Bacitracin is not a specific inhibitor of protein disulfide
isomerase
Anna-Riikka Karala and Lloyd W. Ruddock
Biocenter Oulu and Department of Biochemistry, University of Oulu, Finland
Introduction
Protein disulfide isomerase (PDI) is an endoplasmic
reticulum (ER)-resident protein catalyst that helps
newly translated polypeptide chains to fold and form
native disulfide bonds [1]. PDI can catalyze the oxida-
tion of two cysteines to form a disulfide bond, as well
as the reduction and isomerization of disulfide bonds
in peptides and proteins. PDI has four structural thior-
edoxin-like domains, a, b, b¢, and a¢, a linker region x
between the b¢ and a¢ domains, and a C-terminal acidic
extension. The a and a¢ domains contain the CGHC
active site motif, and are sufficient alone to perform
thiol–disulfide exchange reactions in simple substrates
[2]. The b¢ domain has the principal peptide and non-
native protein-binding site, and is required for isomeri-
zation reactions [2–4], whereas the b domain is of
unknown function.
PDI is one of a family of  20 PDI-like proteins
identified in the ER [1]. These proteins contain one or
more domains that are similar to the domains of PDI,
and many have been shown to catalyze thiol–disulfide
exchange reactions. However, their specific roles,
substrate specificities and mechanisms of cooperation
with other catalysts and chaperones in the cell are not
yet clear.
Besides PDI being abundant in the ER, several stud-
ies have shown non-ER locations for PDI family mem-

tein disulfide isomerase (PDI) in vivo. However, the specificity of action of
an inhibitor for a protein-folding catalyst cannot be determined in vivo.
Furthermore, in vitro evidence for the specificity of bacitracin for PDI is
scarce, and the mechanism of inhibition is unknown. Here, we present
in vitro data showing that 1 mm bacitracin has no significant effect on the
ability of PDI to introduce or isomerize disulfide bonds in a folding protein
or on its ability to act as a chaperone. Where bacitracin has an effect on
PDI activity, the effect is relatively minor and appears to be via competition
of substrate binding. Whereas 1 mm bacitracin has minimal effects on PDI,
it has significant effects on both noncatalyzed protein folding and on other
molecular chaperones. These results suggest that the use of bacitracin as a
specific inhibitor of PDI in cellular systems requires urgent re-evaluation.
Abbreviations
BPTI, bovine pancreatic trypsin inhibitor; CM, carboxymethyl; ER, endoplasmic reticulum; PDI, protein disulfide isomerase.
2454 FEBS Journal 277 (2010) 2454–2462 ª 2010 The Authors Journal compilation ª 2010 FEBS
bacitracin contains at least nine different peptides, of
which bacitracin A is the most abundant, and it is
mainly used as an antibiotic against infections caused
by Gram-positive bacteria [9]. The antibiotic effect is
based on the inhibition of bacterial cell wall synthesis
by a variety of mechanisms.
Bacitracin has been used as a specific PDI inhibitor in
a very wide range of studies. These include studying the
mechanisms of virus entry [10–12], the reductive activa-
tion of diphtheria and cholera toxins [7,13], gamete
fusion [14], platelet adhesion [15,16], melanoma cell
death [17], glioma cell invasion [18], the regulation of
transcriptional activity of nuclear factor kappaB [19],
the regulation of NAD(P)H oxidase [20], the shedding
of human thyrotropin receptor ectodomain [21], the

Results
Bacitracin does not inhibit the catalysis of
disulfide bond formation and isomerization by PDI
PDI is a catalyst of thiol–disulfide exchange reactions,
including oxidation, reduction and isomerization [1].
The simplest in vitro assays for catalysis of thiol–disul-
fide exchange are based on small peptides. To examine
whether bacitracin is able to inhibit the ability of PDI
to introduce disulfide bonds into a substrate in the
absence of the concomitant formation of secondary
structure, a fluorescent decapeptide PDI substrate [32]
was used. In a glutathione buffer at pH 7.0, a time-
dependent decrease in fluorescence was observed that
could be fitted to a first-order process (Fig. 1A), con-
sistent with the formation of a disulfide bond in the
substrate [32]. The rate constant for oxidation of
3.4 lm peptide in the presence of 0.7 lm PDI was
0.85 ± 0.05 min
)1
(n = 6). Bacitracin contains a mix-
ture of peptides, with the most abundant, bacitracin A,
containing an aromatic phenylalanine moiety. Hence,
at 1 mm, there are two opposing effects on the fluores-
cence of the system in the presence of bacitracin. First,
there is a net increase in fluorescence due to the baci-
tracin. However, with excitation at 280 nm and emis-
sion at 350 nm, bacitracin is much less fluorescent on
a per molar basis than the PDI peptide substrate,
which contains a tryptophan. Second, there is a net
decrease in the fluorescence due to the inner filter

BPTI has to undergo isomerization reactions. Noncat-
alyzed glutathione-based refolding of BPTI is slow,
A R. Karala and L. W. Ruddock Bacitracin is not specific for PDI
FEBS Journal 277 (2010) 2454–2462 ª 2010 The Authors Journal compilation ª 2010 FEBS 2455
with only around one-quarter of the BPTI being able
to achieve the native 3S state within 2 h [33]. However,
all the steps of BPTI refolding are catalyzed by PDI,
and within 40 min BPTI was refolded to 94% ± 3%
native 3S form (Fig. 1B). When 1 mm bacitracin is
added to the PDI-catalyzed BPTI refolding system, the
MS analysis becomes significantly less accurate, so an
additional step to remove excess bacitracin after
quenching of the reaction but prior to analysis is
required. With this, the refolding of BPTI followed
very similar kinetics in the presence or absence of baci-
tracin, and after 40 min of refolding with 1 mm baci-
tracin present, 90% ± 5% of BPTI was in the native
3S state (Fig. 1C). These results imply that bacitracin
does not inhibit the ability of PDI to introduce or
isomerize disulfide bonds in a folding protein.
Bacitracin inhibits rhodanese aggregation and
the chaperone activity of BiP
In addition to disulfide bond formation, PDI has been
shown to have chaperone-like activity [34]. As rhoda-
nese contains no disulfide bonds, and is prone to aggre-
gation during refolding, it can be used as a model with
which to study chaperone activity in folding. Analysis
of the nonassisted refolding of rhodanese showed the
expected aggregation of the folding intermediates. The
addition of PDI or the noncatalytic PDI family member

Table 1. Analysis of the aggregation rate during rhodanese refold-
ing at pH 7.2. The rate of aggregation relative to the negative con-
trol in the absence of bacitracin is presented as mean ± standard
deviation (number of samples). Statistical significance between
each pair of samples with and without bacitracin present was
determined using Student’s t-test (two-tailed, two-sample unequal
variance). Note that the effects of bacitracin on PDI and ERp27
inhibition of aggregation are equivalent to those on the noncata-
lyzed reaction.
Sample
No
bacitracin
1m
M
bacitracin
t-test for an effect
of bacitracin
Negative control 100 ± 19 (8) 69 ± 9 (8) P < 0.05
+4.5 l
M PDI 82 ± 12 (6) 56 ± 15 (5) P < 0.05
+4.5 l
M BiP 6 ± 11 (6) 18 ± 6 (5) P < 0.05
+4.5 l
M ERp27 71 ± 11 (4) 54 ± 2 (3) P < 0.05
Bacitracin is not specific for PDI A R. Karala and L. W. Ruddock
2456 FEBS Journal 277 (2010) 2454–2462 ª 2010 The Authors Journal compilation ª 2010 FEBS
imply that bacitracin interacts with rhodanese, decreas-
ing the aggregation of its folding intermediates, and that
it has no observable effect on the chaperone activity of
PDI family members. In parallel studies, the aggrega-

ance), this effect was found to be significant (P < 0.05),
even with the addition of 0.1 mm bacitracin. Unlike in
the rhodanese assay, in this assay 1 mm bacitracin had
no significant effect on the lag-phase of the reaction or
on the subsequent gradient for aggregation (Fig. 2A),
implying that the effects of bacitracin addition observed
on PDI were due directly to inhibition of PDI-catalyzed
insulin reduction.
Fig. 2. Effects of bacitracin and other compounds on the relative rate of reduction of the B-chain of bovine insulin. Insulin was reduced at
1mgÆmL
)1
in the presence of 10 mM dithiothreitol and 1 mM EDTA at pH 7. When present, PDI, PDI a domain (PDIa), DsbA and DsbC were
used at 1 l
M. Bacitracin (Bac) was used at 1 mM, if not indicated otherwise in the figure. Triton X-100 (TX) was used at 0.05% (v ⁄ v) and 2-
propylphenol (2PP) at 1 m
M. The reduction of the B-chain of insulin causes precipitation that can be followed as an absorbance increase at
540 nm. (A) Representative changes in absorbance as a function of time. From left to right, the traces are: PDI, PDI + bacitracin, noncata-
lyzed reaction, noncatalyzed reaction + bacitracin. (B–D) Lag times for precipitation of insulin under different conditions. The relative activity
is presented as mean ± standard deviation; n = 2–7, with the value given in parentheses. (B) With PDI present. (C) Noncatalyzed reactions.
(D) With PDI a domain, DsbA or DsbC present.
A R. Karala and L. W. Ruddock Bacitracin is not specific for PDI
FEBS Journal 277 (2010) 2454–2462 ª 2010 The Authors Journal compilation ª 2010 FEBS 2457
To study the potential mechanism of action of baci-
tracin, the insulin reduction assay was also performed
in the presence of other two other thiol–disulfide
exchange enzymes, Escherichia coli DsbA and DsbC,
as well as the isolated catalytic a domain of PDI. Both
the PDI a domain and DsbA have a catalytic site, with
an associated substrate-binding site, but lack an inde-
pendent substrate-binding site, which is present in full-

Triton X-100 or 1 mm 2-propylphenol. In the noncata-
lyzed reaction, the addition of Triton X-100 had a
minimal effect on the system, whereas the addition of
2-propylphenol increased the rate of insulin aggre-
gation (Fig. 2C; P < 0.05). In contrast, the addition
of either Triton X-100 or 2-propylphenol decreased the
rate of PDI-catalyzed insulin aggregation (Fig. 2B;
P < 0.05), showing that inhibition of the primary
substrate-binding site in the b¢ domain decreases the
insulin-reducing activity of PDI.
Discussion
Inhibitors are widely used to study the physiological
functions of proteins in vivo. Bacitracin is a metallo-
peptide antibiotic that has been widely used as a spe-
cific PDI inhibitor [7,10–29]. However, neither the
specificity of bacitracin for PDI nor the detailed mech-
anisms of inhibition of PDI have been investigated.
Furthermore, since the original reporting of the inhibi-
tion of PDI by bacitracin [8], concerns have been
raised about protease contamination of some commer-
cially available bacitracin preparations [31].
Here, we have tested the effect of bacitracin in a
variety of in vitro assays for PDI activity. On the basis
of the BPTI refolding assay and peptide oxidation
assay 1 mm bacitracin does not have a significant
effect on the oxidative or isomerization activity of
PDI. In addition, the chaperone activity of PDI in the
rhodanese-refolding assay was not significantly chan-
ged in the presence of bacitracin. In the insulin reduc-
tion assay, bacitracin was able to decrease the activity

strongly imply that bacitracin is not a selective inhibi-
tor of PDI. Instead, bacitracin can also interact with
folding polypeptide chains and other molecular chaper-
ones and folding catalysts. Bacitracin probably inter-
acts with these, and with PDI, via its hydrophobic side
chains, which could interact with exposed hydrophobic
Bacitracin is not specific for PDI A R. Karala and L. W. Ruddock
2458 FEBS Journal 277 (2010) 2454–2462 ª 2010 The Authors Journal compilation ª 2010 FEBS
side chains of a folding protein or with the hydropho-
bic binding site of molecular chaperones such as BiP,
PDI and DsbC.
To further study the mechanism of action of bacitra-
cin, the reduction of insulin was analyzed in the pres-
ence of the PDI a domain and DsbA, which are
capable of catalyzing the oxidation and reduction reac-
tions, but lack the independent substrate-binding site
that is present in PDI and DsbC. Bacitracin had no
effect on the activity of the PDI a domain or DsbA,
implying that the active site is probably not the target
of inhibition by bacitracin. This result was confirmed
by assays in which EDTA was omitted from the assay.
Molecules that are known to interfere with substrate
binding by PDI were also used in the insulin reduction
assay. Triton X-100 at 0.05% (v ⁄ v) (equivalent to
0.8 mm), which is known to affect substrate binding by
the noncatalytic b¢ domain [3], reduced PDI activity
in the assay to a slightly greater extent than 1 mm
bacitracin.
Although pathways for cellular metabolism are
unknown, and it is possible that in vivo processing of

cant protease activity in the material that we used, reduced,
denatured BPTI was incubated with 1 mm bacitracin for
1 h at room temperature in 0.1 m sodium phosphate buffer
(pH 7.0) containing 1 mm EDTA. Analysis by SDS ⁄ PAGE
showed no evidence of degradation of the denatured BPTI
over this time period. In addition, ESI-MS analysis of BPTI
refolding (see below) showed no evidence of BPTI degrada-
tion products.
Generation of expression vectors
N-terminally histidine-tagged mature PDI, PDI a domain
and mature BPTI with an additional initiating methionine
expression vector were generated for previous studies
[39,40]. Mature human BiP (Glu19–Leu653) was generated
by PCR from a human liver cDNA library (Clontech,
Mountain View, CA, USA) in two parts. BiP Glu19–
Arg323 was constructed as an NdeI–SacI fragment, and
BiP Ala324–Leu653 as a SacI–XhoI fragment. Mature
human ERp27 (Glu26–Leu273) was generated by PCR
from IMAGE clone 5207225 as an NdeI–BamHI fragment.
Mature E. coli DsbA (Ala20–Leu208) and mature E. coli
DsbC (Asp21–Lys236) were constructed as NdeI–BamHI
fragments by PCR from an E. coli colony. All inserts were
cloned into a modified pET23b (Novagen, Madison, WI,
USA), which codes for an N-terminal hexahistidine tag
before the first amino acid of the protein sequence.
Protein expression and purification
PDI (EC 5.3.4.1; UniProt ID P07237), PDI a domain, BiP
(EC 3.6.4.10; UniProt ID P11021), ERp27 (Uni-
Prot ID Q96DN0), DsbA (UniProt ID P0AEG4) and
DsbC (UniProt ID P0AEG6) were expressed in E. coli

)1
Æcm
)1
, M
r
= 56386; BPTI, 5680 m
)1
Æcm
)1
,
M
r
= 6648; BiP, 29 660 m
)1
Æcm
)1
, M
r
= 71 356; ERp27,
18 450 m
)1
Æcm
)1
, M
r
= 28 837; DsbA, 22 560 m
)1
Æcm
)1
,

thermal equilibration of the solution; 6.3 lL of substrate
peptide (539 lm stock solution in 30% acetonitrile ⁄ 0.1%
trifluoroacetic acid) was added and mixed, and the change
in fluorescence intensity (excitation at 280 nm, emission at
350 nm, slits of 5 ⁄ 5 nm) was monitored over an appropri-
ate time period (15 min–1 h), with 600–1800 data points
being collected.
Refolding of reduced and denatured BPTI
A modified version of the methods that we have previously
reported [40] was used to analyze BPTI refolding. In partic-
ular, this method has additional steps to remove excess bac-
itracin prior to the MS analysis. The refolding of BPTI was
initiated by the addition of denatured reduced protein to
the refolding buffer (2 mm reduced glutathione, 0.5 mm
oxidized glutathione, 0.1 m sodium phosphate, 1 mm
EDTA, pH 7.0). In the catalyzed refolding, PDI was pre-
equilibrated in the refolding buffer for 5 min before BPTI
was added. BPTI was used at 50 lm and, when present in
the refolding reaction, PDI was used at 7 lm and bacitracin
at 1 mm. The folding reaction was stopped by the addition
of 1.1 m iodoacetamide (Sigma-Aldrich), and BPTI and its
folding intermediates were purified with a PepClean C-18
spin column (Pierce, Rockford, IL, USA) before ESI-MS
analysis (Micromass, Manchester, UK). Bacitracin-contain-
ing BPTI samples were additionally purified by cation
exchange chromatography and with a PepClean C-18 spin
column. The cation exchange resin carboxymethyl (CM)
cellulose 32 (Whatman, Maidstone, UK) was first pretreat-
ed by suspending 3 g of resin in 30 mL of 0.5 m sodium
hydroxide and stirring for 30 min. The cellulose was then

thiosulfate). The aggregation of rhodanese during refolding
was followed spectrophotometrically at 320 nm over 5 min.
PDI, BiP and ERp27 were used at 4.5 lm, and bacitracin
at 1 mm, when present. ATP was used at 2 mm when BiP
was present. Proteins were equilibrated in the refolding
buffer for 3 min before the addition of rhodanese.
Insulin precipitation assay
A modified version of the insulin turbidity assay reported
by Holmgren [41] was used. The precipitation reaction of
the B-chain of bovine insulin (Sigma-Aldrich) was initiated
by adding the insulin to 0.1 m sodium phosphate buffer
Bacitracin is not specific for PDI A R. Karala and L. W. Ruddock
2460 FEBS Journal 277 (2010) 2454–2462 ª 2010 The Authors Journal compilation ª 2010 FEBS
(pH 7.0) containing 1 mm EDTA and 10 mm dithiothreitol.
PDI, DsbC, PDI a domain and DsbA were used at 1 lm,
and bacitracin and 2-propylphenol at 1 mm and Triton X-
100 at 0.05% (v ⁄ v), if included in the reaction. Insulin was
used at 1 mgÆmL
)1
. Before the insulin addition, protein cat-
alysts and bacitracin were equilibrated in the reaction buf-
fer for 5 min. The precipitation of the B-chain of bovine
insulin was monitored spectrophotometrically at 540 nm.
Acknowledgements
This work was supported by the University of Oulu
and Biocenter Oulu.
References
1 Hatahet F & Ruddock LW (2009) Protein disulfide
isomerase: a critical evaluation of its function in disul-
fide bond formation. Antioxid Redox Signal 11, 2807–

9 Ming LJ & Epperson JD (2002) Metal binding and
structure–activity relationship of the metalloantibiotic
peptide bacitracin. J Inorg Biochem 91, 46–58.
10 Markovic I, Stantchev TS, Fields KH, Tiffany LJ,
Tomic M, Weiss CD, Broder CC, Strebel K & Clouse
KA (2004) Thiol ⁄ disulfide exchange is a prerequisite for
CXCR4-tropic HIV-1 envelope-mediated T-cell fusion
during viral entry. Blood 103, 1586–1594.
11 Ryser HJ, Levy EM, Mandel R & DiSciullo GJ (1994)
Inhibition of human immunodeficiency virus infection
by agents that interfere with thiol–disulfide interchange
upon virus-receptor interaction. Proc Natl Acad Sci
USA 91, 4559–4563.
12 Jain S, McGinnes LW & Morrison TG (2007)
Thiol ⁄ disulfide exchange is required for membrane
fusion directed by the newcastle disease virus fusion
protein. J Virol 81, 2328–2339.
13 Orlandi PA (1997) Protein-disulfide isomerase-mediated
reduction of the A subunit of cholera toxin in a human
intestinal cell line. J Biol Chem 272, 4591–4599.
14 Ellerman DA, Myles DG & Primakoff P (2006) A role
for sperm surface protein disulfide isomerase activity in
gamete fusion: evidence for the participation of ERp57.
Dev Cell 10, 831–837.
15 Lahav J, Wijnen EM, Hess O, Hamaia SW, Griffiths D,
Makris M, Knight CG, Essex DW & Farndale RW
(2003) Enzymatically catalyzed disulfide exchange is
required for platelet adhesion to collagen via integrin
alpha2beta1. Blood 102, 2085–2092.
16 Essex DW & Li M (1999) Protein disulphide isomerase

A R. Karala and L. W. Ruddock Bacitracin is not specific for PDI
FEBS Journal 277 (2010) 2454–2462 ª 2010 The Authors Journal compilation ª 2010 FEBS 2461
superoxide dismutase 1. J Biol Chem 281, 30152–30165.
23 Wajih N, Hutson SM & Wallin R (2007) Disulfide-
dependent protein folding is linked to operation of the
vitamin K cycle in the endoplasmic reticulum. A protein
disulfide isomerase–VKORC1 redox enzyme complex
appears to be responsible for vitamin K1 2,3-epoxide
reduction. J Biol Chem 282, 2626–2635.
24 Descamps E, Petrault-Laprais M, Maurois P, Pages N,
Bac P, Bordet R & Vamecq J (2009) Experimental
stroke protection induced by 4-hydroxybenzyl alcohol is
cancelled by bacitracin. Neurosci Res 64, 137–142.
25 Cho J, Furie BC, Coughlin SR & Furie B (2008) A crit-
ical role for extracellular protein disulfide isomerase
during thrombus formation in mice. J Clin Invest 118 ,
1123–1131.
26 Lahav J, Karniel E, Bagoly Z, Sheptovitsky V, Dardik
R & Inbal A (2009) Coagulation factor XIII serves as
protein disulfide isomerase. Thromb Haemost 101,
840–844.
27 Ahamed J, Versteeg HH, Kerver M, Chen VM, Mueller
BM, Hogg PJ & Ruf W (2006) Disulfide isomerization
switches tissue factor from coagulation to cell signaling.
Proc Natl Acad Sci USA 103, 13932–13937.
28 Reinhardt C, von Bruhl ML, Manukyan D, Grahl L,
Lorenz M, Altmann B, Dlugai S, Hess S, Konrad I,
Orschiedt L et al. (2008) Protein disulfide isomerase acts
as an injury response signal that enhances fibrin genera-
tion via tissue factor activation. J Clin Invest 118,

binding b¢ domain. J Mol Biol 383, 1144–1155.
36 McCarthy AA, Haebel PW, Torronen A, Rybin V,
Baker EN & Metcalf P (2000) Crystal structure of the
protein disulfide bond isomerase, DsbC, from Escheri-
chia coli. Nat Struct Biol 7, 196–199.
37 Solovyov A & Gilbert HF (2004) Zinc-dependent
dimerization of the folding catalyst, protein disulfide
isomerase. Protein Sci 13, 1902–1907.
38 Klappa P, Freedman RB, Langenbuch M, Lan MS,
Robinson GK & Ruddock LW (2001) The pancreas-
specific protein disulphide-isomerase PDIp interacts
with a hydroxyaryl group in ligands. Biochem J 354,
553–559.
39 Alanen HI, Salo KE, Pekkala M, Siekkinen HM,
Pirneskoski A & Ruddock LW (2003) Defining the
domain boundaries of the human protein disulfide
isomerases. Antioxid Redox Signal 5, 367–374.
40 Karala AR, Psarrakos P, Ruddock LW & Klappa P
(2007) Protein disulfide isomerases from C. elegans are
equally efficient at thiol–disulfide exchange in simple
peptide-based systems but show differences in reactivity
towards protein substrates. Antioxid Redox Signal 9,
1815–1823.
41 Holmgren A (1979) Thioredoxin catalyzes the reduction
of insulin disulfides by dithiothreitol and dihydrolipoa-
mide. J Biol Chem 254 , 9627–9632.
Bacitracin is not specific for PDI A R. Karala and L. W. Ruddock
2462 FEBS Journal 277 (2010) 2454–2462 ª 2010 The Authors Journal compilation ª 2010 FEBS