A protein engineering approach differentiates the functional importance of carbohydrate moieties of interleukin-5 receptor α.

Abstract

Human interleukin-5 receptor α (IL5Rα) is a glycoprotein that contains four N-glycosylation sites in the extracellular region. Previously, we found that enzymatic deglycosylation of IL5Rα resulted in complete loss of IL5 binding. To localize the functionally important carbohydrate moieties, we employed site-directed mutagenesis at the N-glycosylation sites (Asn(15), Asn(111), Asn(196), and Asn(224)). Because Asn-to-Gln mutagenesis caused a significant loss of structural integrity, we used diverse mutations to identify stability-preserving changes. We also rationally designed mutations at and around the N-glycosylation sites based on sequence alignment with mouse IL5Rα and other cytokine receptors. These approaches were most successful at Asn(15), Asn(111), and Asn(224). In contrast, any replacement at Asn(196) severely reduced stability, with the N196T mutant having a reduced binding affinity for IL5 and diminished biological activity because of the lack of cell surface expression. Lectin inhibition analysis suggested that the carbohydrate at Asn(196) is unlikely involved in direct ligand binding. Taking this into account, we constructed a stable variant, with triple mutational deglycosylation (N15D, I109V/V110T/N111D, and L223R/N224Q). The re-engineered protein retained Asn(196) while the other three glycosylation sites were eliminated. This mostly deglycosylated variant had the same ligand binding affinity and biological activity as fully glycosylated IL5Rα, thus demonstrating a unique role for Asn(196) glycosylation in IL5Rα function. The results suggest that unique carbohydrate groups in multiglycosylated receptors can be utilized asymmetrically for function.

Full text

Page 1

A Protein Engineering Approach Differentiates the Functional
Importance of Carbohydrate Moieties of Interleukin-5 Receptor α
Tetsuya Ishino,†,⊥Nicoleta J. Economou,†Karyn McFadden,†Meirav Zaks-Zilberman,†Monika Jost,‡
Sabine Baxter,†Mark R. Contarino,†Adrian E. Harrington,†Patrick J. Loll,†Gianfranco Pasut,§
Sam Lievens,∥Jan Tavernier,∥and Irwin Chaiken*,†
†Department of Biochemistry and Molecular Biology, Drexel University College of Medicine, 11102 New College Building, 245
North 15th Street, Philadelphia, Pennsylvania 19102, United States
‡Department of Radiation Oncology, Drexel University College of Medicine, 11102 New College Building, 245 North 15th Street,
Philadelphia, Pennsylvania 19102, United States
§Department of Pharmaceutical Sciences, University of Padua, Via F. Marzolo 5, Padua 35131, Italy
∥Department of Medical Protein Research, Flanders Interuniversity Institute for Biotechnology, VIB09-Faculty of Medicine and
Health Sciences, Ghent University, Ghent, Belgium
*
S Supporting Information
ABSTRACT: Human interleukin-5 receptor α (IL5Rα) is a glycoprotein that contains four N-
glycosylation sites in the extracellular region. Previously, we found that enzymatic
deglycosylation of IL5Rα resulted in complete loss of IL5 binding. To localize the functionally
important carbohydrate moieties, we employed site-directed mutagenesis at the N-glycosylation
sites (Asn15, Asn111, Asn196, and Asn224). Because Asn-to-Gln mutagenesis caused a significant
loss of structural integrity, we used diverse mutations to identify stability-preserving changes. We
also rationally designed mutations at and around the N-glycosylation sites based on sequence
alignment with mouse IL5Rα and other cytokine receptors. These approaches were most
successful at Asn15, Asn111, and Asn224. In contrast, any replacement at Asn196severely reduced
stability, with the N196T mutant having a reduced binding affinity for IL5 and diminished
biological activity because of the lack of cell surface expression. Lectin inhibition analysis
suggested that the carbohydrate at Asn196is unlikely involved in direct ligand binding. Taking
this into account, we constructed a stable variant, with triple mutational deglycosylation (N15D,
I109V/V110T/N111D, and L223R/N224Q). The re-engineered protein retained Asn196while
the other three glycosylation sites were eliminated. This mostly deglycosylated variant had the same ligand binding affinity and
biological activity as fully glycosylated IL5Rα, thus demonstrating a unique role for Asn196glycosylation in IL5Rα function. The
results suggest that unique carbohydrate groups in multiglycosylated receptors can be utilized asymmetrically for function.
I
eosinophils.1Although eosinophils play a major role in host
defense by combating parasites, they are also central to tissue
damage in several allergic disorders, including asthma and
hypereosinophilic syndrome.2−4Hence, IL5 has been impli-
cated in the pathogenesis of these inflammatory diseases.
IL5 exerts its effects through its cognate receptor on the cell
surface of eosinophils. The IL5 receptor is composed of an α
subunit that binds IL5 specifically and a βc subunit that shares
affinity for IL3 and GM-CSF and is responsible for cytoplasmic
signal transduction. Data from cell FRET and other
biochemical analyses suggest an activation model in which
IL5 triggers cellular responses by sequentially binding first to
IL5Rα followed by binding of the complex to a preassembled
βc subunit.5−8
We and others have investigated the structural elements
underpinning molecular recognition in the complex between
IL5 and IL5Rα and have identified important characteristics of
nterleukin-5 (IL5) is a hematopoietic growth factor that
promotes maturation, proliferation, and activation of
epitope usage. The receptor binding elements in IL5 are
dominated by charged residues in helix B (His38, Lys39, and
His41), the C−D turn (Glu88, Glu89, Arg90, and Arg91), and
helix D (Glu110).9,10For βc binding and signal transduction,
Glu13in helix A is a key residue.10The ligand binding residues
in IL5Rα also are charged. These are located in domain 1
(Asp55, Asp56, and Glu58), domain 2 (Lys186and Arg188), and
domain 3 (Arg297) of the three fibronectin-type III domains of
the extracellular region of IL5Rα.11,12A homology model of the
IL5−IL5Rα complex provided supporting evidence that a pair
of complementary charged interfaces plays an important role in
the specific interaction between IL5 and IL5Rα.12Moreover,
thermodynamic studies suggest that IL5Rα is likely to undergo
significant conformational rearrangement upon ligand binding,
which might be important for βc recruitment and subsequent
Received:
Revised:
Published: July 19, 2011
June 13, 2011
July 17, 2011
Article
pubs.acs.org/biochemistry
© 2011 American Chemical Society
7546
dx.doi.org/10.1021/bi2009135|Biochemistry 2011, 50, 7546−7556

Page 2

receptor activation.13Such complementary charge−charge
interactions and ensuing conformational rearrangement are
likely the key steps in receptor activation.
While the biological function of a protein is primarily
determined by its polypeptide sequence, glycosylation can have
an impact on many aspects of its activity, such as intracellular
trafficking, membrane secretion, folding, structure, stability,
solubility, and interactions with binding partners.14−16IL5Rα is
a glycoprotein that contains four N-glycosylation sites (Asn15,
Asn111, Asn196, and Asn224) in the extracellular region, and it
was shown that ∼14% of the total mass of the extracellular
region of IL5Rα is carbohydrate.13Previously, it was found that
IL5Rα deglycosylated by PNGase F could not bind to IL5,
suggesting the importance of receptor glycosylation for
stabilizing the structure of the receptor itself and/or
receptor−ligand interaction.13However, it has remained
unclear how glycosylation affects the ligand binding activity
of IL5Rα and whether individual glycosylation sites play a more
important role than others. In this study, we investigated the
functional roles of each of the four glycosylation sites in ligand
binding activity, biological function, and cell surface expression
of IL5Rα. Because our initial trial of Asn-to-Gln mutagenesis
caused a significant loss of stability judging from its binding to a
conformationally sensitive antibody, we employed a more
sophisticated protein engineering approach by combining
apolar-to-polar mutations and homology-based mutagene-
sis.17,18This work led to a stable variant in which three of
four glycosylation sites are eliminated (N15D, I109V/V110T/
N111D, and L223R/N224Q). This mostly deglycosylated
variant retained the same ligand binding affinity and biological
activity as fully glycosylated IL5Rα. We demonstrated in this
work that protein engineering is a useful tool for identifying
functionally critical glycosylation sites and deriving tools for
structural analysis of other glycoproteins.
■EXPERIMENTAL PROCEDURES
Materials. Concanavalin A (ConA), bovine serum albumin
(BSA), and mouse interleukin-3 (IL3) were purchased from
Sigma-Aldrich (St. Louis, MO). Human interleukin-5 (IL5) and
cyanovirin-N (CV−N) were produced and purified as
previously described.19,20All the enzymes were purchased
from New England Biolabs (Beverly, MA). All oligo DNA
primers, Drosophila Schneider 2 (S2) cells, cell culture medium,
L-glutamine solution, and RPMI 1640 medium were purchased
from Invitrogen (Carlsbad, CA). The BaF3 cell line was
purchased from German Collection of Microorganisms and
Cell Cultures (DSMZ, Braunschweig, Germany). For surface
plasmon resonance measurements (SPR), sensor chip CM5,
surfactant P20, N-ethyl-N-[3-(dimethylamino)propyl]-
carbodiimide (EDC), N-hydroxysuccinimide (NHS), 1 M
ethanolamine (pH 8.5), and 10 mM glycine hydrochloride
(glycine-HCl) (pH 1.5 and 2.0) were purchased from Biacore
(Piscataway, NJ).
S2 Cell Transient Expression of Soluble IL5Rα. For
mutational deglycosylation of soluble IL5Rα, site-directed
mutagenesis was introduced into pMT-IL5Rα-V5-His12using
a QuikChange site-directed mutagenesis kit (Stratagene). The
presence of the desired mutations was verified by DNA
sequencing. Drosophila S2 cells were transiently transfected
with vector pMT-IL5Rα-V5-His together with pMT-GFP
[expression vector for green fluorescent protein (Invitrogen)]
using Cellfectin reagent (Invitrogen) and grown in serum-free
medium supplemented with 20 mM
expression was induced by addition of 650 μM copper sulfate
3 days after transfection. Cell-free supernatant was collected
after 2 days and stored at −20 °C for the following binding
analysis.
Protein Purification.
Soluble IL5Rα. The soluble forms
of human IL5Rα and its deglycosylated variant with N15D,
I109V/V110T/N111D, and L223R/N224Q mutations (de-
noted IL5Rα[ΔN15/111/224]) were produced in stably
transfected Drosophila S2 cells. The Drosophila expression
vector and blasticidin-resistant vector pCoBlast (Invitrogen)
were cotransfected into S2 cells using Cellfectin reagent
(Invitrogen). Blasticidin-resistant cells were selected as a stable
polyclonal population and grown in serum-free medium
supplemented with 20 mM
blasticidin S (Invitrogen). For expression, the cell density was
expanded to approximately 1 × 107cells/mL and protein
expression was induced by addition of 0.6 mM copper sulfate.
The culture was harvested after 72 h and centrifuged at 1000g
for 10 min. The soluble forms of IL5Rα and the deglycosylated
variant were purified as described previously.7Briefly, the
cleared supernatant was loaded onto a 2B6R-conjugated affinity
column, and bound proteins were eluted with 0.1 M glycine
(pH 2.8). The resulting protein solution was neutralized with a
1 M Tris-HCl solution (pH 9.0), buffer exchanged to PBS,
concentrated, and stored below −80 °C. The ligand binding
activity was confirmed by an SPR binding assay as described
previously.12
Anti-IL5Rα Monoclonal Antibody. 2B6R and α16 are
mouse monoclonal antibodies raised against soluble
IL5Rα.7,21The hybridoma cells were maintained in RPMI
1640 medium with 10% FCS (Hyclone). Antibodies were
purified by using protein G affinity column chromatography
according to a standard procedure.
Protein Characterization. The purity of the protein was
confirmed by Western blot analyses using polyacrylamide gel
electrophoresis in the presence of 0.1% sodium dodecyl sulfate
(SDS−PAGE). The integrity of proteins was verified by matrix-
assisted laser desorption ionization mass spectrometry
(MALDI-MS) (Wistar Institute, Philadelphia, PA). Protein
quantitation was achieved by measuring the absorbance at 280
nm and calculating the concentration using the molar
absorption coefficient using the method of Pace et al.22
SPR Biosensor Binding Analysis. The kinetic interaction
assay was conducted using an SPR biosensor, Biacore 3000
(Biacore Inc., Uppsala, Sweden). All the experiments were
conducted at 25 °C in PBS buffer containing 0.005% P20.
Screening Assay for the Structural Integrity of IL5Rα
Deglycosylation Mutants. We used four different antibodies,
namely, 17b (anti-HIV gp120 mAb23), anti-GFP pAb
(Clontech), anti-V5-tag pAb (Invitrogen), and α16 (anti-
human IL5Rα mAb21). For immobilization of the antibodies on
a CM5 sensor chip, a 1 mg/mL protein stock solution was
diluted 50-fold in 10 mM acetate (pH 4.5) and injected onto a
biosensor surface that had been preactivated with a 1:1 mixture
of 200 mM EDC and 50 mM NHS, followed by the injection of
1 M ethanolamine hydrochloride (pH 8.5). Transiently
expressed IL5Rα mutational variants were injected over the
antibody-immobilized surface. After the surface had been
washed, the amount of bound protein was measured for each
antibody surface. To regenerate chip surfaces, bound proteins
L-glutamine. Protein
L-glutamine and 25 μg/mL
Biochemistry
Article
dx.doi.org/10.1021/bi2009135|Biochemistry 2011, 50, 7546−7556
7547

Page 3

were removed from the antibody surfaces with 10 mM glycine-
HCl (pH 1.6).
Interaction between IL5 and IL5Rα Deglycosylation
Mutants. To determine the binding affinity for IL5, we
employed a previously established “on-chip purification”
method,12in which culture medium containing the V5-tagged
receptor protein was injected over an anti-V5 antibody surface.
Analogous to affinity column chromatography, only the tagged
receptor was precaptured on the chip surface, while all of the
nonspecific components were washed away. Subsequently,
various concentrations of IL5 were injected into the flow cell
containing a particular precaptured IL5Rα deglycosylated
variant as well as a control surface. To regenerate the chip
surfaces, both captured and bound proteins were removed from
the antibody surfaces by double injections of a 10 mM glycine-
HCl solution (pH 1.6). For this sensor assay, all procedures
were automated to create repetitive cycles of injection of cell-
free culture (10 μL/min), 0−40 nM IL5 (50 μL/min), and
regeneration buffer (100 μL/min). The injection time for the
variants was varied to achieve a similar extent of capturing
(approximately 200 RU). For the IL5Rα[ΔN196] variant, the
binding assay was conducted with a higher-density receptor
surface (up to 800 RU).
Lectin Binding and Inhibition Assay. Immobilization of
lectins (ConA and CV-N) on a CM5 sensor chip was
conducted by the amine coupling method as described above.
The real-time interaction was measured by injecting purified
IL5Rα or IL5Rα[ΔN15/111/224] onto these surfaces. Bound
proteins were removed from the surfaces by triple injections of
10 mM glycine-HCl (pH 1.6) after each cycle. All the
procedures were automated to create repetitive cycles of
injection of various concentrations of purified IL5Rα or
IL5Rα[ΔN15/111/224] (flow rate of 50 μL/min), and
regeneration buffer (flow rate of 100 μL/min). To examine
whether CV-N affects the IL5−IL5Rα[ΔN15/111/224]
interaction, different concentrations of IL5 were passed over
the IL5Rα[ΔN15/111/224] captured by CV-N or α16
(control).
Data Analysis. Nonlinear least-squares analysis was used to
calculate the association and dissociation rate constants (kon
and koff, respectively). Prior to the calculation, the binding data
were corrected for nonspecific interaction by subtracting the
reference surface data from the reaction surface data and further
corrected for the buffer effect by subtracting the signal due to
buffer injections from those of protein sample injections.24The
interaction curves thus obtained were globally fit using a model
for 1:1 Langmuir binding (BIAevaluation, Biacore). Individual
kinetic parameters were obtained from at least three separate
experiments. The equilibrium dissociation constant (Kd) was
calculated with the relationship Kd= koff/kon.
Construction of BaF3 Cell Lines Expressing Full-
Length IL5Rα. Mammalian expression vector pcDNA3.1
(Invitrogen) encoding full-length IL5Rα was subcloned by
using pSV-SPORT-IL5Rα25as a template. For full-length
IL5Rα[ΔN196] and full-length IL5Rα[ΔN15/111/224], mu-
tations were introduced into the pcDNA for full-length IL5Rα
Figure 1. Sequence alignment of the extracellular domain of IL5Rα and other cytokine receptors. Multiple-sequence alignments of human and
mouse IL5Rα, human prolactin receptor, and human growth hormone receptor were performed by using ClustalW version 1.6049and further
adjusted manually. Three fibronectin type III domains of IL5Rα are denoted D1, D2, and D3.
Biochemistry
Article
dx.doi.org/10.1021/bi2009135|Biochemistry 2011, 50, 7546−7556
7548

Page 4

using QuikChange site-directed mutagenesis kit (Stratagene).
The mouse pro B lymphocyte IL3-dependent cell line, BaF3,
was maintained in RPMI 1640 medium with 10% FCS
(Hyclone) supplemented with 10 nM mouse IL3. For
transfection, cells (2 × 106) were washed with PBS buffer
and then incubated in 0.8 mL of PBS buffer containing 50 μg of
DNA for 10 min at 4 °C. Electroporation was performed using
Gene Pulser II (Bio-Rad) at 300 V and 975 μF. Subsequently,
the cells were allowed to recover from transfection in the IL3-
supplemented growth medium. Selection was initiated 48 h
after transfection in the IL3-supplemented growth medium
containing 600 μg/mL G418 (Invitrogen). After 8−10 days,
healthy populations of cells were obtained and maintained.
Cell Proliferation Assay. The assay was conducted using
BaF3 cells expressing full-length IL5Rα. These cells were
washed four times in RPMI 1640 medium with 10% FCS
(Hyclone) to remove residual mouse IL3 from the culture
medium. Plates with 96 wells were seeded with 5000 cells/well
(50 μL) and incubated for 48 h (37 °C and 5% CO2) in the
presence of different concentrations of human IL5 (50 μL).
Following the 48 h incubation, 10 μL of WST-1 {4-[3-(4-
iodophenyl)-2-(4-nitrophenyl)]-2H-5-tetrazolio-1,3-benzene
disulfonate (Roche Diagnostics)} was added to each well and
incubated for 4 h (37 °C). The plate was then read using a
microplate reader at an absorbance of 450 nm with a 650 nm
reference. Individual values were obtained from three experi-
ments, and the average values with the standard deviation were
plotted using Excel (Microsoft). To obtain the ED50value, the
average values were fitted to a sigmoidal dose−response model
using Prism (GraphPad).
Flow Cytometry. For staining, cells were suspended in
staining buffer (PBS with 10% BSA) and incubated with 20 μg/
mL α16 or 2B6R as the primary antibody or 17b as an isotype-
matched control, for 30 min at 4 °C, followed by incubation
with 20 μg/mL Alexa Fluor 488-conjugated goat anti-mouse
secondary antibody (Invitrogen) for 45 min at 4 °C. Cells were
washed with staining buffer between antibody incubations.
Flow cytometry analysis was performed using a FACSORT
flow cytometer (BD Biosciences, San Diego, CA) with 488 nm
excitation from an argon ion laser at 15 mW. The forward
scatter threshold was set to exclude small debris. Alexa Fluor
488 log fluorescence was captured on the FL1 channel
equipped with a 530 nm wavelength filter with a 30 nm
bandwidth. Data were acquired using Lysis II (version 2.0, BD
Biosciences). At least 10000 events were acquired per sample.
Data analysis was performed using WinMDI (version 2.8, J.
Trotter, Scripps Research Institute, La Jolla, CA, available at
http://www.facs.scripps.edu).
Homology Modeling. The three-dimensional structure of
IL5Rα[ΔN15/111/224] was modeled using MODELLER
6v2.26The previously reported modeled structure of wild-
type IL5Rα12was used as a template. Basic steric principles
such as close contacts or violation of stereochemistry were
examined using PROCHECK,27and the three-dimensional
profiles of the models were examined using ProsaII.28On the
basis of these quality checks, the best modeled structure was
chosen.
■RESULTS
Assessment of the Structural Integrity of IL5Rα. To
define the functional roles of each N-linked carbohydrate
moiety at Asn15, Asn111, Asn196, and Asn224of IL5Rα (Figure
1), we performed site-directed mutagenesis at the N-
glycosylation sites and investigated both the ligand binding
activity and the structural integrity of the mutants. The
extracellular region of IL5Rα was fused with the V5-tag peptide
at its C-terminus and was expressed in S2 cells as described
previously.12Prior to functional analyses of deglycosylation
mutants of IL5Rα, we first tested whether the mutations could
affect the structural integrity of IL5Rα. We previously found
that an anti-IL5Rα mAb α1621recognizes a three-dimensional
epitope of IL5Rα.12In this study, we demonstrated that the
soluble form of IL5Rα protein denatured via SDS−PAGE was
detected by anti-IL5Rα mAb, 2B6R, and anti-V5-tag antibody
while it was not detected by α16 (Figure 2A). These findings
suggest that α16 is a conformationally sensitive antibody. We
developed a high-throughput assay in which α16 was
immobilized on a biosensor surface to examine the stability
of transiently expressed proteins. The anti-V5-tag antibody was
also immobilized onto a different biosensor chip surface to
normalize the expression level of the protein. As a proof of
concept, we demonstrated that heat-denatured purified IL5Rα
was bound to the anti-V5-tag antibody but failed to bind to α16
(Figure 2B), whereas native IL5Rα was bound to both α16 and
the anti-V5-tag antibody (Figure 2C). Thus, we employed this
method to verify the structural integrity of various deglycosy-
lated mutants of IL5Rα.
Design of Mutationally Deglycosylated IL5Rα Var-
iants. We first used diverse mutations to identify conforma-
tion-preserving changes at Asn15, Asn111, Asn196, and Asn224,
the four glycosylation sites in IL5Rα. For replacement of
asparagine, we chose four charged residues (aspartic acid,
glutamic acid, lysine, and arginine), one neutral residue
(glutamine), and one hydrophobic residue (valine), all of
which have molecular volumes similar to that of asparagine.
Deglycosylated mutants were transiently expressed and
subjected to the structural integrity assay described above.
The data are shown in Figure 3. Mutations such as N15D,
N111D, and N224K displayed significantly increased stabilities
relative to those of Asn-to-Gln mutations (Figure 3A). Of note,
replacement with valine at any of the glycosylation sites largely
reduced the level of α16 binding, indicating that the four N-
glycosylation sites are exposed to the solvent, and thus, the
hydrophobic residue introduced at these positions might
destabilize the structure of IL5Rα. Because replacement of
Asn196with any of these six residues severely reduced stability,
we expanded mutations of Asn196to all other amino acid
residues except cysteine, which could cause dimerization of
IL5Rα. We found that the Asn-to-Thr mutation was the best
among those replacements (Figure 3B).
Second, we designed several mutations on the basis of a
hypothesis that apolar-to-polar mutations on the protein
surface should increase the stability of the protein. We chose
the hydrophobic residues that are exposed to solvent and
adjacent to the N-glycosylation sites in the modeled structure
of IL5Rα12and generated double or triple mutations in those
regions based on sequence alignment with mouse IL5Rα,
human prolactin receptor, and human growth hormone
receptor (Figure 1). The extracellular domains of prolactin
and growth hormone receptors were produced in Escherichia
coli, and their crystal structures in complex with their ligand
have been determined,29,30indicating the structural stability of
the deglycosylated forms. We found that mutants such as
I109V/V110T/N111Q, I109V/V110T/N111D, and L223R/
Biochemistry
Article
dx.doi.org/10.1021/bi2009135|Biochemistry 2011, 50, 7546−7556
7549

Page 5

N224Q had almost the same stability as the fully glycosylated
form (Figure 3C).
Ligand Binding Activities of Deglycosylated Mu-
tants. For the ligand binding assay, we chose the most stable
deglycosylated mutants at each site. These were N15D
(denoted IL5Rα[ΔN15]), I109V/V110T/N111D (denoted
IL5Rα[ΔN111]), N196T (denoted IL5Rα[ΔN196]), and
L223R/N224Q (denoted IL5Rα[ΔN224]). To validate the
kinetic interaction with IL5, we employed a previously
established SPR biosensor assay.12Briefly, different concen-
trations of IL5 were injected over the IL5Rα variants that had
been precaptured via their C-terminal V5-tag by the anti-V5-tag
antibody immobilized covalently on the biosensor surface
(Figure 4). Kinetic analysis showed that deglycosylation
variants such as IL5Rα[ΔN15], IL5Rα[ΔN111], and IL5Rα-
[ΔN224] had almost the same kinetic profiles as fully
glycosylated IL5Rα, whereas IL5Rα[ΔN196] showed a 5-fold
increase in its Kd value because of the combination of a
decreased konand an increased koff(Table 1). Our attempt to
obtain a completely deglycosylated IL5Rα by using these four
mutations resulted in the complete loss of ligand binding and
α16 binding (data not shown). Because the N196T mutation is
most likely the mutation that destabilized the structure of
IL5Rα (Figure 3), we performed triple-mutational deglycosy-
lation with N15D, I109V/V110T/N111D, and L223R/N224Q
but maintained Asn196glycosylation to obtain the mostly
deglycosylated form of IL5Rα. This variant, which had three of
four glycosylation sites eliminated (denoted IL5Rα[ΔN15/
111/224]), showed the same ligand binding affinity and kinetic
parameters as fully glycosylated IL5Rα (Figure 4 and Table 1).
Effect of Lectin on the IL5−IL5Rα Interaction. We
conducted further biochemical characterizations of purified
IL5Rα[ΔN15/111/224] to elucidate the functional role of the
carbohydrate moiety at Asn196. MALDI-MS analysis of purified
IL5Rα showed that the mass of total carbohydrates was 5873
Da, that is, 14% of the total mass of IL5Rα (Table 2). On the
other hand, the mass of total carbohydrates of IL5Rα[ΔN15/
111/224] was 2344 Da, that is, 6% of the total mass of
IL5Rα[ΔN15/111/224]. These results indicate that a single
carbohydrate moiety at Asn196occupies as much as 40% of the
total carbohydrate content of IL5Rα. To examine the
glycosylation at Asn196, we characterized the lectin binding
activity of IL5Rα[ΔN15/111/224] using an SPR biosensor. In
this assay, either concanavalin A (ConA) or cyanovirin-N (CV-
N) was immobilized on the biosensor surface, and then either
purified IL5Rα[ΔN15/111/224] or purified IL5Rα was
injected over the lectin surface. We found that the lectin
binding affinity of IL5Rα[ΔN15/111/224] was lower than that
of fully glycosylated IL5Rα (Figure S1 of the Supporting
Information), probably because of the reduced carbohydrate
content of IL5Rα[ΔN15/111/224]. Because CV-N selectively
binds to high-mannose oligosaccharides, the carbohydrate
moiety at Asn196might be high-mannose, which is consistent
with the previous observations on the carbohydrate structure of
IL5Rα.13
To test whether the carbohydrate at Asn196is involved in the
IL5−IL5Rα interaction, we examined binding of IL5 to
IL5Rα[ΔN15/111/224] in the presence of CV-N. In this
assay, purified IL5Rα[ΔN15/111/224] was captured by CV-N
immobilized on the biosensor surface and then nonglycosylated
IL5 was injected over the receptor surface. We found that
IL5Rα[ΔN15/111/224] binds IL5 with high affinity even when
the single carbohydrate group at Asn196is occupied by binding
to CV-N (Figure S2 of the Supporting Information). This IL5
binding activity suggests that the carbohydrate at Asn196of
IL5Rα is not involved in a direct contact with IL5.
Biological Activity and Cell Surface Expression. We
asked whether N-glycosylation plays a role in the cellular
function of IL5Rα. Plasmids encoding the full length of the two
deglycosylated variants (IL5Rα[ΔN196] and IL5Rα[ΔN15/
111/224]) as well as wild-type IL5Rα were stably transfected
into BaF3 cells, and IL5-induced cell proliferation in these cells
Figure 2. Characterization of a conformationally sensitive anti-IL5Rα
antibody. (A) Western blotting of the expressed soluble domain of
IL5Rα. The conditioned medium containing soluble IL5Rα was mixed
with SDS buffer, boiled for 10 min, and loaded onto an SDS−PAGE
gel: lane 1, band stained with anti-IL5Rα mAb (R&D Systems); lane 2,
anti-V5-tag pAb (Invitrogen); lane 3, anti-IL5Rα mAb α16; lane 4,
anti-IL5Rα mAb 2B6R. (B and C) Injection of the heat-deactivated
form of purified IL5Rα (B) or the native form of purified IL5Rα (C)
at 0 s, followed by injection of running buffer at 120 s. In this assay,
either form of IL5Rα protein was injected over the antibody surfaces,
and binding of IL5Rα to each antibody was monitored. Black lines
show the sensorgrams for anti-V5-tag pAb and red lines those for α16.
Biochemistry
Article
dx.doi.org/10.1021/bi2009135|Biochemistry 2011, 50, 7546−7556
7550

Page 6

was investigated. BaF3 is an IL3-dependent mouse pro B cell
line that intrinsically expresses the βc receptor subunit. Because
human IL5 and IL5Rα can cross-react with mouse βc, this cell
line has been used for the functional analyses of IL5Rα
mutational variants.25Cells transfected with full-length IL5Rα-
[ΔN15/111/224] exhibited IL5-induced cell proliferation
similar to that found with fully glycosylated IL5Rα (Figure
5), indicating that N-glycosylations at Asn15, Asn111, and Asn222
of IL5Rα are not important for the cell proliferation activity of
the receptor. On the other hand, cells transfected with full-
length IL5Rα[ΔN196] exhibited no proliferation (Figure 5A).
To determine whether the N196T mutation completely
suppressed the biological function because of the absence of
the receptor on the cell surface, we assessed the cell surface
expression of the deglycosylated mutant with anti-IL5Rα
antibodies. Flow cytometry analysis revealed no signal shift in
BaF3 cells transfected with full-length IL5Rα[ΔN196], whereas
signal shifts were detected in BaF3 cells transfected with either
fully glycosylated IL5Rα or IL5Rα[ΔN15/111/224] (Figure
5B). These data indicate that IL5Rα[ΔN196] was not
expressed on the cell. Hence, N-glycosylation at Asn196might
be required for transport of IL5Rα to the cell surface.
■DISCUSSION
We have previously shown the effects of amino acid
replacement of human IL5Rα on ligand binding and receptor
antagonism.12,31,32The main purpose of this study was to
provide insights into the specific roles of N-glycosylation of
IL5Rα in its structure and function. IL5Rα contains four
consensus N-glycosylation sites (Asn-X-Ser/Thr) at Asn15,
Asn111, Asn196, and Asn224in the extracellular region (Figure
1). In general, N-glycosylation of proteins can play an
important role in a variety of aspects, such as ligand binding,
protein folding, trafficking, and stability.14−16,33,34We have
previously found that enzymatic deglycosylation of IL5Rα
weakened its ability to bind IL5.13However, it has remained
unclear whether all of the N-glycosylations are involved in
ligand binding or if only a specific subset of N-glycosylation
sites is necessary. It also was unclear what caused such a drastic
loss of ligand binding. To answer these questions, we
investigated the structural and functional roles of N-
glycosylation sites by site-directed mutagenesis. We initially
introduced an Asn-to-Gln mutation as one of the general
approaches to mutational deglycosylation. However, we found
that such a mutation caused significant decreases in the level of
protein expression (data not shown) as well as losses of binding
to its conformationally sensitive antibody (Figure 3), suggesting
that introduction of glutamine at these points destabilizes the
structure of IL5Rα. Therefore, we evaluated the ability of a
series of combinatorial and rationally designed mutations to
preserve the stability of deglycosylated mutants of IL5Rα. The
former approach is based on a hypothesis that asparagine
residues at the N-glycosylation sites can be replaced with
certain charged residues while maintaining the stability of the
protein.18The latter approach is based on a finding that the
Figure 3. Screening assay of the structural integrity of soluble forms of deglycosylation mutants. In this assay, culture supernatant containing IL5Rα
mutational variants was injected over the immobilized antibody α16 SPR chip surface. This antibody surface was used to examine the stability of
transiently expressed proteins, while the anti-V5-tag antibody was used to normalize the expression level of the protein. Thus, the ratio of α16
binding activity to anti-V5 activity denotes the extent of conformational intactness of the expressed proteins. The percent α16 binding level was
calculated by dividing amounts of α16-binding proteins by those of anti-V5-tag antibody and then by dividing the resulting values of mutants by that
of wild-type IL5Rα. Individual values were obtained from three experiments. The average values are shown as bars, and standard deviations are
shown as lines.
Biochemistry
Article
dx.doi.org/10.1021/bi2009135|Biochemistry 2011, 50, 7546−7556
7551

Page 7

stability and solubility of glycan-free forms of glycoproteins can
be improved by using apolar-to-polar mutations of any surface
residue in functionally irrelevant epitopes.17We utilized a
conformationally sensitive anti-IL5Rα mAb as a marker to
verify the structural integrity of IL5Rα and performed a
medium-throughput assay of all the deglycosylated mutants
with a surface plasmon resonance (SPR) biosensor (Figure 2).
From these two protein engineering approaches, we
generated a total of 54 deglycosylation mutations (Figure 3).
We found that mutations such as N15D, N111D, and N224R
displayed significantly increased structural stability relative to
the Asn-to-Gln mutation, while any replacement at Asn196
severely reduced stability (Figure 3A,B). Interestingly, Asn-to-
Asp mutations were more stable than Asn-to-Glu or Asn-to-Gln
mutations at Asn15, Asn111, and Asn224, suggesting that not only
the charge of an amino acid but also its molecular volume
might be important for the stability of local structure. These
data imply that the Asn-to-Gln mutation, which is generally
chosen for deglycosylation because of the retention of
electrostatic properties, may not be the best strategy for all
glycoproteins. Although any replacement at Asn196with
charged residues severely reduced stability, substitution with
threonine gave the best result (Figure 3B). It should be noted
that threonine is the residue at the corresponding position of
the human prolactin receptor in our sequence alignment
(Figure 1). From rationally designed mutations, we found that
IL5Rα[ΔN111] (I109V/V110T/N111D) and IL5Rα[ΔN224]
(L223R/N224Q) had almost the same structural stability as the
fully glycosylated form of IL5Rα (Figure 3C).
Because IL5Rα[ΔN15], IL5Rα[ΔN111], and IL5Rα-
[ΔN224] mutants have almost the same stability and binding
affinity for IL5 (Figure 4 and Table 1), we engineered a
deglycosylated variant from which three of four glycosylation
sites have been eliminated. This mostly deglycosylated variant
(denoted IL5Rα[ΔN15/111/224]) also exhibited the same
binding affinity for IL5 as fully glycosylated IL5Rα. We sought
a rationale for these stability-preserving mutations in a
homology-modeled structure of IL5Rα[ΔN15/111/224]. We
found that mutations such as N15D and L223R/N224Q might
stabilize the structure through interactions with neighboring
charged residues. In the model, the side chain of Asp15appears
to interact with the side chain of Lys19(Figure 6A), and
therefore, the Asn-to-Gln mutation at this position would be
expected to lead to a loss of a favorable charge interaction. The
side chains of Arg223and Gln224appear to stabilize the
interstrand conformation by interacting with the side chain of
Glu239in the neighboring strand (Figure 6B). In contrast, no
favorable interactions of mutated residues in the N111 region
(I109V/V110T/N111Q) were observed with neighboring
residues in the modeled structure, even though the I109V/
V110T/N111Q mutant preserved the stability to the same
extent as fully glycosylated IL5Rα. Perhaps the combination of
apolar and polar groups in this locus is sufficient to contribute
entropically to the stabilization of the protein. One could
imagine that Asn-to-Gln mutation alone at positions 111 and
Figure 4. Ligand binding of deglycosylated variants. The soluble forms
of wild-type IL5Rα and five deglycosylation mutants were transiently
expressed in S2 cells, and the supernatants were injected over the anti-
V5-tag antibody biosensor chip surface on the biosensor (“on-chip
purification”). Subsequently, various concentrations (5, 10, 20, and 40
nM) of IL5 were injected at 0 s, followed by injection of running buffer
alone at 120 s. The rate constants were calculated by globally fitting
the association phase (0−120 s) and the dissociation phase (120−360
s) to a model for 1:1 Langmuir binding. Red lines show calculated
sensorgrams. Individual kinetic parameters were obtained from three
separate experiments. The average and standard deviation are listed in
Table 1.
Table 1. Kinetic Interaction between IL5 and Soluble Forms of IL5Rα Deglycosylated Variants
mutationskon(×10−6M−1s−1)
1.2 ± 0.3
1.0 ± 0.8
1.1 ± 0.7
0.47 ± 0.3
1.0 ± 0.4
1.0 ± 0.3
koff (×103s−1)
3.1 ± 0.1
2.5 ± 0.1
3.0 ± 0.2
7.1 ± 0.1
3.3 ± 0.2
3.4 ± 0.2
Kd(×109M)
2.7 ± 0.1
2.5 ± 0.2
2.7 ± 0.1
15 ± 1
3.3 ± 0.1
3.2 ± 0.2
wild-type IL5Rα
IL5Rα[ΔN15]
IL5Rα[ΔN111]
IL5Rα[ΔN196]
IL5Rα[ΔN224]
IL5Rα[ΔN15/111/224]
N15D
I109V/V110T/N111D
N196T
L223R/N224Q
N15D, I109V/V110T/N111D, and L223R/N224Q
Table 2. Comparison of Carbohydrate Contents of Wild-
Type and Mostly Deglycosylated IL5Rα Determined by
MALDI-TOF MS
MS(calc) MS(exp)
ΔMS(exp−
calc)
ΔMS(wt
mutant)
−
3529
IL5Rα
IL5Rα[ΔN15/111/
224]
35877
36447
41750
38791
5873
2344
Biochemistry
Article
dx.doi.org/10.1021/bi2009135|Biochemistry 2011, 50, 7546−7556
7552

Page 8

224 might not be sufficient to stabilize the local conformation
without other favorable interactions or increased hydrophilicity
at nearby mutational sites. In any case, the model suggests that
some carbohydrate moieties of IL5Rα could be replaced
without compromising the stability and function of IL5Rα
protein.
IL5Rα is a member of the class I cytokine receptor
superfamily, members of which are characterized by the
presence of the so-called cytokine recognition motif.35
Extensive structural studies have been performed for various
cytokine−receptor complexes in this superfamily (reviewed in
references 36 and 37). In some cases (e.g., G-CSF38), crystal
structures of the cytokine−receptor complex have been
determined without deglycosylation. However, early trials to
obtain crystals of the fully glycosylated IL5Rα in complex with
deglycosylated IL5 proved unsuccessful.13Because carbohy-
drate moieties of glycoproteins affect crystal formation,
deglycosylation prior to crystallization could improve the
chances for obtaining high-quality crystals. For deglycosylation
of the cytokine receptors, the following four methods have been
used: bacterial expression (e.g., growth hormone,29prolactin,30
and IL439), enzymatic deglycosylation (e.g., IL1040and
IL1241), mutational deglycosylation (e.g., erythropoietin42and
IL643), and inhibition of glycosidase (e.g., IL244). Most
recently, Hansen et al. have successfully determined the crystal
structure of the receptor complex of GM-CSF, which is the
closest homologue of IL5, by using GM-CSF from bacterial
expression, a fully glycosylated α subunit, and a βc subunit with
Asn-to-Gln mutational deglycosylation.45Here, we utilized site-
directed mutagenesis and engineered IL5Rα[ΔN15/111/224],
which retained Asn196, while eliminating the other three
glycosylation sites of IL5Rα. This mostly deglycosylated variant
might facilitate crystallization of the IL5−receptor complex, and
the stability-preserving deglycosylation strategies presented
here could be applied to crystallographic analyses of other
glycoproteins.
Our data demonstrated that deglycosylation at Asn196of
IL5Rα severely reduced its stability and binding affinity for IL5
Figure 5. Cell proliferation activity and cell surface expression of full-
length IL5Rα deglycosylation mutants. (A) The IL3-dependent mouse
BaF3 cell line was used to test the bioactivity of exogenously expressed
full-length IL5Rα (▲), full-length IL5Rα[ΔN15/111/224] (◼), full-
length IL5Rα[ΔN196] (◆), and empty vector (○). Five thousand
cells per well were incubated with various dilutions (0.01 fM to 40
nM) of IL5. After incubation for 48 h, proliferation was evaluated as
described in Experimental Procedures. The average values of triplicate
wells are plotted with standard deviations. (B) Cytofluorometric
analysis of cellular surface expression of the wild type and mutant of
full-length IL5Rα on BaF3 transfectants. Cells were incubated with
anti-IL5Rα antibodies or negative control (17b) as the primary
antibody and Alexa-conjugated goat anti-mouse pAb (Invitrogen) as
the secondary antibody. Histogram plots from 2B6R (gray area) show
the expression of full-length forms of wild-type IL5Rα (wt) and
IL5Rα[ΔN15/111/224] (ΔN15/111/224) compared to no expres-
sion of empty vector (mock) and IL5Rα[ΔN196] (ΔN196). The
negative control indicates staining with the secondary antibody alone
(black lines).
Figure 6. Potential charged interactions and hydrogen bonds
introduced by the mutations at N-glycosylation sites. The homol-
ogy-modeled structure of IL5Rα[ΔN15/111/224], based on a
previously reported model of IL5Rα, suggests potential favorable
interactions upon deglycosylation mutation at Asn15(A) and Asn224
(B). The residues of interest are shown as a CPK model. Oxygen and
nitrogen atoms are colored red and blue, respectively. All the
molecular graphics in this article were prepared with RasMol.50
Biochemistry
Article
dx.doi.org/10.1021/bi2009135|Biochemistry 2011, 50, 7546−7556
7553

Page 9

(Figure 3 and Figure 4). We also observed that IL5Rα[ΔN196]
lost ligand binding activity over time (data not shown). These
findings inspired us to test whether deglycosylation at Asn196of
IL5Rα caused a reduction in the ligand binding activity not
because the carbohydrate is involved in the direct interaction
with the ligand but as a secondary effect of the partially
destabilized structure of IL5Rα upon deglycosylation. We have
previously shown that IL5-binding residues are located in the
D1 domain (Asp55, Asp56, and Glu58) as well as in the D2
domain (Lys186and Arg188) and the D3 domain (Arg297).12
The binding interface of IL5Rα comprises a cluster of
negatively charged residues [site I epitope (Figure 7A)] from
the D1 domain and a cluster of positively charged residues [site
II epitope (Figure 7A)] from the D2D3 tandem domain. The
homology-deduced IL5Rα structure shows that Asn196in the
D2 domain is not located in the proximity of either site I or II
epitopes (Figure 7A). The four N-glycosylation sites of IL5Rα
are conserved in GM-CSF receptor α, which shares a high
degree of amino acid sequence similarity with IL5Rα as well as
signaling receptor βc (sequence alignment available in reference
32). The crystal structure of the GM-CSF receptor complex
showed no direct interaction of the receptor carbohydrate and
ligand.45Moreover, our lectin inhibition assay showed that
neither cyanovirin-N nor concanavalin-A had any impact on the
interaction of IL5Rα[ΔN15/111/224] with deglycosylated IL5
(Figure S2 of the Supporting Information). From these
observations, we conclude that the carbohydrate at Asn196is
less likely involved in the direct interaction between IL5 and
IL5Rα and instead likely plays a critical role in maintaining the
correct structure of the receptor. Interestingly, Yoon et al.41
found that a carbohydrate extending from Asn200of the D2
domain lies in the interface of the D1 and D2 domains in the
crystal structure of the p35−p40 heterodimer of IL12 (Figure
7B). They have proposed that the carbohydrate stabilizes the
structure by fixing the relative orientation of these domains
through hydrogen bonds and van der Waals contacts. It is
noteworthy that the spatial position of Asn196of IL5Rα is
similar to that of Asn200of IL12 (Figure 7). Although the
detailed structural role of the carbohydrate at Asn196of IL5Rα
must await high-resolution structure determination of the IL5−
IL5Rα[ΔN15/111/224] complex by X-ray crystallography, it is
tempting to speculate that the carbohydrate at Asn196might
stabilize the overall structure of the receptor by bridging the D1
and D2 domains as in the IL12 system.
We developed a medium-throughput screening assay to
characterize the structural integrity and ligand binding affinity
of the extracellular domain of IL5Rα by using the transient
expression system with Drosophila S2 insect cells. Although
insect expression facilitates overexpression and screening of a
large number of recombinant proteins, it is known that the
carbohydrate components of N-glycosylation of insect cells are
different from those of mammalian cells.46,47To verify the role
of N-glycosylation of IL5Rα in mammalian cells, we used the
BaF3 cell line, a mouse pro B cell line that has been used for
functional analyses of human IL5Rα.25,48Using BaF3 cells
expressing full-length IL5Rα and its mutational variants, we
found that deglycosylation at Asn196resulted in the failure of
cell surface expression of IL5Rα (Figure 5B) and caused a
complete loss of the biological response to IL5 (Figure 5A). In
light of the finding that the carbohydrate at Asn196is likely
indispensable for the correct structure of the receptor (see
above), it is plausible that Asn196deglycosylation causes the
failure of cell surface expression of IL5Rα[ΔN196] by
destabilizing the structure of IL5Rα.
In conclusion, this study reveals the essential role of a single
carbohydrate moiety at Asn196in the structure and function of
IL5Rα. Our unique protein engineering approach also
demonstrated that most, but not all, N-linked carbohydrates
can be eliminated without compromising the function of the
protein via introduction of targeted mutations at and around N-
glycosylation sites. We obtained a stable and functional variant
that has three of four glycosylation sites eliminated. The mostly
Figure 7. Maps of the N-glycosylation sites and ligand binding epitopes on the homology-modeled structure of IL5Rα compared with the crystal
structure of IL12. (A) Modeled structure of IL5Rα.12(B) Crystal structure of IL12 p40.41Three fibronectin type III domains (D1, D2, and D3) are
denoted from the N- to C-terminus. Asn196of IL5Rα and Asn200of IL12 are colored green, and other asparagine residues for N-glycosylation are
colored yellow. The IL5-binding residues at site I and II epitopes, which have been found previously,12,32are shown as a CPK model. The loop
regions are represented as coils and β strands as ribbons. Carbohydrate groups are shown as stick models. GlcNAc and Man stand for N-
acetylglucosamine and mannose, respectively.
Biochemistry
Article
dx.doi.org/10.1021/bi2009135|Biochemistry 2011, 50, 7546−7556
7554

Page 10

deglycosylated form we obtained in this study could open the
way to improving high-resolution structural analysis of the IL5
system and help in the design of new compounds for
therapeutic treatment of IL5-related allergic inflammatory
diseases.
■ASSOCIATED CONTENT
*
Biosensor data of kinetic interaction of lectin and IL5Rα
deglycosylation variants. This material is available free of charge
via the Internet at http://pubs.acs.org.
■AUTHOR INFORMATION
Corresponding Author
*E-mail: Irwin.Chaiken@DrexelMed.edu. Phone: (215) 762-
4197. Fax: (215) 762-4452.
Present Address
⊥Pfizer Global Research & Development, 87 Cambridgepark
Dr., Cambridge, MA 02140.
Funding
This work was supported by National Institutes of Health
Grant AI40462.
■ACKNOWLEDGMENTS
We thank Dr. Keith A. Vosseller (Drexel University College of
Medicine) for the useful discussion.
■ABBREVIATIONS
IL5Rα, human interleukin-5 receptor α; GM-CSF, granulocyte/
macrophage colony-stimulating factor; PBS, phosphate-buffered
saline; SPR, surface plasmon resonance; RU, resonance unit;
GFP, green fluorescent protein; ConA, concanavalin A; CV-N,
cyanovirin-N; mAb, monoclonal antibody; pAb, polyclonal
antibody; ΔN15, N15D mutant of IL5Rα; ΔN111, I109V/
V110T/N111D mutant of IL5Rα; ΔN196, N196T mutant of
IL5Rα; ΔN224, L223R/N224Q mutant of IL5Rα; ΔN15/111/
224, N15/I109V/V110T/N111D/L223R/N224Q mutant of
IL5Rα.
■REFERENCES
(1) Karlen, S., De Boer, M. L., Lipscombe, R. J., Lutz, W.,
Mordvinov, V. A., and Sanderson, C. J. (1998) Biological and
molecular characteristics of interleukin-5 and its receptor. Int. Rev.
Immunol. 16, 227−247.
(2) Sanderson, C. J., and Urwin, D. (2000) Interleukin-5: A drug
target for allergic diseases. Curr. Opin. Invest. Drugs 1, 435−441.
(3) Foster, P. S., Hogan, S. P., Yang, M., Mattes, J., Young, I. G.,
Matthaei, K. I., Kumar, R. K., Mahalingam, S., and Webb, D. C. (2002)
Interleukin-5 and eosinophils as therapeutic targets for asthma. Trends
Mol. Med. 8, 162−167.
(4) Rothenberg, M. E. (2004) Eosinophilic gastrointestinal disorders
(EGID). J. Allergy Clin. Immunol. 113, 11−28.
(5) Zaks-Zilberman, M., Harrington, A. E., Ishino, T., and Chaiken, I.
M. (2008) Interleukin-5 receptor subunit oligomerization and
rearrangement revealed by fluorescence resonance energy transfer
imaging. J. Biol. Chem. 283, 13398−13406.
(6) Ishino, T., Harrington, A. E., Zaks-Zilberman, M., Scibek, J. J.,
and Chaiken, I. (2008) Slow-dissociation effect of common signaling
subunit βc on IL5 and GM-CSF receptor assembly. Cytokine 42, 179−
190.
(7) Scibek, J. J., Evergren, E., Zahn, S., Canziani, G. A., Van Ryk, D.,
and Chaiken, I. M. (2002) Biosensor analysis of dynamics of
S Supporting Information
interleukin 5 receptor subunit βc interaction with IL5:IL5Rα
complexes. Anal. Biochem. 307, 258−265.
(8) Carr, P. D., Gustin, S. E., Church, A. P., Murphy, J. M., Ford, S.
C., Mann, D. A., Woltring, D. M., Walker, I., Ollis, D. L., and Young, I.
G. (2001) Structure of the complete extracellular domain of the
common β subunit of the human GM-CSF, IL-3, and IL-5 receptors
reveals a novel dimer configuration. Cell 104, 291−300.
(9) Morton, T., Li, J., Cook, R., and Chaiken, I. (1995) Mutagenesis
in the C-terminal region of human interleukin 5 reveals a central patch
for receptor α chain recognition. Proc. Natl. Acad. Sci. U.S.A. 92,
10879−10883.
(10) Tavernier, J., Tuypens, T., Verhee, A., Plaetinck, G., Devos, R.,
Van der Heyden, J., Guisez, Y., and Oefner, C. (1995) Identification of
receptor-binding domains on human interleukin 5 and design of an
interleukin 5-derived receptor antagonist. Proc. Natl. Acad. Sci. U.S.A.
92, 5194−5198.
(11) Cornelis, S., Plaetinck, G., Devos, R., Van der Heyden, J.,
Tavernier, J., Sanderson, C. J., Guisez, Y., and Fiers, W. (1995)
Detailed analysis of the IL-5-IL-5Rα interaction: Characterization of
crucial residues on the ligand and the receptor. EMBO J. 14, 3395−
3402.
(12) Ishino, T., Pasut, G., Scibek, J., and Chaiken, I. (2004) Kinetic
interaction analysis of human interleukin 5 receptor α mutants reveals
a unique binding topology and charge distribution for cytokine
recognition. J. Biol. Chem. 279, 9547−9556.
(13) Johanson, K., Appelbaum, E., Doyle, M., Hensley, P., Zhao, B.,
Abdel-Meguid, S. S., Young, P., Cook, R., Carr, S., Matico, R., et al.
(1995) Binding interactions of human interleukin 5 with its receptor α
subunit. Large scale production, structural, and functional studies of
Drosophila-expressed recombinant proteins. J. Biol. Chem. 270, 9459−
9471.
(14) West, C. M. (1986) Current ideas on the significance of protein
glycosylation. Mol. Cell. Biochem. 72, 3−20.
(15) Helenius, A. (1994) How N-linked oligosaccharides affect
glycoprotein folding in the endoplasmic reticulum. Mol. Biol. Cell 5,
253−265.
(16) Niu, L., Heaney, M. L., Vera, J. C., and Golde, D. W. (2000)
High-affinity binding to the GM-CSF receptor requires intact N-
glycosylation sites in the extracellular domain of the β subunit. Blood
95, 3357−3362.
(17) Sun, Z. Y., Dotsch, V., Kim, M., Li, J., Reinherz, E. L., and
Wagner, G. (1999) Functional glycan-free adhesion domain of human
cell surface receptor CD58: Design, production and NMR studies.
EMBO J. 18, 2941−2949.
(18) Narhi, L. O., Arakawa, T., Aoki, K., Wen, J., Elliott, S., Boone,
T., and Cheetham, J. (2001) Asn to Lys mutations at three sites which
are N-glycosylated in the mammalian protein decrease the aggregation
of Escherichia coli-derived erythropoietin. Protein Eng. 14, 135−140.
(19) Wu, S. J., Tambyraja, R., Zhang, W., Zahn, S., Godillot, A. P.,
and Chaiken, I. (2000) Epitope randomization redefines the functional
role of glutamic acid 110 in interleukin-5 receptor activation. J. Biol.
Chem. 275, 7351−7358.
(20) McFadden, K., Cocklin, S., Gopi, H., Baxter, S., Ajith, S.,
Mahmood, N., Shattock, R., and Chaiken, I. (2007) A recombinant
allosteric lectin antagonist of HIV-1 envelope gp120 interactions.
Proteins 67, 617−629.
(21) Tavernier, J., Van der Heyden, J., Verhee, A., Brusselle, G., Van
Ostade, X., Vandekerckhove, J., North, J., Rankin, S. M., Kay, A. B.,
and Robinson, D. S. (2000) Interleukin 5 regulates the isoform
expression of its own receptor α-subunit. Blood 95, 1600−1607.
(22) Pace, C. N., Vajdos, F., Fee, L., Grimsley, G., and Gray, T.
(1995) How to measure and predict the molar absorption coefficient
of a protein. Protein Sci. 4, 2411−2423.
(23) Dowd, C. S., Leavitt, S., Babcock, G., Godillot, A. P., Van Ryk,
D., Canziani, G. A., Sodroski, J., Freire, E., and Chaiken, I. M. (2002)
β-turn Phe in HIV-1 Env binding site of CD4 and CD4 mimetic
Biochemistry
Article
dx.doi.org/10.1021/bi2009135|Biochemistry 2011, 50, 7546−7556
7555

Page 11

miniprotein enhances Env binding affinity but is not required for
activation of co-receptor/17b site. Biochemistry 41, 7038−7046.
(24) Myszka, D. G. (2000) Kinetic, equilibrium, and thermodynamic
analysis of macromolecular interactions with BIACORE. Methods
Enzymol. 323, 325−340.
(25) Cornelis, S., Fache, I., Van derHeyden, J., Guisez, Y., Tavernier,
J., Devos, R., Fiers, W., and Plaetinck, G. (1995) Characterization of
critical residues in the cytoplasmic domain of the human interleukin-5
receptor α chain required for growth signal transduction. Eur. J.
Immunol. 25, 1857−1864.
(26) Sali, A., and Blundell, T. L. (1993) Comparative protein
modelling by satisfaction of spatial restraints. J. Mol. Biol. 234, 779−
815.
(27) Laskowski, R., MacArthur, M., Moss, D., and Thornton, J.
(1993) PROCHECK: A program to check the stereochemical quality
of protein structures. J. Appl. Crystallogr. 26, 283−291.
(28) Sippl, M. J. (1993) Recognition of errors in three-dimensional
structures of proteins. Proteins 17, 355−362.
(29) de Vos, A. M., Ultsch, M., and Kossiakoff, A. A. (1992) Human
growth hormone and extracellular domain of its receptor: Crystal
structure of the complex. Science 255, 306−312.
(30) Somers, W., Ultsch, M., De Vos, A. M., and Kossiakoff, A. A.
(1994) The X-ray structure of a growth hormone-prolactin receptor
complex. Nature 372, 478−481.
(31) Ishino, T., Pillalamarri, U., Panarello, D., Bhattacharya, M.,
Urbina, C., Horvat, S., Sarkhel, S., Jameson, B., and Chaiken, I. (2006)
Asymmetric usage of antagonist charged residues drives interleukin-5
receptor recruitment but is insufficient for receptor activation.
Biochemistry 45, 1106−1115.
(32) Ishino, T., Urbina, C., Bhattacharya, M., Panarello, D., and
Chaiken, I. (2005) Receptor epitope usage by interleukin 5 mimetic
peptide. J. Biol. Chem. 280, 22951−22961.
(33) Waetzig, G. H., Chalaris, A., Rosenstiel, P., Suthaus, J., Holland,
C., Karl, N., Valles Uriarte, L., Till, A., Scheller, J., Grotzinger, J.,
Schreiber, S., Rose-John, S., and Seegert, D. (2010) N-linked
glycosylation is essential for the stability but not the signaling function
of the interleukin-6 signal transducer glycoprotein 130. J. Biol. Chem.
285, 1781−1789.
(34) McElroy, C. A., Dohm, J. A., and Walsh, S. T. (2009) Structural
and biophysical studies of the human IL-7/IL-7Rα complex. Structure
17, 54−65.
(35) Taniguchi, T. (1995) Cytokine signaling through nonreceptor
protein tyrosine kinases. Science 268, 251−255.
(36) Walter, M.R. (2002) Crystal structures of α-helical cytokine-
receptor complexes: We’ve only scratched the surface. Biotechniques
46−48 (Suppl.), 50−47.
(37) Wang, X., Lupardus, P., Laporte, S. L., and Garcia, K. C. (2009)
Structural Biology of Shared Cytokine Receptors. Annu. Rev. Immunol.
27, 29−60.
(38) Aritomi, M., Kunishima, N., Okamoto, T., Kuroki, R., Ota, Y.,
and Morikawa, K. (1999) Atomic structure of the GCSF-receptor
complex showing a new cytokine-receptor recognition scheme. Nature
401, 713−717.
(39) Hage, T., Sebald, W., and Reinemer, P. (1999) Crystal structure
of the interleukin-4/receptor α chain complex reveals a mosaic binding
interface. Cell 97, 271−281.
(40) Josephson, K., Logsdon, N. J., and Walter, M. R. (2001) Crystal
structure of the IL-10/IL-10R1 complex reveals a shared receptor
binding site. Immunity 15, 35−46.
(41) Yoon, C., Johnston, S. C., Tang, J., Stahl, M., Tobin, J. F., and
Somers, W. S. (2000) Charged residues dominate a unique
interlocking topography in the heterodimeric cytokine interleukin-
12. EMBO J. 19, 3530−3541.
(42) Syed, R. S., Reid, S. W., Li, C., Cheetham, J. C., Aoki, K. H., Liu,
B., Zhan, H., Osslund, T. D., Chirino, A. J., Zhang, J., Finer-Moore, J.,
Elliott, S., Sitney, K., Katz, B. A., Matthews, D. J., Wendoloski, J. J.,
Egrie, J., and Stroud, R. M. (1998) Efficiency of signalling through
cytokine receptors depends critically on receptor orientation. Nature
395, 511−516.
(43) Boulanger, M. J., Chow, D. C., Brevnova, E. E., and Garcia, K.
C. (2003) Hexameric structure and assembly of the interleukin-6/IL-6
α-receptor/gp130 complex. Science 300, 2101−2104.
(44) Wang, X., Rickert, M., and Garcia, K. C. (2005) Structure of the
quaternary complex of interleukin-2 with its α, β, and γ receptors.
Science 310, 1159−1163.
(45) Hansen, G., Hercus, T. R., McClure, B. J., Stomski, F. C.,
Dottore, M., Powell, J., Ramshaw, H., Woodcock, J. M., Xu, Y.,
Guthridge, M., McKinstry, W. J., Lopez, A. F., and Parker, M. W.
(2008) The structure of the GM-CSF receptor complex reveals a
distinct mode of cytokine receptor activation. Cell 134, 496−507.
(46) Davidson, D. J., and Castellino, F. J. (1991) Structures of the
asparagine-289-linked oligosaccharides assembled on recombinant
human plasminogen expressed in a Mamestra brassicae cell line
(IZD-MBO503). Biochemistry 30, 6689−6696.
(47) Benting, J., Lecat, S., Zacchetti, D., and Simons, K. (2000)
Protein expression in Drosophila Schneider cells. Anal. Biochem. 278,
59−68.
(48) Czabotar, P. E., Holland, J., and Sanderson, C. J. (2000)
Identification of regions within the third FnIII-like domain of the IL-
5Rα involved in IL-5 interaction. Cytokine 12, 867−873.
(49) Thompson, J. D., Higgins, D. G., and Gibson, T. J. (1994)
CLUSTAL W: Improving the sensitivity of progressive multiple
sequence alignment through sequence weighting, position-specific gap
penalties and weight matrix choice. Nucleic Acids Res. 22, 4673−4680.
(50) Sayle, R. A., and Milner-White, E. J. (1995) RASMOL:
Biomolecular graphics for all. Trends Biochem. Sci. 20, 374.
Biochemistry
Article
dx.doi.org/10.1021/bi2009135|Biochemistry 2011, 50, 7546−7556
7556