Molecular cloning, expression and characterization of
protein disulfide isomerase from Conus marmoreus
Zhi-Qiang Wang
1
, Yu-Hong Han
1,2
, Xiao-Xia Shao
1
, Cheng-Wu Chi
1,2
and Zhan-Yun Guo
1
1 Institute of Protein Research, Tongji University, Shanghai, China
2 Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Graduate School of
the Chinese Academy of Sciences, Shanghai, China
In eukaryotes, all proteins that travel along or reside
in the secretory pathway are folded in the endoplas-
mic reticulum (ER). As one of the most important
post-translational modifications, the disulfide bonds
are formed in the ER lumen, where oxidoreductases
catalyze the reaction and serve as disulfide donors [1].
The archetypical oxidoreductase in the ER lumen is
protein disulfide isomerase (PDI). In the oxidized
state, PDI functions as a disulfide donor for its client
proteins. In the reduced state, PDI catalyzes reduction
and isomerization of pre-existing disulfides. The abil-
ity of PDI to function as a reductase, an oxidase and
an isomerase ensures PDI’s ability to serve as a major
catalyst for disulfide formation in vivo [2,3]. Moreover,
PDI also acts as a chaperone for substrates during
catalysis [4].

by protein disulfide isomerase. To elucidate the physiologic roles of pro-
tein disulfide isomerase in the folding of conotoxins, we have cloned a
novel full-length protein disulfide isomerase from Conus marmoreus. Its
ORF encodes a 500 amino acid protein that shares sequence homology
with protein disulfide isomerases from other species, and 70% homology
with human protein disulfide isomerase. Enzymatic analyses of recombi-
nant C. marmoreus protein disulfide isomerase showed that it shared
functional similarities with human protein disulfide isomerase. Using
conotoxins tx3a and sTx3.1 as substrate, we analyzed the oxidase and
isomerase activities of the C. marmoreus protein disulfide isomerase and
found that it was much more efficient than glutathione in catalyzing oxi-
dative folding and disulfide isomerization of conotoxins. We further dem-
onstrated that macromolecular crowding had little effect on the protein
disulfide isomerase-catalyzed oxidative folding and disulfide isomerization
of conotoxins. On the basis of these data, we propose that the C. mar-
moreus protein disulfide isomerase plays a key role during in vivo folding
of conotoxins.
Abbreviations
cPDI, Conus marmoreus protein disulfide isomerase; ER, endoplasmic reticulum; hPDI, human protein disulfide isomerase; GSH, reduced
glutathione; GSSG, oxidized glutathione; nTx3.1, native Tx3.1; PDI, protein disulfide isomerase; RNase A, bovine pancreatic RNase A;
sTx3.1, swap Tx3.1.
4778 FEBS Journal 274 (2007) 4778–4787 ª 2007 The Authors Journal compilation ª 2007 FEBS
One mechanism exploited by cone snails to ensure
efficient folding of conotoxins is adding necessary
folding information to the mature polypeptides. For
example, the C-terminal Gly that is used for amidation
of the mature form of x-conotoxin containing an ami-
dated C-terminus, as isolated from the venom of Conus
magnus (x-MVIIA) can significantly increase the folding
yield [6]. The carboxylation of Glu residues can also

Molecular cloning of a novel PDI from
C. marmoreus
PDI is an abundant protein in the venom ducts of
C. textile, from which two PDIs have been cloned
[5,11]. In present work, we cloned a novel PDI
from C. marmoreus (GenBank accession number
DQ486867). The 1742 bp full-length cDNA includes a
3¢-UTR and a polyadenylation consensus sequence
(AATTATAA) located 12 nucleotides upstream of the
polyA tail. Its 1500 bp ORF encodes a 500 amino acid
protein (Fig. 1A) that shares sequence homology with
PDIs from other species. A signal peptide (17 amino
acids) predicted by the signalp program [12] is present
at its N-terminus, and a typical ER retention signal,
RDEL, is present at its C-terminus. The mature
cPDI protein has a calculated molecular mass of
54 913.7 Da and an isoelectric point of 4.6. The cPDI
contains four thioredoxin domains and an acidic C-ter-
minal tail (a, b, b¢, a¢ and c). Two thioredoxin active
sites (WCGHCK) are found in the a and a¢ domains,
respectively. The cPDI shares 94% amino acid
sequence identity with its homologs C. textile PDI 1
and C. textile PDI 2 [5], 70% identity with human
PDI (hPDI), and 42% with yeast PDI.
An unrooted neighbor-joining phylogenetic tree was
obtained by comparing the deduced amino acid
sequences of different PDIs from fungi to mammals
using the mega (Molecular Evolutionary Genetic
Analysis Software, Version 3.1) program, bootstrap:
1000 replications (Fig. 1B).

During oxidative folding, oxidized PDI catalyzes disul-
fide formation through transferring its active site’s
disulfide to dithiols of reduced polypeptide. We ana-
lyzed the oxidase activity of cPDI using reduced tx3a
(20 lm) as substrate. As shown in Fig. 3, the conotoxin
Z Q. Wang et al. Characterization of PDI from Conus marmoreus
FEBS Journal 274 (2007) 4778–4787 ª 2007 The Authors Journal compilation ª 2007 FEBS 4779
A
B
Fig. 1. Comparison of amino acid sequences of PDIs from C. marmoreus and other species. (A) Multiple sequence alignment of PDIs from
human, yeast, C. marmoreus,andC. textile. Identical or similar residues are shaded in black or gray. The potential N-terminal signal peptides are
boxed; the thioredoxin active sites are underlined; and the ER-retention motif is indicated by underlined dashes. (B) A phylogenetic tree con-
structed on the basis of the amino acid sequences of different PDIs, listed with GenBank accession numbers. The scale bar represents 0.1 units.
Table 1. Enzymatic activities of cPDI compared with those of hPDI.
PDI
Reduction activity
a
[10
2
· (DAÆmin
)1
ÆlM PDI
)1
)]
%of
cPDI
Oxidase activity
b
[10
2

bonds [14,15]. Figure 4A shows the HPLC profiles of
the cPDI-catalyzed (1 lm) oxidative refolding of
reduced tx3a at different refolding stages. The refold-
ing was finished after 2 h, and the folding yield was
over 90%. As shown in Fig. 4B, when the concentra-
tion of cPDI increased, the refolding of reduced tx3a
accelerated accordingly. Thus cPDI can catalyze the
oxidative folding of reduced conotoxins in vitro.Itis
logical to expect that this process also occurs in vivo.
The calculated oxidase activity of cPDI was measured
as the initial rate of decrease of reduced tx3a
(Table 2).
Besides cPDI, both oxidized glutathione (GSSG) and
molecular oxygen can also oxidize dithiols to form
disulfide bonds [16], and both of them are present in
the ER lumen. To compare the roles of these different
oxidants during the folding of conotoxins, the refolding
of reduced tx3a was carried out in three different sys-
tems (Fig. 5). In all these systems, the final refolding
yields were approximately 90 ± 5%. When molecular
oxygen dissolved in the buffer was used as an oxidant,
the folding of reduced tx3a was barely detectable at
the start and finally reached equilibrium 10 h later
(Fig. 5A). As with air oxidation of reduced bovine pan-
creatic RNase A (RNase A) [17], a significant lag time
( 60 min) caused by formation of partially and⁄ or
fully oxidized intermediates was observed when molec-
ular oxygen was used as an oxidant (Fig. 5A). The
refolding intermediates were collected, alkylated by
N-ethylmaleimide or 4-vinylpyridine, and analyzed by

(data no shown). When GSSG was used as an oxidant,
the folding process was expedited, and the lag time was
diminished ( 6 min), as shown in Fig. 5B. When cPDI
(at a final concentration of 2 lm) was added to the
GSSG refolding system, the refolding process further
accelerated (Fig. 5B). The calculated molar specific oxi-
dase activities of cPDI and GSSG are listed in Table 2:
cPDI was about 268-fold more effective than GSSG in
promoting disulfide formation.
cPDI-catalyzed disulfide isomerization
of swap Tx3.1
Besides having an oxidase function, PDI can also cata-
lyze the isomerization of non-native disulfide bonds
[2,3]. To date, there are no reports of the PDI-cata-
lyzed disulfide isomerization of conotoxins. Here, we
used a homogeneous non-native conotoxin isomer,
swap Tx3.1 (sTx3.1), to study the isomerase activity of
cPDI. As shown in Fig. 3, Tx3.1 is an 18 amino acid
conotoxin with three disulfide bonds [14]. During the
oxidative refolding of reduced Tx3.1, two major fold-
ing products, native Tx3.1 (nTx3.1) and sTx3.1, were
formed at the final folding stage (first trace of
Fig. 6A). The molecular mass of sTx3.1 as measured
by MS was identical to that of nTx3.1. After modifica-
tion by N-ethylmaleimide under denatured condition
(6 m urea), its molecular mass did not change. Thus,
sTx3.1 was a fully oxidized isomer that had similar
thermodynamic stability to the native form. We
purified sTx3.1 and used it as the cPDI substrate in
the isomerase activity assay. The traces b–e in Fig. 6A

M cPDI). Filled circles, native tx3a in the presence of PDI; open circles, linear tx3a in the presence of PDI;
filled squares, native tx3a in the absence of PDI; open squares, linear tx3a in the absence of PDI. At different reaction times, the refolding
mixture was acidified and immediately analyzed by RP-HPLC. The amounts of native and linear tx3a were calculated from their elution peak
areas. The data are the average of three independent experiments. For the rate of decrease of the linear form, the original data were fitted
by Y(t) ¼ e
–kt
· 100%, where Y is the percentage of the linear form, and t is the refolding time. For the rate of increase of the native form
in the presence of cPDI, the original data were fitted by Y(t) ¼ Y
max
(1 ) e
–kt
) · 100%, where Y is the percentage of the native form, and
t is the refolding time. For the rate of increase of the native form in the absence of cPDI, the original data were fitted by
Y(t) ¼ Y
max
⁄ [1 + e
–k(t ) a)
] · 100%, where Y is the percentage of the native form, and t is the refolding time.
Characterization of PDI from Conus marmoreus Z Q. Wang et al.
4782 FEBS Journal 274 (2007) 4778–4787 ª 2007 The Authors Journal compilation ª 2007 FEBS
represent the disulfide isomerization process of sTx3.1
(20 lm) catalyzed by cPDI (at a final concentration of
1 lm). cPDI could accelerate disulfide reshuffle, but it
could not shift the equilibrium between nTx3.1 and
sTx3.1, which had similar thermodynamic stability.
Thus, cPDI could not convert all of the sTx3.1 to the
native form. When the concentration of cPDI was
increased, the disulfide isomerization process signifi-
cantly accelerated (Fig. 6B). The calculated molar spe-
cific isomerase activity of cPDI was expressed as the

to those obtained with hirudin, which is a 65 amino
acid peptide containing three disulfide bonds [18].
Discussion
In the present work, we cloned a novel PDI from
C. marmoreus. The cPDI shares high sequence homol-
ogy with PDIs from C. textile and other organisms. It
also has similar biological functions as hPDI, including
disulfide reductase, oxidase and isomerase activities, as
well as chaperone and antichaperone activities. The
high sequence and function conservations support the
hypothesis that all of the current PDIs evolved from a
common ancestral enzyme [19].
We further analyzed the enzymatic activities of cPDI
using its potential endogenous substrates, namely
reduced tx3a and sTx3.1. Both tx3a and Tx3.1 belong
to the M-1 branch of the M-superfamily. The different
branches in the M-superfamily possess different disul-
fide linkages [14]. The oxidative folding properties of
AB
Fig. 6. The isomerase activity of cPDI determined by using sTx3.1 as substrate. (A) (a) The HPLC profile of the refolding mixture of reduced
Tx3.1. The refolding was carried out in the refolding buffer (50 m
M NH
4
CO
3
, pH 8.0, 5 mM GSH, 0.5 mM GSSG) for 8 h. nTx3.1 and sx3.1
are designated as N and S, respectively. (b–e) The HPLC profiles of sTx3.1 refolding mixtures. The disulfide isomerization of sTx3.1 (20 l
M)
was carried out in the refolding buffer (0.1
M Tris ⁄ Cl, pH 7.5, 1 mM EDTA, 0.1 mM GSH) and catalyzed by 1 lM cPDI. At different reaction

M cPDI) (filled circles). At different reaction times, the refolding mixture was acidified and
immediately analyzed by HPLC. The amounts of nTx3.1 and sTx3.1 were calculated from their elution peak areas. The data are the average
of three independent experiments. For the rate of increase of the native form, the original data were fitted by Y(t) ¼ Y
max
(1 ) e
–kt
) · 100%,
where Y is the percentage of the native form, and t is the refolding time. For the rate of decrease of the linear form, the original data were
fitted by Y(t) ¼ [Y
max
+(1) Y
max
)e
–kt
] · 100%, where Y is the percentage of the linear form, and t is the refolding time.
Fig. 8. Effects of macromolecular crowding on the PDI-catalyzed folding of conotoxins. (A) The oxidative folding of reduced tx3a in the
absence (open squares) or presence (filled squares) of crowding agent. The refolding was performed in refolding buffer (0.1
M Tris ⁄ Cl, pH 7.5,
1m
M EDTA, 1 mM GSH, 1 mM GSSG, 2 lM cPDI) in the presence or absence of 200 gÆL
)1
Ficoll 70. At different reaction times, the refolding
mixture was acidified and immediately analyzed by HPLC. The amounts of native and linear tx3a were calculated from their elution peak areas.
(B) The isomerization of sTx3.1 in the absence (open squares) or presence (filled squares) of crowding agent. The isomerization was per-
formed in refolding buffer (0.1
M Tris ⁄ Cl, pH 7.5, 1 mM EDTA, 0.1 mM GSH, 2 lM cPDI) in the presence or absence of 200 gÆL
)1
Ficoll 70.
The amounts of nTx3.1 and sTx3.1 were calculated from their elution peak areas. The data are the average of three independent experiments.
For the rate of increase of the native form in the presence of cPDI, the original data were fitted by Y(t) ¼ Y

Finland). Ni
2+
-chelating Sepharose Fast Flow resin and
Q-Sepharose Fast Flow resin were obtained from Amer-
sham Biosciences (Arlington Heights, IL, USA). The
RACE kit was obtained from Invitrogen (Carlsbad, CA,
USA). Lysozyme, Micrococcus lysodeikticus and RNase A
were products of Sigma (St Louis, MO, USA). Other
reagents were of analytical grade.
Gene cloning of cPDI
The full-length cDNA of cPDI was amplified by RT-PCR
from total RNAs isolated from the venom ducts of
C. marmoreus. The 3¢-end fragment was amplified using a
3¢-RACE adapter primer and a degenerate primer based
on the conserved amino acid sequence (WCGHCK) of
the thioredoxin-like active site found in other PDIs. The
3¢-RACE product was gel-purified, cloned into pGEM-T
easy vector, and sequenced. The nested PCR primers for
5¢-RACE amplification were based on the 3¢-end sequence,
and the 5¢-end fragment was amplified using the nested
primer and the 5¢-RACE adapter primer. The 5¢-RACE
product was also gel-purified, cloned into pGEM-T easy
vector, and sequenced. Primers for amplifying the full-
length cDNA were designed on the basis of these RACE
products. The full-length cDNA of the cPDI was inserted
into an expression vector, pET24a, which contains an
N-terminal His
6
tag.
Expression and purification of cPDI and hPDI

The thiol-protein oxidoreductase activity of PDI was mea-
sured as described previously, using insulin as substrate
[22]. The assay was performed in 0.2 m sodium phosphate
buffer (pH 7.5) containing 8 mm GSH, 30 lm insulin,
120 lm NADPH, 0.5 units of glutathione reductase, 5 mm
EDTA, and 0.7 lm PDI. The assay mixture (without insu-
lin and PDI) was equilibrated at 25 °C, and the NADPH
oxidation rate was recorded against a reference cuvette con-
taining NADPH, EDTA and buffer only. Subsequently,
insulin was added, and a stable nonenzymatic rate was
recorded. Finally, PDI was added, and the total NADPH
oxidation rate was recorded.
The oxidase and isomerase activities of PDI were mea-
sured using the refolding assay of fully reduced RNase A as
previously described [17,23,24]. Briefly, it was performed in
the assay solution (0.1 m Tris ⁄ Cl, pH 8.0, 0.2 mm GSSG,
1mm GSH, 2 mm EDTA, 4.5 mm cCMP) at 25 °C. After
preincubation, the fully reduced RNase A (8.4 lm) and dif-
ferent concentrations of PDI (0–10 lm) were added to the
assay solution to initiate refolding. The formation of active
RNase A was measured spectrophotometrically by monitor-
ing cCMP hydrolysis at 296 nm. During the oxidative
refolding, the reduced RNase A was quickly converted to
inactive oxidized forms by the oxidase activity of PDI, and
these inactive oxidized forms were then slowly converted to
active native form by the isomerase activity of PDI [23].
The lag time before appearance of the active RNase A
indicates the oxidase activity, which corresponds to the
x-intercept of the RNase activity plot. The oxidase activity
matches the slope of the linear plot of reciprocal of lag times

4
CO
3
,5mm GSH, 0.5 mm GSSG,
pH 8.0) at 25 °C for 8 h, and the refolding mixture was
analyzed by RP-HPLC. Two major disulfide isomers,
nTx3.1 and sTx3.1, were collected and lyophilized.
PDI-catalyzed oxidative folding and isomerization
of conotoxins
The oxidative folding or isomerization of conotoxins was
performed in the refolding buffer (0.1 m Tris ⁄ Cl, pH 7.5,
plus appropriate concentrations of GSH, GSSG and cPDI
as indicated in the figure legends) at ambient temperature
(23–25 °C). The folding reaction was initiated by adding
the peptide stock solution to the final concentration of
20 lm. At different reaction times, the folding reaction was
quenched by adding formic acid to the final concentration
of 8%. The reaction mixture was immediately analyzed
using C
18
analytical RP-HPLC. The amounts of linear
form, native form and swap form were calculated from
their integrated elution peak areas, and the PDI’s oxidase
and isomerase activities were expressed as the initial rate of
decrease of linear tx3a and the initial rate of increase of
nTx3.1, respectively. To investigate the effect of macromo-
lecular crowding on the PDI-catalyzed oxidative folding or
disulfide isomerization of conotoxins, the folding reaction
was carried out as described above, except that 200 gÆL
)1

venoms. Eur J Biochem 264, 271–275.
8 van den Berg B, Ellis RJ & Dobson CM (1999) Effects
of macromolecular crowding on protein folding and
aggregation. EMBO J 18, 6927–6933.
9 Gething MJ & Sambrook J (1992) Protein folding in the
cell. Nature 355, 33–45.
10 Ellis J (1987) Proteins as molecular chaperones. Nature
328, 378–379.
11 Garrett JE, Buczek O, Watkins M, Olivera BM & Bulaj
G (2005) Biochemical and gene expression analyses of
conotoxins in Conus textile venom ducts. Biochem
Biophys Res Commun 328, 362–367.
12 Nielsen H, Engelbrecht J, Brunak S & von Heijne G
(1997) Identification of prokaryotic and eukaryotic
signal peptides and prediction of their cleavage sites.
Protein Eng 10, 1–6.
13 Song J, Quan H & Wang C (1997) Dependence of the
anti-chaperone activity of protein disulphide isomerase
on its chaperone activity. Biochem J 328 (3), 841–846.
14 Han YH, Wang Q, Jiang H, Liu L, Xiao C, Yuan DD,
Shao XX, Dai QY, Cheng JS & Chi CW (2006) Charac-
terization of novel M-superfamily conotoxins with new
disulfide linkage. FEBS J 273, 4972–4982.
15 Corpuz GP, Jacobsen RB, Jimenez EC, Watkins M,
Walker C, Colledge C, Garrett JE, McDougal O, Li W,
Gray WR et al. (2005) Definition of the M-conotoxin
Characterization of PDI from Conus marmoreus Z Q. Wang et al.
4786 FEBS Journal 274 (2007) 4778–4787 ª 2007 The Authors Journal compilation ª 2007 FEBS
superfamily: characterization of novel peptides from mol-
luscivorous Conus venoms. Biochemistry 44, 8176–8186.

24 Wilkinson B, Xiao R & Gilbert HF (2005) A structural
disulfide of yeast protein-disulfide isomerase destabilizes
the active site disulfide of the N-terminal thioredoxin
domain. J Biol Chem 280, 11483–11487.
Z Q. Wang et al. Characterization of PDI from Conus marmoreus
FEBS Journal 274 (2007) 4778–4787 ª 2007 The Authors Journal compilation ª 2007 FEBS 4787