Advances in Bioscience and Biotechnology, 2011, 2, 97-102 ABB
doi:10.4236/abb.2011.22015 Published Online April 2011 (
Published Online April 2011 in SciRes.
Isolation of mimotopes to the anti-human VEGF antibody
Bevacizumab by mRNA display using random peptide
libraries and the vaccination of a rabbit
Teruaki Kobayashi, Tatsuro Shibui*
Molecuence Corporation, Mitsubishi Chemical Group Yokohama Research Center, Yokohama, Japan.
Received 8 February 2011; revised 20 March 2011; accepted 1 April 2011.
An mRNA display syst em using synthetic DNA coding
for random 10- and 20-amino-acid peptide libraries
was employed to isolate mimotopes that could substi-
tute for the anti-human VEGF antibody Bevacizumab.
After six rounds of affinity selection, three clones that
bound to the antibody were isolated. Their ran-
dom-peptide portions were chemically synthesized,
and further characterized. All of the peptides showed
clear specific binding to the antibody. Two of them
were further conjugated with Keyhole limpet hemo-
cyanin (KLH) to immunize a rabbit. After five immu-
nizations biweekly, antibodies to the peptides were
purified with a column conjugated with the peptides.
The purified antibodies reacted specifically to the an-
tibody’s original antigen, human VEGF. mRNA dis-
plays could be useful for the isolation of mimotopes
for vaccines to substitute for therapeutic antibodies.
Keywords: Random Peptide; mRNA Display; Mimotope;
Antigen; Antibody; Vaccine
In vitro peptide or protein display systems [1,2] have
developed with advances in cell-free translation sys-
tems such as E. coli cell lysate [3,4] and wheat embryo
extract [5,6]. Compared with in vivo systems such as
phage display [7] and yeast surface [8] display, the in
vitro systems can utilize larger libraries since they do
not require the transformation of living cells. Theoreti-
cally, 1 1013 independent clones can be used for
screening. Ribosome and mRNA displays are com-
monly employed as in vitro protein display systems
[1,9,10]. Ribosome displays are simpler but less stable
than mRNA displays. The peptide portion in ribosome
displays is attached to the mRNA through a ribosome,
whereas in mRNA displays, in vitro-synthesized pep-
tides are chemically conjugated to the mRNAs via
puromycin, a peptidyl acceptor antibiotic [1,9,11]. A
covalent bond can be generated between the peptide
and mRNA through translation in vitro with synthetic
mRNA that carrying puromycin at the 3’ end. During
selection, the covalent bond stabilizes the pep-
tide-mRNA complex, although the process of making
the display molecules is more complicated.
Bevacizumab [12], a humanized monoclonal anti-
body produced by recombinant Chinese Ovary Cells
[13], recognizes and blocks vascular end o thelia l g ro w th
factor A (VEGF-A). VEGF-A is a protein that triggers a
chemical signal stimulating the growth of new blood
vessels (angiogenesis), especially in cases of cancer,
retinal proliferation of diabetes in the eye, and other
diseases. Bevacizumab is used to treat cancers that are
metastatic. However, as it is produced by mammalian
cells as a recombinant protein, it is very expensive.
Cost-effective substitutes for th e an tibod y ar e need ed.
In many monoclonal antibodies, phage display sys-
tems with random peptide libraries (RPL) have been
used for epitope mapping [7]. In cases of antibodies
that recognize conformational epitopes, the systems
have often provided binding peptides with no homol-
ogy to the amino acid sequences of the original anti-
gens [14]. Such peptides are called mimotopes [15].
Several mimotopes have been shown to induce the
production of antibodies against the original antigen
[15,16]. These new antibodies seemed to recognize a
similar or the same epitope as the original antibody
[16]. Thus mimotopes could potentially be used as sub-
stitutes for expensive therapeutic antibodies [17].
Compared to phage disp lays, mRNA displays provide a
larger library, and would increase the chance of identi-
fying mimotopes.
In this paper, we describe the successful application
T. Kobay ashi et al. / Advances in Bioscience and Biotechnology 2 (2011) 97-102
Copyright © 2011 SciRes. ABB
of an mRNA-display system with random linear 10-and
20-amino-acid peptide libraries to the isolation of mi-
motope peptides to a therapeutic antibody, Bavacizu-
mab, and their possible use for a vaccine.
2.1. Antibodies
Abciximab (ReoPro), Cetuximab (Erbitux), Adalimumab
(Humira) and Omalizumab (Xolair) were purchased
from Eli Lilly (Switzerland), Merck (Switzerland), Ab-
bott (Switzerland) and Novartis (Switzerland), respec-
tively. Tocilizumab (Actermura) and Bevacizumab
(Avastin) were obtained from Roche (Switzerland). Hu-
man IgGs was obtained from Sigma (USA).
2.2. Construction of mRNA Display Libraries
DNA fragments for the linear random 10- and 20- amino
acid libraries, RL10aPL and RL20aPL, were synthesized
by BEX (Tokyo, Japan). The mRNA display libraries
were constructed separately using the DNA templates as
described previously [14]. The sequences of the peptide
portion of mRNA display molecules are as follows.
in which X is a random amino acid.
2.3. Affinity Selection
To prevent non-specific bindings to the antibody, the
mRNA display molecules (in 200 µl of a washing buffer
(50 mM Tris/HCl pH 7.5, 150 mM NaCl, and 0.1%
Tween 20) containing a final concentration of 0.05%
bovine serum albumin and 25 µg/ml tRNA) were incu-
bated with 10 µl of human normal IgG (Sigma)-bound
Protein G-Sepharose beads (GE Healthcare) for 30 min-
utes at room temperature with rotation. The supernatant,
the non-bound fraction, was transferred to 10 µl of the
Bevacizumab -bound Protein G-Sepharose beads and
rotated for 1 hour at room temperature. The beads were
washed three times with 0.1 ml of the washing buffer
(50 mM Tris/HCl pH 7.5, 150 mM NaCl, and 0.1%
Tween 20), and the bound molecules were eluted two
times with 40 µl of 0.1 M Glycin-HCl pH2.5, and im-
mediately neutralized with 8 µl of 1 M Tris-HCl pH8.0.
The eluted molecules were re-amplified by PCR for the
next round of selection. Subsequent rounds were per-
formed in a similar manner.
2.4. Cloning and Sequencing
After six selection cycles, PCR-amplified fragments
coding for bound peptides were cloned into a vector
(pGEM T Easy Vector, Promega), and sequenced. The
sequence of each clone was sorted using the CLUS-
TALW Program.
2.5. Synthetic Peptides
Random peptide-portions from isolated clones were
chemically synthesized, and purified with HPLC by To-
ray Research Center (Kanagawa, Ja pan)
2.6. Immunization of Peptides
The peptides of two isolated clones, M074_F02 and
M074_D12 (see Table 1), were conjugated with keyhole
limpet hemocyanin (KLH), and injected into a rabbit in
every two weeks (five times in total) by Scrum (Tokyo,
2.7. Purification of Antibodies against Peptides
C-terminally biotinylated M074_F02 and M074_D12
were synthesized by Scrum. First, 50 µl of NeutrAvidin
Agarose (Thermo Scientific) was placed in a spin col-
umn and washed three times with 200 µl of PBST (1
PBS and 0.05% Tween 20). Then, 100 µl of 10 µM bio-
tinylated M074_F02 and M074_D12 in PBS was added
and the column was incubated at 4C overnight to make
peptide-immobilized beads. The beads were washed
three times with 200 µl of PBST. Next, 100 µl of serum
from the immunized rabbit was diluted in 400 µl of
PBST. The diluted serum was mixed with the M074_F02
and M074_D12-immobilized beads for 30 minutes at
room temperature. The beads were washed three times
with PBST. The bound antibodies were eluted with 100
µl of 10 mM Glycin-HCl (pH 2.7), and immediately
neutralized by adding 11 µl of 1 M Tris-HCl (pH 8.9).
These fractions were used to monitor the antibodies
against the original antigen, human VEGF.
Total serum was collected one week after the final
immunization, and its IgG fraction was purified with a
Protein A column by Scrum. Then the aliquots of ap-
proximately 1 ml were further purified separately with
the M074_F02 and M074_D12-immobilized beads as
described above. Th e bound fractio n of each p eptide was
used to analyze reactivity to the original antigen by
Table 1. A list of clones isolated.
Clone # Amino acid seque nce Appearance
rate (%)
T. Kobay ashi et al. / Advances in Bioscience and Biotechnology 2 (2011) 97-102
Copyright © 2011 SciRes. ABB
3.1. Amino Acid Sequence of Bound Molecules
Amino acid sequences of the random peptide portion of
isolated molecules are listed in Table 1. Of the 92 clones
sequenced, the most dominant was M074_D12. Al-
though M075_A10 was isolated from the 20-aa linear
peptide library, it contains two cysteine residues. These
clones di d not show any a pparent h omology in sequence.
3.2. Binding to Bevacizumab
To clarify whether the polypeptide portions of the
mRNA-display molecules bound to the antibody, the
polypeptides were translated in vitro and labeled at the
C-terminus with Cy3 instead of mRNA using puromycin.
The binding of the labeled polypeptides to Bevacizumab
was then examined by pull-down assay as described in
the legend to Figure 1. All of the clones showed specific
binding to the antibody. Unexpectedly, the most domi-
nant clone, M074_D12, did not show the highest inten-
sity or specificity of binding.
3.3. Binding Properties of the Peptides
Peptides from random portions in isolated clones were
Figure 1. Pull-down assay of isolated clones
with Bevacizumab. A Clone M074_F02, B
Clone M074_D12, and C Clone M075_A10.
C-terminally Cy3-labeled polypeptides were
mixed with Bevacizumab-bound Protein G
beads and normal IgG-bound beads as a nega-
tive control. The beads were spun down by
brief centrifugation and washed three times
with binding buffer. The bound polypeptides
were eluted with 0.1M Glycin-HCl (pH 2.5)
and neutralized by 1M Tris-buffer (pH 8.0)
(See details in Materials and Methods). La-
beled polypeptides were separated on a Tricin
polyacrylamide gel by electrophoresis and
detected through Cy3 fluorescence. Lane 1:
ten percent of the labeled polypeptide was
loaded on the gel. Lane 2: the labeled poly-
peptide eluted from the Bevacizumab-bound
beads. Lane 3: negative control. The labeled
polypeptide eluted from the normal human
IgG-bound beads.
synthesized chemically and purified with HPLC. Figure
2 shows the inhibition of antigen binding to the antibody.
All of them showed similar patterns. The results did not
indicate higher affinity of M074_D12. The specificity of
the peptides was examined further using other therapeu-
tic antibodies (Figure 3). All of them bound specifically
to Bevacizumab. M074_D12 showed slightly higher
binding than the others.
3.4. Influence of ReducingConditions on Binding
Next, we examined whether the two cysteine residues in
M075_A10 formed a di-sulfide bond and were essential
for binding. Under reducing conditions, M075_A10 lost
binding activity (Figure 4). It seemed to form a circular
structure, however, this remains to be confirmed. To
avoid this structural complication, we chose M074_F02
and M074_D12 for immunization.
3.5. Immunization of the Rabbit and
Characteristics of its Serum
To minimize number of animals to be sacrificed,
M074_F02-KLH and M074_D12-KLH conjugates were
mixed together in adjuvants, and injected into a rabbit
five times at two-week intervals.
Figure 2. Inhibition of antigen-antibody binding by syn-
thetic peptides of isolated clones. : M074_F02, :
F074_D12, : M075_A10, and : control. The random
peptide portion of each clone was chemically synthesized by
Toray Research Center (Japan). Assays were conducted es-
sentially as described previously [10]. Aliquots (100 µl) of
Bevacizumab solution (3 µg/ml in PBS) were added to wells
of ELISA plates (COSTAR 9018), and incubated at 4C
overnight. The plates were washed 3 times with PBS. The
wells were blocked with 1% BSA in PBS at 37C for 2 hours.
The plates were washed three times with PBST. Peptides of
the indicated concentration (100 µl) were incubated with
human VEGF solution (70 ng/ml in PBS, PeproTech) in the
Bevacizumab-coated wells at room temperature for 1 hour.
Bound human VEGF was detected with a rabbit anti-human
VEGF, and then bound anti-human VEGF was detected with
anti-rabbit IgG conjugated with HRP (Sigma). Inhibition of
binging (%) was calculated from absorbance at 490 nm us-
ing 1.2-Phenylenediamine in Citrate buffer containing H2O2.
T. Kobay ashi et al. / Advances in Bioscience and Biotechnology 2 (2011) 97-102
Copyright © 2011 SciRes. ABB
Figure 3. Specificity of the peptides; M074_F02;
M074_ D12; M075_A10; Negative control. Binding to
therapeutic antibodies was assayed by ELISA. All antibodies
were purchased (see Materials and Methods). Aliquots (100 µl)
of each antibody solution (3 µg/ml in PBS) were added to
wells of ELISA plates (COSTAR 9018), and incubated at 4C
overnight. The plates were washed three times with PBS. The
wells were blocked with 1% BSA in PBS at 37C for 2 hours.
The plates were washed 3 times with PBST. Aliquots (100 µl)
of C-terminally biotin-labeled peptides (10 µM in PBS) were
put into the wells, and incubated for 1 hour at room tempera-
ture. The plates were again washed three times with PBST.
Next, 100 µl of 200-fold diluted streptavidin conjugated with
HRP (RD systems) was added to the wells and incubation
was continued for 1 hour at room temperature. The plates were
again washed three times with PBST. Bound peptides were
detected by measuring absorbance at 490 nm using
1.2-Phenylenediamine in Citrate buffer containing H2O2.
Figure 4. Effect of a reducing condition on binding to Bevaci-
zumab. (a) M074_D12. (b) M075_A10. M074_D12 and
M075_A10 were incubated with () or without () 100
mM dithiothreitol (DTT) in PBS for 3 hours at room tempera-
ture prior the assay. Indicated amounts of peptides were incu-
bated with Bevacizumab-coated plates (See Figure 2 legend).
Assays were performed essentially as described in Figure 3.
Figure 5. Serum titers against each peptide. Serum
titers against each peptide in a rabbit immunized
with peptide-KLH conjugates were assayed with
ELISA. Sera were collected prior to immunization
( ) and one week after the 3rd ( ), 4th ( ) and
final ( ) immunization. Aliquots (100 µl) of pep-
tide solution (1 µM in PBS) were added to wells of
a NeutrAvidin coated plate (Thermo Scientific), and
incubated at 4C overnight. The plate was washed 3
times with PBS. The wells were blocked with 1%
BSA in PBS at 37C for 2 hours. The plate was
washed 3 times with PBST. 100 µl of rabbit serums
diluted 103-fold in PBS was put into each well, and
incubation was continued for 1 hour at room tem-
perature. The plate was again washed 3 times with
PBST. Next, 100 µl of 104-fold diluted anti-rabbit
IgG conjugated with HRP (Sigma) was added to
each well and incubated for 1 hour at room tem-
perature. The plate was washed 3 more times with
PBST. Bound rabbit IgGs were detected by meas-
uring absorbance at 490 nm using 1.2-Phenylene-
diamine in Citrate buffer containing H2O2.
Serum titers were monitored one week after the 3rd,
4th and 5th (final) immunizations (Figure 5). Titers
against immunized peptides were specifically raised al-
most the same manner.
The serum samples from the rabbit one week after the
3rd and 4th immunizations were further affinity-purified
on a column packed with both M074_F02 and M074_
D12-conjugated beads to monitor their reactivity to hu-
man VEGF and other antigens (Figure 6). Specific an-
tibodies against human VEGF were successfully raised
in the rabbit.
Blood was collected one week after the final immuni-
zation. Titers against each peptide in the final serum
sample are shown in Figure 5. The final serum sample
was firstly purified with a Protein A column, and further
affinity purified on columns separately packed with
beads conjugated with M074_F02 and M074_ D12. The
T. Kobay ashi et al. / Advances in Bioscience and Biotechnology 2 (2011) 97-102
Copyright © 2011 SciRes. ABB
Figure 6. Monitoring of the binding of antibodies to
human VEGF during immunization; Pre-immune
serum; One week after the 3rd immunization;
One week after the 4th immunization. Serum from a
rabbit immunized with the peptide-KLH conjugates was
affinity-purified with a peptide (M074_D12 and M075_
A10)-conjugated column as described in Materials and
Methods. Purified fractions were diluted 10-fold, and
assayed for reactivity to human VEGF with ELISA as
described in Figure 5 except the plate was coated with
the proteins indicated instead of peptides
Figure 7. Binding activity of antibodies purified sepa-
rately with M074_D12 and M075_A10-conjugated col-
umns. (a) Antibodies purified with the M074_D12-con-
jugated column. (b) Antibodies purified with the M075_
A10-conjugated column; 10-fold dilution of the purified
fraction; 50-fold dilution of the purified fraction. Assays
were done essentially as described in Figure 6.
purified fractions were examined for reactivity to human
VEGF and other antigens (Figure 7). They both bound
specifically to human VEGF. The antibodies purified
with M074_ D12 showed more reactivity to the antigen
than those purified with M074_F02 (Figure 7).
M074_D12 seems to be more similar to Bevacizumab’s
epitope than the other mimotope.
Three mimotopes against Bvacizumab were isolated
display system with random peptide libraries. Chemi-
cally synthesized peptides of the mimotopes reacted si-
milarly to the antibody in ELISAs. One of them
(M074_D12) generated antibodies to Bevacizumab’s
original antigen, human VEGF, in a rabbit. The results
indicate that mRNA displays can be used to isolate pep-
tide mimotopes to sub s titute for therapeutic antibodies.
[1] Lipovsek, D. and Pluckthun, A. (2004) In-vitro protein
evolution by ribosome display and mRNA display.
Journal of Immunological Methods, 290, 51-67.
[2] Shibui, T., Kobayashi, T. and Shiratori, M. (2008)
Isolation of cross-reacting antigen candidates by mRNA-
display using a mixed cDNA library. Biotechnology
Letters, 30, 2037-2043. doi:10.1007/s10529-008-9803-5
[3] Shimizu, Y., et al. (2001) Cell-free translation recons-
tituted with purified components. Nature Biotechnology,
19, 751-755. doi:10.1038/90802
[4] Kim, D.M. et al. (1996) A highly efficient cell-free
protein synthesis system from Escherichia coli. European
Journal of Biochemistry, 239, 881-886.
[5] Koga, H., Misawa, S. and Shibui, T. (2009) A wheat
embryo cell-free protein synthesis system not requiring
an exogenous supply of GTP. Biotechnology Progress,
25, 1322-1327. doi:10.1002/btpr.230
[6] Madin, K., et al. (2000) A highly efficient and robust
cell-free protein synthesis system prepared from wheat
embryos: plants apparently contain a suicide system
directed at ribosomes. Proceedings of the National
Academy of Sciences of the United States of America, 97,
559-564. doi:10.1073/pnas.97.2.559
[7] Smith, G.P. and Scott, J.K. (1993) Libraries of peptides
and proteins displayed on filamentous phage. Methods in
Enzymology, 217, 228-257.
[8] Furukawa, H., et al. (2006) Development of novel yeast
cell surface display system for homo-oligomeric protein
by coexpression of native and anchored subunits.
Biotechnology Progress, 22, 994-997.
[9] Nemoto, N., et al. (1997) In vitro virus: Bonding of
mRNA bearing puromycin at the 3'-terminal end to the
C-terminal end of its encoded protein on the ribosome in
T. Kobay ashi et al. / Advances in Bioscience and Biotechnology 2 (2011) 97-102
Copyright © 2011 SciRes. ABB
vitro. FEBS Letters, 414, 405-408.
[10] Shibui, T., et al. (2009) In vitro selection of scFv and its
production: An application of mRNA display and wheat
embryo cell-free and E. coli cell production system.
Applied Microbiology and Biotechnology, 84, 725-732.
[11] Miyamoto-Sato, E., et al. (2000) Specific bonding of
puromycin to full-length protein at the C-terminus.
Nucleic Aci d s Research, 28, 1176-1182.
[12] Chen, H.X., Gore-Langton, R.E. and Cheson, B.D. (2001)
Clinical trials referral resource: Current clinical trials of
the anti-VEGF monoclonal antibody bevacizumab. On-
cology (Williston Park), 15, 1023-1026.
[13] Kito, M. et al. (2002) Construction of engineered CHO
strains for high-level production of recombinant proteins.
Applied Microbiology and Biotechnology, 60, 442-448.
[14] Shiratori, M., Kobayashi, T. and Shibui, T. (2009) Iden-
tification of amino acids essential for antibody binding
by mRNA-display using a random peptide library: an
anti-human tumor protein p53 antibody as a model.
Molecular Biotechnology, 41, 99-105.
[15] Riemer, A.B., et al. (2004) Generation of Peptide mimics
of the epitope recognized by trastuzumab on the onco-
genic protein Her-2/neu. Journal of Immunology, 173,
[16] Saphire, E.O., et al. (2007) Structure of a high-affinity
“mimotope” peptide bound to HIV-1-neutralizing anti-
body b12 explains its inability to elicit gp120 cross-
reactive antibodies. Journal of Molecular Biology, 369,
696-709. doi:10.1016/j.jmb.2007.01.060
[17] Aires da Silva, F., Corte-Real, S. and Goncalves, J. (2008)
Recombinant antibodies as therapeutic agents: Pathways
for modeling new biodrugs. BioDrugs, 22, 301-314.