Dept. of Biomedical Sciences, Graduate School, Ajou University, 206 World cup-ro, Yeongtong-gu, Suwon 443-749, South KoreaDept. of Microbiology, Ajou University School of Medicine, 206 World cup-ro, Yeongtong-gu, Suwon 443-749, South KoreaAnimal Biotechnology Division, National Institute of Animal Science, Rural Development Administration, Suwon 441-706, South KoreaDepartment of Biochemistry and Molecular Biology, Ajou University School of Medicine, 206 World cup-ro, Yeongtong-gu, Suwon 443-749, South KoreaDivision of Applied Life Science (BK21 Plus Program), Systems and Synthetic Agrobiotech Center (SSAC), Plant Molecular Biology and Biotechnologyesearch Center (PMBBRC), Research Institute of Natural Science (RINS), Gyeongsang National University (GNU), 501 Jinju-daero, Gazha-dong, Jinju60-701, South KoreaDept. of Microbiology, Changwon National University, 9 Sarim, Changwon 641-773, South Korea
r t i c l e i n f o
rticle history:eceived 18 August 2014eceived in revised form 7 October 2014ccepted 11 October 2014vailable online 28 October 2014
eywords:DR grafting
a b s t r a c t
In contrast to a number of studies on the humanization of non-human antibodies, the reshaping of a non-human antibody into a chicken antibody has never been attempted. Therefore, nothing is known aboutthe animal species-dependent compatibility of the framework regions (FRs) that sustain the appropri-ate conformation of the complementarity-determining regions (CDRs). In this study, we attempted thereshaping of the variable domains of the mouse catalytic anti-nucleic acid antibody 3D8 (m3D8) intothe FRs of a chicken antibody (“chickenization”) by CDR grafting, which is a common method for thehumanization of antibodies. CDRs of the acceptor chicken antibody that showed a high homology to the
FRs of m3D8 were replaced with those of m3D8, resulting in the chickenized antibody (ck3D8). ck3D8retained the biochemical properties (DNA binding, DNA hydrolysis, and cellular internalizing activities)and three-dimensional structure of m3D8 and showed reduced immunogenicity in chickens. Our studydemonstrates that CDR grafting can be applied to the chickenization of a mouse antibody, probably dueto the interspecies compatibility of the FRs.
. Introduction
The reshaping, or “humanization,” of the variable regionsf an antibody is required for reducing the immunogenic-ty of monoclonal antibodies derived from non-human sources
hat occurs when these antibodies are repeatedly adminis-ered to humans during antibody-based therapy for humaniseases (Kim et al., 2005). In most antibodies conventionally
Abbreviations: CDR, complementarity-determining region; scFv, single chainariable fragment; FR, framework region; VH , variable region of heavy chain; VL ,ariable regions of light chain; FRET, fluorescence resonance energy transfer; ELISA,nzyme-linked immunosorbent assay; RMSD, root mean square deviations; RMSF,oot mean square fluctuations.∗ Corresponding authors at: Dept. of Microbiology, Ajou University School ofedicine. Tel.: +82 31 219 5074; fax: +82 31 219 5079 or Dept. of Microbiology,
humanized from non-human sources (most commonly mouse), sixcomplementarity-determining regions (CDRs) of the heavy chainvariable region (VH) and light chain variable region (VL) are graftedinto the framework regions (FRs) of a human antibody scaffold byantibody engineering (Ahmadzadeh et al., 2014). Along with therapid development of humanization methods for variable regions(Safdari et al., 2013), a large number of non-human antibodies withspecificity for antigens of therapeutic interest have been human-ized from antibodies of mouse, rat, rabbit, and camelid origin(Finlay et al., 2009; Gorman and Clark, 1990; Vincke et al., 2009;Yu et al., 2010). So far, all of these approaches have been appliedto the humanization of antibodies for therapeutic uses, and thereshaping of a non-human antibody into other non-human anti-bodies has been limited to camelization (Davies and Riechmann,1995) and caninization [offered by the company Creative Biolabs
(Shirley, NY)]. There have not been any reports of the chickenizationof human/non-human antibodies.
We previously humanized the mouse antibody 3D8 (m3D8)by CDR grafting (Kim et al., 2009). This antibody, a catalytic
nti-nucleic acid antibody that contains a single chain variableragment (scFv) and can induce DNA/RNA hydrolysis and cellularnternalization (Jang et al., 2009; Kim et al., 2006), was humanizedo determine whether its DNA-binding and hydrolyzing activitiesre derived from an intrinsic property of the catalytic antibodyDRs or whether the FRs are involved. In terms of the antibodyngineering, studies on the reshaping of an antibody from onepecies into another may increase our understanding of the species-ependent compatibility of FRs, which sustain the appropriateonformation of the CDRs, and will reveal whether the strategiessed for antibody humanization can be easily applied to the reshap-
ng of antibodies derived from non-humans.In the present study, we attempted the reshaping of the m3D8
ntibody into a chicken antibody (“chickenization”) by CDR graftingased on FR homology. The 3-dimensional structure and structuraltability by computational analysis, the DNA-binding/hydrolyzingctivities, the cellular internalization, and the antibody responsesnduced by administration to chickens of the parental m3D8 and thehickenized 3D8 (ck3D8) were compared. The ck3D8 was showno retain the structural/biochemical properties of m3D8 and toesult in reduced immunogenicity in chickens, as expected. Thistudy suggests that it may be possible to reshape antibody variableomains between other non-human species using various human-
zation strategies.
. Material and methods
.1. Preparation of scFv proteins
Approximately 800 bp of the ck3D8 scFv gene, in which the VH
nd VL sequences are connected by a (Gly4/Ser1)3 linker, was syn-hesized by Bioneer Co. (Korea). The ck3D8 gene with a stop codont the 3′ end was subcloned into the pIg20 expression vector (Kimt al., 2006) using the XmaI and NcoI restriction sites. ck3D8 and3D8 scFv proteins with both a (His)5 tag and a Protein A tag at
he C-terminus, or scFv proteins lacking the Protein A tag, werexpressed in Escherichia coli BL21DE3 (pLysE) and prepared througholubilization of the inclusion body and refolding. Briefly, proteinsere expressed in BL21DE3 (pLysE) by inducing a 1 L culture in
B broth plus 100 �g/ml ampicillin and 25 �g/ml chloramphenicolith 1 mM isopropyl �-d-1-thiogalactopyranoside (IPTG) for 6 h at
7 ◦C once it had reached an A600 (the absorbance at 600 nm) of.0. After preparation of inclusion bodies from the cells (Zhao et al.,010), the denatured scFv protein was purified from the solubilized
nclusion bodies by chromatography on a HisPur Cobalt columnThermo Scientific) according to the manufacturer’s instructions.ubsequently, the denatured/purified protein was renatured bylowly diluting the purified protein 80-fold in 8 M urea/150 mMmidazole at 4 ◦C over a period of 30 h. Finally, the urea was com-letely removed from the protein by dialysis in cold PBS (pH 7.4).nless otherwise specified, scFv proteins with both a His tag and arotein A tag at the C-terminus were used.
.2. Enzyme-linked immunosorbent assay (ELISA)
To assay DNA-binding activity, scFv was incubated on wells of 96-well ELISA plate that had been coated with pUC57 plasmidNA (2 �g/ml). Bound scFv proteins were detected using rabbit IgG
Sigma) followed by alkaline phosphatase (AP)-conjugated anti-abbit IgG (Pierce), as described previously (Kim et al., 2006). Toetermine the immunogenicity of scFv in chickens, scFv lacking
he C-terminal Protein A tag was used as the antigen to excludemmune responses against the tag. In an indirect ELISA, wells coated
ith m3D8 or ck3D8 (2 �g/ml) were incubated with either seri-lly diluted pre-immune sera or immune sera from the immunized
logy 63 (2015) 513–520
chickens. Bound chicken IgY was detected using AP-conjugatedanti-chicken IgY (Abcam. 1/500 dilution). In a competitive ELISA,wells coated with m3D8 (2 �g/ml) were incubated with a mixtureof sera from mice immunized with m3D8 scFv diluted at 1:1,000and various concentrations of competitors (m3D8 or ck3D8 pro-tein), followed by incubation with AP-conjugated anti-chicken IgY.
2.3. Fluorescence resonance energy transfer (FRET)-based DNAcleavage assay
A DNA substrate comprised of 21 nucleotides labeled with6-carboxyfluorescein (FAM) at the 5′-terminus and a black holequencher (BHQ) at the 3′-terminus (5′FAM-CGATGAGTGCCATGGATATAC-BHQ 3′) was generated by M-Biotech Inc. scFv (1 �M) orDNase I (1 U) were either pre-incubated or not with heparin (a com-petitor molecule; 10 �g/ml) for 10 min at room temperature andthen loaded onto the wells of a 96-well black plate containing DNAsubstrate (250 nM). Immediately after the addition of the DNA sub-strate, the fluorescence intensity was read in real time over a periodof 6 h at 37 ◦C in a fluorescence detector (molecular devices). Eachreaction was carried in a final volume of 100 �l with TBSM buffer[100 mM Tris–HCl (pH 7.4), 150 mM NaCl, 2 mM MgCl2].
2.4. Flow cytometry
HeLa cells seeded in 6-well plates at a density of 5 × 105
cells/well were incubated with 5 �M scFv in serum-free mediumfor 6 h at 37 ◦C. The cells were washed three times with ice-coldPBS and fixed with 4% paraformaldehyde in PBS for 10 min at 4 ◦C.After washing with PBS, cell membranes were permeabilized withP buffer (1% BSA, 0.1% saponin, and 0.1% sodium azide in PBS) andincubated with rabbit IgG (10 �g/ml) followed by Alexa Fluor 647-conjugated anti-rabbit IgG (Life Technologies) at a dilution of 1:500.P buffer was used for the dilution of all antibodies. Each incuba-tion step was performed for 1 h at 4 ◦C, followed by washing withice-cold PBS. Finally, the cells were suspended in 4% paraformalde-hyde and analyzed using a FACSCanto II flow cytometer (BectonDickinson).
2.5. Confocal microscopy
HeLa cells were seeded on glass coverslips in 24-well platesat a density of 4 × 104 cells/well and incubated with 5 �M scFvin serum-free medium for 6 h at 37 ◦C. Cells were washed, fixed,permeabilized, and labeled with antibodies as described for flowcytometry experiments. The cell nuclei were stained with Hoechst33342 (Vector Laboratories) for 30 min at room temperature. Then,cells on the coverslips were mounted with Vectashield mountingmedium (Vector Laboratories). Images were obtained using a laserscanning confocal fluorescence microscope (model LSM710, CarlZeiss).
2.6. Computational analysis of ck3D8
The 3-dimensional structure of ck3D8 cFv was generated usingthe “Build homology models” algorithm, available in Discovery Stu-dio (DS) 3.5 software (Accelrys). The crystal structure of m3D8 scFv[Protein Databank (PDB) ID: 2GKI, 2.88 A] was used as the template.The molecular dynamics (MD) simulations were performed usingthe GROMACS program (version 4.5.3) (Berendsen et al., 1995;Pronk et al., 2013) with an assisted model building with energyrefinement (AMBER) 03 force field (Duan et al., 2003). All systems
were neutralized by the addition of Cl− counterions, and the parti-cle mesh Ewald (PME) method was used to calculate long-rangeelectrostatic interactions (Darden et al., 1993). Each system hasapproximately 39,200 atoms, including approximately 3600 atoms
muno
fsrca3bbtpancu
2
at2wtwtFtwcwui
3
3g
hb2wazfc2cSofapddtCffoobip
J. Roh et al. / Molecular Im
or protein alone and 11,900 water molecules. The systems wereubjected to a steepest descent energy minimization process toemove possible bad contacts. In the system equilibration pro-ess, the heavy atoms were restrained and the solvent moleculesnd counterions were allowed to move under NPT conditions of00 K and 1 bar. The bonds between heavy atoms and hydrogenond atoms were constrained by the Library of Integrated network-ased cellular signatures (LINCS) algorithm (Hess et al., 1997). Theime step was set to 2 fs and the coordinate data were saved everyicosecond (ps). The representative structures for m3D8 (4090 ps)nd ck3D8 (4210 ps) were selected from snapshots taken in the lastanosecond (ns). The refined structures were utilized for structuralomparison. All analyses of the MD simulations were performedsing the GROMACS package and DS3.5 software.
.7. Chicken immunization
All animal experiments were performed using a protocolpproved by the Institutional Animal Care and Use Committee ofhe National Institute of Animal Science of Korea (permission No.014-066). Two groups of two female 28-week-old single-combhite Leghorn (SCWL) chickens were immunized with 200 �l of
he purified m3D8 or ck3D8 scFv protein. The proteins were mixedith complete Freund’s Adjuvant (CFA, Sigma) for primary injec-
ion, followed by boosters at 2, 4, and 5 weeks with incompletereund’s Adjuvant (IFA, Sigma). For each injection, 150 �g of pro-ein was diluted in sterile PBS to a volume of 100 �l and mixedith an equal volume of adjuvant. Chickens were inoculated sub-
utaneously on the leg and intramuscularly at the back of the neck,ith 100 �l per injection site. Sera were collected from individ-al chickens prior to the first immunization and after the fourth
mmunization.
. Results
.1. The chickenized 3D8 antibody, ck3D8, was generated by CDRrafting based on FR homology
To select suitable chicken variable domains that had FRs withigh homology to the FRs of m3D8, FR sequences of chicken anti-odies from the literature (Abi-Ghanem et al., 2008; Foord et al.,007; Lee et al., 2007; Sapats et al., 2006; Yamanaka et al., 1996)ere compared with that of m3D8. Variable regions (VH and VL) of
chicken antibody, S8L, which is specific for Eimeria tenella sporo-oites (Abi-Ghanem et al., 2008), was chosen as the acceptor FRor CDR grafting. Finally, a chickenized version of m3D8, calledk3D8, was designed using a CDR grafting method (Safdari et al.,013) in which six CDRs from m3D8 were grafted into the S8Lhicken FR template. Along with CDR grafting, nine FR residues of8L were replaced with residues that are commonly observed inther chicken antibodies or in the m3D8 FR for favorable CDR con-ormation. Eight FR residues of S8L were replaced with residues thatre most commonly observed in other chicken antibodies at thoseositions (Thr → Ala, Ile → Phe, Phe → Trp and Ile → Val in the VH
omain; Glu → Lys, Ser → Ala, Asp → Glu, and Phe → Tyr in the VL
omain), and one FR residue in the VH-FR3 of S8L was changed tohe corresponding residue of m3D8 FR3 (Lys → Arg) (Fig. 1). VH-DR2 and VH-CDR3 were also modified in generating ck3D8, as
ollows. With respect to VH-CDR2, ten amino acids (YINPYNDGTK)rom m3D8 VH-CDR2 were transferred to the S8L scaffold, insteadf grafting the entire m3D8 VH-CDR2 sequence, which is composed
f 17 amino acids (YINPYNDGTKYNEKFKG). This decision was madeecause grafting the entire VH-CDR2 region caused a 50% reduction
n DNA-binding and hydrolyzing compared to m3D8, with a com-lete loss of the cellular internalizing ability (data not shown). It
logy 63 (2015) 513–520 515
was reported that grafting six residues (NPYNDG) of VH-CDR2 ontoa human antibody scaffold resulted in the generation of a human-ized 3D8 that retains the activity of m3D8 (Kim et al., 2009). Withrespect to VH-CDR3, the residues DA, which are found in VH-CDR3in most chicken antibodies, were retained in ck3D8 for favorableCDR conformation because another version of ck3D8 lacking the DAresidues showed an approximately 20-fold lower protein expres-sion in bacterial culture compared to m3D8 (data not shown).
3.2. ck3D8 retains the biochemical properties of m3D8
ck3D8 scFv protein was purified to homogeneity with a molecu-lar weight of 34 kDa (Fig. 2A). Approximately 5 mg of ck3D8 proteinwas obtained from 100 ml of the cultured E. coli, comparable tom3D8. Using the purified scFv protein, the DNA binding, DNAhydrolysis, and cellular internalization activities were assayed. AnELISA using a plasmid DNA antigen showed that ck3D8 retains theDNA-binding activity of m3D8. A negative control, the scFv of theHW6 antibody, which recognizes TRAIL receptor 2, did not showDNA-binding activity (Fig. 2B). Next, the DNA hydrolyzing activityof ck3D8 was analyzed using a FRET-based DNA cleavage assay inwhich the hydrolysis of a DNA substrate double-labeled with a fluo-rophore at the 5′-terminus and its quencher at the 3′-terminus wasmeasured according to the increase in fluorescence intensity. Theassay was performed in the presence or absence of heparin. Heparincompetes with the m3D8 for DNA-binding sites and inhibits its DNAhydrolyzing activity (Lee et al., 2013); if the DNA hydrolyzing activ-ities of m3D8 and ck3D8 are inhibited by heparin, false-positivescaused by contamination during the protein preparation can beruled out. The DNA hydrolyzing activity of ck3D8 was compara-ble to that of m3D8 and was completely blocked by the solublecompetitor, heparin (Fig. 2C). By contrast, the DNA hydrolyzingactivity of DNase I (a negative control that does not bind to heparin)was not affected by heparin, as expected. Another negative control,HW6 scFv, did not show DNA hydrolyzing activity regardless of thepresence of heparin, as expected (Fig. 2C).
The cellular internalization of ck3D8 was analyzed by flowcytometry (Fig. 2D) and confocal microscopy (Fig. 2E) in HeLa cellstreated with antibody. The ck3D8 was internalized by the HeLa cellsto the same extent as the m3D8, whereas internalization of thenegative control, HW6 scFv, was not detected.
3.3. The structures of ck3D8 and m3D8 are similar
A molecular modeling study was carried out to compare thestructures of m3D8 scFv and ck3D8 scFv by homology modeling andan MD simulation (Fig. 3). The 3-dimensional structure of ck3D8was built using the homology modeling method, in which the crys-tal structure of m3D8 (PDB ID: 2GKI) (Kim et al., 2006) was used asa template. The built ck3D8 scFv structure was fairly well super-imposed onto the crystal structure of m3D8 scFv (Fig. 3A). The5 ns MD simulations were carried out to (1) refine the generated3-dimensional structure, and (2) compare structural differencesbetween ck3D8 and m3D8. To investigate structural stability, thebackbone root mean square deviations (RMSDs) and root meansquare fluctuations (RMSFs) were calculated for both structures.The RMSD values for each protein were stabilized after 1000 ps andthe values were maintained within approximately 0.25 nm (25 A)during the 5 ns MD simulation time. The average RMSD value ofm3D8 was only slightly lower than that of ck3D8 (0.14 nm and0.22 nm, respectively) (Fig. 3B), indicating that there is no signifi-cant structural difference between the structures. The RMSFs of the
VL and VH chains for the proteins were measured during the last 2 nsto determine their flexibility (Fig. 3C and D). In a comparative RMSFanalysis, although slightly higher flexibilities were observed in theCDR-H2, CDR-L1, and FR-L3 regions of ck3D8 compared to those
516 J. Roh et al. / Molecular Immunology 63 (2015) 513–520
Fig. 1. Design and production of ck3D8. Amino acid sequences for the m3D8, ck3D8, and S8L antibodies. CDR residues in the VH and VL regions are underlined. The absenceof corresponding residues is indicated with a dash (-). The nine residues that are not identical between ck3D8 and S8L are indicated in the ck3D8 amino acid sequence. Eightof these, FR residues of S8L (four residues in VH and four residues in VL) that were replaced with the residues that are most commonly observed in other chicken antibodies atthose positions, are shaded in black. One FR residue of S8L that was replaced with the corresponding residue of m3D8 is boxed. Changes in CDRs are denoted with asterisks.”.
Fig. 2. Biochemical properties of ck3D8. (A) SDS-PAGE gels showing the purified ck3D8 single chain variable fragment (scFv) with (His)5 and Protein A tags. The protein(10 �g) was separated on a 12% SDS-PAGE gel and visualized by staining with Coomassie Blue. (B) Enzyme-linked immunosorbent assay (ELISA) for the binding of ck3D8 toDNA. scFv was incubated in wells coated with 2 �g/ml of plasmid DNA. Bound protein was detected using rabbit IgG followed by alkaline phosphatase-conjugated anti-rabbitIgG. The irrelevant HW6 scFv was used as a negative control. Data represent the mean ± S.D. of triplicate wells from two independent experiments. (C) Fluorescence resonanceenergy transfer (FRET)-based DNA cleavage assay. scFv (1 �M) or a mixture of scFv (1 �M) and heparin (10 �g/ml) were incubated with a double-labeled DNA substrate(250 nM) as described in Section 2. The fluorescence intensity was then measured in real time over 6 h. RFU, relative fluorescence units; hp, heparin. Data are representativeof four independent experiments. (D and E) Cellular internalizing activity. HeLa cells were incubated with scFv (5 �M) for 6 h at 37 ◦C and then permeabilized. The cells werethen incubated with rabbit IgG followed by an Alexa Fluor 647-labeled anti-rabbit IgG and analyzed by flow cytometry (D) and confocal microscopy (E). Cell nuclei werestained with Hoechst 33342 (blue). Scale bar = 10 �m. Data are representative of three independent experiments. (For interpretation of the references to color in this figurelegend, the reader is referred to the web version of this article.)
J. Roh et al. / Molecular Immunology 63 (2015) 513–520 517
Fig. 3. Analysis of the ck3D8 structure. (A) Superposition of the structure of the homology modeled ck3D8 single chain variable fragment (scFv) with the crystal structureof m3D8 scFv, showing the variable regions of light chain (VL) (left) and variable region of heavy chain (VH) (right) with the complementarity-determining regions (CDRs)of the light chain (green) and the CDRs of the heavy chain (red). Upper panel: top view; lower panel: side view. (B) Root mean square deviations (RMSD) plot of ck3D8 andm3D8. (C) Root mean square fluctuations (RMSF) analysis of ck3D8 VH and m3D8 VH . (D) RMSF analysis of ck3D8 VL and m3D8 VL . Results for m3D8 and ck3D8 are shown inblue and violet lines, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
otisa
3
abpbfewApcwmaile
f m3D8 (Fig. 3C and D), overall, RMSF differences between thesewo proteins ranged from 0.1 to 0.2, suggesting stable fluctuationsn the VH and VL regions of the two antibodies. Taken together, noignificant difference was found between the structures of ck3D8nd m3D8.
.3.1. ck3D8 is less immunogenic than m3D8 in chickensIt is highly expected that a successful reshaping to a chicken
ntibody would decrease the immunogenicity of the reshaped anti-ody in chickens. Therefore, IgY responses against m3D8 and ck3D8rotein were assayed in chickens immunized with each proteiny ELISA using dilutions (10−3, 10−4, and 10−5) of sera obtainedrom two chickens immunized four times (highly boosted) withach scFv. The IgY antibody responses against m3D8 and ck3D8ere compared to pre-immune sera (Table 1). The ratio of the405 in sera from immunized chickens to the A405 in sera obtainedrior to immunization was calculated; the ratio when sera fromk3D8-immunized chickens were incubated with ck3D8-coatedells was similar to or slightly lower than the ratio of sera from3D8-immunized chickens incubated on m3D8-coated wells, at
ll dilutions tested. Furthermore, the ratio when sera from m3D8-mmunized chickens were incubated on ck3D8-coated wells wasower than that obtained when sera from ck3D8-immunized chick-ns were incubated on m3D8-coated wells, at all dilutions (Table 1).
This result suggests that ck3D8 has reduced immunogenicity inchickens compared to m3D8.
Next, a competitive ELISA was performed to further verify thereduced immunogenicity of ck3D8 in chickens (Fig. 4). Serum sam-ples from chickens immunized four times with m3D8 were dilutedat 1:1000 and incubated with various concentrations of competi-tors, m3D8 or ck3D8. The chicken serum IgY bound to m3D8 wasdetected with anti-chicken IgY. As expected, the ck3D8 competitordid not significantly decrease IgY binding to m3D8 even at a concen-tration of 50 �g/ml. By contrast, serum IgY binding to m3D8 wasdramatically reduced in a dose-dependent manner by the m3D8competitor in sera from two chickens (Fig. 4A), showing, approx-imately, a 65% reduction at 2 �g/ml and, approximately, an 80%reduction at 50 �g/ml of ck3D8 competitor. These results demon-strate that ck3D8, unlike m3D8, is not well recognized by chickenIgY raised against m3D8. This is possibly due to the fact that chickenIgY raised against m3D8 recognizes the CDRs and FRs of m3D8,which are of mouse origin, but only recognizes the CDRs of ck3D8, inwhich the FRs are of chicken origin. This conclusion is supported byanother competitive ELISA in which sera from ck3D8-immunized
chickens was mixed with various concentrations of competitors(ck3D8 and m3D8) and added to wells coated with ck3D8 (Fig. 4B).Binding of IgY in the sera of ck3D8-immunized chickens to ck3D8was strongly inhibited (60–80%) at 50 �g/ml of both competitors
518 J. Roh et al. / Molecular Immunology 63 (2015) 513–520
Table 1Comparison of IgY responses in chickens immunized with single chain variable fragment (scFv) proteins.
Antigen used for well coating Antigen used for immunizationa Relative absorbance of immune serum to pre-immune serumb
a Sera obtained from chickens immunized four times with the indicated scFv protein were used for this enzyme-linked immunosorbent assay (ELISA).b Ratio of immune serum absorbance at 405 nm to pre-immune serum absorbance at 405 nm. Data are the average of two independent ELISAs with samples run in duplicate.c Serum dilutions.
Fig. 4. Competitive ELISA to compare the difference in IgY response to the scFv proteins in chickens. Wells coated with 2 �g/ml of m3D8 (A) or 2 �g/ml of ck3D8 (B) wereincubated with sera (diluted 1:1000) from chickens that had been immunized four times with m3D8 (A) or ck3D8 (B), respectively, in the presence of various concentrationso as detf ls andw
(tti
4
cCtiamaa
f competitor m3D8 or ck3D8. The chicken serum IgY bound to m3D8 or ck3D8 wrom immunized chicken number 1 and chicken number 2 are shown in the left paneells from two independent experiments.
ck3D8 and m3D8), suggesting that more IgY was induced againsthe CDRs of the chickenized antibody than against the FRs. Takenogether, these results verify our expectation that ck3D8 is lessmmunogenic than m3D8 in chickens.
. Discussion
In this study, we first report the successful generation of ahickenized antibody from a mouse antibody using a traditionalDR grafting method. The chickenized antibody, ck3D8, retainedhe biochemical properties (DNA binding/hydrolysis and cellularnternalization) and 3-dimensional structure of the parental mouse
ntibody, m3D8. Moreover, ck3D8 was less immunogenic than3D8 in chickens, as expected. Our study is the first to report that
CDR grafting approach is applicable to reshaping from a mouse ton avian antibody.
ected with an alkaline phosphatase-conjugated anti-chicken IgY. Data with serum right panels of (A) and (B), respectively. Data represent the mean ± S.D. of triplicate
Several humanization methods for variable domains of non-human antibodies have been developed, including (1) CDR graftingbased on FR homology, in which six CDRs from non-human antibod-ies are inserted into FRs of appropriate acceptor human antibodiesthat have high homology to the FRs of the non-human antibody(Cheung et al., 2012); (2) CDR grafting based on CDR homology, inwhich human FRs from a set of human germline genes are chosenas an FR source based on the structural similarity of the CDRs of thehuman and the non-human antibody, regardless of the homologyof the FRs (Hwang et al., 2005); (3) CDR grafting based on frequencyof gene usage and structural stability, in which a human antibody,which is derived from a family of the most frequently used genesegments in the germline and forms the most stable structure for
VH and VL domains, is used as an acceptor antibody, without regardfor the homology of the FRs or CDRs (Kim et al., 2009); (4) germlinehumanization, in which germline genes of human antibodies areused as the FR source (Pelat et al., 2008; Rosok et al., 1996; Tan
muno
ebF(nmeawucawoV
cctuttbltCieFba
vihaU(cftVecfVita
mas1atl1tat
vmhs
J. Roh et al. / Molecular Im
t al., 2002); and (5) the resurfacing method, which is accomplishedy replacing FR residues that are not common in human antibodyRs with common human residues at the corresponding positionsChiu et al., 2011). Besides these humanization strategies by ratio-al design, several groups have utilized empirical humanizationethods such as a combinatorial library design strategy (Rosok
t al., 1996), guided selection by a serial transition process from rodent to human antibody (Osbourn et al., 2005), and frame-ork shuffling (Dall’Acqua et al., 2005). Among these methods, wesed CDR grafting based on FR homology for the generation of thehickenized antibody (ck3D8), with a few additional modifications,s follows: several FR residues in the scaffold chicken antibodyere changed to the residues that are most commonly found in
ther chicken antibodies, and two CDR sequences (VH-CDR2 andH-CDR3) were also empirically modified (Fig. 1).
Mouse antibody residues in the so-called “Vernier zone,” whichomprises the FRs underlying the CDRs, are thought to affect theonformation of CDR loops and are generally retained in the FRs ofhe humanized antibody when mouse antibodies are humanizedsing the CDR grafting method to avoid distortion of CDR conforma-ion (Makabe et al., 2008); thus we tried grafting both the CDRs andhe Vernier zone of m3D8 for the generation of a chickenized anti-ody. Interestingly, however, this chickenized antibody had much
ower DNA-binding/hydrolyzing activities and a greater tendencyo aggregate than ck3D8 that was chickenized by straightforwardDR grafting without the Vernier zone (data not shown). Moreover,
n CDR grafting based on FR homology, FRs derived from two differ-nt antibodies can be chosen for CDR grafting, with the heavy chainRs from one antibody and the light chain FRs from the other anti-ody. In generating ck3D8, heavy and light chain FRs derived from
single chicken antibody were used as a scaffold for CDR grafting.Antibody diversity in chickens is achieved through gene con-
ersion of the VH1 and VL1 gene by many VH and VL genes,nstead of the V(D)J recombination used by mammals. Chickensave a single functional rearranged gene coding for the VH (VH1)nd a single functional rearranged gene coding for the VL (VL1).pstream of the functional VH1 and VL1 genes, many pseudo-VH
VH) and pseudo-VL ( VL) genes exist. In the gene conversion pro-ess, sequence blocks from the VH or the VL gene (ranging in sizerom approximately 50 nucleotides to the entire pseudogene) areranslocated (overwritten) into a single functional rearranged orH and VL gene (Hecker, 2008; McCormack et al., 1991; Reynaudt al., 1989). Due to the mechanism of gene conversion, the FRs ofhicken antibodies are not variable in amino acid sequence; there-ore, only two primer sets are needed for the amplification of theH and VL gene repertoires, unlike in human and mouse antibod-
es. The limited FR variability makes it fairly simple and conveniento select the proper chicken FR template for the chickenization ofntibodies.
Chickenization of mouse antibodies can be an alternativeethod for obtaining a chicken antibody. Chicken hybridomas
re very unstable and often do not grow or lose their ability toecrete antibody after several passages in culture (Nishinaka et al.,991, 1996), because the chicken myeloma cells that are used as
fusion partner are rare and their genetics are not well charac-erized. Therefore, the production of chicken antibodies is quiteimited (Asaoka et al., 1992; Matsuda et al., 1999; Matsushita et al.,998; Nakamura et al., 2004), although phage display is an alterna-ive technique for it. Chickenization of a mouse antibody could bepplied to the development of chicken antibodies to overcome thisechnological difficulty.
It is noteworthy that m3D8 scFv has anti-viral activity against
iruses including classical swine fever virus (CSFV), vesicular sto-atitis virus (VSV) (Joung et al., 2012; Jun et al., 2010), and human
erpes simplex virus-1 (HSV-1) (Lee et al., 2014). The engineeredingle domain of 3D8 (3D8 VL) was able to inhibit the replication
logy 63 (2015) 513–520 519
of influenza (H9N2) virus (Kim et al., 2012). Based on the similarstructures and biochemical properties of m3D8 and ck3D8, ck3D8should have anti-viral activity as well and thus we expect thatck3D8 could be useful for scientific applications against chickeninfectious viruses.
Consequently, our study demonstrates that well-establishedmethods for antibody humanization can be used for reshapingantibodies between non-human species, probably due to the inter-species compatibility of the FRs.
Conflict of interest statement
The authors have no conflicts of interest to declare.
Acknowledgments
This work was supported by the Science Research Center Pro-gram (2011-0030043) of the National Research Foundation of Korea(NRF) and the Cooperative Research Program (PJ010201032014) ofthe Rural Development Administration of Korea.
References
Abi-Ghanem, D., Waghela, S.D., Caldwell, D.J., Danforth, H.D., Berghman, L.R., 2008.Phage display selection and characterization of single-chain recombinant anti-bodies against Eimeria tenella sporozoites. Vet. Immunol. Immunopathol. 121,58–67.
Ahmadzadeh, V., Farajnia, S., Feizi, M.A., Nejad, R.A., 2014. Antibody humaniza-tion methods for development of therapeutic applications. Monoclon. Antib.Immunodiagn. Immunother. 33, 67–73.
Asaoka, H., Nishinaka, S., Wakamiya, N., Matsuda, H., Murata, M., 1992. Two chickenmonoclonal antibodies specific for heterophil Hanganutziu-Deicher antigens.Immunol. Lett. 32, 91–96.
Berendsen, H.J., van der, Spoel D., van Drunen, R., 1995. GROMACS: a message-passing parallel molecular dynamics implementation. Comput. Phys. Commun.91, 43–56.
Darden, T., York, D., Pedersen, L., 1993. Particle mesh Ewald: an N log (N) methodfor Ewald sums in large systems. J. Chem. Phys. 98, 10089–10092.
Davies, J., Riechmann, L., 1995. Antibody VH domains as small recognition units.Biotechnology 13, 475–479.
Duan, Y., Wu, C., Chowdhury, S., Lee, M.C., Xiong, G., Zhang, W., Yang, R., Cieplak, P.,Luo, R., Lee, T., 2003. A point-charge force field for molecular mechanics simula-tions of proteins based on condensed-phase quantum mechanical calculations.J. Comput. Chem. 24, 1999–2012.
Finlay, W.J., Cunningham, O., Lambert, M.A., Darmanin-Sheehan, A., Liu, X., Fennell,B.J., Mahon, C.M., Cummins, E., Wade, J.M., O’Sullivan, C.M., Tan, X.Y., Piche, N.,Pittman, D.D., Paulsen, J., Tchistiakova, L., Kodangattil, S., Gill, D., Hufton, S.E.,2009. Affinity maturation of a humanized rat antibody for anti-RAGE therapy:comprehensive mutagenesis reveals a high level of mutational plasticity bothinside and outside the complementarity-determining regions. J. Mol. Biol. 388,541–558.
Foord, A.J., Muller, J.D., Yu, M., Wang, L.F., Heine, H.G., 2007. Production and applica-tion of recombinant antibodies to foot-and-mouth disease virus non-structuralprotein 3ABC. J. Immunol. Methods 321, 142–151.
Gorman, S., Clark, M., 1990. Humanisation of monoclonal antibodies for therapy.Semin. Immunol. 2, 457–466.
Hecker, N., 2008. Primary Diversity Mechanisms of the Antibody Synthesis inHumans, Mice and Chickens, Universität zu Lübeck (http://www.tcs.uni-luebeck.de/downloads/papers/2008/diversity.pdf).
Hess, B., Bekker, H., Berendsen, H.J., Fraaije, J.G., 1997. LINCS: a linear constraintsolver for molecular simulations. J. Comput. Chem. 18, 1463–1472.
Hwang, W.Y., Almagro, J.C., Buss, T.N., Tan, P., Foote, J., 2005. Use of human germlinegenes in a CDR homology-based approach to antibody humanization. Methods36, 35–42.
Jang, J.Y., Jeong, J.G., Jun, H.R., Lee, S.C., Kim, J.S., Kim, Y.S., Kwon, M.H., 2009. Anucleic acid-hydrolyzing antibody penetrates into cells via caveolae-mediated
endocytosis, localizes in the cytosol and exhibits cytotoxicity. Cell. Mol. Life Sci.66, 1985–1997.
Joung, Y.K., Son, S., Jang, J.Y., Kwon, M.H., Park, K.D., 2012. Sustained cytoplasmicdelivery and anti-viral effect of PLGA nanoparticles carrying a nucleic acid-hydrolyzing monoclonal antibody. Pharm. Res. 29, 932–942.
un, H.R., Pham, C.D., Lim, S.I., Lee, S.C., Kim, Y.S., Park, S., Kwon, M.H., 2010. An RNA-hydrolyzing recombinant antibody exhibits an antiviral activity against classicalswine fever virus. Biochem. Biophys. Res. Commun. 395, 484–489.
im, A., Lee, J.Y., Byun, S.J., Kwon, M.H., Kim, Y.S., 2012. Viral genome RNA degra-dation by sequence-selective, nucleic-acid hydrolyzing antibody inhibits thereplication of influenza H9N2 virus without significant cytotoxicity to host cells.Antiviral Res. 94, 157–167.
im, D.S., Lee, S.H., Kim, J.S., Lee, S.C., Kwon, M.H., Kim, Y.S., 2009. Generationof humanized anti-DNA hydrolyzing catalytic antibodies by complementaritydetermining region grafting. Biochem. Biophys. Res. Commun. 379, 314–318.
im, S.J., Park, Y., Hong, H.J., 2005. Antibody engineering for the development oftherapeutic antibodies. Mol. Cells 20, 17–29.
im, Y.R., Kim, J.S., Lee, S.H., Lee, W.R., Sohn, J.N., Chung, Y.C., Shim, H.K., Lee, S.C.,Kwon, M.H., Kim, Y.S., 2006. Heavy and light chain variable single domains of ananti-DNA binding antibody hydrolyze both double- and single-stranded DNAswithout sequence specificity. J. Biol. Chem. 281, 15287–15295.
ee, G., Yu, J., Cho, S., Byun, S.J., Kim, D.H., Lee, T.K., Kwon, M.H., Lee, S., 2014. Anucleic-acid hydrolyzing single chain antibody confers resistance to DNA virusinfection in hela cells and C57BL/6 mice. PLoS Pathog. 10, e1004208.
ee, J., Kim, H.J., Roh, J., Seo, Y., Kim, M., Jun, H.R., Pham, C.D., Kwon, M.H., 2013.Functional consequences of complementarity-determining region deactivationin a multifunctional anti-nucleic acid antibody. J. Biol. Chem. 288, 35877–35885.
akabe, K., Nakanishi, T., Tsumoto, K., Tanaka, Y., Kondo, H., Umetsu, M., Sone,Y., Asano, R., Kumagai, I., 2008. Thermodynamic consequences of mutationsin vernier zone residues of a humanized anti-human epidermal growth factorreceptor murine antibody, 528. J. Biol. Chem. 283, 1156–1166.
atsuda, H., Mitsuda, H., Nakamura, N., Furusawa, S., Mohri, S., Kitamoto, T., 1999.A chicken monoclonal antibody with specificity for the N-terminal of humanprion protein. FEMS Immunol. Med. Microbiol. 23, 189–194.
atsushita, K., Horiuchi, H., Furusawa, S., Horiuchi, M., Shinagawa, M., Matsuda,H., 1998. Chicken monoclonal antibodies against synthetic bovine prion proteinpeptide. J. Vet. Med. Sci. 60, 777–779.
cCormack, W.T., Tjoelker, L.W., Thompson, C.B., 1991. Avian B-cell development:generation of an immunoglobulin repertoire by gene conversion. Annu. Rev.Immunol. 9, 219–241.
akamura, N., Shuyama, A., Hojyo, S., Shimokawa, M., Miyamoto, K., Kawashima,T., Aosasa, M., Horiuchi, H., Furusawa, S., Matsuda, H., 2004. Establishment ofa chicken monoclonal antibody panel against mammalian prion protein. J. Vet.
Med. Sci. 66, 807–814.
ishinaka, S., Akiba, H., Nakamura, M., Suzuki, K., Suzuki, T., Tsubokura, K., Hori-uchi, H., Furusawa, S., Matsuda, H., 1996. Two chicken B cell lines resistant toouabain for the production of chicken monoclonal antibodies. J. Vet. Med. Sci.58, 1053–1056.
logy 63 (2015) 513–520
Nishinaka, S., Suzuki, T., Matsuda, H., Murata, M., 1991. A new cell line for the pro-duction of chicken monoclonal antibody by hybridoma technology. J. Immunol.Methods 139, 217–222.
Osbourn, J., Groves, M., Vaughan, T., 2005. From rodent reagents to human thera-peutics using antibody guided selection. Methods 36, 61–68.
Pelat, T., Bedouelle, H., Rees, A.R., Crennell, S.J., Lefranc, M.P., Thullier, P., 2008.Germline humanization of a non-human primate antibody that neutralizesthe anthrax toxin, by in vitro and in silico engineering. J. Mol. Biol. 384,1400–1407.
Pronk, S., Páll, S., Schulz, R., Larsson, P., Bjelkmar, P., Apostolov, R., Shirts, M.R., Smith,J.C., Kasson, P.M., van der Spoel, D., 2013. GROMACS 4.5: a high-throughputand highly parallel open source molecular simulation toolkit. Bioinformatics29, 845–854.
Reynaud, C.A., Dahan, A., Anquez, V., Weill, J.C., 1989. Somatic hyperconversiondiversifies the single Vh gene of the chicken with a high incidence in the Dregion. Cell 59, 171–183.
Rosok, M.J., Yelton, D.E., Harris, L.J., Bajorath, J., Hellstrom, K.E., Hellstrom, I., Cruz,G.A., Kristensson, K., Lin, H., Huse, W.D., Glaser, S.M., 1996. A combinatoriallibrary strategy for the rapid humanization of anticarcinoma BR96 Fab. J. Biol.Chem. 271, 22611–22618.
Safdari, Y., Farajnia, S., Asgharzadeh, M., Khalili, M., 2013. Antibody human-ization methods—a review and update. Biotechnol. Genet. Eng. Rev. 29,175–186.
Sapats, S.I., Trinidad, L., Gould, G., Heine, H.G., van den Berg, T.P., Eterradossi, N.,Jackwood, D., Parede, L., Toquin, D., Ignjatovic, J., 2006. Chicken recombinantantibodies specific for very virulent infectious bursal disease virus. Arch. Virol.151, 1551–1566.
Tan, P., Mitchell, D.A., Buss, T.N., Holmes, M.A., Anasetti, C., Foote, J., 2002. “Superhu-manized” antibodies: reduction of immunogenic potential by complementarity-determining region grafting with human germline sequences: application to ananti-CD28. J. Immunol. 169, 1119–1125.
Vincke, C., Loris, R., Saerens, D., Martinez-Rodriguez, S., Muyldermans, S., Conrath,K., 2009. General strategy to humanize a camelid single-domain antibody andidentification of a universal humanized nanobody scaffold. J. Biol. Chem. 284,3273–3284.
Yamanaka, H.I., Inoue, T., Ikeda-Tanaka, O., 1996. Chicken monoclonal antibody iso-lated by a phage display system. J. Immunol. 157, 1156–1162.
Yu, Y., Lee, P., Ke, Y., Zhang, Y., Yu, Q., Lee, J., Li, M., Song, J., Chen, J., Dai, J., Do Couto,F.J., An, Z., Zhu, W., Yu, G.L., 2010. A humanized anti-VEGF rabbit monoclonalantibody inhibits angiogenesis and blocks tumor growth in xenograft models.
Zhao, M., Wu, M., Guo, L., Jiang, J., Huang, W., Lin, X., Zhang, Z., Xiang, D., Lu, H.,Zhu, S., Yu, Y., Moldenhauer, A., Han, W., 2010. Expression, purification, andcharacterization of a novel soluble form of human Delta-like-1. Appl. Biochem.Biotechnol. 160, 1415–1427.
![DEPARTMENT OFMICROBIOLOGY N]MS MEDICAL COLLEGE …nationalinstituteofmedicalsciences.com/wp-content/uploads/2016/11/... · DEPARTMENT OFMICROBIOLOGY N]MS MEDICAL COLLEGE ... 12.11.2016](https://static.documents.pub/doc/80x56/5abec8e07f8b9ad8278d98aa/department-ofmicrobiology-nms-medical-college-nationalinstit-ofmicrobiology.jpg)
