Targeting a homogeneously glycosylated antibody Fc to bind cancer cells using a synthetic receptor ligand

Article abstract of DOI:10.1021/ja9045179

(Chemical Equation Presented) The targeting of a glycosylated antibody Fc fragment to bind to cancer cells by site-selective incorporation of a synthetic ligand is described. Homogeneously glycosylated immunoglobulin G subclass 1 fragment crystallizable (

Full text of DOI:10.1021/ja9045179

Published on Web 09/03/2009
Targeting a Homogeneously Glycosylated Antibody Fc To Bind Cancer Cells
Using a Synthetic Receptor Ligand
Junpeng Xiao,† Rui Chen,‡ Mark A. Pawlicki,† and Thomas J. Tolbert*,†,‡
Interdisciplinary Biochemistry Graduate Program and Department of Chemistry, Indiana UniVersity,
800 East Kirkwood AVenue, Bloomington, Indiana 47405
Received June 3, 2009; E-mail: tolbert@indiana.edu
Site-selective chemical modification of proteins has wide
application in biochemistry and biotechnology, including the
incorporation of labels and defined post-translational modifica-
tions onto proteins to facilitate biochemical studies and incor-
poration of synthetic modifications onto proteins to alter
bioactivity and physical properties.1 As part of our research into
the role of N-linked glycosylation in antibody-dependent immune
responses, we have needed to develop methods for targeting
immunoglobulin G subclass 1 (IgG1)-type antibody fragment
crystallizable (Fc) regions to bind to cancer cells in the absence
of antibody fragment antigen binding (Fab) regions (Figure 1a).
To accomplish this, we have developed methods to apply native
chemical ligation (NCL)2 to the site-selective modification of
an expressed glycoprotein fragment. Here we report the use of
these methods for the site-selective modification of a homoge-
neously glycosylated 225 amino acid IgG1 Fc glycoprotein with
a cyclic RGD integrin antagonist as well as the results of
subsequent binding and bioactivity studies.
Figure 1. Antibody fragments and RGD peptides. (a) Antibody
fragments obtained upon protease digestion. (b) Cyclic-RGDfK Rvꢀ3
integrin antagonist3 1. (c) Solid-phase synthesis of cyclic RGDfK
thioester derivative 2: (i) 1% TFA in CH2Cl2; (ii) succinic anhydride
and N-methylmorpholine; (iii) methyl 3-mercaptopropionate, DIC, HOBt,
DMAP; (iv) TFA with TIS.
Our strategy for targeting antibody Fc regions to bind to cancer
cells is to link ligands of cell-surface receptors highly expressed
on cancer cells to the N-terminus of IgG1 Fc. In this way, the
ligands are displayed on the IgG1 Fc in a manner similar to
how the Fab’s would be displayed on a full-length antibody.
Ligand binding to cellular receptors should then mimic antibody
binding to antigen, displaying the Fc region in a way that should
allow antibody effector functions such as antibody-dependent
cell-mediated cytotoxicity (ADCC) and complement-dependent
cytotoxicity (CDC) to function. As a receptor ligand to attach
to IgG1 Fc, we chose the cyclic RGD peptide 1 (Figure 1b), a
selective antagonist of the Rvꢀ3 integrin receptor.3 The Rvꢀ3
integrin receptor is overexpressed in many types of cancer and
is also involved in angiogenesis, making it a good target for
delivering therapeutics to cancer cells and for imaging.4 To
achieve attachment of receptor ligands such as 1 site-selectively
on the N-terminus of IgG1 Fc, we chose to use NCL because it
is a chemoselective reaction that specifically modifies the
N-terminus of proteins containing N-terminal cysteines, linking
proteins to synthetic molecules via stable peptide bonds.2 To
use NCL for this application, it was necessary to develop
methods for expressing the IgG1 Fc glycoprotein with a
N-terminal cysteine and producing cyclic RGD thioester deriva-
tive 2 (Figure 1c). Also, to overcome the natural heterogeneity
of glycoproteins produced in most cell lines, we utilized a
glycosylation-deficient yeast strain and enzymatic digestion to
produce a homogeneously glycosylated IgG1 Fc, which greatly
simplified the monitoring of glycoprotein modification reactions
with mass spectrometry.
NCL is a chemical reaction between N-terminal cysteines and
thioesters that forms native peptide bonds.2 Methods for expres-
sion of glycoproteins containing N-terminal cysteines have not
been reported, so a method for expressing the IgG1 Fc
glycoprotein with a N-terminal cysteine was developed. As a
N-linked glycoprotein, IgG1 Fc must pass through the secretory
pathway in order to be glycosylated, which necessitates the
presence of a N-terminal signal peptide in the protein sequence
to direct IgG1 Fc into the secretory pathway.5 In addition, for
efficient secretion of proteins from yeast cells, a N-terminal
secretion signal, the R-factor, is required. As a result, a
N-terminal cysteine-containing IgG1 Fc cannot be directly
expressed but must also contain these additional peptide
sequences. Since the N-terminal signal peptide and R-factor
secretion signal are proteolytically removed during protein
expression, we explored the possibility of using the endogenous
yeast protease involved in the last step of this process, the Kex2
protease, to generate a N-terminal cysteine form of the IgG1 Fc
glycoprotein. IgG1 Fc was expressed with a cysteine residue
immediately following the Kex2 protease cleavage site (Figure
S1 in the Supporting Information).6 This approach worked well,
and expression of this construct in Pichia pastoris resulted in
the production of ∼30 mg/L of N-terminal cysteine-containing
IgG1 Fc, designated as C-IgG1 Fc (3) (Figure 2a), as character-
ized by electrospray ionization mass spectrometry (ESI-MS) after
PNGase F treatment to remove the heterogeneous N-linked
† Interdisciplinary Biochemistry Graduate Program.
‡ Department of Chemistry.
9
13616 J. AM. CHEM. SOC. 2009, 131, 13616–13618
10.1021/ja9045179 CCC: $40.75  2009 American Chemical Society

C O M M U N I C A T I O N S
Figure 2. Synthesis of RGD-Man5-IgG1 Fc. (a) Conversion of C-IgG1 Fc (3) to RGD-Man5-IgG1 Fc (5) by ManIA digestion and NCL with thioester
2. (b) ESI-MS of heterogeneously glycosylated C-IgG1 Fc (3). Calcd molecular weights for glycoforms containing 8-12 mannose residues: 26 872,
27 035, 27 198, 27 361, and 27 524, respectively. (c) ESI-MS of 4: calcd, 26 386; obsd, 26 387. (d) ESI-MS of 5: calcd, 27 071; obsd, 27 070.
oligosaccharide (Figure S3). One concern with this approach is
that generation of reactive N-terminal cysteines during yeast
secretory expression has the potential to form thiazolidine
adducts through the reaction of the N-terminal cysteines with
aldehyde- or ketone-containing metabolites.7 Little thiazolidine
formation was observed by mass spectrometry of the PNGase
F-treated C-IgG1 Fc, but to ensure free N-terminal cysteines
for NCL reactions, the glycoproteins used in this research were
treated with hydroxylamine to hydrolyze potential thiazolidines
prior to NCL reactions (Figure S3). To produce homogeneously
glycosylated, N-terminal cysteine-containing IgG1 Fc, the C-
IgG1 Fc glycoprotein 3 was expressed in a glycosylation-
deficient yeast strain produced in our laboratory using the
methods of Nett and Gerngross.8 Expression in the glycosylation-
deficient yeast strain resulted in a mixture of high-mannose
C-IgG1 Fc glycoforms containing 8-12 mannoses (Figure 2b).
Figure 3. Cell adhesion and fluorescence microscopy. (a) WM-115 cell
adhesion to microplate wells coated with 5 µg/mL of each specified
protein. (b) Merged image of WM-115 cells stained with FITC-labeled
RGD-Man5-IgG1 Fc 5 and DAPI nuclear staining. (c) Merged image of
Treatment of heterogeneously glycosylated 3 with R-1,2-
mannosidase IA9 (ManIA) to remove R-1,2 linked mannose
residues produced a glycoform of C-IgG1 Fc containing five
mannose residues [C-Man5-IgG1 Fc (4)], with a conversion of
>95% as estimated by mass spectrometry (Figure 2b,c). This
homogeneous C-Man5-IgG1 Fc 4 was then used in the NCL
reactions.
WM-115 cells stained with FITC-labeled C-IgG1 Fc 3 and DAPI nuclear
staining.
complete after 24 h, as estimated by ESI-MS. Site-selective
modification of IgG1 Fc prevents important amino acid residues
from being modified, which should improve its in vivo behavior,
as has been reported for selective modification of full-length
antibodies.11 N-terminal modification of IgG1 Fc did not interfere
with Fc receptor binding, as determined by ELISA binding assays
(Figure S10).
The synthesis of thioester 2 was based upon the solid-phase
synthesis of cyclic RGD derivatives reported by McCusker et
al.,10 except that a monomethoxytrityl (Mmt) side-chain-
protected lysine was utilized (Figure 1c). After on-resin con-
struction of the cyclic-(RGDfK) peptide, the Mmt protecting
group of the lysine was selectively deprotected with 1%
trifluoroacetic acid (TFA) in CH2Cl2; the lysine amine was then
acylated with succinic anhydride. Thioester formation was
effected on-resin by addition of methyl 3-mercaptopropionate,
4-dimethylaminopyridine (DMAP), and N-hydroxybenzotriazole
(HOBt) using N,N′-diisopropylcarbodiimide (DIC) as a coupling
reagent. Cleavage from the resin and removal of the remaining
protecting groups were accomplished using TFA with triisopro-
pylsilane (TIS) as a scavenger. The resulting peptide thioester
2 was precipitated from diethyl ether and purified by reversed-
phase HPLC.
To determine whether RGD-Man5-IgG1 Fc 5 was targeted to
bind Rvꢀ3 integrin receptor-expressing cells, experiments were
conducted using WM-115 melanoma cells, which express the
Rvꢀ3 integrin receptor at a high level.12 First, an adhesion assay
was conducted, taking advantage of the fact that antibodies and
antibody fragments such as IgG1 Fc coat the hydrophobic
surfaces of ELISA plates, allowing ELISA-type binding assays
to be conducted. The ability of ELISA microplate wells coated
with 5 to promote adhesion of WM-115 cells was compared with
those of wells coated with fibrinogen (a natural ligand for the
Rvꢀ3 integrin receptor), C-IgG1 Fc 3, and bovine serum albumin
(BSA) (Figure 3a). Wells containing 3 alone had few cells
adhering to them and were similar to the negative-control BSA
wells. Interestingly, the wells coated with 5 had the highest
amounts of adhering cells, ∼14-fold more than IgG1 Fc alone
and also more than twice as much as adhered to the fibrinogen-
coated wells (Figure 3a). This indicates strong binding interac-
tions between RGD-Man5-IgG1 Fc 5 on the wells and the Rvꢀ3
integrin receptors on WM-115 cells. Additional cell lines were
Ligation of C-Man5-IgG1 Fc 4 to thioester 2 was carried out
by mixing 4 (2 mg/mL, 76 µM) with 2.6 mM 2 in 20 mM sodium
phosphate buffer (pH 7.5) containing 5 mM betaine and 30 mM
sodium 2-mercaptoethanesulfonic acid (MESNA). The reaction
mixture was incubated for 24 h, after which an aliquot of the
ligation reaction mixture was analyzed by ESI-MS. As Figure 2
shows, the ligation proceeded well and was greater than 95%
9
J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009 13617

C O M M U N I C A T I O N S
also tested in this adhesion assay to confirm RGD specificity
(see Figure S5).
tion. The techniques developed here for NCL modification of
expressed glycoproteins are general and may be applied to the
modification of other glycoproteins or for the production of
glycosylated protein fragments for the chemoenzymatic synthesis
of homogeneously glycosylated glycoproteins.14
To further assess the ability of the RGD-modified IgG1 Fc 5
to bind to the Rvꢀ3 integrin receptor, adhesion inhibition assays
comparing 5 with the free RGD peptide antagonist 1 were
conducted using the methods of Wu et al.12 Interestingly, 5 had
an IC50 value nearly identical to that of 1 (104 and 102 nM,
respectively). This indicates that the cyclic RGD modification
of IgG1 Fc targets the antibody fragment to bind to Rvꢀ3 integrin
receptor-expressing cancer cells in a manner similar to the free
peptide.
Fluorescence microscopy experiments were also conducted
to observe the binding of the RGD-modified IgG1 Fc to WM-
115 cells. Both C-IgG1 Fc 3 and RGD-IgG1 Fc 5 were labeled
with fluorescein isothiocyanate (FITC) and utilized in this
experiment. WM-115 cells grown on glass coverslips were
incubated with 1 µM FITC-labeled 3 or 5 for 15 min at 37 °C.
The cells were then washed and fixed and the cell nuclei stained
with 4′,6-diamidino-2-phenylindole (DAPI), after which the cells
were mounted for fluorescence microscopy. As Figure 3 shows,
there is a significant difference between the binding of FITC-
labeled 5 and 3, confirming that the RGD ligand is active for
binding to the Rvꢀ3 integrin receptor. These results were
quantified (see Figure S8 and Table S2) and found to be similar
to those observed for RGD dendrimers binding to the Rvꢀ3
integrin receptor, which may indicate endocytosis of the RGD-
IgG1 Fc 5.13
Acknowledgment. The authors gratefully acknowledge Dr.
Vasily Gelfanov and Dr. Richard DiMarchi for cell-culture
discussions and facilities, Dr. Kelley Moremen for providing a
ManIA-expressing yeast strain, Dr. J. Shaun Lott for providing
a PNGase F-overexpressing E. coli strain,15 and Dr. Shang Cai
and Dr. Claire Walczak for fluorescence microscopy advice and
facilities.
Supporting Information Available: Experimental procedures and
characterization data for expression of the C-IgG1 Fc 3; synthesis
of 2, 4, and 5; WM-115 adhesion assay; WM-115 adhesion inhibition
assay; fluorescence microscopy data; and complete ref 11b. This
material is available free of charge via the Internet at http://
pubs.acs.org.
References
(1) (a) Hackenberger, C. P. R.; Schwarzer, D. Angew. Chem., Int. Ed. 2008,
47, 10030. (b) Flavell, R. R.; Muir, T. W. Acc. Chem. Res. 2009, 42, 107.
(c) Xiao, J.; Burn, A.; Tolbert, T. J. Bioconjugate Chem. 2008, 19, 1113.
(d) Kochendoerfer, G. G. Curr. Opin. Chem. Biol. 2005, 9, 555. (e) Tolbert,
T. J.; Wong, C.-H. J. Am. Chem. Soc. 2000, 122, 5421. (f) Tolbert, T. J.;
Wong, C.-H. Angew. Chem., Int. Ed. 2002, 41, 2171.
(2) Dawson, P. E.; Muir, T. W.; Clark-Lewis, I.; Kent, S. B. H. Science 1994,
266, 776.
In summary, methods to N-terminally modify glycosylated
IgG1 Fc produced in yeast have been developed and applied to
the N-terminal attachment of a bioactive receptor ligand, a cyclic
RGD Rvꢀ3 integrin receptor antagonist. Homogenously glyco-
sylated C-Man5-IgG1 Fc 4 with a N-terminal cysteine was
produced by expression in a glycosylation-deficient yeast strain,
utilizing Kex2 proteolytic processing to generate the N-terminal
cysteine on the glycoprotein in the secretory pathway and in
vitro ManIA digestion to produce the homogeneous Man5
glycoform. A thioester-containing cyclic RGD integrin antagonist
2 was produced by solid-phase peptide synthesis and attached
to the N-terminus of C-Man5-IgG1 Fc 4 utilizing NCL. The
resulting RGD-Man5-IgG1 Fc glycoprotein 5 retained its ability
to bind to and antagonize the Rvꢀ3 integrin receptor on WM-
115 melanoma cells, as demonstrated by adhesion assays,
adhesion inhibition assays, and fluorescence microscopy. Anti-
body fragments such as IgG1 Fc have significant bioactivity in
their own right, with long circulatory half-lives and the ability
to direct antibody-dependent immune responses such as ADCC
and CDC. Thus, attachment of cell-receptor ligands to IgG1 Fc
in this manner may augment the biological activity of the
attached ligands, and this possibility is under further investiga-
(3) Haubner, R.; Gratias, R.; Diefenbach, B.; Goodman, S. L.; Jonczyk, A.;
Kessler, H. J. Am. Chem. Soc. 1996, 118, 7461.
(4) Garanger, E.; Boturyn, D.; Dumy, P. Anti-Cancer Agents Med. Chem. 2007,
7, 552.
(5) (a) Shuster, J. R. Curr. Opin. Biotechnol. 1991, 2, 685. (b) Brake, A. J.;
Merryweather, J. P.; Coit, D. G.; Heberlein, U. A.; Masiarz, F. R.;
Mullenbach, G. T.; Urdea, M. S.; Valenzuela, P.; Barr, P. J. Proc. Natl.
Acad. Sci. U.S.A. 1984, 81, 4642.
(6) Wheatley, J. L.; Holyoak, T. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 6626.
(7) (a) Zhang, L.; Tam, J. P. Anal. Biochem. 1996, 233, 87. (b) Gentle, I. E.;
De Souza, D. P.; Baca, M. Bioconjugate Chem. 2004, 15, 658.
(8) Nett, J. H.; Gerngross, T. U. Yeast 2003, 20, 1279.
(9) Lal, A.; Pang, P.; Kalelkar, S.; Romero, P. A.; Herscovics, A.; Moremen,
K. W. Glycobiology 1998, 8, 981.
(10) McCusker, C. F.; Kocienski, P. J.; Boyle, F. T.; Schatzlein, A. G. Bioorg.
Med. Chem. Lett. 2002, 12, 547.
(11) (a) Sesay, M. A. BioPharm Int. 2003, 16, 32. (b) Junutula, J. R.; et al. Nat.
Biotechnol. 2008, 26, 925. (c) Hamblett, K. J.; Senter, P. D.; Chace, D. F.;
Sun, M. M. C.; Lenox, J.; Cerveny, C. G.; Kissler, K. M.; Bernhardt, S. X.;
Kopcha, A. K.; Zabinski, R. F.; Meyer, D. L.; Francisco, J. A. Clin. Cancer
Res. 2004, 10, 7063.
(12) Wu, J. M.; Rosser, M. P.; Howlett, A. R.; Feldman, R. I. Methods Mol.
Biol. 1999, 129, 211.
(13) Boturyn, D.; Coll, J.-L.; Garanger, E.; Favrot, M.-C.; Dumy, P. J. Am.
Chem. Soc. 2004, 126, 5730.
(14) (a) Bennett, C. S.; Wong, C.-H. Chem. Soc. ReV. 2007, 36, 1227. (b)
Gamblin, D. P.; Scanlan, E. M.; Davis, B. G. Chem. ReV. 2009, 109, 131.
(15) Loo, T.; Patchett, M. L.; Norris, G. E.; Lott, J. S. Protein Expression Purif.
2002, 24, 90.
JA9045179
9
13618 J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009