Molecular Immunology 67 (2015) 350–356

Contents lists available at ScienceDirect

Molecular Immunology

journa l homepage: www.e lsev ier .com/ locate /mol imm

Structural consequences of aglycosylated IgG Fc variants evolved forFc�RI binding

Man-Seok Jua, Jung-Hyun Naa,b, Yeon Gyu Yua, Jae-Yeol Kimc, Cherlhyun Jeongd,∗∗,Sang Taek Junga,∗

a Department of Bio and Nano Chemistry, Kookmin University, Seoul 136-702, Republic of Koreab Marine Biotechnology Research Division, Korea Institute of Ocean Science and Technology, Ansan 426-744, Republic of Koreac Department of Physics, Pohang University of Science and Technology (POSTECH), Pohang 790-784, Republic of Koread Center for Theragnosis, Biomedical Research Institute, Korea Institute of Science and Technology (KIST), Seoul 136-791, Republic of Korea

a r t i c l e i n f o

Article history:Received 13 April 2015Received in revised form 4 June 2015Accepted 12 June 2015Available online 4 July 2015

Keywords:Antibody engineeringAglycosylated antibodyEffector functionSingle-molecule analysisFörster resonance energy transferAlternative laser excitation

a b s t r a c t

In contrast to the glycosylated IgG antibodies secreted by human plasma cells, the aglycosylated IgGantibodies produced by bacteria are unable to bind Fc�Rs expressed on the surface of immune effec-tor cells and cannot trigger immune effector functions. To avoid glycan heterogeneity problems, elicitnovel effector functions, and produce therapeutic antibodies with effector function using a simple bac-terial expression system, Fc�RI-specific Fc-engineered aglycosylated antibodies, Fc11 (E382V) and Fc(E382V/M428I), containing mutations in the CH3 region, were isolated in a previous study. To elucidatethe relationship between Fc�RI binding affinity and the structural dynamics of the upper CH2 region ofFc induced by the CH3 mutations, the conformational variation of Fc variants was observed by single-molecule Förster resonance energy transfer (FRET) analysis using alternating-laser excitation (ALEX). Insharp contrast to wild-type Fc, which exhibits a highly dynamic upper CH2 region, the mutations in theCH3 region significantly stabilized the upper CH2 region. The results indicate that conformational plas-ticity, as well as the openness of the upper CH2 region, is critical for Fc�R binding and therapeutic effectorfunctions of IgG antibodies.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Abnormalities in immune surveillance mechanisms and imbal-ances in immune cell activation states result in serious acute orchronic immunological disorders including infection, tumors, orautoimmune diseases. To reinforce the human body’s immuneregulatory system and to effectively remove pathogens or defec-tive cells, over 30 highly specific monoclonal IgG antibodies have

Abbreviations: Fc, fragment crystallizable; IgG, immunoglobulin G; Fab, frag-ment antigen binding; Fc�R, Fc gamma receptor; ADCC, antibody-dependentcell-mediated cytotoxicity; ADCP, antibody-dependent cell-mediated phagocyto-sis; CDC, complement-dependent cytotoxicity; FRET, Förster resonance energytransfer; smFRET, single-molecule FRET; ALEX, alternating-laser excitation; PBS,phosphate buffered saline; TCEP, tris-carboxyethylphosphine; ELISA, enzyme-linked immunosorbent assay; SAXS, small-angle X-ray scattering; PDB, Protein DataBank.

∗ Corresponding author. Tel.: +82 2 910 5769; fax: +82 2 910 4415.∗∗ Corresponding author. Tel.: +82 2 958 5943; fax: +82 2 958 5909.

E-mail addresses: che.jeong@kist.re.kr (C. Jeong), sjung@kookmin.ac.kr(S.T. Jung).

been approved by the US Food and Drug Administration (US FDA)and the European Medicines Agency (EMA) and successfully com-mercialized for targeted immunotherapy (Elvin et al., 2013). Ofthe ten best-selling biopharmaceutical products of 2013, six weremonoclonal antibody products, and the second best-selling bio-pharmaceutical product was an antibody Fc fusion protein (Walsh,2014).

The human IgG antibody, a complex tetrameric molecule com-posed of two identical heavy chains and two identical light chains,can be divided into a Fab region, which binds to highly variablepathogenic antigens, and an Fc portion, which recruits and acti-vates immune effector leukocytes such as macrophages, dendriticcells, and natural killer (NK) cells. Each chain of a homodimeric Fcis N-linked glycosylated at an Asn297 residue and the appendedbiantennary glycans play critical roles in the structure and func-tion of IgG antibodies. The N-linked glycans are known to causethe CH2 domain to protrude from the internal hydrophobic region(Jefferis, 2005; Krapp et al., 2003), allowing the IgG antibody Fcregion to associate with various activating Fc�Rs (Fc�RI, IIa, andIIIa) and inhibitory Fc�RIIb, which are expressed on the surfaceof various immune cells. The engagement of antibody Fc with the

http://dx.doi.org/10.1016/j.molimm.2015.06.0200161-5890/© 2015 Elsevier Ltd. All rights reserved.

M.-S. Ju et al. / Molecular Immunology 67 (2015) 350–356 351

activating Fc�Rs promotes the recruitment of effector cells andtriggers effector functions for the removal of defective cells, suchas cancer cells or infected cells, by bridging humoral and cellu-lar immune functions (Jefferis, 2009a,b; Schwab and Nimmerjahn,2013).

Mammalian cells have been exhaustively exploited as primaryhosts for the production of full-length IgG antibodies. However,production of glycosylated IgG using mammalian expression sys-tems requires a long time for cell line development and a highcapital investment in the facility. To express IgG antibody in bacte-ria, increase the ease of downstream processing, and bypass glycanheterogeneity issues, which can result in side effects and incon-sistent efficacy, an aglycosylated IgG format, lacking the N-linkedglycan, would possess many advantages over the conventional gly-cosylated format, from a bioprocessing standpoint. AglycosylatedIgGs display antigen binding, solubility, stability at physiologicaltemperature, and serum persistence in vivo similar to those of gly-cosylated IgGs produced in mammalian cells (Hristodorov et al.,2013), therefore, several aglycosylated antibodies have been devel-oped for clinical trials (Ju and Jung, 2014).

Despite their many favorable properties, aglycosylated antibod-ies show almost no binding to Fc�Rs, resulting in the loss of effectorfunctions such as antibody-dependent cell-mediated cytotoxicity(ADCC), antibody-dependent cell-mediated phagocytosis (ADCP),and complement-dependent cytotoxicity (CDC) for the clearanceof defective cells (Jefferis, 2005; Simmons et al., 2002). Recently, toovercome the limitations of wild-type aglycosylated IgG, compre-hensive directed evolution of an aglycosylated Fc region has beenused to generate aglycosylated IgG Fc variants with effector func-tions. Two mutations in the upper CH2 region of aglycosylated Fc,which is the Fc�R-binding region of the Fc domain, restored theFc�RIIa binding affinity and the Fc�RIIa-mediated effector func-tions shown by glycosylated IgG Fc (Sazinsky et al., 2008). Usingbacterial display and more comprehensive aglycosylated Fc libraryscreening by flow cytometry, an aglycosylated Fc mutant was iso-lated that displayed high selectivity for activating Fc�RIIa overinhibitory Fc�RIIb and greatly enhanced macrophage-mediatedADCP (Jung et al., 2013).

In a previous study, aglycosylated Fc variants exhibiting selec-tive binding to Fc�RI and enhanced dendritic cell-mediated killingof tumor cells were isolated (Jung et al., 2010). Crystal struc-tures of Fc–Fc�R complexes (Radaev et al., 2001; Ramsland et al.,2011; Sondermann et al., 2000) have revealed that Fc�Rs directlycontact the upper CH2 region of the Fc portion. However, the ben-eficial mutations conferring specific Fc�RI binding were found tobe located in the CH3 region, which is far from the Fc�R-bindingregion, and it remains unclear how the two mutations induce struc-tural changes and alter conformational dynamics to induce highlyspecific Fc�RI binding.

X-ray crystallographic studies provide high-resolution, three-dimensional structural information. However, these results showonly a snapshot of a single conformer among the numerouspossible conformers that an aglycosylated Fc can adopt. Otherbiochemical techniques, such as circular dichroism, differential

scanning calorimetry (DSC), fluorescence spectroscopic analysis,and isothermal titration calorimetry (ITC), have also been used forstructural analysis of large proteins, such as antibody Fc. However,these techniques show only an ensemble average, representing theoverall bulk behavior of the molecules, and do not provide data onthe complete range of conformational variability.

Recently, single-molecule Förster resonance energy transfer(smFRET) between a single donor and a single acceptor fluorophorehas been widely used to study the structure and dynamics ofbiological molecules (Deniz et al., 2000; Ha et al., 1999; Jeonget al., 2011). The smFRET assay enables high-throughput measure-ment of conformational changes because it acts as a spectroscopic“ruler” for measuring the distance between the two fluorophoreprobes (Stryer, 1978). Thus, it can provide powerful insight intothe dynamic conformational changes of proteins, and reveal sub-populations in a heterogeneous mixture (Deniz et al., 2000; Kelliheret al., 2014; Kim et al., 2012a,b). The flexibility of a given proteinconformer can be inferred from the width and position of the peaksobserved in the histogram of a subpopulation. Further, combinationof an alternating-laser excitation (ALEX) assay with smFRET hasbeen shown to be useful for studying subpopulations in solution,without surface immobilization (Kim et al., 2012a,b).

Here, we analyzed the impact of isolated mutations on the con-formational dynamics of aglycosylated antibody Fc variants at thesingle-molecule level. By observing the conformational variabil-ity of the aglycosylated Fc domain, we monitored the structuralchanges induced during a directed evolution process for highlyselective Fc�RI binding.

2. Materials and methods

2.1. Chemicals, reagents, and other materials

Restriction enzymes and Vent polymerase were purchasedfrom New England BioLabs (Ipswich, MA). Oligonucleotide primerswere obtained from Integrated DNA Technologies (Coralville, IA).Ampicillin and 1-StepTM Ultra TMB-ELISA substrate solution werepurchased from Thermo Fisher Scientific, Inc. (Waltham, MA).Protein A agarose resin and Terrific Broth (TB) medium wereobtained from Genscript (Piscataway, NJ) and Becton Dickin-son Diagnostic Systems (Sparks, MD), respectively. Recombinanthuman Fc�RI purified from mouse myeloma NS0 cell line was pur-chased from R&D Systems (Minneapolis, MN). Cy3-maleimide andCy5-maleimide were from GE Healthcare Life Sciences (Piscataway,NJ), and the horseradish peroxidase (HRP) conjugate of mouse anti-polyHistidine antibody and all other reagents were purchased fromSigma-Aldrich (St. Louis, MO).

2.2. Construction of plasmids for S267C Fc variants

All of the plasmids and primers used in this study are describedin Tables 1 and 2. The pDsbA-NheI plasmid was constructed usinga QuikChange Site-Directed Mutagenesis Kit (Agilent Technolo-gies, Santa Clara, CA) using two primers (STJ#664 and STJ#665)

Table 1Plasmids used in this study.

Plasmids Relevant characteristics Reference or source

pTrc99A Apr, Trc promoter, lacIq Amersham Biosci.pTrc99A-DsbA DsbA signal sequence gene in pTrc99A Jung et al. (2010)pTrc99A-DsbA-Fc Fc gene in pTrc99A-DsbA Jung et al. (2010)pTrc99A-DsbA-Fc5 Fc5 gene in pTrc99A-DsbA Jung et al. (2010)pTrc99A-DsbA-Fc11 Fc11 gene in pTrc99A-DsbA Jung et al. (2010)pDsbA-NheI NheI restriction enzyme site in pTrc99A-DsbA This studypDsbA-Fc-S267C Fc-S267C mutant gene in pDsbA This studypDsbA-Fc5-S267C Fc5-S267C mutant gene in pDsbA This studypDsbA-Fc11-S267C Fc11-S267C mutant gene in pDsbA This study

352 M.-S. Ju et al. / Molecular Immunology 67 (2015) 350–356

Table 2Oligonucleotides used in this study.

Primers Primer nucleotide sequence (5′ → 3′)

STJ#488 TTTAAGGGAAGCTTTTTACCCGGGGACAGGGAGAGGSTJ#664 GTTTAGTTTTAGCGTTTAGCGCTAGCGCGGTCGACGACAAAACTCACACSTJ#665 GTGTGAGTTTTGTCGTCGACCGCGCTAGCGCTAAACGCTAAAACTAAACSTJ#666 CGCAGCGAGCTAGC GCG GACAAAACTCACACATGCCCACCGSTJ#796 GTGGTGGACGTGTGCCACGAAGACCSTJ#797 GGTCTTCGTGGCACACGTCCACCACSTJ#850 CGCAGCGAGCTAGCGCGGASTJ#851 TTAGGGAAGCTTCTATTATTTTACCCGGGGACAGGGAGA

and pTrc99A-DsbA-Fc (Jung et al., 2010), a plasmid for the expres-sion of wild type IgG1 Fc fragment (Gene Bank Accession No.AF237583), as a template. To obtain Fc variant genes contain-ing the S267C mutation, Fc (S267C), Fc11 (S267C/E382V), and Fc5(S267C/E382V/M428I), the first fragment, amplified with primerset STJ#850 and STJ#797, and the second fragment, amplifiedwith primer set STJ#796 and STJ#851, were assembled usingtheir respective templates, pTrc99A-DsbA-Fc, pTrc99A-DsbA-Fc11,and pTrc99A-DsbA-Fc5. Subsequently, the generated Fc variantscontaining the S267C mutation were amplified with the primersSTJ#850 and STJ#851, digested with NheI/HindIII, ligated into thepDsbA-NheI plasmid, and then digested using the same restrictionenzyme sites to generate pDsbA-Fc-S267C, pDsbA-Fc11-S267C, andpDsbA-Fc5-S267C. The resulting plasmids were transformed intoEscherichia coli Jude-1 (F′ [Tn10(Tetr) proAB+ lacIq �(lacZ)M15]mcrA �(mrr-hsdRMS-mcrBC) 80dlacZ�M15 �lacX74 deoR recA1araD139 �(ara leu)7697 galU galK rpsL endA1 nupG) (Kawarasakiet al., 2003).

2.3. Culture conditions and preparation of Fc variants

The transformants were cultured overnight at 37 ◦C with shak-ing at 250 rpm in TB media supplemented with 2% (w/v) glucose andampicillin (50 �g/mL). The resulting culture broths were diluted1:100 in fresh TB broth containing ampicillin (50 �g/mL), incu-bated at 37 ◦C for 3.5 h, and then cooled at 30 ◦C for 20 min. Toinduce the expression of Fc variants, 1 mM isopropyl-1-thio-�-d-galactopyranoside (IPTG) was added to each broth and incubatedat 30 ◦C with shaking at 250 rpm. Sixteen hours after induction,culture supernatants were recovered by centrifuging at 7,000 × gfor 1 h. The supernatants were filtered with a 0.45 �m membranebottle-top filter (Millipore, Billerica, MA), and the filtrates wererecovered for loading onto a Protein A agarose column equilibratedwith 1× phosphate buffered saline (PBS). After extensive washingwith 50 mL of 1× PBS, Fc variants were eluted from the column bythe addition of 100 mM glycine and collected in a tube containing1 M Tris (pH 8.0) for immediate neutralization. The sample bufferwas changed to PBS (pH 7.4) and concentrated to approximately1 mg/mL using Amicon Ultra Centrifugal Filter Units (Millipore, Bil-lerica, MA) with a molecular weight cutoff of 3 kDa.

2.4. Fluorescent dye labeling for FRET analysis

To label Fc-S267C, Fc11-S267C (S267C/E382V), and Fc5-S267C(S267C/E382S267CV/M428I) with fluorescent Cy3-maleimide andCy5-maleimide for FRET analysis, 0.25 mg of each Fc variant wasincubated with 1 mM TCEP (Tris(2-carboxyethyl)phosphine) atroom temperature for 1 h, 0.1 mg each of Cy3- and Cy5-maleimidepowder was added for a 15:10 molar ratio of dye to Fc mutant, andthen the mixture was incubated in the dark at 4 ◦C for 16 h. Unla-beled excess dyes were removed by loading onto a PD MiniTrapG-25 gel filtration column (GE Healthcare Life Sciences, Piscat-away, NJ) developed in 1× PBS (pH 7.4). Labeling efficiency wasdetermined by measuring the absorbance at 280 nm (for proteins),

550 nm (for Cy3), and 649 nm (for Cy5), using a Beckman CoulterDU 730 spectrophotometer (Beckman, DU Series) as described bythe manufacturer.

2.5. ELISA assay

Polystyrene 96-well ELISA plates (Corning, Corning, NY) werecoated with 50 �L of 4 �g/mL Fc-S267C, Fc11-S267C, and Fc5-S267C diluted in 0.05 M Na2CO3 (pH 9.6) by incubating the platesat 4 ◦C for 16 h. Following four times of washing in a solution of0.05% Tween20 in 1× PBS (pH 7.4), plates were incubated withblocking solution (0.5% bovine serum albumin in 1× PBS, pH 7.4)at room temperature for 2 h, washed four times in washing solu-tion, and incubated for 1 h with serially diluted recombinant humanFc�RI at room temperature. After four times of washing, plateswere incubated with 1:10,000 diluted anti-polyhistidine antibodyHRP conjugate at room temperature for 1 h. Plates were devel-oped by adding 50 �L of Ultra-TMB substrate to each well, andthe binding signal was monitored by measuring the absorbance at450 nm using an EpochTM Multi-Volume Spectrophotometer Sys-tem (BioTek, Winooski, VT).

2.6. smFRET measurement and analysis

Single-molecule measurements were performed using an inhouse-built alternating-laser excitation (ALEX) system, based ona commercial confocal microscopy system (Kim et al., 2012a,b; Leeet al., 2005). A 532 nm solid-state green laser (TECGL-30, WorldStar Tech, Toronto, Canada) and a 633 nm He–Ne laser (25-LHP-925,Melles-Griot, Carlsbad, CA) were used as alternating light sources.The 100 �s period was generated using acousto-optic modulators(AOM, 23080-1, Gooch & Housego, Ilminster, UK), and the excita-tion powers of the 532 nm and 633 nm lasers were 120 �W and40 �W, respectively. The coupled laser lights were directed to aninverted microscope Ti-U (Nikon Corporation, Tokyo, Japan) usinga dichroic mirror (Z532bcm, Chroma, Bellows Falls, VT). The lightreflected by a dichroic mirror was focused at 20 �m from the cov-erslip, using a 60× objective with a numerical aperture (NA) of1.2 (CFI Plan Apochromat) (Nikon Corporation, Tokyo, Japan). Theemissions from the dye molecules were collected by the objective,directed into two avalanche photodiode detectors, APD and SPCMAQR-13 (PerkinElmer, Waltham, MA), and then split by a dichroicmirror (624DCLP, Chroma). The Cy3 and Cy5 emissions were fil-tered using band-pass filters. The donor intensity (ID) of the Cy3emissions and the acceptor intensity (IA) of the Cy5 emissions werecorrected for cross-talk between the donor and acceptor channels,for fluorescence emission of acceptor dyes excited by the donor-excitation laser (532 nm), and for background fluorescence. TheFRET efficiencies were then calculated as the ratio of the correcteddonor intensity to the total intensity of the donor and acceptor(Kim et al., 2012a,b). The FRET data for IgG proteins double-labeledwith Cy3 and Cy5 in the intermediate stoichiometry (0.3–0.8) wereselected to exclude donor-only or acceptor-only data. Fluorescencetime traces were obtained by binning the photons detected at 1 ms.All data were analyzed using LABVIEW software (National Instru-ments, Austin, TX).

3. Results and discussion

3.1. Single point mutation for structural analysis of aglycosylatedFc variants

Based on X-ray crystallographic data, glycosylated IgG Fc(Fig. 1A, PDB: 1FC1) (Deisenhofer, 1981) shows a more open CH2conformation than aglycosylated Fc (Fig. 1B, PDB: 3S7G) (Borroket al., 2012), and truncation of a glycan in the C′E loop of Fc

M.-S. Ju et al. / Molecular Immunology 67 (2015) 350–356 353

Fig. 1. Schematic structure of wild-type Fc and Fc variants. (A–C) X-ray crystal structure of wild-type glycosylated IgG1 Fc (PDB: 1FC1) (A), wild-type aglycosylated IgG1 Fc(PDB: 3S7G) (B), and enzymatically deglycosylated wild-type IgG1 Fc (PDB: 3DNK) (C). (D–F) Förster resonance energy transfer (FRET) dye-labeled aglycosylated Fc variants,Fc-S267C (D), Fc11-S267C (E), and Fc5-S267C (F). Cy3 and Cy5 molecules of Cys267 are shown in blue and red. Amino acid (a.a.) substitutions in Fc variants (a.a. 382 and 428)are represented in yellow and purple, respectively, on the 3D structure of wild-type aglycosylated IgG1 Fc (PDB: 3S7G). (For interpretation of the references to color in thisfigure legend, the reader is referred to the web version of the article.)

leads to progressive closure of the upper CH2 structure (Mimuraet al., 2000, 2001). Therefore, the distance between the two P329residues located at the top of the CH2 region, indicating its degreeof openness, is considered to be highly correlated with glycan sta-tus, Fc�R binding affinity, and therapeutic effector functions (Krappet al., 2003). However, the hypothesis that the removal of a glycaninduces a more closed upper CH2 region has recently been refuted.The distance between P329 residues is greater in the crystal struc-ture of an enzymatically deglycosylated Fc (Fig. 1C, PDB: 3DNK)than in its glycosylated counterpart, but deglycosylated murine Fcshows an extremely closed CH2 conformation (Feige et al., 2009).In addition, a recent small-angle X-ray scattering (SAXS) experi-ment showed that the radius of gyration (Rg) of aglycosylated Fc islarger than that of glycosylated Fc (Borrok et al., 2012). Consideringthat X-ray crystallography data provide only a snapshot of Fc struc-tures, that structures can be affected by crystal packing in solution,and that high B-factors (the temperature factor) of the upper CH2region have been observed (Borrok et al., 2012), we proposed thatthis discrepancy might be caused by high conformational plasticityof the CH2 region (Jung et al., 2011).

Aglycosylated Fc variants (Fc11, Fc5) previously isolated forselective Fc�RI binding and monocyte-derived dendritic cell-mediated killing of tumor cells contain mutations in the CH3 region,which is distant from the Fc�R-binding region, the lower hinge, andthe upper CH2 region. In a previous study, in an attempt to explainthe structure–function relationship in this region, we hypothesizedthat the Fc5 mutant might show a more stable conformation, whichimproves its capacity for Fc�RI binding, in contrast to the wild-typeaglycosylated IgG Fc, which might show high CH2 conformationalplasticity, resulting in a loss of Fc�R binding affinity (Jung et al.,2011). To validate this hypothesis and to gain structural insight,we investigated how mutations in the CH3 region allow selectiveFc�RI binding using smFRET analysis, which can be used to deter-mine the distance between the two CH2 regions. For site-specificconjugation with the maleimide group-linked FRET donor (Cy3)

and acceptor (Cy5) molecules in the upper CH2 region, the surfaceSer267 residue was mutated to Cys (Fig. 1D–F).

3.2. Bacterial expression, purification, and Fc�RI bindingcharacteristics of S267C-containing Fc variants

Enzymatic deglycosylation has routinely been used to under-stand the effect of N-linked glycans on Fc structure and function(Alsenaidy et al., 2013; Mimura et al., 2000, 2001; Wang et al.,2013). However, many glycosidases do not completely remove gly-cans and append one or more carbohydrate residues to Asn297.Although treatment with glycosidases such as PNGase F completelytruncates oligosaccharides, it also converts Asn residues to Asp(Tarentino et al., 1985), resulting in the addition of a negative chargeto the C′E loop of the IgG CH2 region. In this study, we expressedintact aglycosylated homodimeric Fc fragments in E. coli. To preparesoluble aglycosylated Fc variants containing an S267C mutation(Fc-S267C, Fc11-S267C, and Fc5-S267C) in an E. coli periplasm, theFc genes were fused to the C-terminus of the signal sequence ofDsbA (a periplasmic disulfide bond oxidoreductase), which directstarget proteins to the signal recognition particle (SRP) pathway(Schierle et al., 2003).

Compared to wild-type Fc, the introduction of a Cys mutation atthe Ser267 residue of Fc did not significantly decrease the expres-sion level of homodimeric Fc domains. Under optimized cultivationconditions (Jung et al., 2010), homodimeric Fc variants containingthe S267C mutation were secreted into the media and efficientlypurified to >90% homogeneity from the media fraction of the cul-ture broth, without the necessity of cell disruption, using simpleProtein A affinity chromatography. More than 1 mg of purified Fc-S267C, Fc11-S267C, and Fc5-S267C was routinely obtained from500 mL cultures (Fig. 2A and B). To investigate the effect of theS267C mutation on Fc�RI binding, we performed an ELISA assay.The purified aglycosylated Fc variants, Fc5-S267C and Fc11-S267C,showed an excellent Fc�RI binding profile, suggesting that S267C,

354 M.-S. Ju et al. / Molecular Immunology 67 (2015) 350–356

Fig. 2. Expression and purification of S267C-containing aglycosylated Fc variants. (A and B) SDS-PAGE analysis showing purified Fc-S267C, Fc11-S267C, and Fc5-S267C undernon-reducing conditions (A) and reducing conditions (B), Lane 1: Lane 1 Fc-S267C; Lane 2 Fc5-S267C; and Lane 3 Fc11-S267C.

located in the BC loop of the CH2 region, does not significantlyalter the secondary structure of the upper CH2 region (Fig. 3).This suggested that the structure of Fc domains containing S267C(Fc-S267C, Fc11-S267C, and Fc5-S267C) recapitulates the structureof wild-type aglycosylated Fc and Fc�RI-binding aglycosylated Fcvariants (Fc11 and Fc5).

3.3. Highly flexible CH2 conformation of aglycosylated Fc-S267C

To investigate the dynamics of the upper CH2 region, weapplied a diffusion-based single-molecule method for FRET anal-ysis, alternating-laser excitation (ALEX), which allows accuratemeasurement of FRET efficiencies and distances without the needfor surface immobilization (Lee et al., 2005). For the FRET analysis,the S267C Fc variants were labeled with Cy3 and Cy5 maleimidemolecules and developed in a desalting column to remove unla-beled dye molecules. The typical labeling efficiency for Cy3 andCy5, measured by spectrophotometry, was approximately 30% and20%, respectively. In the analysis of Fc-S267C and its representa-tive time traces, both low and high FRET efficiencies, which showedtwo Gaussian distributions, with peaks of 0.01 ± 0.24 (mean ± SD)

and 0.94 ± 0.07, were observed (Fig. 4A). This result clearly showsthat the Fc� receptor-binding regions were intrinsically unstable,explaining the discrepancy in the degree of CH2 openness amongvarious crystal structures of previously reported IgG Fcs.

The major population showed a low and broad FRET efficiency,suggesting that the preferred conformers were more open. Thisresult is in agreement with the previous SAXS experiment showingthat the Rg of an aglycosylated Fc was unexpectedly large comparedto the Rg of a glycosylated Fc (Borrok et al., 2012). In addition, thisfinding of high conformational plasticity is similar to results froma previous smFRET analysis of an enzymatically deglycosylated Fccontaining an N297D mutation in the C′E loop of the CH2 region andan S254C mutation in the CH2-CH3 interface region for FRET dyeconjugation (Kelliher et al., 2014). Results from biophysical anal-yses using differential scanning calorimetry (DSC) and differentialscanning fluorimetry (DSF) indicated that the deglycosylated IgGCH2 region showed reduced stability at elevated temperatures (Haet al., 2011; Zheng et al., 2011) and that the thermostability of theFab and CH3 regions is not significantly affected by the presenceof N-linked glycosylation (Alsenaidy et al., 2013). A highly dynamicCH2 region might contribute to the somewhat low thermostabilityof aglycosylated IgG Fc at high temperatures.

Fig. 3. Binding of purified Fc-S267C, Fc5-S267C, and Fc11-S267C to Fc�RIa. Serially diluted Fc�RIa was incubated in plates containing immobilized Fc-S267C, Fc5-S267C, andFc11-S267C. The amount of bound Fc�RI was monitored using an HRP conjugate of mouse anti-His antibody.

M.-S. Ju et al. / Molecular Immunology 67 (2015) 350–356 355

Fig. 4. Single-molecule FRET analysis of the Fc� receptor-binding regions of IgG molecules. (A–C) Schematic representation of the dynamic structures of Fc-S267C (A), Fc11-S267C (B), and Fc5-S267C (C). (D–F) Single-molecule FRET efficiency histogram of Fc-S267C (D), Fc11-S267C (E), and Fc5-S267C. ID

Green, IAGreen, and IA

Red show the fluorescenceemissions of Cy3 excited by a 532 nm laser, the FRET emission of Cy5 excited by a 532 nm laser, and the fluorescence emission of Cy5 excited by a 633 nm laser, respectively.Each symbol, *, §, and #, in the FRET efficiency histogram represents the corresponding FRET signal at the time trace denoted by the same symbol.

3.4. Stabilized conformation of engineered Fc�RI bindingaglycosylated Fc variants

Using directed evolution to improve the Fc�RI binding affinityand flow cytometry-based Fc library screening, several aglycosy-lated Fc variants were isolated (Jung et al., 2010). Fc11, which hasa single mutation (E382V) in the CH3 region, showed significantlyincreased Fc�RI binding affinity. One additional mutation, M428Iin Fc5 (E382V/M428I), further enhanced Fc�RI binding affinity, toa level similar to that of glycosylated IgG Fc (Jung et al., 2010).

In sharp contrast to wild-type Fc, the Fc11 population moved toa higher FRET efficiency, and multiple-fitted Gaussian peaks wereobserved at 0.19 ± 0.12, 0.61 ± 0.16 and 0.88 ± 0.07 (Fig. 4). Clearly,this result illustrates that the upper CH2 regions, the Fc�R-bindingregions of F11, were more stable, with lower degrees of FRET effi-ciency variation than the highly dynamic wild-type Fc conformers.

The F5 population, which contained one more mutation(M428I) than Fc11, showed a dramatically higher FRET efficiency(0.93 ± 0.07) (Fig. 4C). The FRET efficiency histogram shows that theisolated mutations (E382V and M428I) in the CH3 region restrictdynamic movement of the upper CH2 region of aglycosylated Fc;the Fc�RI binding affinity is inversely proportional to the degree ofconformational flexibility in the CH2 region. This result is in agree-ment with that of the previous SAXS experiment, indicating thatFc5 has a relatively lower Rg and a more closed CH2 conformationin solution than wild-type Fc (Borrok et al., 2012).

4. Conclusions

In this study, the dynamics of aglycosylated Fc and Fc vari-ants produced in E. coli were analyzed. The S267C-containing Fcvariants were labeled with FRET donor and acceptor molecules,and the conformational dynamics of the upper CH2 region wereobserved using single-molecule spectroscopy. In sharp contrast to

wild-type aglycosylated Fc, the molecules with the isolated muta-tions showed significantly altered conformational flexibility andoverall conformation. Protein dynamics and flexibility are cru-cial for protein–protein interactions (Henzler-Wildman and Kern,2007; Marsh and Teichmann, 2014). In accordance with previousresults, we found the rigidity of the protein structure to be criticalfor enhanced binding of receptor molecules. It has been proposedthat an IgG CH2 domain with sialylated glycans shows greater con-formational flexibility than a nonsialylated CH2 domain (Pinceticet al., 2014). Similar to this modification of terminal glycans, substi-tution of amino acids in the CH3 region, in a location remote fromthe Fc�R-binding epitopes, changed the dynamics of the upper CH2region and subsequently increased the Fc�RI binding affinity. In thisstudy, directed evolution of aglycosylated Fc to obtain higher Fc�RIbinding affinities was monitored using protein structural analy-sis; we found that accumulation of beneficial mutations for Fc�RIbinding in the CH3 region induced a change in the conformationalvariability of the upper CH2 region. Generally, highly evolvableproteins are also highly promiscuous, and are composed of manyconformers (Tokuriki and Tawfik, 2009). Given the relationshipbetween protein dynamism and evolvability, highly flexible wild-type aglycosylated Fc domains can be evolved for different affinitiesand specificities to target a variety of Fc binding ligands, such asFc�Rs, FcRn, C1q, and rheumatoid factor (RF), for novel therapeu-tic functions. Understanding the relationship between antibody Fcconformational dynamics and antibody Fc-mediated effector func-tion will provide insight to design improved immunotherapeuticswith enhanced efficacy.

Acknowledgements

This work was supported by a grant from the National R&D Pro-gram for Cancer Control, Ministry of Health and Welfare, Republic ofKorea (1420160), the Basic Science Research Program through the

356 M.-S. Ju et al. / Molecular Immunology 67 (2015) 350–356

National Research Foundation of Korea (NRF) funded by the Min-istry of Science, ICT and Future Planning (2013R1A1A1004576), theBio & Medical Technology Development Program of the NationalResearch Foundation (NRF) funded by the Korean Government(MSIP) (2014M3A9D9069609), and the KIST Institutional Program(Project No. 2E25270).

References

Alsenaidy, M.A., Kim, J.H., Majumdar, R., Weis, D.D., Joshi, S.B., Tolbert, T.J., Mid-daugh, C.R., Volkin, D.B., 2013. High-throughput biophysical analysis and datavisualization of conformational stability of an IgG1 monoclonal antibody afterdeglycosylation. J. Pharm. Sci. 102, 3942–3956.

Borrok, M.J., Jung, S.T., Kang, T.H., Monzingo, A.F., Georgiou, G., 2012. Revisitingthe role of glycosylation in the structure of human IgG Fc. ACS Chem. Biol. 7,1596–1602.

Deisenhofer, J., 1981. Crystallographic refinement and atomic models of a humanFc fragment and its complex with fragment B of protein A from Staphylococcusaureus at 2.9- and 2.8-ANG. resolution. Biochemistry 20, 2361–2370.

Deniz, A.A., Laurence, T.A., Beligere, G.S., Dahan, M., Martin, A.B., Chemla, D.S.,Dawson, P.E., Schultz, P.G., Weiss, S., 2000. Single-molecule protein folding: dif-fusion fluorescence resonance energy transfer studies of the denaturation ofchymotrypsin inhibitor 2. Proc. Natl. Acad. Sci. U. S. A. 97, 5179–5184.

Elvin, J.G., Couston, R.G., van der Walle, C.F., 2013. Therapeutic antibodies: marketconsiderations, disease targets and bioprocessing. Int. J. Pharm. 440, 83–98.

Feige, M.J., Nath, S., Catharino, S.R., Weinfurtner, D., Steinbacher, S., Buchner, J.,2009. Structure of the murine unglycosylated IgG1 Fc fragment. J. Mol. Biol.391, 599–608.

Ha, S., Ou, Y., Vlasak, J., Li, Y., Wang, S., Vo, K., Du, Y., Mach, A., Fang, Y., Zhang, N.,2011. Isolation and characterization of IgG1 with asymmetrical Fc glycosylation.Glycobiology 21, 1087–1096.

Ha, T., Ting, A.Y., Liang, J., Caldwell, W.B., Deniz, A.A., Chemla, D.S., Schultz, P.G.,Weiss, S., 1999. Single-molecule fluorescence spectroscopy of enzyme confor-mational dynamics and cleavage mechanism. Proc. Natl. Acad. Sci. U. S. A. 96,893–898.

Henzler-Wildman, K., Kern, D., 2007. Dynamic personalities of proteins. Nature 450,964–972.

Hristodorov, D., Fischer, R., Joerissen, H., Muller-Tiemann, B., Apeler, H., Linden, L.,2013. Generation and comparative characterization of glycosylated and aglyco-sylated human IgG1 antibodies. Mol. Biotechnol. 53, 326–335.

Jefferis, R., 2005. Glycosylation of recombinant antibody therapeutics. Biotechnol.Prog. 21, 11–16.

Jefferis, R., 2009a. Glycosylation as a strategy to improve antibody-based therapeu-tics. Nat. Rev. Drug Discov. 8, 226–234.

Jefferis, R., 2009b. Recombinant antibody therapeutics: the impact of glycosylationon mechanisms of action. Trends. Pharmacol. Sci. 30, 356–362.

Jeong, C., Cho, W.K., Song, K.M., Cook, C., Yoon, T.Y., Ban, C., Fishel, R., Lee, J.B., 2011.MutS switches between two fundamentally distinct clamps during mismatchrepair. Nat. Struct. Mol. Biol. 18, 379–385.

Ju, M.S., Jung, S.T., 2014. Aglycosylated full-length IgG antibodies: steps toward next-generation immunotherapeutics. Curr. Opin. Biotechnol. 30C, 128–139.

Jung, S.T., Reddy, S.T., Kang, T.H., Borrok, M.J., Sandlie, I., Tucker, P.W., Georgiou,G., 2010. Aglycosylated IgG variants expressed in bacteria that selectively bindFcgammaRI potentiate tumor cell killing by monocyte-dendritic cells. Proc. Natl.Acad. Sci. U. S. A. 107, 604–609.

Jung, S.T., Kang, T.H., Kelton, W., Georgiou, G., 2011. Bypassing glycosylation: engi-neering aglycosylated full-length IgG antibodies for human therapy. Curr. Opin.Biotechnol. 22, 858–867.

Jung, S.T., Kelton, W., Kang, T.H., Ng, D.T., Andersen, J.T., Sandlie, I., Sarkar, C.A.,Georgiou, G., 2013. Effective phagocytosis of low Her2 tumor cell lines with engi-neered, aglycosylated IgG displaying high FcgammaRIIa affinity and selectivity.ACS Chem. Biol. 8, 368–375.

Kawarasaki, Y., Griswold, K.E., Stevenson, J.D., Selzer, T., Benkovic, S.J., Iverson,B.L., Georgiou, G., 2003. Enhanced crossover SCRATCHY: construction and

high-throughput screening of a combinatorial library containing multiple non-homologous crossovers. Nucleic Acids Res. 31, e126.

Kelliher, M.T., Jacks, R.D., Piraino, M.S., Southern, C.A., 2014. The effect of sugarremoval on the structure of the Fc region of an IgG antibody as observedwith single molecule Forster resonance energy transfer. Mol. Immunol. 60,103–108.

Kim, C., Kim, J.Y., Kim, S.H., Lee, B.I., Lee, N.K., 2012a. Direct characterization of pro-tein oligomers and their quaternary structures by single-molecule FRET. Chem.Commun. (Camb.) 48, 1138–1140.

Kim, J.Y., Choi, B.K., Choi, M.G., Kim, S.A., Lai, Y., Shin, Y.K., Lee, N.K., 2012b. Solutionsingle-vesicle assay reveals PIP2-mediated sequential actions of synaptotagmin-1 on SNAREs. EMBO J. 31, 2144–2155.

Krapp, S., Mimura, Y., Jefferis, R., Huber, R., Sondermann, P., 2003. Structural analysisof human IgG-Fc glycoforms reveals a correlation between glycosylation andstructural integrity. J. Mol. Biol. 325, 979–989.

Lee, N.K., Kapanidis, A.N., Wang, Y., Michalet, X., Mukhopadhyay, J., Ebright,R.H., Weiss, S., 2005. Accurate FRET measurements within single diffusingbiomolecules using alternating-laser excitation. Biophys. J. 88, 2939–2953.

Marsh, J.A., Teichmann, S.A., 2014. Protein flexibility facilitates quaternary structureassembly and evolution. PLoS Biol. 12, e1001870.

Mimura, Y., Church, S., Ghirlando, R., Ashton, P.R., Dong, S., Goodall, M., Lund, J.,Jefferis, R., 2000. The influence of glycosylation on the thermal stability and effec-tor function expression of human IgG1-Fc: properties of a series of truncatedglycoforms. Mol. Immunol. 37, 697–706.

Mimura, Y., Sondermann, P., Ghirlando, R., Lund, J., Young, S.P., Goodall, M., Jefferis,R., 2001. Role of oligosaccharide residues of IgG1-Fc in Fc gamma RIIb binding.J. Biol. Chem. 276, 45539–45547.

Pincetic, A., Bournazos, S., DiLillo, D.J., Maamary, J., Wang, T.T., Dahan, R., Fiebiger,B.M., Ravetch, J.V., 2014. Type I and type II Fc receptors regulate innate andadaptive immunity. Nat. Immunol. 15, 707–716.

Radaev, S., Motyka, S., Fridman, W.H., Sautes-Fridman, C., Sun, P.D., 2001. The struc-ture of a human type III Fcgamma receptor in complex with Fc. J. Biol. Chem.276, 16469–16477.

Ramsland, P.A., Farrugia, W., Bradford, T.M., Sardjono, C.T., Esparon, S., Trist, H.M.,Powell, M.S., Tan, P.S., Cendron, A.C., Wines, B.D., Scott, A.M., Hogarth, P.M., 2011.Structural basis for Fc gammaRIIa recognition of human IgG and formation ofinflammatory signaling complexes. J. Immunol. 187, 3208–3217.

Sazinsky, S.L., Ott, R.G., Silver, N.W., Tidor, B., Ravetch, J.V., Wittrup, K.D., 2008.Aglycosylated immunoglobulin G1 variants productively engage activating Fcreceptors. Proc. Natl. Acad. Sci. U. S. A. 105, 20167–20172.

Schierle, C.F., Berkmen, M., Huber, D., Kumamoto, C., Boyd, D., Beckwith, J., 2003. TheDsbA signal sequence directs efficient, cotranslational export of passenger pro-teins to the Escherichia coli periplasm via the signal recognition particle pathway.J. Bacteriol. 185, 5706–5713.

Schwab, I., Nimmerjahn, F., 2013. Intravenous immunoglobulin therapy: how doesIgG modulate the immune system? Nat. Rev. Immunol. 13, 176–189.

Simmons, L.C., Reilly, D., Klimowski, L., Raju, T.S., Meng, G., Sims, P., Hong, K., Shields,R.L., Damico, L.A., Rancatore, P., Yansura, D.G., 2002. Expression of full-lengthimmunoglobulins in Escherichia coli: rapid and efficient production of aglycosy-lated antibodies. J. Immunol. Methods 263, 133–147.

Sondermann, P., Huber, R., Oosthuizen, V., Jacob, U., 2000. The 3.2-A crystal structureof the human IgG1 Fc fragment-Fc gammaRIII complex. Nature 406, 267–273.

Stryer, L., 1978. Fluorescence energy transfer as a spectroscopic ruler. Annu. Rev.Biochem. 47, 819–846.

Tarentino, A.L., Gomez, C.M., Plummer, T.H., 1985. Deglycosylation of asparagine-linked glycans by peptide:N-glycosidase F. Biochemistry 24, 4665–4671.

Tokuriki, N., Tawfik, D.S., 2009. Protein dynamism and evolvability. Science 324,203–207.

Walsh, G., 2014. Biopharmaceutical benchmarks 2014. Nat. Biotechnol. 32,992–1000.

Wang, X., Kumar, S., Buck, P.M., Singh, S.K., 2013. Impact of deglycosylation andthermal stress on conformational stability of a full length murine IgG2a mono-clonal antibody: observations from molecular dynamics simulations. Proteins81, 443–460.

Zheng, K., Bantog, C., Bayer, R., 2011. The impact of glycosylation on monoclonalantibody conformation and stability. MAbs 3, 568–576.