1032 BioTechniques Vol. 35, No. 5 (2003)
Genetically engineered antibod-
ies and their fragments produced in
Escherichia coli have a great potential
as specific targeting reagents, both in
scientific research and clinical trials
(1,2). There has been growing interest
especially in the single-chain antibody
(scAb) format of engineered antibod-
ies, particularly for in vivo diagnostic
and therapeutic applications because
of its small size. However, one major
factor that can limit the applicabil-
ity of scAbs is the ability to produce
them in large amounts as soluble and
functional proteins. Actually, scAbs are
often prone to form insoluble aggre-
gates known as inclusion bodies when
expressed in E. coli, especially at high
level, resulting in low active yields. Al-
though the expression of proteins as in-
clusion bodies leads to larger amounts
of target proteins, the production of
biologically functional products is sty-
mied by the laborious procedures (3,4).
Many strategies have been proposed to
obtain improved solubility in the past
years (5–14). However, the application
of signal sequences or different culture
conditions to a variety of engineered
antibodies within our laboratory has
met with little success (15). Therefore,
as the optimization of engineered anti-
body expression still tends to proceed
on a case-by-case basis, a generic sys-
tem is desirable to improve engineered
antibody expression as soluble and ac-
tive products in E. coli.
FkpA, a periplasmic peptidyl-prolyl
cis, trans-isomerase (16), is one of the
heat shock proteins that are inducible in
response to accumulation of misfolded
proteins in E. coli (16–18). Missiakas
et al. found that FkpA decreased the
sigma E-dependent response constitu-
tively induced by misfolded proteins
and proposed that FkpA plays an ac-
tive role either as a folding catalyst or
as a chaperone in E. coli (18). Further
investigation showed that FkpA has
two activities, the cis, trans-isomerase
activity and the chaperone-like activity
(19–21). A poorly folded single-chain
variable fragment (scFv) displayed on
filamentous phage and a defective-
folding variant of the maltose-bind-
ing protein, MalE31, were produced
in biologically active form in E. coli
functionally assisted by co-expression
of FkpA (21,22).
Because all the engineered antibod-
ies constructed in our laboratory were
insoluble when expressed in E. coli, we
sought to identify whether overexpres-
sion of FkpA would be of benefit to the
expression of engineered antibodies
of interest as soluble and functional
forms. We have constructed various
vectors for the fusion expression or
co-expression of FkpA with engineered
antibodies anti-human bladder carci-
noma scAb (PG) and its fusion protein
cellulose-binding domain from Cellu-
lomonas fimi (CBD)-PG, anti-human
CD3×anti-human ovarian carcinoma-
bispecific scAb (BH1), and anti-human
ovarian carcinoma-trispecific scAb
(TRI). Our results demonstrate con-
clusively that fusion expression or
co-expression of FkpA is sufficient for
soluble and functional expression of
engineered antibodies in E. coli.
Production of soluble and functional engineered
antibodies in Escherichia coli improved by FkpA
Zhong Zhang, Li-ping Song, Min Fang, Fei Wang, Dan He, Rui Zhao, Jing Liu,
Zhi-yong Zhou, Chang-cheng Yin, Qing Lin, and Hua-liang Huang
BioTechniques 35:1032-1042 (November 2003)
Overproduction of genetically engineered antibodies, such as single-chain antibodies (scAbs) in Escherichia coli often results in in-
soluble and inactive products known as inclusion bodies. We now report that fusion or co-expression of FkpA, the E. coli periplasmic
peptidyl-prolyl-isomerase with chaperone activity, substantially improves soluble and functional expression of scAbs. Anti-human blad-
der carcinoma scAb (PG) and anti-human CD3×anti-human ovarian carcinoma-bispecific scAb (BH1) were fused with FkpA on the
pTMF-based plasmid and expressed in E. coli. More than half of the amount of each expressed fusion protein FkpA-PG or FkpA-BH1
was soluble. In addition, the fusion protein cellulose-binding domain from Cellulomonas fimi (CBD)-PG and anti-human CD3×anti-
human CD28×anti-human ovarian carcinoma-trispecific scAb (TRI) fused to the pelB (a signal peptide from pectate lysase B of a Ba-
cillus sp.) signal sequence were co-expressed with FkpA under the control of the T7 promoter. A substantial portion of the co-expressed
CBD-PG or TRI was soluble. Furthermore, PG, BH1, and TRI were biologically active as judged by ELISA and in vitro cytotoxicity
assay. These results suggest that overexpression of FkpA should be useful in expressing heterologous proteins in E. coli.
Academia Sinica, Beijing, China
Vol. 35, No. 5 (2003) BioTechniques 1033
MATERIALS AND METHODS
Mouse anti-c-myc monoclonal an-
tibody (cell line 9E10) was purchased
from Santa Cruz Biotechnology (Santa
Cruz, CA, USA). Goat anti-mouse im-
munoglobulin G horseradish peroxi-
dase (IgG-HRP) and protein markers
were purchased from Sino-American
Biotechnology Company (Beijing,
China). Anti-CD3 monoclonal antibody
was purchased from Wuhan Institute of
Biological Products (Wuhan, China).
Anti-CD28 monoclonal antibody and
rhCD28/FcChimera, recombinant hu-
man CD28 extracellular domain fused
to human IgG1 Fc, were products of
R&D System (Minneapolis, MN,
USA). Anti-ovarian carcinoma scFv
was prepared by renaturing and puri-
fying the inclusion body expressed in
E. coli. Restriction endonucleases, T4
DNA ligase, LA Taq DNA polymerase,
and other modification enzymes were
purchased from TaKaRa Biotechnol-
ogy (Dalian), Ltd. (Dalian, China).
pfu DNA polymerase was a product of
Bioasia Company (Shanghai, China).
Prestained protein markers were pur-
chased from New England Biolabs
(Beverly, MA, USA).
Human bladder carcinoma cell line
BIU87, human ovarian carcinoma cell
line SKOV3, and human T lymphoma
cell line Jurkat, on the surface of which
CD3 was overexpressed, were from our
MTT was purchased from Sigma (St.
Louis, MO, USA). All other chemicals
were of analytical grade.
Bacterial Strains and Plasmids
E. coli DH5α and Top10 were used
as the host strains for cloning and main-
tenance of plasmids throughout the
experiments. E. coli BL21 Star (DE3)
was purchased from Invitrogen (Carls-
bad, CA, USA) and used as the host for
expression of all the T7 promoter-based
vectors constructed in this study.
The plasmid pTMF was constructed
in our laboratory (15) and used as the
basic vector for construction of all the
expression plasmids in this study. The
plasmid pTCM, constructed previ-
ously by inserting the complete coding
sequence of c-myc (33 bp) into the mul-
ticloning site of pTMF (15), was used as
a control in ELISA of expressed target
proteins against specific antigens. Plas-
mids pTP1 (15) and pTBNP1 (unpub-
lished) were constructed previously and
used for expression of PG and its fusion
protein CBD-PG. Plasmids pALMBH1
and pTTRI were constructed in our lab-
oratory (unpublished data) and used for
expression of BH1 and TRI, respective-
ly. All plasmids constructed and used in
this study were diagrammatically shown
in Figure 1 and summarized in Table 1.
For the facile detection of the ex-
pressed scAbs against specific antigens,
the C terminus of each of the scAbs
mentioned in this study contains a c-myc
Table 1. Plasmids Used in this Study
pTMF 5.3 kb, Kmr, T7 promoter Our laboratory
pTCM5.3 kb, Kmr, c-myc tag, T7 promoterOur laboratory
p TP16.0 kb, pg-c-myc Our laboratory
pTFP16.8 kb, pg-c-myc fused to fkpAThis study
pTMFBH16.9 kb, bh1-c-mycOur laboratory
pTFBH17.7 kb, bh1-c-myc fused to fkpAThis study
pTBNP16.5 kb, cbd fused to pg-c-myc Our laboratory
7.3 kb, cbd fused to pg-c-myc and
co-expressed with fkpA
pTTRI7.4 kb, triOur laboratory
pTTRIF8.2 kb, tri, co-expressed with fkpAThis study
Figure 1. pTMF-based expression plasmids. Organization of fkpA and pg, bh1, cbd-pg, and tri
is schematically shown here. pg, bh1, cbd-pg, and tri genes were joined with fkpA, respectively, on
the pTMF under the control of the T7 promoter and the lac operator. T7, the T7 promoter; O, the lac
operator; R, ribosome binding site (RBS); T, the T7 transcription terminator; taa, stop codon; c-myc, a
short 11-amino acid tag.
tag, which can be detected by mouse
anti-c-myc monoclonal antibody.
Isolation of the fkpA Open Reading
fkpA genes [complete coding se-
quence with or without a ribosome
binding site (RBS)] were isolated by
PCR with pfu polymerase, using E.
coli strain K12 chromosomal DNA as
templates. The complete fkpA was am-
plified as follows: the forward primer
ACG-3′. The reverse primer, 5′-GC-
GCGG-3′, was designed to contain a
SacI site (underlined). The PCR prod-
uct was digested with SacI, purified,
and cloned into the NcoI (blunted by
the Klenow fragment of E. coli DNA
polymerase I)-SacI site of pTMF, yield-
ing plasmid pTFA (Figure 1). Similarly,
rbs-fkpA was amplified using the fol-
lowing primers. The forward primer
AGTAAC-3′ and was designed to
contain a NdeI site (underlined) and an
RBS site (bolded). The reverse primer
CAGAATCTGC-3′ containing a BamHI
site (underlined). The PCR product was
digested with NdeI and BamHI, puri-
fied, and cloned into the same sites of
pTMF, yielding pRFA (Figure 1).
The sequences of all the amplified
genes were verified by DNA sequencing.
Construction of the FkpA Fusion
and Co-Expression Vectors
For construction of the FkpA-PG fu-
sion vector, the BglII-SacI fragment of
pTFA was inserted into the BglII-SacI
site of pTP1 to yield pTFP1. For con-
struction of FkpA-BH1 fusion vector,
the bispecific scAb BH1 was cloned
from plasmid pALMBH1 by PCR using
specific primers. The forward primer
CAGGAGTCTGG-3′ designed to con-
tain an XhoI site (underlined), and the
reverse primer was 5′-TTACTGCAGT-
pfu polymerase was used in the PCR,
and the products were recovered, di-
gested with XhoI, and ligated into the
XhoI-NdeI (blunted by the Klenow
fragment of E. coli DNA polymerase I)
site of pTFA to yield pTFBH1.
For construction of FkpA co-expres-
sion vector to co-express CBD-PG,
the BglII-XhoI fragment of pTBNP1
was inserted into the BglII-XhoI site
of pTRFA, and the resulting plasmid
was designated pTBNP1F and used to
co-express FkpA with fusion protein
CBD-PG. For construction of FkpA
co-expression vector to co-express tri-
specific scAb TRI, two specific primers
were used to isolate the trispecific scAb
gene by PCR from plasmid pTRI, which
was constructed previously (Song et al.,
Vol. 35, No. 5 (2003) BioTechniques 1035
unpublished). The forward primer was
and the reverse primer was 5′-TTAC-
ensure the efficiency of PCR in cloning
the long fragment TRI, LA Tag DNA
polymerase was used. The PCR product
was recovered and ligated into the NcoI
site (blunted by the Klenow fragment of
E. coli DNA polymerase I) of pTRFA.
The resulting plasmid was designated
pTTRIF. Each subset of the genes
obtained above was placed under the
control of the T7 promoter and the lac
operator on pTMF-based vectors. The
sequences of all the amplified genes
were verified by DNA sequencing. All
the pTMF-based vectors constructed in
this study are diagramed in Figure 1.
Culture Conditions, Protein
Expression, and Analysis
All the expression plasmids were
used to transform E. coli BL21 Star
(DE3). BL21 Star (DE3) transformed
cells were grown in LB medium (23)
supplemented with antibiotics. Cul-
tures were induced at an A600 = 0.8
using 0.4 mM isopropyl-β-D-thiogalac-
topyranoside (IPTG). Typically, after
several hours of induction, the induced
cells from 1 mL of culture broth were
harvested by centrifugation, resuspend-
ed in phosphate-buffered saline (PBS),
and then treated by sonication. After
centrifugation at 8000× g for 10 min
at 4°C, the supernatant was taken as
soluble fraction. The pellet was taken
as insoluble fraction and resuspended
in PBS. If needed, whole-cell proteins
were prepared separately by harvesting
1 mL of culture broth directly with cen-
trifugation. Thirty microliters of each
of the above fractions and the whole-
cell proteins were analyzed by sodium
dodecyl sulfate polyacrylamide gel
electrophoresis (SDS-PAGE), followed
by visualization with Coomassie® Bril-
liant Blue staining, and scanning quan-
titative analysis by spot density using
the marker proteins as control.
ELISA for scFvs Activity and
Western Blot Analysis
ELISAs were carried out in order to
evaluate the antigen-binding activities
of the soluble engineered antibodies
PG, BH1, and TRI expressed in E. coli
under the assistance of FkpA. Mem-
brane antigens of BIU87 cells, Jurkat
cells and SKOV3 cells, and CD28 pure
antigen (rhCD28/FcChimera) were di-
luted in 0.1 M NaCO3-NaHCO3 buffer,
pH 9.6, and the final concentration of
each antigen was adjusted to 10 µg/
mL. ELISA plates (NUNC, Roskilde,
Denmark) were coated overnight at
4°C with each of the above membrane
antigens, respectively. After washing
three times with PBS with Tween® 20
(PBST) and three times with PBS, the
plates were blocked for 2 h at 37°C
with 200 µL of 1% bovine serum al-
bumin (BSA)-PBS and then washed
with PBST and PBS, respectively. A
total of 100 µL supernatant containing
expressed target proteins (PG, BH1,
or TRI) was applied to the blocked
ELISA wells and incubated for 2 h at
37°C. PBS or supernatants from E. coli
cells harboring empty vector pTMF or
pTCM as negative controls, were also
Figure 2. FkpA markedly improves the expression of scAb PG and its fusion protein CBD-PG
as the soluble and biologically active forms in Escherichia coli. (A and B) Sodium dodecyl sulfate
polyacrylamide gel electrophoresis (SDS-PAGE) analysis of PG and CBD-PG (A) without or (B) with
fusion or co-expression of FkpA in E. coli. (A) Lane 1, the whole-cell proteins (W) of E. coli cells
harboring pTMF (empty vector); lanes 2 and 3, supernatant (S) and pellet (P), respectively, of E. coli
cells harboring pTP1 induced at 37°C; lanes 4 and 5, supernatant and pellet, respectively, of E. coli
cells harboring pTP1 induced at 30°C; lanes 6 and 7, supernatant and pellet, respectively, of E. coli
cells harboring pTBNP1 induced at 37°C; lanes 8 and 9, supernatant and pellet, respectively, of E. coli
cells harboring pTBNP1 induced at 30°C; lane 10, molecular mass markers (M; kDa). (B) Lanes 1 and
2, supernatant and pellet, respectively, of E. coli cells harboring pTFP1; lanes 3 and 4, supernatant and
pellet, respectively, of E. coli cells harboring pTBNP1F; lane 5, molecular mass markers; lane 6, the
whole-cell proteins of E. coli cells harboring pTMF. (C) ELISA results of FkpA-PG and CBD-PG co-
expressed with FkpA against the membrane antigens of BIU87. Phosphate-buffered saline (PBS) and
nonrelated scAb-c-myc were used as controls. The sonication supernatants of E. coli cells carrying
respective plasmid pTCM, pTFP1, and pTBNP1F were tested. CBD, cellulose-binding domain from
Cellulomonas fimi; PG, anti-human bladder carcinoma scAb; scAb, single-chain antibody.
1036 BioTechniques Vol. 35, No. 5 (2003)
added and incubated for 2 h at 37°C.
After the plates were washed, 100 µL
mouse anti-c-myc monoclonal anti-
body were added and incubated for 1 h
at 37°C. For detection, 100 µL of goat
anti-mouse IgG-HRP antibody were
used and incubated for another 1 h at
37°C. The development was carried out
with a total of 100 µL o-phenylenedi-
amine (OPD) solution (6 mg/mL OPD,
1% H2O2) for 15–20 min at room
temperature, and signals were read
at 490 nm with Universal Microplate
Reader Elx800 (Bio-Tek Instruments,
Winooski, VT, USA) after stopping the
reaction with 50 µL 2 M H2SO4.
Immunodetection by Western blot
analysis was performed as described
(23). Briefly, proteins from sonica-
tion supernatants were subjected to
gel). Prestained protein markers were
used as the standard. Electrotransfer
of proteins to nitrocellulose filters
(0.45 µm; Pall-Gelman Sciences, Ann
Arbor, MI, USA) was performed in the
transfer buffer (39 mM glycine, 40 mM
Tris-base, 0.03% SDS, 20% methanol)
by using Hoefer™ TE 22 Mighty Small
Transphor Tank Transfer Unit (Am-
ersham Biosciences, Piscataway, NJ,
USA) at constant current 400 mA for 1
h, and then the filter was blocked with
2% low-fat dry milk in TBS (TBS: 100
mol/L Tris-HCL, pH 7.5, 0.9% NaCl)
at room temperature for about 1 h. Then
the filter was washed with TBST (TBS
containing 0.1% Tween 20). Mouse
anti-c-myc monoclonal antibody was
added, and the filter was incubated for
1 h at room temperature. Next, goat
anti-mouse IgG-HRP antibody was
added, and the filter was incubated for
another hour at room temperature. The
peroxidase activity was developed with
a solution of diaminobenzidine (DAB;
0.6 mg/mL) in 20 mM Tris-base, 0.5 M
NaCl-buffered saline, and 0.1% H2O2.
In Vitro Cytotoxicity Assay for
Trispecific scAb Co-Expressed
Human peripheral blood mono-
nuclear cells (PBMCs) from healthy
donors were isolated using Ficoll®
density gradient centrifugation, and
monocytes/macrophages were depleted
using glass adherence incubated at
37°C, 5% CO2 for 2 h. A 96-well flat-
bottom plate was coated with target
tumor cells SKOV3 and incubated
overnight to prepare cell monolay-
ers. The effector cells (PMBCs) were
added to the tumor cell monolayers
at the appropriate effector cells/tumor
cells (E/T) ratios. Different dilutions
of supernatants containing TRI were
added to the plate, and then the plate
was incubated at 37°C, 5% CO2 for
72 h. After PBMCs were removed by
washing, MTT solution (0.5 mg/mL)
was added, and the plate was incubated
for another 4 h. The MTT solution
was then removed, and the crystals of
formazan were dissolved in dimethyl
sulfoxide (DMSO) at room tempera-
ture. The results were calculated based
on the means of absorbance obtained
from three wells according to the fol-
% lysis = 100 × (C-E)/(C-B)
where C is the absorbance reading of
the cells with target cells (control), B
is the background, and E is the absor-
bance reading of adherence tumor cells
remaining in the wells after co-incuba-
tion with PBMCs (24).
PG, CBD-PG, BH1, and TRI Always
Aggregate as Inclusion Bodies When
Expressed in E. coli
Although the induction conditions
were optimized for efficient expression
of the scAbs constructed in our labora-
tory, scAb PG and its fusion protein
CBD-PG were totally insoluble in E.
coli BL21 Star (DE3) (Figure 2A). For
facilitating the secretion of expressed
proteins, the mature pg gene was fused
to a signal peptide stII (a signal peptide
from E. coli thermostable enterotoxinII)
(25) or CBD (26), which contains a pow-
Figure 3. The fusion protein FkpA-BH1 was expressed as the soluble and biologically active form in Escherichia coli. (A) Sodium dodecyl sulfate
polyacrylamide gel electrophoresis (SDS-PAGE) analysis of FkpA-BH1 expressed in E. coli. Lane 1, molecular mass markers (M); lane 2, the whole-cell
proteins (W) of E. coli cells harboring pTMF; lanes 3, 4, and 5, the whole-cell proteins, supernatant (S) and pellet (P), respectively, of E. coli cells harboring
pTFBH1. (B) ELISA results of FkpA-BH1 against both CD3 antigen from Jurkat and tumor cell antigens from SKOV3. pTCM vs. Jurkat and pTCM vs.
SKOV3: the sonication supernatant from E. coli cells carrying empty vector pTCM was used against both the antigen from Jurkat (CD3) and SKOV3. pTFBH1
vs. Jurkat and pTFBH1 vs. SKOV3: the sonication supernatant from E. coli cells carrying pTFBH1 was used against both the antigen from Jurkat (CD3) and
SKOV3. BH1, anti-human CD3×anti-human ovarian carcinoma-bispecific scAb; scAb, single-chain antibody.
1038 BioTechniques Vol. 35, No. 5 (2003)
erful signal peptide, to yield expression
vector pTP1 (15) and pTBNP1 (unpub-
lished). However, no soluble products
were observed as shown in Figure 2.
Under the same conditions as mentioned
above, the expressed BH1 existed as
inclusion bodies (data not shown), and
most of the expressed TRI existed as
insoluble form (see Figure 4A).
Effects of FkpA Fusion or
Co-Expression on the Soluble
Production of Engineered
Antibodies in E. coli
To achieve large amounts of soluble
and biologically active scAbs, the
FkpA-based fusion and co-expression
vectors pTFP1, pTFBH1, pTBNP1F,
and pTTRIF were constructed. In
contrast to separate expression, most
part of each scAb or its fusion protein
FkpA-PG, FkpA-BH1, CBD-PG, or
TRI was soluble after sonication of the
whole cells analyzed by SDS-PAGE
(Figures 2B, 3A, and 4A), and the sol-
uble portion of FkpA-PG, FkpA-BH1,
or CBD-PG accounts for 50%, 70%,
Figure 4. FkpA markedly improves the expression of trispecific scAb TRI as the soluble and biologically active form in Escherichia coli. (A) Results of
Western blot analysis of TRI. Lanes 1 and 7, the whole-cell proteins (W) of E. coli cells harboring pTMF, and the uninduced E. coli cells harboring pTTRIF,
respectively; lanes 2 and 3, supernatant (S) and pellet (P), respectively, of E. coli cells harboring pTTRI; lane 4, molecular mass markers (M); lanes 5 and 6,
supernatant and pellet, respectively, of E. coli cells harboring pTTRIF. (B) ELISA results of TRI co-expressed with FkpA against CD3 molecules on Jurkat
cells and tumor antigens on SKOV3, respectively. pTMF vs. SKOV3 and pTMF vs. Jurkat: supernatant of E. coli cells harboring pTMF against tumor antigens
from SKOV3 and CD3 molecules from Jurkat, respectively; TRI vs. SKOV3 and TRI vs. Jurkat: supernatant of E. coli cells harboring pTTRIF against tumor
antigens from SKOV3 and CD3 molecules from Jurkat, respectively. (C) ELISA results of TRI co-expressed with FkpA against CD28 antigen. TRI and control
stand for supernatants of induced E. coli cells harboring pTTRIF and induced E. coli cells harboring pTMF, respectively. (D) TRI-mediated in vitro cytotoxicity
against tumor cell SKOV3. OCCD3CD28: anti-ovarian scAb, anti-CD3 monoclonal antibody, and anti-CD28 monoclonal antibody were used simultaneously
as positive control; OCCD3: anti-ovarian scAb and anti-CD3 monoclonal antibody were used simultaneously; control: RPM1640 culture medium was used as
negative control; TRI: the sonication supernatant of E. coli cells carrying pTTRIF was used to evaluate the in vitro cytotoxicity of TRI co-expressed with FkpA;
pTMF: the sonication supernatant of E. coli cells harboring empty vector pTMF was used as a control. scAb, single-chain antibody; TRI, anti-human CD3×anti-
human CD28×anti-human ovarian carcinoma-trispecific scAb.
Vol. 35, No. 5 (2003) BioTechniques 1041
and 81% of the total amount of each
target protein, respectively, expressed
at 37°C (Table 2).
ELISA and Western Blot
Analysis Results for scAbs
The ultimate object of FkpA-as-
sisted expression of scAbs is to achieve
biologically active products. Therefore,
the antigen-binding activities of solu-
ble FkpA-PG, FkpA-BH1, CBD-PG,
and TRI, with their specific membrane
antigens at different concentrations,
were assayed by ELISA, and the re-
sults are shown in Figures 2C, 3B, and
4B and C, respectively. These results
demonstrate that the soluble FkpA-PG,
FkpA-BH1, CBD-PG, and TRI have
relatively high binding affinities for
their specific antigens, and the binding
of the soluble target proteins with their
respective antigens is specific. The
immunoblotting results of the soluble
TRI from co-expression with FkpA are
shown in Figure 4A.
In Vitro Cytotoxicity for Trispecific
scAb TRI Co-Expressed with FkpA
All the scAbs constructed previously
in our laboratory were against vari-
ous human carcinomas. Therefore, the
in vitro cytotoxicity test is one of the
most valuable evaluations. In this study,
TRI-mediated in vitro cytotoxicity was
measured by a colorimetric MTT-based
assay, and the results are shown in Fig-
ure 4D. Varying E/T ratios were carried
out with a constant amount of 60 µg/mL
of sonication supernatant containing
TRI. Anti-human ovarian carcinoma
scAb, anti-CD3 monoclonal antibody,
and anti-CD28 monoclonal antibody
were used as controls. In addition, 60
µg/mL of sonication supernatant of E.
coli cells harboring empty vector pTMF
and PBS were used as negative controls.
These in vitro cytotoxicity tests indicate
that scAb TRI co-expressed with FkpA
is biologically functional in vitro.
In the present work, we constructed
a series of plasmids for induced expres-
sion of FkpA to assess the effects of
fusion and/or co-expression on soluble
and functionally active production of
scAbs and their fusion proteins in E.
coli. Our results demonstrated that
FkpA is very effective for producing
soluble and functional engineered an-
tibodies in E. coli. Many attempts to
improve functional expression of het-
erologous proteins by overexpression
with folding catalysts were carried out,
but only limited success was reported
(27–29). We previously tested the ef-
fects of E. coli Dsb proteins, a group of
disulfide-forming proteins with chaper-
one activities, on soluble and functional
expression of scAbs in E. coli and ob-
tained excellent results (15).
As described in the Results section, a
substantial portion of each fusion protein
FkpA-PG, CBD-PG, FkpA-BH1, or TRI
in E. coli was soluble and functionally
active. In all ELISAs, the antigen-bind-
ing affinities of each tested scAb or its
fusion protein were much higher than
those of controls used. Very promis-
ingly, the in vitro cytotoxicity assay
revealed that the trispecific scAb TRI
folding assisted by FkpA was able to
efficiently mediate the cell lysis against
specific tumor cells. Our results clearly
indicate that the scAbs overexpressed
with FkpA are not only soluble but also
biologically active in vitro. FkpA fused
to or co-expressed with scAbs provides
enough molecular activity and isom-
erase activity for the correct folding of
target proteins of interest. Therefore, it
is presumed that FkpA is the main factor
leading to the change of scAbs from the
insoluble forms to soluble forms when
expressed in E. coli. The overproduced
FkpA is most likely to efficiently en-
hance the correct folding by preventing
aggregation of folding intermediates into
inclusion bodies, thus yielding correctly
folded proteins. There have been reports
on demonstrating the chaperone-like ac-
tivity of FkpA (21). The chaperone-like
activity of FkpA is probably responsible
for the functional production of scAbs
as a crucial factor. It is well known
that some essential steps, such as the
formation of correct configuration, the
formation and isomerization of disul-
fide bonds, and the peptidyl-prolyl cis,
trans isomerization, are involved in the
production of functional scAbs in vivo.
Actually, all the tested antibodies in our
work contain several proline residues.
Therefore, we proposed that the enzy-
matic activities of FkpA identified and
still unidentified are also indispensable
for the expression of biologically active
scabs, since all the tested scAbs retained
full antigen-binding affinities for their
respective antigens and are able to kill
specific tumor cells through in vitro
Table 2. Comparison of Induced Soluble Protein Contents in E. coli BL21 Star (DE3) Harboring
Various Constructs Used in this Studya
in Total Target
TRI, co-expressed with
aNormally, cells were cultured at 37°C and induced with 0.4 mM isopropyl-β-D-
thiogalactopyranoside (IPTG) for 4 h.
bTarget proteins refer to PG, FkpA-PG, FkpA-BH1, CBD-PG, and TRI.
BH1, anti-human CD3×anti-human ovarian carcinoma-bispecific scA; CBD,
cellulose-binding domain from Cellulomonas fimi; PG, anti-human bladder
carcinoma scAb; scAb, single-chain antibody; TRI, anti-human CD3×anti-human
CD28×anti-human ovarian carcinoma-trispecific scAb.
1042 BioTechniques Vol. 35, No. 5 (2003) Download full-text
cytotoxicity. In our previous work, we
tested extensively the effects of the
general molecular chaperones GroEL/
GroES, the excellent fusion expres-
sion partners glutathione-S-transferase
(GST) and thioredoxin, and the disulfide
bond-forming related proteins DsbC
and DsbG, on functional production of
scAbs (15). Only DsbC and DsbG mark-
edly enhanced the soluble and functional
expression of scAbs, most probably
due to their enzymatic activities as well
as chaperone-like activities. In addi-
tion, high-level expression of FkpA
and its fusion proteins with scAbs are
not detrimental to the host cells in our
experiments, suggesting that expression
systems based on FkpA are applicable
as well as sufficient for assisted folding
expression of scAbs in E. coli.
In this report, we have described the
expression with high efficiency and the
marked increase of protein solubilization
by overproduction of FkpA under stan-
dard culture conditions. The successful
use of FkpA for the soluble and func-
tional expression of engineered antibod-
ies should be useful for the development
of strategies for the efficient production
of recombinant proteins, in general, in
soluble and functional form in E. coli.
1.Ward, E.S. 1992. Antibody engineering: the
use of Escherichia coli as an expression host.
FASEB J. 6:2422-2427.
2.Raag, R. and M. Whitlow. 1995. Single-chain
Fvs. FASEB J. 9:73-80.
3.Cheadle, C., L.E. Hook, D. Givol, and G.A.
Ricca. 1992. Cloning and expression of the
variable regions of mouse myeloma protein
MOPC315 in E. coli: recovery of active Fv
fragments. Mol. Immunol. 29:21-30.
4.Pantoliano, M.W., R.E. Bird, S. Johnson,
E.D. Asel, S.W. Dodd, J.F. Wood, and K.D.
Hardman. 1991. Conformational stability,
folding, and ligand-binding affinity of single-
chain Fv immunoglobulin fragments expressed
in Escherichia coli. Biochemistry 30:10117-
5.Hudson, P.J. 1998. Recombinant antibody
fragments. Curr. Opin. Biotechnol. 9:395-402.
6.Knappik, A. and A. Pluckthun. 1995. Engi-
neered turns of a recombinant antibody im-
prove its in vivo folding. Protein Eng. 8:81-89.
7.Kipriyanov, S.M., G. Moldenhauer, and M.
Little. 1997. High level production of soluble
single chain antibodies in small-scale Esch-
erichia coli cultures. J. Immunol. Methods
8.Makrides, S.C. 1996. Strategies for achieving
high-level expression of genes in Escherichia
coli. Microbiol. Rev. 60:512-538.
9.Verma, R., E. Boleti, and A.J. George. 1998.
Antibody engineering: comparison of bacte-
rial, yeast, insect and mammalian expression
systems. J. Immunol. Methods 216:165-181.
10.Humphreys, D.P., N. Weir, A. Lawson, A.
Mountain, and P.A. Lund. 1996. Co-expres-
sion of human disulfide isomerase (PDI) can
increase the yield of an antibody Fab′ fragment
expressed in Escherichia coli. FEBS Lett. 380:
11.Knappik, A., C. Krebber, and A. Pluckthun.
1993. The effect of folding catalysts on the in
vivo folding process of different antibody frag-
ments expressed in Escherichia coli. Biotech-
nology (NY) 11:77-83.
12.Ward, E.S., D. Gussow, A.D. Griffiths, P.T.
Jones, and G. Winter. 1989. Binding activities
of a repertoire of single immunoglobulin vari-
able domains secreted from Escherichia coli.
13.Skerra, A. and A. Pluckthun. 1988. Assembly
of a functional immunoglobulin Fv fragment in
Escherichia coli. Science 240:1038-1040.
14.Better, M., C.P. Chang, R.R. Robinson, and
A.H. Horwitz. 1988. Escherichia coli secre-
tion of an active chimeric antibody fragment.
15.Zhang, Z., Z.H. Li, F. Wang, M. Fang, C.C.
Yin, Z.Y. Zhou, Q. Lin, and H.L. Huang.
2002. Overexpression of DsbC and DsbG
Markedly improves soluble and functional ex-
pression of single-chain fv antibodies in Esch-
erichia coli. Protein Expr. Purif. 26:218-228.
16.Danese, P.N. and T.J. Silhavy. 1997. The
sigma (E) and the Cpx signal transduction
systems control the synthesis of periplasmic
protein-folding enzymes in Escherichia coli.
Genes Dev. 11:1183-1193.
17.Dartigalongue, C., D. Missiakas, and S.
Raina. 2001. Characterization of the Esch-
erichia coli sigma E regulon. J. Biol. Chem.
18.Missiakas, D., J.M. Betton, and S. Raina.
1996. New components of protein folding in
extracytoplasmic compartments of Escherichia
coli SurA, FkpA and Skp/OmpH. Mol. Micro-
19.Ramm, K. and A. Pluckthun. 2000. The
periplasmic Escherichia coli peptidylprolyl cis,
trans-isomerase FkpA. II. Isomerase-indepen-
dent chaperone activity in vitro. J. Biol. Chem.
20.Ramm, K. and A. Pluckthun. 2001. High
enzymatic activity and chaperone function are
mechanistically related features of the dimeric
E. coli peptidyl-prolyl-isomerase FkpA. J. Mol.
21.Arie, J.P., N. Sassoon, and J.M. Betton. 2001.
Chaperone function of FkpA, a heat shock pro-
lyl isomerase, in the periplasm of Escherichia
coli. Mol. Microbiol. 39:199-210.
22.Bothmann, H. and A. Pluckthun. 2000. The
periplasmic Escherichia coli peptidylprolyl
cis, trans-isomerase FkpA. I. Increased func-
tional expression of antibody fragments with
and without cis-prolines. J. Biol. Chem. 275:
23.Sambrook, J., E.F. Fritsch, and T. Maniatis.
1989. Molecular Cloning: A Laboratory Manu-
al, 2nd ed. CSH Laboratory Press, Cold Spring
24.Heo, D.S., J.G. Park, K. Hata, R. Day, R.B.
Herberman, and T.L. Whiteside. 1990.
Evaluation of tetrazolium-based semiauto-
matic colorimetric assay for measurement of
human antitumor cytotoxicity. Cancer Res. 50:
25.Zhu, Z., G. Zapata, R. Shalaby, B. Snedecor,
H. Chen, and P. Carter. 1996. High level
secretion of a humanized bispecific diabody
from Escherichia coli. Biotechnology (NY)
26.Greenwood, J.M., E. Ong, N.R. Gilkes, R.A.
Warren, R.C. Miller, Jr., and D.G. Kilburn.
1992. Cellulose-binding domains: potential for
purification of complex proteins. Protein Eng.
27.Jeong, K.J. and S.Y. Lee. 2000. Secretory
production of human leptin in Escherichia coli.
Biotechnol. Bioeng. 67:398-407.
28.Kurokawa, Y., H. Yanagi, and T. Yura. 2000.
Overexpression of protein disulfide isomerase
DsbC stabilizes multiple-disulfide-bonded
recombinant protein produced and transported
to the periplasm in Escherichia coli. Appl. En-
viron. Microbiol. 66:3960-3965.
29.Kurokawa, Y., H. Yanagi, and T. Yura. 2001.
Overproduction of bacterial protein disulfide
isomerase (DsbC) and its modulator (DsbD)
markedly enhances periplasmic production of
human nerve growth factor in Escherichia coli.
J. Biol. Chem. 276:14393-14399.
Received 30 December 2002; accepted
22 July 2003.
Address correspondence to:
Institute of Genetics and Developmental Biology
No. 3 Datun Rd.
Beijing 100101, China