onto the OvGST3 ⁄ 5 iso-
form. Deletions in the splice isoform are shown in green. It is likely
that splicing will cause the structures (A–D) to fold in a substantially
different fashion.
12345
12 3 45
1
- 47.5
- 32.5
- 25.0
- 16.5
23
r
OvGST3
- rOvGST3/5
A
B
C
Fig. 4. Characterization of recombinant OvGST3 ⁄ 5 and affinity puri-
fication of anti-OvGST3 serum. (A) Bottom panel: Coomassie-stained
SDS ⁄ PAGE [12.5% (w ⁄ v) gel]; top panel: corresponding western
blot probed with affinity-purified anti-OvGST3 ⁄ 5. Supernatant- (lane
1) and pellet-fraction (lane 2) of Escherichia coli BLR DE3 containing
pJC40-OvGST3 ⁄ 5. Lane 3, flow-through from the nickel-affinity
chromatography after loading the E. coli supernatant, followed by
NTA-purification step (lane 4) and gelfiltration (lane 5). (B) The
obtained anti-OvGST3 antibody was purified by affinity chromatogra-

phy using OvGST3 ⁄ 5 immobilized on CNBr-activated Sepharose 4B.
Western blot of extract of E. coli overexpressing OvGST3 ⁄ 5. Lane
1, anti-OvGST3 prior to affinity purification; lanes 2–5, eluted anti-
body fractions; only fractions 6 ⁄ 7 (lane 5) were used for western
blot and immunolocalization experiments. (C) Immunoblot showing
the abundance of OvGST3 in male and female O. volvulus homo-
genate. Lanes 1 and 2, 100 lg of female and male worms,
respectively; lane 3, lysate of E. coli Origami DE3 containing pJC40-
OvGST3 ⁄ 5 as a control. Immunodetection was carried out using
fractions 6 ⁄ 7 of the affinity-purified OvGST3-antibody.
E. Liebau et al. Omega-class glutathione transferase from O. volvulus
FEBS Journal 275 (2008) 3438–3453 ª 2008 The Authors Journal compilation ª 2008 FEBS 3443
Expression of the recombinant OvGST3 ⁄ 5 and
substrate specificities
To investigate the enzymatic characteristics of the
OvGST3 ⁄ 5, the enzyme was expressed in Escherichia
coli using various vectors containing different con-
ventional affinity tags and fusion partners. In all host
systems used, the recombinant protein accumulated
intracellularly in insoluble aggregates (Fig. 4A, lane 2).
The addition of 1% Triton X-100 to the lysis buffer
improved the extraction of soluble rOvGST3 ⁄ 5. Due
to the insolubility of the enzyme, purification of the
recombinant OvGST3 ⁄ 5(rOvGST3 ⁄ 5) was difficult.
Even though expression of rOvGST3 ⁄ 5 with the fusion
partner maltose-binding protein resulted in enhanced
yields and increased solubility, subsequent site specific
proteolysis to remove the fusion partner resulted in
almost immediate protein aggregation and loss of yield
(data not shown). Therefore, a modified protocol for

autoinduction of protein expression was used and
approximately 0.4 mg of rOvGST3 ⁄ 5 was purified
from 1 L liquid culture, using conventional nickel–
nitrilotriacetic acid (Ni-NTA) affinity purification
(Fig. 4A, lane 4). Unfortunately, due to their insolubil-
ity, purification of the other recombinant OvGST3
isoforms has not been achieved under native conditions,
and their biological function remains speculative.
To identify catalytic activities that may reveal the
biological function of the OvGST3 ⁄ 5, the substrate
specificity of the recombinant enzyme with a broad
range of substrates was determined. Elimination of the
His-tag by factor Xa did not influence enzyme activity.
The purified enzyme was able to use GSH as an elec-
tron donor to reduce hydroxyethyl disulfide
(57.9 ± 11.7 nmolÆmin
)1
Æmg
)1
) and showed rather low
GSH conjugating activity towards 1-chloro-2,4-dinitro-
benzene (CDNB) (113.8 ± 22.1 nmolÆmin
)1
Æmg
)1
).
There was no detectable activity with the substrates
dimethylarsenic acid, S-(4-nitrophenacyl)glutathione
and cumene hydroperoxid (data not shown).
The omega-class GST has a cysteine residue in the

active site that can form a mixed disulfide bond with
GSH. Therefore, conjugating reactions with GSH can
only be performed if the disulfide bond is not formed
or broken down in the catalytic mechanism. The low
CDNB-conjugating activity observed for the OvGST3
should thus be interpreted with caution because it
might also be due to the active site cysteine rapidly
reacting with CDNB. The enzymatic activities
observed in the present study are in contrast to the
findings of Kampko
¨
tter et al. [22] who designed a
recombinant protein short of seven amino acids at the
N-terminus. Furthermore, Kampko
¨
tter et al. [22] dem-
onstrated that the OvGST3 reacts with trans-2-none-
nal, possibly indicating an involvement in the
elimination of end products of lipid peroxidation.
The thiol oxidoreductase activity is reminiscent of
glutaredoxins and also characteristic for the omega-
class, where dethiolation of specific S-glutathionylated
proteins that accumulate under stress conditions has
been proposed as a possible function, with the open
and not particularly hydrophobic H-site being large
enough to accommodate protein substrates [33].
Because the OvGST3 is dramatically up-regulated at
the steady-state transcription level in response to oxi-
dative stress and reacts sensitively to alterations in
redox status [21,22], a role of the enzyme in reversible

S-glutathionylation and glutathione-mediated redox
regulation of proteins is feasible.
Antibody response to the secretory OvGST3 ⁄ 5
The mechanism by which helminths down-regulate
host immunity at the molecular level is the subject of
intense research. Immunologists have focused on excre-
tory–secretory products and surface molecules because
these have the capacity to actively shape the immuno-
logical environment. In the present study, we investi-
gated whether the secretory OvGST3 is recognized by
antibodies generated in patients infected with O. vol-
vulus. We studied the reactivities of IgG1 and IgG4 by
ELISA applying sera from 117 patients with onchocer-
ciasis, including 77 patients with the hyporeactive gen-
eralized form and 40 patients with the chronic
hyperreactive form (also designated as sowda). Signifi-
cantly elevated IgG1 and IgG4 titers (P < 0.001) were
found on comparing the reactivitity of the patient sera
with those from 20 healthy Europeans as a control
(Fig. 5A). As a positive control for OvGST3, we
included another O. volvulus antigen, the fatty acid-
and retinol-binding protein Ov20, which is strongly
immunogenic [34]. In comparison to the very high
IgG1 and IgG4 reactivities with the Ov20 antigen, the
responses against OvGST3 were significantly lower
(P < 0.0001).
With regard to the IgG1 and IgG4 reactivities in
subgroups of the onchocerciasis patients, we found
modest higher IgG1 titers in sera from generalized
patients with high microfilaria (mf) density as well as

with the hyperreactive form compared to patients with
the generalized form and low mf numbers (P < 0.017
and P < 0.033, respectively) (Fig. 5B). The IgG4 titer
for the patients with the generalized form and high mf
density showed significantly higher reactivity
(P < 0.007) compared to the hyperreactive form of
onchocerciasis.
Omega-class glutathione transferase from O. volvulus E. Liebau et al.
3444 FEBS Journal 275 (2008) 3438–3453 ª 2008 The Authors Journal compilation ª 2008 FEBS
These results correspond to earlier observations,
where high IgG1 levels to O. volvulus antigens were
found predominantly in patients with high mf densities
who were exposed to higher levels of filarial antigens
and in patients with the chronic hyperreactive form; in
the present study, IgG4 levels were lower compared to
patients with a high mf load [34–36]. These findings
indicate an exposure of the human immune system to
the secretory OvGST3 antigen. The resulting antibody
profile is characteristic of the varying forms of oncho-
cerciasis that reflect different immune states. In the
present study, OvGST3 was shown to be an antigen of
low immunogenicity, comparable to the results obtained
for other enzymatic antioxidants from O. volvulus such
as the superoxide dismutase 1 (OvSOD1) or the
OvGST2 [37].
Immunolocalization studies clearly show a short
developmental stage-specific expression of the OvGST3
and a major localization in the egg shell. O. volvulus
completes embryogenesis and the larvae hatch and leave
the egg shell before leaving the maternal uterus. How-

ever, uterus fluid is continuously released by female
worms. Furthermore, there is a turnover in adult worm
populations and proteins are exposed when the adult
worm dies and degenerates. The restricted antibody
response to the OvGST3 might therefore be due to the
limited presence of the OvGST3 in the external environ-
ment of the parasite or due to low immunogenicity.
Immunohistological localization by light and
electron microscopy
We used immunohistochemistry to determine the
stage- and tissue-specific distribution of the unusual
secretory omega-class OvGST3. Using the 1 : 100 or
1 : 250 diluted yolk collected before immunization, no
staining of any tissue of female or male O. volvulus
was detected. The preimmune yolk did not contain any
antibodies against O. volvulus. Following immuniza-
tion, strong staining of the egg shells around morulae
was seen. This staining was almost completely removed
following absorption of the antibodies using rOvGST3.
This indicated the high specificity of the antibodies for
OvGST3 [38]. For further analyses, the pooled frac-
tions 6 ⁄ 7 of the affinity purified antibodies were used
(Fig. 4B). Strong staining was observed in the egg
shells surrounding several stages of the developing
embryos in the uterus of worms (Fig. 6). Oocytes in
the ovary and oocytes or zygotes in the uterus were
negative (Fig. 6A,B). Weak staining was first seen in
young morulae (i.e. the stage where the egg shell first
appears) (Fig. 6B). The staining intensity increased
Fig. 5. IgG1 and IgG4 responses of patients with generalized and

hyperreactive onchocerciasis to recombinantly expressed
OvGST3 ⁄ 5. (A) Endpoint titers for IgG1 and IgG4 reactivities in sera
from 117 patients with onchocerciasis (Ov) with OvGST3 ⁄ 5 and
Ov20 compared to 20 healthy European controls (EC). Significant
differences (P < 0.0001) in the titers were found for all patients
groups compared to the control sera as well as between the titers
for OvGST3 and Ov20 in the respective groups. (B) Comparison of
the serum titers found for the patients with the generalized form of
onchocerciasis and low mf density (1Mf l), high Mf density (Mf h),
the hyperreactive (sowda) form (Sow) and healthy controls (EC) in
response to OvGST3 ⁄ 5. The P-values for IgG1 were between
0.017 comparing patients with high and low mf densities and 0.033
comparing patients with low mf density and chronic hyperreactive
onchocerciasis, respectively, indicating weak differences (P < 0.05
when corrected for multi-comparison). When comparing the IgG4
response of the generalized form showing high mf densities with
the sowda form, P = 0.007.
E. Liebau et al. Omega-class glutathione transferase from O. volvulus
FEBS Journal 275 (2008) 3438–3453 ª 2008 The Authors Journal compilation ª 2008 FEBS 3445
with the development of the morulae as long as the
shell was attached to the embryo (Figs 6C and 7A–D).
The egg shells of coiled and stretched mf and those
from which the mf had hatched, were distinctly but
less intensively stained (Figs 6D,E and 7E,F). This
staining pattern was also observed in female worms
from four other species of the genus Onchocerca but
not in five species belonging to other genera of the
family Onchocercinae [38].
Degenerating embryos showed stronger staining of
the egg shell than normal mf (Fig. 6D). This is best

observed in the degenerated embryos following antifi-
A
B
C
D
E
G
H
F
Fig. 6. Lightmicroscopic immunolocalization of OvGST3 within the egg shell of embryos in the uterus of O. volvulus. (A–E) Untreated
patients. (A) Oocytes in the ovary are not labeled (arrow). (B) Oocytes or zygotes in the uterus are negative (arrow), whereas the egg shells
of young morulae are weakly labeled (arrowheads). (C) Mature morulae show strongly labeled egg shells (arrowheads). (D) The shells of
coiled microfilaria (mf) are still slightly labeled (arrow) and those of degenerating embryos are more strongly labeled (arrowheads). The mf
are negative. (E) The mf are negative (arrow) but some still show well labeled shells (arrowheads). (F) Whereas degenerated morulae pres-
ent strongly labeled shells (arrowheads), oocytes or zygotes are negative (arrows). Ten months after 4 weeks of doxycycline treatment. (G)
Labeling of the shells of young morulae (arrow) and stronger labeling of degenerating mature morulae (arrowheads). Six weeks after suramin
treatment. (H) The shells of normal coiled mf are slightly labeled and the mf are negative (arrow), whereas the degenerated stretched mf
are strongly labeled (arrow heads). Typical finding 2 months after ivermectin treatment. The hypodermis and the epithelia of ovary and uterus
are negative. Immunostaining using fraction 6 ⁄ 7 (diluted 1 : 20) of the purified antibody against OvGST3. Scale bar = 40 lm.
Omega-class glutathione transferase from O. volvulus E. Liebau et al.
3446 FEBS Journal 275 (2008) 3438–3453 ª 2008 The Authors Journal compilation ª 2008 FEBS
larial treatment with doxycycline or suramin
(Figs 6F,G) or following a single dose of ivermectin,
mainly causing degeneration of the stretched mf
(Fig. 6H).
The tissues of male and female worms were usually
not, or only weakly, stained (Figs 6 and 7), and the
sperms never labeled. Using light microscopy, some
worms showed staining of the hypodermis, the epithe-
lia of uterus and intestine and the afibrillar inner

portions of the muscles [38]. Using electron micro-
scopy, we did not find labeling of the morulae
(Figs 7B,D) or the uterus epithelium adjacent to the
egg shell, making prediction of the production site of
the OvGST3 impossible. Using light microscopy, we
observed distinct labeling of the outer cells of the
morulae (Fig. 6C); however, because this finding is not
supported by electron microscopy, it may also be an
artifact.
In conclusion, the immunohistological examinations
showed specific labeling of the OvGST3 in the egg shell
of developing embryos of O. volvulus. The staining
appeared to be stronger in the shells of degenerating
untreated and drug-treated embryos.
The extracellular environment is highly oxidizing
and, unsurprisingly, most secreted surface proteins are
rich in disulfides. The maintenance of a reduced state
of surface thiols requires protein disulfide oxidoreduc-
tase and also GSH [39]. It is conceivable that surface
thiols of the egg shell are early targets of oxidative
stress. This is particularly evident for short-lived oxi-
dants and those that cannot easily permeate into the
cells. Because their location makes them particularly
sensitive to extracellular oxidants, egg shell proteins
might play a key role as sensors that signal any changes
in redox state to the embryo as it moves forward to the
proximal part of the uterus. In this respect, a potential
A
B
D C

F
E
Fig. 7. Electron microscopic localization of
OvGST3 within the egg shell of embryos in
the uterus of O. volvulus. (A, B) Morula with
an egg shell that is well labeled (arrows in
B). (C) Microfilaria with a well labeled egg
shell (arrows). The mf and the epithelium of
the uterus is negative. (D) Degenerated
morula cell with well labeled shell (arrows).
(E, F) Negative uterus epithelium and well
labeled shell (F, arrows) shed by stretched
mf. Immunogold labeling using fraction 6 ⁄ 7
(diluted 1 : 500) of the purified antibody
against OvGST3. ba, endobacterium; mf,
microfilaria; mo, morula; ut, uterus. (A–E)
Scale bar = 1 lm. (F) Scale bar = 0.5 lm.
E. Liebau et al. Omega-class glutathione transferase from O. volvulus
FEBS Journal 275 (2008) 3438–3453 ª 2008 The Authors Journal compilation ª 2008 FEBS 3447
role of the OvGST3 in the regulation of exofacial pro-
tein function is feasible. Changes in the outside envi-
ronment might be caused by drug treatment with
doxycycline, ivermectin and suramin, leading to degen-
erate embryos and a more pronounced staining due to
the induction of the highly stress-responsive OvGST3.
The nematode egg shell is thought to be formed by
secretion of egg shell components shortly after fertil-
ization by the embryo itself. Its composition is highly
variable but a general plan has been outlined descri-
bing the egg shell as a multilamellate structure with an

inner proteolipid layer, a central mechanically resistant
chitin–protein complex and outer layers synthesized in
a co-ordinated way by both the embryo and the uter-
ine epithelium [40,41]. Recently, the interactome net-
work of C. elegans was analyzed with high-throughput
yeast, two-hybrid screens providing functional hypo-
theses for thousands of uncharacterized proteins [42].
An interaction was observed between the homologous
omega-class GSTO3 (K10F12.4) and a putative egg
shell protein Y47D7A.13. Both proteins possess a sig-
nal peptide, indicative of their extracellular localiza-
tion. After the signal peptide, Y47D7A.13 has a region
with clusters of cysteine residues, followed by a highly
repetitive glycine-, alanine, proline-, glutamine- and
tyrosine-rich domain, reminiscent of cuticle compo-
nents of nematodes that are extensively cross-linked
[43]. Investigations of the homologous omega-class
GSTO3 from C. elegans are under way to analyse how
the enzyme is involved in the complex modifications,
folding and processing of the egg shell.
Experimental procedures
Parasites and human sera
Onchocercomas had been previously removed from
untreated onchocerciasis patients and after treatment with
doxycycline, ivermectin and suramin in several studies con-
ducted in Burkina Faso, Ghana, Liberia and Uganda under
local anaesthesia [44–47]. Nodulectomies for research pur-
poses had been approved by the Ethics Commission of the
Medical Board, Hamburg, and by ethic committees and
medical authorities in the respective countries. The studies,

including nodulectomies and collection of sera, conformed
to the principles of the Helsinki declaration of 1975 (and for
the later studies as revised 1983 and 2000). For biochemical
experiments, the filariae were isolated from the nodules by
incubation in RPMI 1640 containing 0.5% collagenase
(Clostridium histolyticum; Roche, Mannheim, Germany) and
checked for purity of host material and viability micro-
scopically. The parasites were snap frozen and stored in
liquid nitrogen. Sera were collected from 77 patients with
the generalized form of onchocercerciasis, living in endemic
villages in Liberia (n = 24) or Uganda (n = 53) and from
40 patients with the chronic hyperreactive form (sowda) of
onchocerciasis living in Liberia (n = 20) and Ghana
(n = 20) [45,48]. The mf density was in the range
0.1–189 mfÆmg skin
)1
. Thirty-four patients with a general-
ized onchoceriasis had a low (< 10 mfÆmg skin
)1
) and 43
patients had a high (> 96 mfÆmg skin
)1
) mf density. As
control sera, twenty uninfected Europeans were included.
Parasite preparation, SDS ⁄ PAGE and western
blotting
For western blots, adult worms were homogenized with a
glass ⁄ glass homogenizer in 0.1 m Tris–HCl (pH 7.5)
containing 0.1 mm phenylmethanesulfonyl fluoride, with or
without the addition of 1% Triton-X100. The homogenate

was centrifuged at 20 000 g for 45 min. After electrophoret-
ical separation, proteins were transferred to nitrocellulose
membrane, then probed with purified anti-rOvGST3
primary serum, anti-(chicken Ig)-alkaline phosphatase-
labelled serum conjugate (Dianova, Hamburg, Germany)
and the chromogenic substrates nitroblue tetrazolium and
5-bromo-4-chloro-3-indolyl phosphate.
Preparation of total RNA and genomic DNA
from O. volvulus
Adult female worms were homogenized in an all-glass
homogenizer in guanidinium thiocyanate and extracts were
layered on a CsCl step gradient. After centrifugation for 18 h
at 27 000 g (TST 41.14 rotor; Kontron, Eching, Germany),
RNA and DNA were prepared as previously described [49].
Isolation of the gene encoding OvGST3
The entire protein-encoding cDNA of the OvGST3 (acces-
sion number AF203814) [19] was used to screen a lambda
Fix II genomic library made from adult female worms. The
DNA from ten positive phages was analyzed by restriction
enzyme fragmentation and Southern blot hybridization
using the same screening probe. Hybridizing fragments were
cloned into conventional plasmid vectors, followed by
sequencing.
5¢ RACE-PCR and cloning of OvGST3 isoforms
Full-length cDNA corresponding to four isoforms were
cloned by rapid amplification of cDNA ends using the
5¢ RACE system (Invitrogen, Karlsruhe, Germany) accord-
ing to manufacturer’s instructions. Total RNA isolated
from adult worms was used for reverse transcription,
followed by 5¢ RACE-PCR (using the gene-specific primers

5¢-GTAAAGAGATCGATGTCGAAT-3¢ and 5¢-TTAGGC
Omega-class glutathione transferase from O. volvulus E. Liebau et al.
3448 FEBS Journal 275 (2008) 3438–3453 ª 2008 The Authors Journal compilation ª 2008 FEBS
ATTGTAATATCCACTGCC-3¢) and subsequent cloning
into the TOPOII TA cloning vector (Invitrogen). One clone
obtained had a spliced leader (SL1) trans-spliced to its
5¢ end. Therefore, subsequent PCRs were performed with
the SL1-primer (5¢-GCAGGATCCGGTTTAATTACCCA
AGCTTGAG-3¢). Following sequence analyses, the ORFs
obtained were subcloned into a variety of vectors for
expression in the heterologous host E. coli.
To synthesize the isoform OvGST3 ⁄ 5 for heterologous
expression, sense primers encoding the first eight amino
acid residues following the signal peptide [5¢-rs-ATCAGTC
TAAATCAATTTATGTCA-3¢; primers with restriction
sites (rs) to simplify directed in-frame cloning] and antisense
primers encoding the last eight residues of the OvGST3
(5¢-rs-TTAGGCATTGTAATATCCACTGCCGTA-3¢) were
used in PCR with respective cDNAs as template.
Modeling of the OvGST3 ⁄ 5
For illustrative purposes, a 3D model of the OvGST3 ⁄ 5
was generated based on the crystal structure of the homo-
logous structure of human omega-class GST (protein data-
bank code 1EEM). The modeling followed a standard
stepwise procedure, starting with an alignment of the target
sequence onto the template structure. The software used
(modeller 9v2) takes the structural features of the parent
structure and constructs the structural model based on
sequence alignment, including gaps, insertions and mutation
of residues [50]. Validation of the model was performed

using errat ( />ERRAT/), which statistically analyzes the nonbonded inter-
actions between different atom types.
Expression and purification of the OvGST3 ⁄ 5
isoform
A strategy of parallel expression of the OvGST3 ⁄ 5 was
adopted using a variety of vectors containing different con-
ventional affinity tags and fusion partners [pJC40, His-tag;
pASK-IBA2 ⁄ IBA-6, Strep-tag (IBA GmbH, Go
¨
ttingen,
Germany); pMALc2x, Maltose binding protein (NEB Inc.,
Ipswich, MA, USA)] and a variety of E. coli host strains
[BL21 (DE3), BL21 (DE3) pLysS, BL21 (DE3) CodonPlus-
RIL, Rosetta (DE3), Origami (DE3) (Novagen, Madison,
WI, USA)]. Additionally, to improve the solubility of
the expressed proteins, various detergents (Triton-X100,
Tween 20 and Chaps) were added during sonification and
purification.
The Studier method for autoinduction of protein expres-
sion in the T7 system was used [51]. The modified protocol
provides for autoinduction of protein expression without
using inducers during mid-log phase; the bacteria initially
use glucose and, when it is exhausted, lactose can enter the
cell and induce expression of the T7 polymerase. Prelimin-
ary results obtained with small-scale cultures demonstrated
the best yields of recombinant OvGST3 ⁄ 5 using the vector
pJC40 and BLR DE3 E. coli. Briefly, a 1 mL seed culture
of the transformed cells was grown overnight and trans-
ferred to 400 mL of fresh LB medium containing
50 lgÆmL

)1
ampicillin, 25 mm Na
2
HPO
4
,25mm KH
2
PO
4
,
50 mm NH
4
Cl, 5 mm Na
2
SO
4
, 0.5% glycerol, 0.05%
glucose and 0.2% a-lactose monohydrate. After overnight
culture at 20 °C, the bacteria were harvested and the pellet
was resuspended in 10 mm Tris–HCl buffer (pH 8.0)
containing 100 mm NaH
2
PO
4
,20mm imidazole and 1%
Triton-X100 (lysis buffer), sonified and then centrifuged at
10 000 g for 30 min. For enzymatic analysis and antibody
production, the supernatant was purified using affinity
chromatography on a Ni-NTA gel matrix (Qiagen, Hilden,
Germany) under native conditions. For antibody affinity

purification, rOvGST3 ⁄ 5 was used, which was purified
under denaturative conditions using 6 m urea. The homo-
geneity of the enzyme preparation was analyzed by
SDS ⁄ PAGE and the proteins were revealed by Coomassie
blue staining.
Enzyme assays
Recombinantly expressed and purified OvGST3 ⁄ 5 was used
for enzyme assays. The assays were performed in triplicate
from three independent enzyme preparations. The enzyme
activity of the recombinant OvGST3 ⁄ 5 isoform was deter-
mined spectrophotometrically with the substrates CDNB,
cumene hydroperoxide and trans-2-nonenal (Sigma Chemi-
cal Co., Munich, Germany) as described previously [52].
Thiol-transferase activity was determined with hydroxyethyl
disulfide (Sigma-Aldrich Co., Munich, Germany) as sub-
strate [53], the enzymatic reduction of dimethylarsinic acid
(Sigma-Aldrich Co.) was determined as described [15].
S-(phenacyl)glutathione was synthesized and a spectropho-
tometric assay was used to determine the S-phenacyl(glutathi-
one) reductase activity of the OvGST3 ⁄ 5[54].
Antibody production and extraction of
IgY-antibodies
Approximately 200 lg of the purified recombinant
OvGST3 ⁄ 5 was resuspended in 200 lL of NaCl ⁄ P
i
and
mixed with an equal volume of complete Freund’s adjuvant
(Boehringer Diagnostics, Mannheim, Germany). Hens
obtained from a local breeder were immunized by injecting
the antigen into the pectoral muscle. Two booster injections

of 100 lg of antigen mixed with incomplete Freund’s adju-
vant were given 2 and 3 weeks later.
Until antibody extraction, eggs were collected daily and
stored at 4 °C. Egg yolk IgY-antibodies were isolated by
poly(ethylene glycol) precipitation. Briefly, the yolk was col-
lected in a 50 mL screw cap tube, NaCl ⁄ P
i
was added at
twice the volume of the yolk, followed by addition of PEG
6000 (Roth, Karlsruhe, Germany) at 3.5% and incubation
E. Liebau et al. Omega-class glutathione transferase from O. volvulus
FEBS Journal 275 (2008) 3438–3453 ª 2008 The Authors Journal compilation ª 2008 FEBS 3449
of the mix for 30 min at 4 °C under rotation. Following
centrifugation at 21 374 g. for 10 min, the supernatant was
filtered and PEG 6000 was added to a final concentration
of 8.5%. Following centrifugation, the sediment was
dissolved in NaCl ⁄ P
i
and PEG 6000 was added to a final
concentration of 12%. After a final centrifugation, the sedi-
ment was dissolved in 5 mL of NaCl ⁄ P
i
and the antibody
extract was dialyzed against NaCl ⁄ P
i
for at least 24 h.
Affinity purification of antibodies
CNBr-activated Sepharose-4B was used to purify the poly-
clonal OvGST3 antibody. Briefly 9 mg of OvGST3 ⁄ 5 dis-
solved in coupling buffer was incubated with 2 mL of

CNBr-Sepharose (0.66 g of freeze-dried beads reconstituted
in 1 mm HCl) overnight at 4 °C with gentle agitation. The
beads were washed in 0.1 m Tris buffer (pH 8.0) and
blocked in 0.2 m glycine (pH 8.0) for 2 h at room tempera-
ture. To remove excess uncoupled ligand, the absorbent
was washed alternately with high- and low-pH buffer solu-
tions [0.1 m Tris buffer (pH 8.0), followed by 0.1 molÆL
)1
acetate buffer (pH 4.0)]. Finally, the beads were placed in a
column and washed with 10 mL of phosphate-buffered sal-
ine and then 9 mg of the extracted IgY-antibodies were
applied to the column, followed by washing with 2 mL of
0.1 m borate-buffer, 0.5 m NaCl (pH 8.0) and 2 · 10 mL of
NaCl ⁄ P
i
. Antibodies were eluted from the column with
2 mL of 0.1 m glycine (pH 2.5). Eluate was collected in a
15 mL conical tube containing 170 lLof1m Tris (pH 9.0)
and the column was washed again with 5 mL of NaCl ⁄ P
i
.
The flow-through was again applied to the column and sub-
jected to this procedure two more times. Two of the eluates
were always pooled and aliquots were stored at )20 °C
(Fig. 4B).
Measurement of antibody reactivities in sera of
patients with onchocerciasis by ELISA
A semi-quantitative analysis of serum antibody levels to the
recombinant OvGST3 ⁄ 5 protein was performed using
ELISA as described previously [35]. The reactivities of sera

from patients with onchocerciasis were compared with
those of sera from Europeans as a negative control. The
optimal coating concentration of the applied recombinant
proteins (OvGST3 ⁄ 5 and Ov20 as a positive control
antigen) and serum dilutions were determined by a checker-
board titration using a pool of sera from O. volvulus-
infected individuals. Following dialysis against 0.1 m
sodium carbonate buffer (pH 9.6), the OvGST3 ⁄ 5 was
coated to microtiter plates (Nunc, Wiesbaden, Germany) at
a concentration of 2 lgÆmL
)1
. The antigen was incubated
overnight at 4 °C, the plates washed and blocked with 5%
(w ⁄ v) bovine serum albumin in NaCl ⁄ P
i
(pH 7.5). All sera
were used at a dilution of 1 : 1000, 1 : 2000 and 1 : 4000
(v ⁄ v). Bound antibodies were detected with peroxidase-
conjugated antibodies against human IgG1 (HP6070, Cal-
biochem, Karlsruhe, Germany; dilution 1 : 1000) and IgG4
(HP6025, Calbiochem; dilution 1 : 1000) using tetramethyl-
benzidine as the substrate. Between each incubation step,
five washing cycles with NaCl ⁄ P
i
⁄ 0.1% Tween 20 were per-
formed. After the substrate reaction was stopped by 0.5 m
H
2
SO
4

, A
450
was measured. Sera from five Europeans were
used to determine the cut-off level. As a positive control,
one test serum of a patient with generalized onchocerciasis
with known high antibody concentration was included on
each test plate, and only intra- and interassay variations
less than 10% were accepted. The results were expressed as
endpoint titers in arbitrary units (U) [35]. Statistical analy-
sis was performed using the Mann–Whitney U-test for com-
parison of the titers obtained for the tested group of
individuals and the Wilcoxon signed-rank test was applied
for the comparison of the titers for the two antigens tested.
The Bonferroni correction was applied for multiple testing.
P < 0.05 was considered statistically significant.
Immunolocalization
We studied the distribution of the OvGST3 in adult
O. volvulus using immunohistochemistry for light and elec-
tron microscopy. For light microscopy, onchocercomas
from previous studies in Burkina Faso, Ghana, Guatemala,
Liberia and Uganda from 29 untreated patients and from
17 ivermectin-, 18 doxycycline- and eight suramin-treated
patients were used. The nodules were fixed in 4% buffered
formaldehyde, 80% ethanol or modified Karnovsky fixative
and embedded in paraffin. For immunohistochemistry, the
alkaline phosphatase-anti-alkaline phosphatase method was
applied according to the manufacturer’s instructions (Dak-
oCytomaton, Hamburg, Germany). Different eluates of the
affinity-purified antibodies against OvGST3 ⁄ 5 were used as
primary sera. The best results were achieved with the frac-

tions 6 ⁄ 7 at dilutions of 1 : 20 to 1 : 100, but 1 : 500 was
also used. Preimmune purified egg yolk extract was used as
a negative control. Following immunization, purified yolk
extracts preabsorbed with recombinant OvGST3 were used
for specificity controls. As secondary antibodies, polyclonal
rabbit anti-(chicken IgG) (DakoCytomation), monoclonal
mouse anti-(rabbit IgG) (DakoCytomation) and polyclonal
rabbit anti-(mouse IgG) (Dianova, Do
¨
rentrup) sera were
applied. Fast Red TR salt was used as the chromogen, and
hematoxylin functioned as the counterstain.
For electron microscopy, three onchocercomas from
untreated patients were fixed for 24 h (4 °C) in 1% parafor-
maldehyde and 0.025% glutaraldehyde in 0.2 m sodium
cacodylate-HCl buffer (pH 7.2), and then processed for
immunoelectron microscopy. Dehydrated specimens were
embedded in medium-hardness LR White acrylic resin.
Ultrathin sections were incubated with the affinity purified
antibodies raised against OvGST3 ⁄ 5 (fractions 6 ⁄ 7 diluted
1 : 500 in NaCl ⁄ P
i
, 1% BSA), followed by incubation with
Omega-class glutathione transferase from O. volvulus E. Liebau et al.
3450 FEBS Journal 275 (2008) 3438–3453 ª 2008 The Authors Journal compilation ª 2008 FEBS
rabbit anti-(chicken IgG) and incubation in a suspension of
10-nm gold particles coated with protein A. After labeling,
sections were counterstained with 2% aqueous uranyl ace-
tate and Reynold’s lead citrate, followed by examination
with a Philips CM 10 transmission electron microscope

(Philips, Eindhoven, The Netherlands). Pre-immune egg
yolk extract and antibodies against heat shock protein 60
were used as negative controls.
Acknowledgements
This work was supported by the Deutsche Forschungs-
gemeinschaft (DFG project Li 793 ⁄ 5-7). We are
obliged to Dr Kwablah Awadzi for the onchocercomas
from suramin-treated patients. We acknowledge the
technical assistance of Ingeborg Albrecht and Frank
Geisinger [55,56].
References
1 McGoldrick S, O’Sullivan SM & Sheehan D (2005)
Glutathione transferase-like proteins encoded in
genomes of yeasts and fungi: insights into evolution of
a multifunctional protein superfamily. FEMS Microbiol
Lett 242, 1–12.
2 Frova C (2006) Glutathione transferases in the genom-
ics era: new insights and perspectives. Biomol Eng 23,
149–169.
3 Hayes JD, Flanagan JU & Jowsey IR (2005) Glutathione
transferases. Annu Rev Pharmacol Toxicol 45, 51–88.
4 Townsend DM & Tew KD (2003) The role of glutathi-
one-S-transferase in anti-cancer drug resistance. Onco-
gene 22, 7369–7375.
5 Sheehan D, Meade G, Foley VM & Dowd CA (2001)
Structure, function and evolution of glutathione trans-
ferases: implications for classification of non-mamma-
lian members of an ancient enzyme superfamily.
Biochem J 360, 1–16.
6 Piatigorsky J (2003) Crystallin genes: specialization by

changes in gene regulation may precede gene duplica-
tion. Struct Funct Genomics 3, 1–7.
7 Jowsey IR, Thomson AM, Flanagan JU, Murdock PR,
Moore GB, Meyer DJ, Murphy GJ, Smith SA & Hayes
JD (2001) Mammalian class Sigma glutathione S-trans-
ferases: catalytic properties and tissue-specific expression
of human and rat GSH-dependent prostaglandin D2
synthases. Biochem J 359, 507–516.
8 Meyer DJ, Muimo R, Thomas M, Coates D & Isaac RE
(1996) Purification and characterization of prostaglan-
din-H E-isomerase, a sigma-class glutathione S-transfer-
ase, from Ascaridia galli. Biochem J 313, 223–227.
9 Sommer A, Rickert R, Fischer P, Steinhart H, Walter
RD & Liebau E (2003) A dominant role for extracellu-
lar glutathione S-transferase from Onchocerca volvulus
is the production of prostaglandin D2. Infect Immun 71,
3603–3606.
10 Perbandt M, Ho
¨
ppner J, Burmeister C, Lu
¨
ersen K,
Betzel B & Liebau E (2008) Structure of the extracellu-
lar glutathione S-transferase OvGST1 from the human
pathogenic parasite Onchocerca volvulus. J Mol Biol
377, 501–511.
11 van Rossum AJ, Jefferies JR, Rijsewijk FA, LaCourse
EJ, Teesdale-Spittle P, Barrett J, Tait A & Brophy PM
(2004) Binding of hematin by a new class of glutathione
transferase from the blood-feeding parasitic nematode

Haemonchus contortus. Infect Immun 72, 2780–2790.
12 Perbandt M, Burmeister C, Walter RD, Betzel C &
Liebau E (2004) Native and inhibited structure of a Mu
class-related glutathione S-transferase from Plasmodium
falciparum. J Biol Chem 279, 1336–1342.
13 Liebau E, De Maria F, Burmeister C, Perbandt M,
Turella P, Antonini G, Federici G, Giansanti F, Stella
L, Lo Bello M et al.
(2005) Cooperativity and pseudo-
cooperativity in the glutathione S-transferase from
Plasmodium falciparum. J Biol Chem 280, 26121–26128.
14 Whitbread AK, Masoumi A, Tetlow N, Schmuck E,
Coggan M & Board PG (2005) Characterization of the
omega class of glutathione transferases. Methods Enzy-
mol 401, 78–99.
15 Schmuck EM, Board PG, Whitbread AK, Tetlow N,
Cavanaugh JA, Blackburn AC & Masoumi A (2005)
Characterization of the monomethylarsonate reductase
and dehydroascorbate reductase activities of Omega class
glutathione transferase variants: implications for arsenic
metabolism and the age-at-onset of Alzheimer’s and Par-
kinson’s diseases. Pharmacogenet Genomics 15, 493–501.
16 Dulhunty A, Gage P, Curtis S, Chelvanayagam G &
Board P (2001) The glutathione transferase structural
family includes a nuclear chloride channel and a ryano-
dine receptor calcium release channel modulator. J Biol
Chem 276, 3319–3323.
17 Laliberte RE, Perregaux DG, Hoth LR, Rosner PJ,
Jordan CK, Peese KM, Eggler JF, Dombroski MA,
Geoghegan KF & Gabel CA (2003) Glutathione

S-transferase omega 1-1 is a target of cytokine release
inhibitory drugs and may be responsible for their effect
on interleukin-1beta posttranslational processing. J Biol
Chem 78, 16567–16578.
18 Kim J, Suh H, Kim S, Kim K, Ahn C & Yim J (2006)
Identification and characteristics of the structural gene
for the Drosophila eye colour mutant sepia, encoding
PDA synthase, a member of the Omega class glutathi-
one S-transferases. Biochem J 398, 451–460.
19 Dixon DP, Davis BG & Edwards R (2002) Functional
divergence in the glutathione transferase superfamily in
plants. Identification of two classes with putative func-
tions in redox homeostasis in Arabidopsis thaliana.
J Biol Chem 277, 30859–30869.
E. Liebau et al. Omega-class glutathione transferase from O. volvulus
FEBS Journal 275 (2008) 3438–3453 ª 2008 The Authors Journal compilation ª 2008 FEBS 3451
20 Burmeister C, Lu
¨
ersen K, Heinick A, Hussein A,
Domagalski M, Walter RD & Liebau E (2007) Oxida-
tive stress in Caenorhabditis elegans: protective effects of
the Omega class glutathione transferase (GSTO-1).
FASEB J 22, 343–354.
21 Liebau E, Eschbach ML, Tawe W, Sommer A, Fischer
P, Walter RD & Henkle-Du
¨
hrsen K (2000) Identifica-
tion of a stress-responsive Onchocerca volvulus glutathi-
one S-transferase (OvGST-3) by RT-PCR differential
display. Mol Biochem Parasitol 109, 101–110.

22 Kampko
¨
tter A, Volkmann TE, de Castro SH, Leiers B,
Klotz LO, Johnson TE, Link CD & Henkle-Du
¨
hrsen K
(2003) Functional analysis of the glutathione S-transfer-
ase 3 from Onchocerca volvulus (Ov-GST-3): a parasite
GST confers increased resistance to oxidative stress in
Caenorhabditis elegans. J Mol Biol 325, 25–37.
23 Kim E, Goren A & Ast G (2008) Alternative splicing:
current perspectives. BioEssays 30, 38–47.
24 Wongsantichon J & Ketterman AJ (2005) Alternative
splicing of glutathione S-transferases. Methods Enzymol
401, 100–116.
25 Ranson H, Claudianos C, Ortelli F, Abgrall C,
Hemingway J, Sharakhova MV, Unger MF, Collins FH
& Feyereisen R (2002) Evolution of supergene families
associated with insecticide resistance. Science 298,
179–181.
26 Ranson H, Collins F & Hemingway J (1998) The role
of alternative mRNA splicing in generating heterogene-
ity within the Anopheles gambiae class I glutathione
S-transferase family. Proc Natl Acad Sci USA 95,
14284–14289.
27 Udomsinprasert R, Bogoyevitch MA & Ketterman AJ
(2004) Reciprocal regulation of glutathione S-transfer-
ase spliceforms and the Drosophila c-Jun N-terminal
kinase pathway components. Biochem J 383, 483–490.
28 Lewis BP, Green RE & Brenner SE (2003) Evidence

fort he widespread coupeling of alternative splicing and
nonsense-mediated mRNA decay in humans. Proc Natl
Acad Sci USA 100, 189–192.
29 Behm-Ansmant I, Kashima I, Rehwinkel J, Saulie
`
re J,
Wittkopp N & Izaurralde E (2007) mRNA quality con-
trol: an ancient machinery recognizes and degrades
mRNAs with nonsense codons. FEBS Lett 581, 2845–
2853.
30 Longman D, Plasterk RH, Johnstone IL & Ca
´
ceres JF
(2007) Mechanistic insights and identification of two
novel factors in the C. elegans NMD pathway. Genes
Dev
21, 1075–1085.
31 Harju TH, Peltoniemi MJ, Rytila
¨
PH, Soini Y,
Salmenkivi KM, Board PG, Ruddock LW & Kinnula
VL (2007) Glutathione S -transferase omega in the lung
and sputum supernatants of COPD patients. Respir Res
8, 48.
32 Calarco JA, Xing Y, Ca
´
ceres M, Calarco JP, Xiao X,
Pan Q, Lee C, Preuss TM & Blencowe BJ (2007) Global
analysis of alternative splicing differences between
humans and chimpanzees. Genes Dev 21, 2963–2975.

33 Board PG, Coggan M, Chelvanayagam G, Easteal S,
Jermiin LS, Schulte GK, Danley DE, Hoth LR, Griffor
MC, Kamath AV et al. (2000) Identification, character-
ization, and crystal structure of the Omega class gluta-
thione transferases. J Biol Chem 275, 24798–24806.
34 Mpagi JL, Bu
¨
ttner DW, Tischendorf FW, Erttmann
KD & Brattig NW (2000) Humoral responses to a
secretory Onchocerca volvulus protein: differences in the
pattern of antibody isotypes to recombinant Ov20 ⁄ OvS1
in generalized and hyperreactive onchocerciasis. Parasite
Immunol 22, 455–460.
35 Brattig NW, Krawietz I, Abakar AZ, Erttmann KD,
Kruppa TF & Massougbodji A (1994) Strong IgG iso-
typic antibody response in sowdah type onchocerciasis.
J Infect Dis 170, 955–961.
36 Brattig NW (2004) Pathogenesis and host responses in
human onchocerciasis: impact of Onchocerca filariae and
Wolbachia endobacteria. Microbes Infect 6, 113–128.
37 Brattig NW, Henkle-Du
¨
hrsen KH, Hounkpatin S, Lie-
bau E, Kruppa TF & Zipfel PF (1997) Characterization
of human immune responses to the cytosolic superoxide
dismutase and glutathione S-transferase from Oncho-
cerca volvulus. Trop Med Int Health 2, 788–798.
38 Ho
¨
ppner J (2006) Molecular Characterization of the

Gluthation S-Transferases OvGST1, OvGST2 and
OvGST3 from Onchocerca volvulus (Leuckart, 1893) [In
German]. Dissertation, Department of Biology, Faculty
of Mathemathics, Information and Science, University
of Hamburg, Hamburg.
39 Laragione T, Gianazza E, Tonelli R, Bigini P, Mennini
T, Casoni F, Massignan T, Bonetto V & Ghezzi P
(2006) Regulation of redox-sensitive exofacial protein
thiols in CHO cells. Biol Chem 387, 1371–1376.
40 Zahner H, Hobom G & Stirm S (1995) The microfilarial
sheath and its proteins. Parasitol Today 11, 116–120.
41 Selkirk ME, Yazdanbakhsh M, Freedman D, Blaxter
ML, Cookson E, Jenkins RE & Williams SA (1991) A
praline-rich structural protein of the surface sheath of
larval Brugia filarial nematode parasites. J Biol Chem
266, 11002–11008.
42 Li S, Armstrong CM, Bertin N, Ge H, Milstein S, Box-
em M, Vidalain PO, Han JD, Chesneau A, Hao T et al.
(2004) A map of the interactome network of the meta-
zoan C. elegans. Science 303, 540–543.
43 Page A & Johnstone IA (2005) The Cuticle WormBook
(The C. elegans Research Community, ed.). http://
www.wormbook.org.
44 Awadzi K, Hero M, Opoku NO, Addy ET, Bu
¨
ttner
DW & Ginger CD (1995) The chemotherapy of oncho-
cerciasis XVIII. Aspects of treatment with suramin.
Trop Med Parasitol 46, 19–26.
45 Bu

¨
ttner DW, Albiez EJ, von Essen J & Erichsen J
(1988) Histological examination of adult Onchocerca
Omega-class glutathione transferase from O. volvulus E. Liebau et al.
3452 FEBS Journal 275 (2008) 3438–3453 ª 2008 The Authors Journal compilation ª 2008 FEBS
volvulus and comparison with the collagenase technique.
Trop Med Parasitol 39, 390–417.
46 Fischer P, Kipp W, Bamuhiga J, Binta-Kahwa J, Kiefer
A&Bu
¨
ttner DW (1993) Parasitological and clinical
characterization of Simulium neavei-transmitted oncho-
cerciasis in western Uganda. Trop Med Parasitol 44,
311–321.
47 Hoerauf A, Mand S, Volkmann L, Bu
¨
ttner M, Marfo-
Debrekyei Y, Taylor M, Adjei O & Bu
¨
ttner DW (2003)
Doxycycline in the treatment of human onchocerciasis:
kinetics of Wolbachia endobacteria reduction and of
inhibition of embryogenesis in female Onchocerca
worms. Microbes Infect 5, 261–273.
48 Brattig NW, Tischendorf FW, Albiez EJ, Bu
¨
ttner DW
& Berger J (1987) Distribution pattern of peripheral
lymphocyte subsets in localized and generalized form of
onchocerciasis. Clin Immunol Immunopathol 44, 149–159.

49 Liebau E, Wildenburg G, Walter RD & Henkle-Du
¨
hr-
sen K (1994) A novel type of glutathione transferase in
Onchocerca volvulus. Infect Immun 62 , 4762–4767.
50 Eswar N, Marti-Renom MA, Webb B, Madhusudhan
MS, Eramian D, Shen M, Pieper U & Sali A. Compar-
ative protein structure modeling with MODELLER.
Curr Protoc Bioinformatics Supplement 15, 5.6.1–
5.6.30.
51 Studier FW (2005) Protein production by auto-induc-
tion in high density shaking cultures. Protein Expr Purif
41, 207–234.
52 Liebau E, Eckelt VH, Wildenburg G, Teesdale-Spittle
P, Brophy PM, Walter RD & Henkle-Du
¨
hrsen K
(1997) Structural and functional analysis of a glutathi-
one S-transferasefrom Ascaris suum. Biochem J 324,
659–666.
53 Denton H, McGregor JC & Coombs GH (2004) Reduc-
tion of anti-leishmanial pentavalent antimonial drugs by
a parasite-specific thiol-dependent reductase, TDR1.
Biochem J 381, 405–412.
54 Board PG, Coggan M, Cappello J, Zhou H, Oakley AJ
& Anders MW (2007) S-(4-Nitrophenacyl)glutathione is
a specific substrate for glutathione transferase omega
1-1. Anal Biochem 374, 25–30.
55 Hall TA (1999) BIOEDIT: a user-friendly biological
sequence alignment, editor and analysis program for

Windows 95 ⁄ 98 ⁄ NT. Nucleic Acids Symp Ser 41,
95–98.
56 Kumar S, Tamura K & Nei M (2004) MEGA3: inte-
grated software for molecular evolutionary genetics
analysis and sequence alignment. Brief Bioinform 5,
150–163.
E. Liebau et al. Omega-class glutathione transferase from O. volvulus
FEBS Journal 275 (2008) 3438–3453 ª 2008 The Authors Journal compilation ª 2008 FEBS 3453