Báo cáo khoa học: Application of the Fc fusion format to generate tag-free bi-specific diabodies doc

Application of the Fc fusion format to generate tag-free
bi-specific diabodies
Ryutaro Asano
1
, Keiko Ikoma
1
, Hiroko Kawaguchi
1
, Yuna Ishiyama
1
, Takeshi Nakanishi
1
,
Mitsuo Umetsu
1
, Hiroki Hayashi
2
, Yu Katayose
2
, Michiaki Unno
2
, Toshio Kudo
3
and Izumi Kumagai
1
1 Department of Biomolecular Engineering, Graduate School of Engineering, Tohoku University, Sendai, Japan
2 Division of Gastroenterological Surgery, Department of Surgery, Graduate School of Medicine, Tohoku University, Sendai, Japan
3 Cell Resource Center for Biomedical Research, Institute of Development, Aging and Cancer, Tohoku University, Sendai, Japan
Introduction
Bi-specific antibodies (BsAbs) are attractive formats
for recombinant antibodies that can bind to two differ-

ent epitopes on antigens. This bi-specificity can be used
in cancer immunotherapy by cross-linking tumor cells
to immune cells such as cytotoxic T cells, natural killer
cells and macrophages. This linkage accelerates the
destruction of the tumor cells by immune cells, so that
the dose of therapeutic antibodies can be reduced
from that required in the case of mono-specific anti-
bodies [1,2].
Keywords
bi-specific diabody; Fc fusion format;
preparation method; small therapeutic
antibody; tag-free protein
Correspondence
I. Kumagai, Aoba 6-6-11-606, Aramaki,
Aoba-ku, Sendai 980-8579, Japan
Fax: +81 22 795 6164
Tel: +81 22 795 7274
E-mail:
(Received 1 May 2009, revised 9
November 2009, accepted 17 November
2009)
doi:10.1111/j.1742-4658.2009.07499.x
We previously reported the use of a humanized bi-specific diabody that
targets epidermal growth factor receptor and CD3 (hEx3-Db) for cancer
immunotherapy. Bacterial expression can be used to express small recombi-
nant antibodies on a large scale; however, their overexpression often results
in the formation of insoluble aggregates, and in most cases artificial affinity
peptide tags need to be fused to the antibodies for purification by affinity
chromatography. Here, we propose a novel method for preparing refined,
functional, tag-free bi-specific diabodies from IgG-like bi-specific antibodies

(BsAbs) in a mammalian expression system. We created an IgG-like BsAb
in which bi-specific diabodies were fused to the human Fc region via a
designed human rhinovirus 3C (HRV3C) protease recognition site. The
BsAb was purified by protein A affinity chromatography, and the refined
tag-free hEx3-Db was efficiently produced from the Fc fusion format by
protease digestion. The tag-free hEx3-Db from the Fc fusion format
showed a greater inhibition of cancer growth than affinity-tagged hEx3-Db
prepared directly from Chinese hamster ovary cells. We also applied our
novel method to another small recombinant antibody fragment, hEx3 sin-
gle-chain diabody (hEx3-scDb), and demonstrated the versatility and
advantages of our proposed method compared with papain digestion of
hEx3-scDb. This approach may be used for industrial-scale production of
functional tag-free small therapeutic antibodies.
Abbreviations
BsAbs, bi-specific antibodies; CHO, Chinese hamster ovary; Db, diabody; EGFR, epidermal growth factor receptor; hEx3-Db, humanized
bi-specific diabody that targets epidermal growth factor receptor and CD3; hEx3-scDb, hEx3 single-chain diabody; HRV3C, human rhinovirus
3C; MTS, 3-(4,5-dimethylthiazole-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium inner salt; scDb, single-chain diabody;
scFv, single chain Fv; T-LAK, lymphokine-activated killer cells with the T-cell phenotype; tanDb, tandem single-chain diabody;
taFv, tandem scFv.
FEBS Journal 277 (2010) 477–487 ª 2009 The Authors Journal compilation ª 2009 FEBS 477
Conventionally, BsAbs are produced by chemical
conjugation or somatic fusion of two hybridomas, form-
ing a quadroma that can produce bi-specific IgG mole-
cules [1,3]. Clinical studies of these BsAbs have been
performed, and some impressive local anti-tumor
responses have been reported; however, these trials have
also been limited by the occurrence of human anti-
mouse antibody and ⁄ or Fc-mediated side-effects such as
the induction of a cytokine storm [4,5]. Furthermore,
these methods cannot be utilized for large-scale produc-

tion, and a quadroma cannot control the heterogeneity
of the antibodies produced; for instance, ten possible
variants of antibodies can be generated when two heavy
and two light chains are randomly associated. There-
fore, steady production of homogeneous BsAbs requires
the use of a host-vector system.
Advances in antibody engineering techniques and
host-vector expression systems have facilitated the gen-
eration of recombinant BsAbs with improved proper-
ties. A variety of recombinant BsAbs have been
developed from two antibody fragments such as single-
chain Fv fragments (scFv; 25 kDa) [6,7], and diabodies
(Db; 55 kDa) [8] that recognize different antigens. The
most common BsAb formats that have been produced
from these fragments are tandem scFv (taFv) [9], tan-
dem single-chain diabodies (tandem scDb, tanDb) [10]
and mini-bodies (dimeric scDb–CH3 fusion protein)
[11]. Compared with classic BsAbs prepared by chemi-
cal conjugation or production of a quadroma, small
antibody molecules, such as diabodies, are of a suit-
able size for rapid tissue penetration, high target reten-
tion and rapid clearance [12,13]. Their smaller size also
enables expression of BsAbs in bacteria, and as the
structure is composed only of antibody variable
regions, this eliminates the Fc-mediated side-effects of
BsAbs. Although the rapid blood clearance and
monovalency of bi-specific diabodies, scDbs and taFv
(all approximately 55 kDa) may limit their therapeutic
application, engineering the length and amino acid
composition of the middle linker in scDb, for example,

may enable them to assemble into multimers, such as
tanDb (114 kDa), with higher molecular weight and
bivalency for each target antigen [14,15].
Small bi-specific antibody fragments prepared in
bacteria are often expressed as insoluble aggregates in
the cytoplasmic or periplasmic space [10,16–18], and
require fusion of artificial affinity peptide tags, such as
a polyhistidine tag, hemagglutinin tag or FLAG tag,
at the N- or C-terminus of the BsAbs to allow com-
plete removal of the vast amount of host-derived
proteins by affinity chromatography [16,19]. The
requirement for such tags raises concerns about immu-
nogenicity. We have previously reported significant
anti-tumor activity in vitro and in vivo for a humanized
bi-specific diabody targeting epidermal growth factor
receptor (EGFR) and CD3 (hEx3-Db) [20]. However,
even though the yield of hEx3-Db was over 10 mgÆL
)1
culture, it was also expressed as insoluble aggregates,
and fusion of an affinity tag was necessary for purifica-
tion before the re-folding process.
We have also reported the construction of a mam-
malian expression system for affinity-tagged bi-specific
diabodies and their Fc fusion formats [21]. Here, we
developed a novel method for the production of highly
purified tag-free diabodies using the mammalian
expression system. Diagrams of the various gene con-
structs are shown in Fig. 1. The tag-free hEx3-Db
alone was expressed sufficiently to be purified by ion-
exchange chromatography. Expression of the hEx3

diabodies fused to the human Fc region via a designed
protease recognition site enabled high-efficiency purifi-
cation by protein A affinity chromatography and
increased the yield of tag-free hEx3-Db. We also used
our method to produce tag-free small BsAbs to hEx3-
scDb. For hEx3-scDb, use of the designed protease
recognition site had advantages over papain digestion,
which caused unwanted degradation. Both tag-free
hEx3-Db and hEx3-scDb prepared by restriction pro-
tease digestion from the Fc fusion format showed a
greater inhibition of cancer growth in vitro than previ-
ously produced affinity-tagged diabodies directly pre-
pared from the supernatant of Chinese hamster ovary
(CHO) transfectants [21]. Thus, this approach appears
to improve both the yield and efficacy of the bi-specific
antibody fragments.
Results
Preparation of tag-free bi-specific diabodies
Tag-free hEx3-Db was directly secreted from mamma-
lian cells and purified by cation-exchange chromatog-
raphy as described in Experimental procedures.
Purified hEx3-Db was applied to a gel filtration col-
umn for further analysis and purification (Fig. 2A).
The first small peak, second large peak and the shoul-
der of the major peak seen in the chromatograph were
identified as the multimeric, dimeric and monomeric
structures of tag-free hEx3-Db, respectively. Equiva-
lent amounts of hOHh5L (humanized OKT3 VH -
linker - humanized 528 VL) and h5HhOL (humanized
528 VH - linker - humanized OKT3 VL) were con-

firmed in the dimeric fraction by SDS–PAGE analysis
(Fig. 2B). Thus, purified tag-free hEx3-Dbs were
obtained without affinity chromatography at a final
yield of approximately 1 mgÆL
)1
culture.
Fc fusion for generation of tag-free diabodies R. Asano et al.
478 FEBS Journal 277 (2010) 477–487 ª 2009 The Authors Journal compilation ª 2009 FEBS
To prepare the high-quality, tag-free bi-specific dia-
bodies, we fused the hEx3-Db to the human IgG1 Fc
region. We inserted a recognition site for HRV3C pro-
tease between the diabody fragments and the Fc por-
tion of hEx3-Fc. A schematic illustration of the
preparation of tag-free hEx3-Db from its Fc fusion
format is shown in Fig. 3A. The expressed IgG-like
BsAbs were purified by protein A affinity chromatog-
raphy and digested using glutathione S-transferase
(GST)-fused HRV3C protease. The treated solution
was loaded onto a glutathione-immobilized column
and then a protein A column to remove added prote-
ase and digested Fc. SDS–PAGE analysis of each puri-
fication step showed the successful preparation of tag-
free hEx3-Db from its Fc fusion format (Fig. 3B). Gel
filtration chromatography showed that tag-free hEx3-
Db predominantly formed dimers, with a small
amount of multimeric forms (Fig. 4A). The homogene-
ity of tag-free hEx3-Db in the eluted fraction was also
confirmed by SDS–PAGE (Fig. 4B). The final yield of
tag-free hEx3-Db from the Fc fusion format was
approximately 5 mgÆL

)1
culture, i.e. five times that of
the secreted tag-free hEx3-Db. Thus, secretion of
BsAbs as the Fc fusion format increased the amount
of prepared tag-free diabodies due to the high produc-
tivity (approximately 10 mgÆL
)1
) and the efficient puri-
fication using protein A.
Mass spectrometry of tag-free bi-specific
diabodies
We previously reported that the strong inter-domain
interaction between cognate V
H
and V
L
domains of
hEx3-Db leads to the spontaneous formation of func-
tional heterodimers [22]. In the present study, the
molecular weight of the monomorphous heterodimer
of the tag-free hEx3-Db prepared from the Fc fusion
format was confirmed by MALDI-TOF mass spec-
trometry (Fig. 4C). The mass spectrum for the diabod-
ies prepared from the Fc fusion format had two peaks,
one at m ⁄ z 26 424 and another at m ⁄ z 25 970, which
correspond to the calculated molecular weights of
Nhe I Xho I
Nhe I Xho I
Nhe I Xho Io
h5H hOL

Neo
r
hOH h5L
Hyg
r
h5H
h5H
CH2 CH3 Neo
r
pcDNA-h5HhOL-3C-Fc
Tag-free hEx3-Db
pcDNA-h5HhOL(–) pcDNA-hOHh5L(–)
hEx3-Db-3C-Fc(tool for tag-free hEx3-Db)
hEx3-scDb-3C-Fc(tool for tag-free hEx3-scDb)
Nhe I
Xho I
h5H hOL
CH2 CH3 NhOH h5L
h5H hOL
CH2 CH3 Neo
r
hOH h5L
5O
pcDNA-hEx3-scDb-3C-FcpcDNA hEx3 scDb 3C Fc
CMV promoter Kozak sequence
Leader peptide
Peptide linker (GGGGS) Peptide linker [(GGGGS)4]
HRV3C protease recognition site (LEVLFQGP) Hinge
Neo Hyg
Neomycin resistance Hygromycin resistance

Neo
r
Hyg
r
hOL
Fig. 1. Schematic illustration of the BsAb
gene constructs in pCDNA3.1. The V
H
and
V
L
regions of humanized 528 Fv are desig-
nated h5H and h5L, and those of humanized
OKT3 Fv are designated hOH and hOL,
respectively. The positions of important
restriction enzyme sites used and the key
components are shown.
R. Asano et al. Fc fusion for generation of tag-free diabodies
FEBS Journal 277 (2010) 477–487 ª 2009 The Authors Journal compilation ª 2009 FEBS 479
hOHh5L digested from the Fc fusion (26 442) and
h5HhOL without the peptide tag (25 991), respectively.
These results indicate that Db–3C–Fc fusion proteins
can serve as a tool for preparing tag-free diabodies
with high yield and purity.
Binding affinity of tag-free bi-specific diabodies
and its effect on growth inhibition
The binding affinity of tag-free hEx3-Dbs for CD3-
positive lymphokine-activated killer cells with the
T-cell phenotype (T-LAK cells) and EGFR-positive
TFK-1 cells was measured by flow cytometry using

polyclonal antibody to hEx3-Db. Tag-free hEx3-Dbs
interacted with each targeted antigen (Fig. 5A), and
the binding profiles were comparable with those previ-
ously reported for affinity-tagged hEx3-Db [20,22].
These results indicate that the diabody prepared by
HRV3C protease digestion from the Fc fusion format
retained sufficient binding activity and bi-specificity.
To evaluate the inhibition of cancer growth by tag-
free hEx3-Db, an MTS assay was performed for TFK-
1 cells by using T-LAK cells at an effector ⁄ target ratio
of 5 : 1. Tag-free hEx3-Db prepared from the Fc
fusion format inhibited cancer cell growth more effec-
tively than did affinity-tagged hEx3-Db (Fig. 5B).
Imperceptible differences in purity and local structural
perturbations that are dependent on the preparation
method might affect these activities.
67 kDa
A
B
25 kDa43 kDa
47.5 –
5 mAU
32.5
–
Tag-free hOHh5L
25 –
Tag-free h5HhOL
16.5 –
150 200 250 300
Elution volume (mL)

Fig. 2. (A) Gel filtration of tag-free hEx3-Db. The elution volume is
shown on the x axis, and the molecular mass (kDa) is shown
above. The eluted fractions containing the bi-specific diabody are
indicated by the two-headed arrow. (B) SDS–PAGE analysis under
reducing conditions of the eluted fraction. Molecular size markers
are shown on the left.
HRV3C protease site
A
B
Tag-free hEx3-Db
hEx3-Db-3C-Fc
hEx3-Db-3C-Fc
Tag-free hEx3-Db
1
175 –
83 –
62
–
47.5
–
h5HhOL-3C-Fc
–32.5
Fc
Tag-free hOHh5L
25 –
Tag-free h5HhOL
16.5 –
234
Fig. 3. (A) Schematic illustration of the hEx3-Db–3C–Fc fusion pro-
tein. The HRV3C protease cleavage site used for preparation of

tag-free hEx3-Db is indicated. (B) Reducing SDS–PAGE of each
purification step for preparation of tag-free hEx3-Db from hEx3-Db–
3C–Fc. Lane 1, protein A chromatography-purified hEx3-Db–3C–Fc;
lane 2, after HRV3C protease digestion; lane 3, after removal of
HRV3C protease by glutathione Sepharose 4B chromatography;
lane 4, purified tag-free hEx3-Db after removal of the Fc region by
protein A chromatography.
Fc fusion for generation of tag-free diabodies R. Asano et al.
480 FEBS Journal 277 (2010) 477–487 ª 2009 The Authors Journal compilation ª 2009 FEBS
Application of method to tag-free bi-specific sin-
gle-chain diabody
To demonstrate the utility of this novel method, we
applied it to the preparation of tag-free hEx3-scDb,
which is a single-chain form of hEx3-Db (Fig 6A). An
HRV3C protease recognition site was inserted between
hEx3-scDb and the Fc portion, and the recognition
sequence for papain was conserved. Papain is a cyste-
ine protease that is generally used in the preparation
of Fab fragments from IgG, because the recognition
site for papain naturally exists around the hinge region
of intact antibody.
When we digested hEx3-scDb–3C–Fc with HRV3C
protease, hEx3-scDb was separated from the Fc por-
tion with no degradation. Similar to the tag-free hEx3-
Db, the Fc portion was completely removed by protein
A affinity chromatography (Fig. 6B). To confirm the
benefit of the design of the HRV3C protease digestion
site, we also followed the time course of papain
digestion of hEx3-scDb–3C–Fc (Fig. 6C). Although
tag-free hEx3-scDb was successfully prepared by

papain digestion, especially with an incubation time
of 1 h, two unexpected bands corresponding to
hOHh5L and h5HhOL caused by a break in the mid-
dle linker from scDb also appeared. This digestion
proceeded as the incubation time increased, and
further degradation of h5HhOL was observed after
incubation for 10 h.
The binding affinity of tag-free hEx3-scDb prepared
from the Fc fusion format for both targeted cells was
confirmed by flow cytometry (Fig. 7A), and its
enhanced cytotoxicity was compared with affinity-
tagged hEx3-scDb [21] in the MTS assay with the use
of T-LAK cells as effector cells (Fig. 7B). These results
strongly support the utility and general applicability of
our method for the preparation of homogeneous tag-
free small recombinant antibodies.
Discussion
Recombinant BsAbs have several advantages over clas-
sic BsAbs prepared by chemical cross-linkage or fusion
of two hybridoma clones [16,23–25]. The IgG-like
BsAbs containing human Fc regions are highly effec-
tive recombinant antibodies [25–27] because of the
antibody-dependent cellular cytotoxicity effect. By
comparison, small bi-specific diabodies without Fc
have the advantages of rapid tissue penetration, high
target retention and a distance between the two anti-
gen-binding sites of the diabodies that is large enough
to bring two cells together for recruitment of immune
cells [1,2,28].
Large-scale production of bi-specific diabodies in

bacterial expression systems would be expected because
of their small size; however, the yield is typically only
a few mg per L in most cases [10,16,17]. We previously
proposed an in vitro refolding system to prepare
150 200 300
20 000 30 000 35 000 m /z25 000
250
AB C
Fig. 4. (A) Gel filtration of purified hEx3-Db after removal of HRV3C protease and the Fc fragment. The elution volume is shown on the x
axis, and the molecular mass (kDa) is shown above. The eluted fractions containing the bi-specific diabody are indicated by the two-headed
arrow. (B) SDS–PAGE analysis under reducing conditions of the eluted fractions. Molecular size markers are shown on the left. (C) MALDI-
TOF mass spectra of the tag-free hEx3-Db prepared from hEx3-Db–3C–Fc.
R. Asano et al. Fc fusion for generation of tag-free diabodies
FEBS Journal 277 (2010) 477–487 ª 2009 The Authors Journal compilation ª 2009 FEBS 481
functional bi-specific diabodies from the insoluble frac-
tion, but solubilizing the expressed proteins from insol-
uble fraction required purification from the vast
amount of host-derived proteins, which forced us to
utilize an artificial tag [20,22,29]. The immunogenicity
of the artificial peptide tag has not been determined,
and preparation of tag-free formats from insoluble
fractions may be difficult to achieve [16]. For these
reasons, a new preparation method for bi-specific dia-
bodies was needed that required minimal artificial
amino acid sequences and produced high yields.
In the present study, we successfully purified tag-free
hEx3-Db from the supernatant of transfected CHO
T-LAK
A
B

TFK-1
a
b
a
b
Counts
Fluorescent intensity
100
*
E:T = 5
*
50
Affinity-tagged
hEx3-Db
Tag-free hEx3-Db
from Fc fusion
0
0 1 10 100 10000 1 10 100 1000
(T LAK)(T-LAK)
Concentration of BsAb (fmol·mL
–1
)
Growth inhibition of cancer cells (%)
10
0
0 40 80 120 160 200
0 40 80 120 160
10
1
10

2
10
3
10
4
10
0
10
1
10
2
10
3
10
4
Fig. 5. (A) Flow cytometry analysis of tag-free hEx3-Db prepared
from hEx3-Db–3C–Fc. Cells were incubated with NaCl ⁄ P
i
as a nega-
tive control (a) and with either OKT3 parental IgG (for T-LAK cells)
or 528 IgG (for TFK-1 cells), followed by staining with fluorescein
isothiocyanate-conjugated anti-mouse IgG as a positive control (b).
The shaded areas correspond to the fluorescence intensity distribu-
tions of the cells incubated with hEx3-Db. Each mixture was
stained with rabbit anti-hEx3-Db serum followed by fluorescein
isothiocyanate-conjugated anti-rabbit IgG. (B) Growth inhibition of
EGFR-positive TFK-1 cells by tag-free and affinity-tagged hEx3
diabodies. Each bi-specific diabody and T-LAK cells (effectors, E)
were added to TFK-1 cells (T) at a ratio of 5 : 1. The tag-free
hEx3-Db inhibited growth significantly better (*P < 0.005) than the

affinity-tagged hEx3-Db did [21]. Data are mean values ± SD and
are representative of at least three independent experiments with
similar results.
HRV3C protease site
A
B
C
Tag-free hEx3-scDb
hEx3-scDb-3C-Fc
Tag-free hEx3-scDbhEx3-scDb-3C-Fc
1
hEx3-scDb-3C-Fc
(monomer)
Tag-free hEx3-scDb
Fc
1 2
175
-
hEx3-scDb-3C-Fc
83
-
(monomer)
62
-
Tag-free hEx3-scDb
47.5-
32.5-
Fc
Tag-free hOHh5L
Tag-free h5HhOL

25
-
-
16.5-
1 h 5 h 10 h
234
3
1 2
3
1 2
3
Fig. 6. (A) Schematic illustration of the hEx3-scDb–3C–Fc fusion
protein. The HRV3C protease cleavage site used for preparation of
tag-free hEx3-scDb is indicated. (B) Reducing SDS–PAGE of each
purification step for preparation of tag-free hEx3-scDb from hEx3-
scDb–3C–Fc. Lane 1, protein A chromatography-purified hEx3-
scDb–3C–Fc; lane 2, after HRV3C protease digestion; lane 3, after
removal of HRV3C protease by glutathione Sepharose 4B chroma-
tography; lane 4, purified tag-free hEx3-scDb after removal of the
Fc region by protein A chromatography. (C) Reducing SDS–PAGE
of hEx3-scDb–3C–Fc incubated with papain for 1, 5 and 10 h. Lane
1, digested hEx3-scDb–3C–Fc; lane 2, flowthrough from protein
A chromatography; lane 3, eluted protein from protein A chroma-
tography.
Fc fusion for generation of tag-free diabodies R. Asano et al.
482 FEBS Journal 277 (2010) 477–487 ª 2009 The Authors Journal compilation ª 2009 FEBS
cells using cation-exchange and gel filtration chroma-
tography (Fig. 2). However, the final yield of this
secreted tag-free hEx3-Db was approximately 1 mgÆL
)1

culture. We thus developed a novel method using IgG-
like BsAb and a restriction protease with high specific-
ity. The fusion of Fc to diabodies resulted in high
productivity and enabled affinity purification using
protein A. The homogeneous dimer structure and
molecular weight of the tag-free hEx3-Db prepared
from the Fc fusion format (hEx3-Db–3C–Fc) were
confirmed by gel filtration and mass spectrometry, and
the yield was approximately five times that of the
directly secreted tag-free hEx3-Db (Figs 3 and 4).
The specific binding affinity and bi-specificity of the
tag-free hEx3-Db for T-LAK and TFK-1 cells were
observed by flow cytometry (Fig. 5A). Interestingly,
the result of the MTS assay showed that growth inhi-
bition by tag-free hEx3-Db from the Fc fusion was
more intense than that by affinity-tagged hEx3-Db
(Fig. 5B). Although it is unclear why the tag-free dia-
bodies prepared from the Fc fusion format had such
high activity, imperceptible differences in purity and
local structural perturbations that are dependent on
the preparation method might have affected the activ-
ity of the diabodies. The reasons for this difference in
activity are now under investigation. Furthermore, we
were able to reproduce our results with tag-free hEx3-
scDb, which indicates the utility and applicability of
our method for the preparation of tag-free small
recombinant antibodies (Figs 6 and 7). The single-
chain format has additional advantages: scDbs can be
produced from a single expression vector and are
expected to have improved stability in vivo because the

two chains in the diabody are connected to each other
via a linker [14,30].
In general, papain and pepsin have been used in the
preparation of antibody fragments from IgG-like anti-
bodies, and successful preparation of scFv from scFv–
Fc has also been reported [31]. However, for hEx3 sin-
gle-chain diabodies fused with Fc, papain digestion led
to undesired degradation (Fig. 6C). Thus, the advanta-
ges of using the designed protease recognition site were
confirmed, especially in recombinant antibodies that
included a number of artificial sequences.
To date, several different small BsAb formats have
been proposed to increase efficacy and availability,
such as scDbs [30], taFv [9,32] and mini-bodies [11].
Further, dimeric scDbs known as tanDbs, with biva-
lency for each target antigen, can be produced by engi-
neering the length and amino acid composition of
middle linker of scDb [15]. Here, we selected diabodies
and scDb monomers with a 20-amino-acid middle lin-
ker [(GGGGS)
4
] as small BsAbs, because they are one
of the simplest construction formats [20,22]. Use of
our preparation method for other BsAbs formats is
currently in progress.
We previously reported for BsAbs with affinity pep-
tide tags that hEx3-scDb has comparable function to
that of hEx3-Db in vitro [22]. In this work, we
have shown that tag-free formats behave quantitatively
T-LAK

A
B
TFK-1
a
b
a
b
Counts
Fluorescent intensity
100
E:T = 5
*
*
50
Affinity-tagged
hEx3-scDb
Tag-free hEx3-scDb
from Fc fusion
0
0 1 10 100 10000
(T-LAK)
Concentration of BsAb (fmol·mL
–1
)
Growth inhibition of cancer cells (%)
10
0
0 40 80 120 160 200
0 40 80 120 160 200
10

1
10
2
10
3
10
4
10
0
10
1
10
2
10
3
10
4
Fig. 7. (A) Flow cytometric analysis of purified tag-free hEx3-scDb.
Cells were incubated with NaCl ⁄ P
i
as a negative control (a) and
with either OKT3 parental IgG (for T-LAK cells) or 528 IgG (for TFK-
1 cells), followed by staining with fluorescein isothiocyanate-conju-
gated anti-mouse IgG as a positive control (b). The shaded areas
correspond to the fluorescence intensity distributions of the cells
incubated with hEx3-Db. Each mixture was stained with rabbit
anti-hEx3-Db serum followed by fluorescein isothiocyanate-
conjugated anti-rabbit IgG. (B) Growth inhibition of EGFR-positive
TFK-1 cells by tag-free and affinity-tagged hEx3 single-chain diabod-
ies. Each bi-specific diabody and T-LAK cells (effectors, E) were

added to TFK-1 cells (T) at a ratio of 5 : 1. The tag-free hEx3-scDb
inhibited growth significantly better (*P < 0.005) than the affinity-
tagged hEx3-scDb did [21]. Data are mean values ± SD and are
representative of at least three independent experiments with
similar results.
R. Asano et al. Fc fusion for generation of tag-free diabodies
FEBS Journal 277 (2010) 477–487 ª 2009 The Authors Journal compilation ª 2009 FEBS 483
similarly in in vitro cell growth inhibition studies
(Figs 5B and 7B). Therefore, regardless of the presence
or absence of an affinity tag, the activity of hEx3-Db
is comparable to that of hEx3-scDb. Several reports
have demonstrated a higher stability of scDb than
other formats such as Db and taFv [14,33–35].
Although hEx3-Db and hEx3-scDb showed similar
activities in vitro, there is a possibility the hEx3-scDb
may exhibit a higher activity in vivo because of higher
stability. Stability tests under physiological conditions
between hEx3-Db and hEx3-scDb are currently in pro-
gress.
Issues such as rapid blood clearance and the rela-
tively low affinity caused by low molecular weight and
monovalent binding may limit the therapeutic applica-
tion of bi-specific diabodies [14]. In such cases, conver-
sion into more effective formats such as tanDb may be
required. The approach described here is also expected
to be applicable for convenient preparation of such
antibody fragments.
In conclusion, we prepared tag-free bi-specific
diabodies in a mammalian expression system and devel-
oped a novel method using IgG-like antibodies and

protease digestion to prepare highly purified, tag-free
bi-specific diabodies. Our method may allow industrial-
scale production of functional tag-free small biological
agents such as small recombinant antibodies.
Experimental procedures
Preparation of secreted Ex3 diabodies
In accordance with the convention used in a previous
report, we describe the two hetero scFvs of hEx3-Db as
h5HhOL and hOHh5L [20]. The gene constructs (Fig. 1)
were inserted into pcDNA3.1 ⁄ Neo or pcDNA3.1 ⁄ Hygro
mammalian expression vectors (both from Invitrogen,
Groningen, Netherlands). The leader peptide sequences for
protein secretion were derived from the mouse OKT3 IgG
[36]. The methods for expression and purification of the
affinity-tagged hEx3-Db and hEx3-scDb have been
described previously [21]. For production of tag-free hEx3-
Db, CHO cells were co-transfected with pcDNA-h5HhOL
()) and pcDNA-hOHh5L()) (Fig. 1), and cell clones
expressing tag-free hEx3-Db were established in the pres-
ence of neomycin (G418) and hygromycin as described pre-
viously [21]. CHO clones that stably expressed tag-free
hEx3-Db were selected by screening for a growth inhibition
effect of each individual clone. The established CHO clone
was cultured as previously described [27]. The secreted tag-
free hEx3-Db was purified from pooled supernatants using
a 5 mL HiTrap SP XL column (GE Healthcare Bio-Science
Corp., Piscataway, NJ, USA) with a 5–250 mm gradient of
NaCl in 50 mm phosphate solution (pH 6.0).
Preparation of tag-free hEx3 diabodies from the
Fc fusion format

To construct the expression vector for preparing tag-free
diabodies by using IgG-like BsAbs, we connected the hEx3
diabodies and the human IgG1 Fc region via a recognition
site (LEVLFQGP) for human rhinovirus 3C (HRV3C) pro-
tease. CHO cells were co-transfected with equal amounts of
the pcDNA-h5HhOL-3C-Fc and pcDNA-hOHh5L()) vec-
tors (Fig. 1), and grown in presence of neomycin (G418)
and hygromycin as described previously [21]. A CHO clone
that stably expressed the hEx3-Db–3C–Fc fusion protein
was selected in a manner similar to that for tag-free hEx3-
Db. For tag-free hEx3-scDb, CHO cells were transfected
with the pcDNA-hEx3-scDb–3C–Fc vector, and selection
for a stably expressed clone was performed in the presence
of 500 lgÆmL
)1
of G418 (Nacalai Tesque, Kyoto, Japan).
IgG-like BsAbs of hEx3–3C–Fc and hEx3-scDb–3C–Fc
were first purified by affinity chromatography on a protein
A column (GE Healthcare) and then digested by HRV3C
protease fused to GST (PreScission protease; GE Health-
care) according to the protocol described by the manufac-
turer. The protease was removed using a glutathione
Sepharose 4B column (GE Healthcare), and the flow-
through was re-loaded onto the protein A column to
remove the digested Fc and undigested hEx3-scDb–3C–Fc
fusion protein. The presence of the BsAbs in each stage of
purification were confirmed by SDS–PAGE under reducing
conditions.
To illustrate the applicability of this novel method,
papain digestion of hEx3-scDb–3C–Fc was performed by

use of an ImmunoPure Fab preparation kit (Thermo Fisher
Scientific Inc., Rockford, IL, USA). The influence of
papain digestion was confirmed by SDS–PAGE analysis
under reducing conditions at 1, 5 and 10 h after digestion.
Gel filtration chromatography
Gel filtration analysis with a Hiload Superdex 200 pg col-
umn (26 ⁄ 60; GE Healthcare) was used to evaluate the
structure of the bi-specific diabodies. The column was equil-
ibrated using NaCl ⁄ P
i
, and then 5 mL of purified recombi-
nant antibodies was applied to the column at a flow rate of
2.5 mLÆmin
)1
.
Mass spectrometry
Mass spectra were measured using a REFLEX III
MALDI-TOF mass spectrometer (Bruker Daltonics Inc.,
Billerica, MA, USA) equipped with a nitrogen laser
(337 nm). Sinapic acid was applied as a matrix, and was
dissolved to saturation in water:acetonitrile (2 : 1 v ⁄ v) con-
taining 0.067% trifluoroacetic acid. Sample solutions
from each stage were mixed with the sinapic acid-saturated
solution in a 1 : 1 v ⁄ v ratio, and then 1 lL of the mixed
Fc fusion for generation of tag-free diabodies R. Asano et al.
484 FEBS Journal 277 (2010) 477–487 ª 2009 The Authors Journal compilation ª 2009 FEBS
solution was loaded onto the sample target. After co-crys-
tallization on the target, the crystals were washed twice
with 2 lL of water containing 0.1% trifluoroacetic acid to
remove residual salts. Analysis was performed in positive

and linear modes with an accelerating voltage of 27 kV,
and 200 scans were averaged. The spectra obtained were
calibrated externally using the [M + H
+
] ions from
two protein standards: cytochrome c from horse heart
(m ⁄ z 12 360.08) and bovine trypsin (m ⁄ z 23 311.53) [37].
Preparation of T-LAK cells
Peripheral blood mononuclear cells were isolated by den-
sity-gradient centrifugation of heparin-containing blood
from healthy volunteers. To induce proliferation of T-LAK
cells, peripheral blood mononuclear cells were cultured for
48 h at a density of 1 · 10
6
cells per mL in medium supple-
mented with 100 IUÆmL
)1
of recombinant human IL-2
(kindly supplied by Shionogi Pharmaceutical Co., Osaka,
Japan) in a culture flask (A ⁄ S Nunc, Roskilde, Denmark)
that had been pre-coated with OKT3 monoclonal antibody
(10 lgÆmL
)1
). Proliferated cells were then transferred to
another flask, and expanded for 2–3 weeks in a culture
medium containing 100 IUÆmL
)1
IL-2, as reported previ-
ously [38].
Flow cytometric analyses

Test cells (1 · 10
6
) were incubated on ice with 200 pmol of
BsAb for 30 min. After washing with NaCl ⁄ P
i
containing
0.1% NaN
3
, they were exposed for 30 min on ice to rabbit
anti-hEx3-Db serum (kindly supplied by Immuno-Biologi-
cal Laboratories Co. Ltd, Gunma, Japan) as the second
antibody, and fluorescein isothiocyanate-conjugated anti-
rabbit IgG (Santa Cruz Biotechnology, Santa Cruz, CA,
USA) as the third antibody. The stained cells were analyzed
by flow cytometry (FACSCalibur, Becton Dickinson, San
Jose, CA, USA) [20].
In vitro growth inhibition assay
In vitro growth inhibition of TFK-1 (human bile duct carci-
noma) was assayed using a 3-(4,5-dimethylthiazole-2-yl)-
5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium
inner salt (MTS) assay kit (CellTiter 96 aqueous non-radio-
active cell proliferation assay; Promega, Madison, WI,
USA) as reported previously [39].
Acknowledgements
This work was supported by Grants-in-Aid for Scien-
tific Research from the Ministry of Education, Science,
Sports, and Culture of Japan (to R.A. and I.K.) and
by grants from the New Energy and Industrial
Technology Development Organization of Japan.
Additional support was provided through the Program

for Promotion of Fundamental Studies in Health Sci-
ences of the National Institute of Biomedical Innova-
tion.
References
1 Cao Y & Lam L (2003) Bispecific antibody conjugates
in therapeutics. Adv Drug Deliv Rev 55, 171–197.
2 Kufer P, Lutterbuse R & Baeuerle PA (2004) A revival
of bispecific antibodies. Trends Biotechnol 22, 238–244.
3 Koelemij R, Kuppen PJ, van de Velde CJ, Fleuren GJ,
Hagenaars M & Eggermont AM (1999) Bispecific anti-
bodies in cancer therapy, from the laboratory to the
clinic. J Immunother 22, 514–524.
4 Segal DM, Weiner GJ & Weiner LM (1999) Bispecific
antibodies in cancer therapy. Curr Opin Immunol 11 ,
558–562.
5 van Spriel AB, van Ojik HH & van De Winkel JG
(2000) Immunotherapeutic perspective for bispecific
antibodies. Immunol Today 21, 391–397.
6 Bird RE, Hardman KD, Jacobson JW, Johnson S,
Kaufman BM, Lee SM, Lee T, Pope SH, Riordan GS
& Whitlow M (1988) Single-chain antigen-binding
proteins. Science 242, 423–426.
7 Huston JS, Levinson D, Mudgett-Hunter M, Tai MS,
Novotny J, Margolies MN, Ridge RJ, Bruccoleri RE,
Haber E, Crea R et al. (1988) Protein engineering of
antibody binding sites: recovery of specific activity in an
anti-digoxin single-chain Fv analogue produced in
Escherichia coli. Proc Natl Acad Sci USA 85, 5879–5883.
8 Holliger P, Prospero T & Winter G (1993) ‘Diabodies’:
small bivalent and bispecific antibody fragments. Proc

Natl Acad Sci USA 90, 6444–6448.
9 Schlereth B, Fichtner I, Lorenczewski G, Kleindienst P,
Brischwein K, da Silva A, Kufer P, Lutterbuese R,
Junghahn I, Kasimir-Bauer S et al. (2005) Eradication
of tumors from a human colon cancer cell line and
from ovarian cancer metastases in immunodeficient
mice by a single-chain Ep-CAM- ⁄ CD3-bispecific anti-
body construct. Cancer Res 65, 2882–2889.
10 Arndt MA, Krauss J, Kipriyanov SM, Pfreundschuh M
& Little M (1999) A bispecific diabody that mediates
natural killer cell cytotoxicity against xenotransplantat-
ed human Hodgkin’s tumors. Blood 94, 2562–2568.
11 Shahied LS, Tang Y, Alpaugh RK, Somer R, Green-
spon D & Weiner LM (2004) Bispecific minibodies tar-
geting HER2 ⁄ neu and CD16 exhibit improved tumor
lysis when placed in a divalent tumor antigen binding
format. J Biol Chem 279, 53907–53914.
12 Robinson MK, Doss M, Shaller C, Narayanan D,
Marks JD, Adler LP, Gonzalez Trotter DE & Adams
GP (2005) Quantitative immuno-positron emission
R. Asano et al. Fc fusion for generation of tag-free diabodies
FEBS Journal 277 (2010) 477–487 ª 2009 The Authors Journal compilation ª 2009 FEBS 485
tomography imaging of HER2-positive tumor xeno-
grafts with an iodine-124 labeled anti-HER2 diabody.
Cancer Res 65, 1471–1478.
13 Sundaresan G, Yazaki PJ, Shively JE, Finn RD,
Larson SM, Raubitschek AA, Williams LE, Chat-
ziioannou AF, Gambhir SS & Wu AM (2003)
124
I-labeled engineered anti-CEA minibodies and

diabodies allow high-contrast, antigen-specific small-
animal PET imaging of xenografts in athymic mice.
J Nucl Med 44, 1962–1969.
14 Kipriyanov SM, Moldenhauer G, Schuhmacher J, Co-
chlovius B, Von der Lieth CW, Matys ER & Little M
(1999) Bispecific tandem diabody for tumor therapy
with improved antigen binding and pharmacokinetics.
J Mol Biol 293, 41–56.
15 Le Gall F, Reusch U, Little M & Kipriyanov SM
(2004) Effect of linker sequences between the antibody
variable domains on the formation, stability and biolog-
ical activity of a bispecific tandem diabody. Protein Eng
Des Sel 17, 357–366.
16 Peipp M & Valerius T (2002) Bispecific antibodies tar-
geting cancer cells. Biochem Soc Trans 30, 507–511.
17 Helfrich W, Kroesen BJ, Roovers RC, Westers L,
Molema G, Hoogenboom HR & de Leij L (1998)
Construction and characterization of a bispecific
diabody for retargeting T cells to human carcinomas.
Int J Cancer 76, 232–239.
18 Tsumoto K, Shinoki K, Kondo H, Uchikawa M, Juji T
& Kumagai I (1998) Highly efficient recovery of func-
tional single-chain Fv fragments from inclusion bodies
overexpressed in Escherichia coli by controlled introduc-
tion of oxidizing reagent – application to a human
single-chain Fv fragment. J Immunol Methods 219,
119–129.
19 Skerra A & Schmidt TG (2000) Use of the Strep-Tag
and streptavidin for detection and purification of
recombinant proteins. Methods Enzymol 326, 271–304.

20 Asano R, Sone Y, Makabe K, Tsumoto K, Hayashi H,
Katayose Y, Unno M, Kudo T & Kumagai I (2006)
Humanization of the bispecific epidermal growth factor
receptor x CD3 diabody and its efficacy as a potential
clinical reagent. Clin Cancer Res 12, 4036–4042.
21 Asano R, Kawaguchi H, Watanabe Y, Nakanishi T,
Umetsu M, Hayashi H, Katayose Y, Unno M, Kudo T
& Kumagai I (2008) Diabody-based recombinant for-
mats of humanized IgG-like bispecific antibody with
effective retargeting of lymphocytes to tumor cells. J Im-
munother 31, 752–761.
22 Asano R, Sone Y, Ikoma K, Hayashi H, Nakanishi T,
Umetsu M, Katayose Y, Unno M, Kudo T & Kumagai
I (2008) Preferential heterodimerization of a bispecific
diabody based on a humanized anti-EGFR antibody
528. Protein Eng Des Sel 21, 597–603.
23 Presta L (2003) Antibody engineering for therapeutics.
Curr Opin Struct Biol 13, 519–525.
24 Kontermann RE (2005) Recombinant bispecific anti-
bodies for cancer therapy. Acta Pharmacol Sin 26, 1–9.
25 Marvin JS & Zhu Z (2005) Recombinant approaches to
IgG-like bispecific antibodies. Acta Pharmacol Sin 26,
649–658.
26 Carter P (2001) Bispecific human IgG by design.
J Immunol Methods 248, 7–15.
27 Asano R, Watanabe Y, Kawaguchi H, Fukazawa H,
Nakanishi T, Umetsu M, Hayashi H, Katayose Y,
Unno M, Kudo T et al. (2007) Highly effective recom-
binant format of a humanized IgG-like bispecific anti-
body for cancer immunotherapy with retargeting

of lymphocytes to tumor cells. J Biol Chem 282,
27659–27665.
28 Perisic O, Webb PA, Holliger P, Winter G & Williams
RL (1994) Crystal structure of a diabody, a bivalent
antibody fragment. Structure 2, 1217–1226.
29 Asano R, Kudo T, Nishimura Y, Makabe K, Hayashi
H, Suzuki M, Tsumoto K & Kumagai I (2002) Efficient
construction of a diabody using a refolding system:
anti-carcinoembryonic antigen recombinant antibody
fragment. J Biochem 132, 903–909.
30 Alt M, Muller R & Kontermann RE (1999) Novel tet-
ravalent and bispecific IgG-like antibody molecules
combining single-chain diabodies with the immunoglob-
ulin gamma1 Fc or CH3 region. FEBS Lett
454, 90–94.
31 Wu AM, Tan GJ, Sherman MA, Clarke P, Olafsen T,
Forman SJ & Raubitschek AA (2001) Multimerization
of a chimeric anti-CD20 single-chain Fv–Fc fusion pro-
tein is mediated through variable domain exchange.
Protein Eng 14, 1025–1033.
32 Goel A, Beresford GW, Colcher D, Pavlinkova G,
Booth BJ, Baranowska-Kortylewicz J & Batra SK
(2000) Divalent forms of CC49 single-chain antibody
constructs in Pichia pastoris: expression, purification,
and characterization. J Biochem 127, 829–836.
33 Brusselbach S, Korn T, Volkel T, Muller R & Konter-
mann RE (1999) Enzyme recruitment and tumor cell
killing in vitro by a secreted bispecific single-chain
diabody. Tumor Target 4, 115–123.
34 Kipriyanov SM, Moldenhauer G, Braunagel M, Reusch

U, Cochlovius B, Le Gall F, Kouprianova OA, Von
der Lieth CW & Little M (2003) Effect of domain order
on the activity of bacterially produced bispecific single-
chain Fv antibodies. J Mol Biol 330, 99–111.
35 Korn T, Nettelbeck DM, Volkel T, Muller R & Konter-
mann RE (2004) Recombinant bispecific antibodies for
the targeting of adenoviruses to CEA-expressing tumour
cells: a comparative analysis of bacterially expressed sin-
gle-chain diabody and tandem scFv. J Gene Med 6,
642–651.
36 Arakawa F, Kuroki M, Kuwahara M, Senba T, Ozaki
H, Matsuoka Y, Misumi Y, Kanda H & Watanabe T
(1996) Cloning and sequencing of the V
H
and V
K
genes
of an anti-CD3 monoclonal antibody, and construction
Fc fusion for generation of tag-free diabodies R. Asano et al.
486 FEBS Journal 277 (2010) 477–487 ª 2009 The Authors Journal compilation ª 2009 FEBS
of a mouse ⁄ human chimeric antibody. J Biochem 120,
657–662.
37 Umetsu M, Tsumoto K, Hara M, Ashish K, Goda S,
Adschiri T & Kumagai I (2003) How additives influence
the refolding of immunoglobulin-folded proteins in a
stepwise dialysis system. Spectroscopic evidence for
highly efficient refolding of a single-chain Fv fragment.
J Biol Chem 278, 8979–8987.
38 Hayashi H, Asano R, Tsumoto K, Katayose Y,
Suzuki M, Unno M, Kodama H, Takemura S,

Yoshida H, Makabe K et al. (2004) A highly effective
and stable bispecific diabody for cancer immunother-
apy: cure of xenografted tumors by bispecific diabody
and T-LAK cells. Cancer Immunol Immunother 53,
497–509.
39 Asano R, Kudo T, Makabe K, Tsumoto K & Kuma-
gai I (2002) Antitumor activity of interleukin-21 pre-
pared by novel refolding procedure from inclusion
bodies expressed in Escherichia coli. FEBS Lett 528,
70–76.
R. Asano et al. Fc fusion for generation of tag-free diabodies
FEBS Journal 277 (2010) 477–487 ª 2009 The Authors Journal compilation ª 2009 FEBS 487