Crystal Structure of Cell Adhesion Molecule
Nectin-2/CD112 and Its Binding to Immune
Receptor DNAM-1/CD226
This information is current as
of June 14, 2017.
Jun Liu, Xiaomin Qian, Zhujun Chen, Xiang Xu, Feng Gao,
Shuijun Zhang, Rongguang Zhang, Jianxun Qi, George F.
Gao and Jinghua Yan
References
Subscription
Permissions
Email Alerts
This article cites 67 articles, 29 of which you can access for free at:
http://www.jimmunol.org/content/188/11/5511.full#ref-list-1
Information about subscribing to The Journal of Immunology is online at:
http://jimmunol.org/subscription
Submit copyright permission requests at:
http://www.aai.org/About/Publications/JI/copyright.html
Receive free email-alerts when new articles cite this article. Sign up at:
http://jimmunol.org/alerts
The Journal of Immunology is published twice each month by
The American Association of Immunologists, Inc.,
1451 Rockville Pike, Suite 650, Rockville, MD 20852
Copyright © 2012 by The American Association of
Immunologists, Inc. All rights reserved.
Print ISSN: 0022-1767 Online ISSN: 1550-6606.
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
J Immunol 2012; 188:5511-5520; Prepublished online 30
April 2012;
doi: 10.4049/jimmunol.1200324
http://www.jimmunol.org/content/188/11/5511
The Journal of Immunology
Crystal Structure of Cell Adhesion Molecule Nectin-2/CD112
and Its Binding to Immune Receptor DNAM-1/CD226
Jun Liu,*,†,1,2 Xiaomin Qian,*,‡,1 Zhujun Chen,*,x Xiang Xu,*,x Feng Gao,{
Shuijun Zhang,*,‡ Rongguang Zhang,{ Jianxun Qi,* George F. Gao,*,†,‡,‖ and
Jinghua Yan*
C
ytotoxic lymphocytes, such as NK cells and CTLs, are
major players in cell-mediated immunity against viral
infection and tumorigenesis (1–3). The recognition of
nonself and abnormal self-Ags and the activation of NK cells and
CTLs are mediated by AgRs, a series of costimulators, and ad-
*CAS Key Laboratory of Pathogenic Microbiology and Immunology, Institute of
Microbiology, Chinese Academy of Sciences, Beijing, 100101 China; †Graduate University, Chinese Academy of Sciences, Beijing, 100049 China; ‡School of Life Sciences, University of Science and Technology of China, Hefei, Anhui Province, 230026
China; xCollege of Life Science, Anhui Agricultural University, Anhui, 230036
China; {National Laboratory of Biomacromolecules, Institute of Biophysics, Chinese
Academy of Sciences, Beijing, 100101 China; and ‖Research Network of Immunity
and Health, Beijing Institutes of Life Science, Chinese Academy of Sciences, Beijing, 100101 China
1
J.L. and X.Q. contributed equally to this study.
2
Current address: Department of Internal Medicine, Yale University, New Haven, CT.
Received for publication January 26, 2012. Accepted for publication March 31, 2012.
This work was supported by the 973 Project of the China Ministry of Science and
Technology (Protein Science Special Project Grant 2010CB911902). G.F.G. is a leading principal investigator of the Innovative Research Group of the National Natural
Science Foundation of China (Grant 81021003). The funders had no role in study
design, data collection and analysis, decision to publish, or preparation of the manuscript.
The atomic coordinates and structure factors for nectin-2v and FAMP mutant presented in this article have been submitted to the Protein Data Bank (http://www.pdb.
org) under accession numbers 4DFH and 4DFI, respectively.
Address correspondence and reprint requests to Dr. Jinghua Yan and Prof. George F.
Gao, CAS Key Laboratory of Pathogenic Microbiology and Immunology, Institute of
Microbiology, Chinese Academy of Sciences, Beijing 100101, China. E-mail addresses: [email protected] (J.Y.) and [email protected] (G.F.G.)
Abbreviations used in this article: CAM, cell adhesion molecule; DC, dendritic cell;
DNAM-1, DNAX accessory molecule 1; gD, glycoprotein D; Necl, Nectin-like molecule; nectin-2v, V-set domain of nectin-2; PDB, Protein Data Bank; RMSD, root
mean square deviation; SynCAM, synaptic cell adhesion molecule; V, variable; WT,
wild type.
Copyright Ó 2012 by The American Association of Immunologists, Inc. 0022-1767/12/$16.00
www.jimmunol.org/cgi/doi/10.4049/jimmunol.1200324
hesion molecules (4–6). Among the immune costimulatory
receptors, attention has been focused on the Ig superfamily
member DNAX accessory molecule 1 (DNAM-1; also called
CD226), which is expressed on the majority of NK cells, T cells,
monocytes, and platelets (7–9). DNAM-1 is involved in the expansion, differentiation, and activation of NK cells and naive
T cells (10, 11). Furthermore, DNAM-1–deficient mice display
increased tumor development after transplantation of tumor cells
(12). A nonsynonymous polymorphism (Gly307Ser) of DNAM-1
was recently identified as a genetic risk factor for a variety of
autoimmune diseases, such as type 1 diabetes, autoimmune thyroid disease, rheumatoid arthritis, multiple sclerosis, systemic sclerosis, systemic lupus erythematosus, and psoriasis (13). Therefore,
studies of DNAM-1’s interactions with its ligands are important
for our understanding of its true physiological functions, especially
in the immune system.
Nectin-2 (also called CD112 or poliovirus receptor-related
protein 2) and Necl-5 (also called CD155, or poliovirus receptor) have been identified as two ligands of DNAM-1 in both humans
and mice (14–16). Evidence from in vitro and in vivo studies
revealed that the interaction of DNAM-1 with its ligands is involved in the functions of a variety of immune cells (12, 17). For
instance, a strong correlation was found between the expression of
nectin-2 and Necl-5 and the lysis susceptibility of myeloid leukemia by NK cells (18). Nectin-2– and Necl-5–transduced tumor
cells are efficiently rejected in mice through the stimulation of
DNAM-1–expressing innate immunity by CD8a+ dendritic cells
(DCs), as well as NK cells (19). NK cell-mediated lysis of DCs
also involves DNAM-1 and its ligands; moreover, the degree of
contribution of DNAM-1 in this process is correlated with the
surface densities of nectin-2 and Necl-5 (20). Furthermore,
monocyte extravasation through the endothelium is also regulated
by DNAM-1 and Necl-5 expressed at endothelial junctions (21).
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
The nectin and nectin-like molecule (Necl) family includes important cell adhesion molecules (CAMs) characterized by their Iglike nature. Such CAMs regulate a broad spectrum of cell–cell interactions, including the interaction between NK cells and
cytotoxic T lymphocytes (CTLs) and their target cells. CAM members nectin-2 (CD112) and Necl-5 (CD155) are believed to form
homodimers (for nectin-2) or heterodimers in their functions for cell adhesion, as well as to interact with immune costimulatory
receptor DNAX accessory molecule 1 (DNAM-1) (CD226) to regulate functions of both NK and CTL cells. However, the structural
basis of the interactive mode of DNAM-1 with nectin-2 or Necl-5 is not yet understood. In this study, a soluble nectin-2 Ig-like Vset domain (nectin-2v) was successfully prepared and demonstrated to bind to both soluble ectodomain and cell surface-expressed
full-length DNAM-1. The 1.85-Å crystal structure of nectin-2v displays a perpendicular homodimer arrangement, revealing the
homodimer characteristics of the nectin and Necls. Further mutational analysis indicated that disruption of the homodimeric
interface of nectin-2v led to a failure of the homodimer formation, as confirmed by crystal structure and biochemical properties of
the mutant protein of nectin-2v. Interestingly, the monomer mutant also loses DNAM-1 binding, as evidenced by cell staining with
tetramers and surface plasmon resonance assays. The data indicate that interaction with DNAM-1 requires either the homodimerization or engagement of the homodimeric interface of nectin-2v. These results have implications for immune intervention of
tumors or autoimmune diseases in the DNAM-1/nectin-2–dependent pathway. The Journal of Immunology, 2012, 188: 5511–5520.
5512
STRUCTURE OF CD112 HOMODIMER AND ITS BINDING TO CD226
The interaction of eosinophil-expressed nectin-2 with DNAM-1
on mast cells contributes to costimulatory responses in allergic
reactions (22). However, few reports have characterized the interaction of DNAM-1 and its ligands at the molecular level (23).
The impact of the binding mode of DNAM-1 and nectin-2/Necl-5
on their functions is necessary for our understanding of NK/CTL
cell functions.
Nectin-2 and Necl-5 belong to the nectin and Necl family, which
consists of nine members (nectin-1–4 and Necl-1–5). These
molecules are novel Ca2+-independent Ig-like cell adhesion molecules (CAMs) (24). Nectin and Necls each contain three Ig-like
domains (one V-set and two C-set domains) in their extracellular
regions (25). The distinction between nectins and Necls is that
Necls lack an afadin-binding motif in their cytoplasmic domains
(25, 26). It is believed that nectins/Necls can form homo- or
heterodimers, contributing to cell–cell adhesion (24, 27–29).
Thus, nectin/Necls play a pivotal role in the formation of cell
polarity and the metastasis of tumor cells (30). Nectin-2 was
initially discovered as a poliovirus receptor-related protein (31). It
is expressed on a variety of cells and is especially overexpressed
on a variety of tumor cells (16). Nectin-2 can stimulate the reaction
of NK cells and CTLs through its interaction with DNAM-1 (32).
Previous studies suggested that nectins/Necls might form cisdimers through their first C-set domain and consequently form
trans-dimers through their V-set domain during cell–cell adhesion
(28, 33–35). This interactive mode of nectins/Necls may be supported by the structural determination of the Necl-1 (also known
as synaptic CAM [SynCAM]3) V-set domain in 2006, in which
the homodimer of Necl-1 was considered a trans-dimer (36). The
crystal structure of the extracellular region of nectin-1, with three
Ig domains (including the V-set and both two C-set Ig-like
domains), also displayed a dimer architecture via its V-set domain, which is proposed as a cis-dimer because of the perpendicular arrangement of the homodimer interface (37). More
recently, the V-set domain of Necl-3 (also known as SynCAM2)
was also determined to be a homodimer (38). Despite these crystal
structures, the true dimerization mode of nectins/Necls remains
debatable.
In this study, we generated a soluble V-set domain of nectin-2
(nectin-2v) and demonstrated that it binds to both DNAM-1–
expressing cells and the soluble ectodomain of DNAM-1. The
crystal structure of nectin-2v at 1.85 Å clearly shows a homodimer
formed by two perpendicularly arranged molecules. Disruption of
the dimeric interface abolishes nectin-2v binding to DNAM-1 and
yields a stable nectin-2v monomer. We propose that there are two
possible models for DNAM-1 binding to nectin-2: DNAM-1
competes against nectin-2 dimer formation in a 1:1 (DNAM-1:
nectin-2)-binding mode and DNAM-1 binds nectin-2 dimer in a
1:2 (DNAM-1:nectin-2)-binding mode.
L-SeMet–labeled nectin-2v was also expressed in E. coli strain BL21
(DE3). Adaptive medium (20% Luria–Bertani medium and 80% M9 medium) was used to culture the cells at 37˚C to an OD600 of 0.4–0.8.
Subsequently, the E. coli cells were resuspended in M9 medium and cultured in restrictive medium (1% glucose) to an OD600 of 0.6–0.8 before
induction. The medium was supplemented with 60 mg/l L-SeMet; 100 mg/l
lysine, threonine, and phenylalanine; 50 mg/l leucine, isoleucine, and valine; and 0.5 mM isopropyl 1-thio-b-D-galactopyranoside. Induction occurred at 16˚C for 20 h. Finally, the L-SeMet–labeled protein was purified
and stored under the same conditions as was the native protein.
The cDNA-encoding extracellular domain of human DNAM-1 from aa
19–250 was inserted into the NdeI and XhoI sites of the pET-21a expression vector. The recombinant DNAM-1 was expressed in E. coli strain
BL21 as inclusion bodies. The DNAM-1 inclusion bodies were isolated, as
previously described (42), and then dissolved in denaturing buffer (100
mM Tris [pH 8], 0.4 M L-arginine, 2 mM EDTA, 4 M urea, 5 mM reduced
glutathione, 0.5 mM glutathione disulfide). The mixture was dialyzed
against an 8-fold volume of water for 8 h and then against an 8-fold
volume of 10 mM Tris [pH 8], 10 mM NaCl for an additional 8 h. The
renatured protein was concentrated and further purified by gel filtration
on a HiLoad 16/60 Superdex 200 PG column (GE Healthcare).
Materials and Methods
Gel-filtration and native gel assays
Ig variable (V) domain of human nectin-2 (nectin-2v) from aa 32–158 (wild
type [WT]) and mutant of the V domain (FAMP mutant = F145S, A143S,
M89S, and P94S) encoding cDNAs were inserted into the NdeI and XhoI
sites of the pET-21a expression vector (Novagen). Recombinant proteins
were expressed in Escherichia coli strain BL21 (DE3) as inclusion bodies.
Preparation, denaturation, and renaturation of the inclusion bodies were
performed, as previously described (39–41). Refolded soluble proteins
were first purified by gel filtration through a Superdex 75 10/300 GL
column with an AKTA FPLC (GE Healthcare), followed by a Resource Q
ion exchange column (GE Healthcare). Subsequently, gel filtration was
used as the final purification step, with a buffer of 20 mM Tris (pH 8) and
50 mM NaCl, and the target protein was concentrated to 5 or 10 mg/ml for
crystallization.
The hanging-drop vapor-diffusion technique was used to screen crystalgrowing conditions (Hampton Research) for WT nectin-2v and its FAMP
mutant at 18˚C. After 5–7 d, crystals of WT nectin-2v were obtained within
a solution 0.2 M ammonium acetate, 0.1 M Bis-Tris (pH 5.5), and 25%
PEG 3350. Similarly, crystals of the FAMP mutant were grown in 0.2 M
ammonium sulfate, 0.1 M Tris (pH 8.5), and 25% (w/v) PEG 3350.
Diffraction data of the crystals were collected at the Shanghai Synchrotron Radiation Facility (SSRF) Beamline BL17U (Shanghai, China),
whereas anomalous x-ray diffraction data were collected at peak (l =
0.9783 Å) wavelengths from the selenomethionyl-derivative protein at the
Advanced Photon Source (Chicago, IL). The collected intensities were
indexed, integrated, corrected for absorption, scaled, and merged using
HKL2000 (43).
Structure determination
Data collection and processing statistics are summarized in Table I. The
peak dataset was used for experimental phasing by the single-wavelength
anomalous dispersion method. A total of three expected heavy atoms was
located by SHELXD (44), and initial phases were calculated using Phaser
(45). Density modification was performed by DM (46). Approximately
90% of the residues were automatically traced by ARP/wARP (47).
The structures of WT nectin-2v and the FAMP mutant were solved by the
molecular-replacement method using Phaser (37) from the CCP4 program
suite with the L-SeMet model. Extensive model building and restrained
refinement were performed with COOT (48) and REFMAC5 (49). Further
rounds of refinement were performed using the phenix.refine program
implemented in the PHENIX package (50) with isotropic ADP refinement
and bulk solvent modeling. The stereochemical quality of the final models
was assessed with the program PROCHECK (51).
Protein structure accession numbers
The atomic coordinates and structure factors for nectin-2v and FAMP
mutant presented in this article have been submitted to the Protein Data
Bank (PDB) under accession numbers 4DFH (http://www.pdb.org/pdb/
search/structidSearch.do?structureId=4DFH) and 4DFI (http://www.pdb.
org/pdb/search/structidSearch.do?structureId=4DFI), respectively.
A calibrated Superdex 200 column (GE Healthcare) was used for molecular
weight estimation for the WT and mutant nectin-2v. Gel-filtration fractions
were subjected to SDS-PAGE or native PAGE in a Bio-Rad Mini-Protean II
or Protean II (20-cm plates) slab cell (Bio-Rad, Richmond, CA), with the
discontinuous buffer system in 7.5% (w/v) polyacrylamide-separating gels
and 5% (w/v) stacking gels. Samples were incubated at 45˚C for 30 min in
sample buffer, with a final concentration of 47 mM Tris-HCl (pH 7.8), 2%
(w/v) SDS, 7.5% (v/v) glycerol, 40 mM DTT as the thiol-reducing agent,
and 0.002% (w/v) bromphenol blue. The supernatants were centrifuged at
15,000 3 g for 2.5 min before loading. Native PAGE was performed in
7.5% (w/v) polyacrylamide-separating gels and 4% (w/v) stacking gels,
with or without 0.1% (w/v) CHAPS. Before fractionation, the samples
were incubated for 30 min at 45˚C in Tris-HCl (pH 7.8) buffer, with or
without 0.1% (w/v) CHAPS.
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
Protein expression and purification
Crystallization and data collection
The Journal of Immunology
5513
Cell culture and transfection
freshly prepared HRP-DAB substrate reagent (Beyotime) for 15 min; the
developing color was stopped by rinsing with PBS. The images were
obtained via laser confocal microscopy (Zeiss LSM 710).
Human 293T cells were cultured in DMEM supplemented with 10% FCS.
The endogenous expression of nectin-2 and DNAM-1 on the cell surface
was evaluated by flow cytometry with anti–nectin-2 mAb (sc-65333; Santa
Cruz) and anti–DNAM-1 mAb (FAB666F; R&D Systems) conjugated to
fluorescein, respectively.
To test the interaction of DNAM-1 with nectin-2, the plasmid pDNAM-1
was generated by inserting cDNA corresponding to full-length DNAM-1
into modified eukaryotic expression vector pCAGGS, which contains a
FLAG tag sequence downstream of the multiple coding site. Recombinant
plasmids were transfected into 293T cells, and the overexpression of FLAGfusion proteins was confirmed by Western blotting with anti-FLAG Ab
(Sigma), according to standard procedures.
Tetramer preparation and cell staining
Histochemistry staining
Twenty-four hours after the transfection of pDNAM-1 plasmids or empty
vectors, 293T cells were washed with PBS three times and fixed for 15 min
with 3% paraformaldehyde, followed by blocking with 1% BSA for 30 min.
Subsequently, the cells were incubated with biotinylated nectin-2v or FAMP
mutant, with a final concentration of 2 mg/ml, at 25˚C for 1 h. Cells were
washed three times with PBS, incubated with HRP-conjugated streptavidin
solution (1:200; Beyotime, Shanghai, China) for 1 h, and incubated with
Table I.
The binding affinity between DNAM-1 and nectin-2v or FAMP mutant
was analyzed at 25˚C on a Biacore 3000 machine with CM5 chips (GE
Healthcare). HBS-EP buffer (10 mM HEPES, 150 mM NaCl, 3 mM
EDTA, 0.005% Tween 20) was used for all measurements. For surface
plasmon resonance (SPR) measurements, both DNAM-1 and nectin-2v
were purified by gel filtration using a Superdex 75 column (GE Healthcare). About 700 response units of nectin-2v were immobilized on the
chip, followed by blockade with EDTA. When the data collection was
finished in each cycle, the sensor surface was regenerated with 10 mM
NaOH. A series of concentrations ranging from 1.56 to 50 mM was
designed for the experiment. Sensograms were fit globally with Biacore
3000 analysis software (BIAevaluation Version 4.1) using 1:1 Langmuir
binding mode. The equilibrium binding constant (KD) values were calculated using a nonlinear curve fit (y = [P1x]/[P2 + x]).
Results
The dimeric structure of nectin-2v
The V-set domain (residue 32–158) of the extracellular portion
of nectin-2 (i.e., nectin-2v) was produced through an in vitro
refolding procedure (52) from E. coli-expressed inclusion bodies. The crystal structure was solved at a resolution of 1.85 Å
(Table I). Two molecules of nectin-2v were found in one asymmetric unit (Fig. 1A), and they form a tight dimer (perpendicular
head–head interaction). As expected, nectin-2v displays a typical
Ig V-set structure, which consists of nine antiparallel b-strands.
Structural superimposition of one molecule of nectin-2v with the
V domains of all of the known structures of nectin and Necl
family members, including nectin-1 (54), Necl-1 (36), Necl-3
(38), and Necl-5 (55), indicated that nectin-2v retains a similar
overall conformation compared with other members (Fig. 1B).
(Structurally determined nectin-1, Necl-1, and Necl-5 are derived
from human, whereas Necl-3 is from mouse; however, the sequence of the visible part of the Necl-3 structure is 100% con-
Data collection, phasing, and refinement statistics
Data collection
Space group
Cell dimensions
a, b, c (Å)
a, b, g (˚)
Nectin-2v (Native)
Nectin-2v (Se-Met)
FAMP Mutant
P21
P41212
P212121
52.7, 43.9, 56.1
90.0, 118.2, 90.0
43.3, 45.8, 52.6
90.0, 90.0, 90.0
43.3, 45.8, 52.6
90.0, 90.0, 90.0
Peak
Wavelength
Resolution (Å)
Rsym or Rmerge (%)
I/sI
Completeness (%)
Redundancy
Refinement
Resolution (Å)
No. reflections
Rwork/Rfree (%)
No. atoms
Protein
Water
B-factors
Protein
Water
RMSD
Bond lengths (Å)
Bond angles (˚)
1.5418
50–1.85 (1.92–1.85)
3.6 (15.1)
41.7 (11.4)
99.8 (98.8)
3.6 (3.4)
0.9793
50–1.8 (1.86–1.80)
8.9 (89.4)
35.3 (1.8)
99.5 (91.8)
16.0 (11.9)
0.9795
50–1.8 (1.86–1.80)
11.4 (46.9)
16.6 (3.2)
96.1 (88.7)
5.9 (3.5)
27.97–1.85
19,468
17.10/19.42
34.54–1.80
9,250
19.74/23.68
1,914
300
952
124
23.2
38.0
22.7
28.8
0.011
0.096
0.016
1.067
Values in parentheses are for highest-resolution shell.
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
To prepare tetramers of nectin-2v and its mutant FAMP, cDNAs corresponding to nectin-2v and the FAMP with C-terminal biotin-binding tags
were subcloned into the NdeI and XhoI sites of the pET30a prokaryotic
expression vector. Recombinant proteins were expressed in BL21 E. coli
cells as inclusion bodies and were subsequently refolded and purified, as
described previously (52, 53). Biotinylation of proteins in vitro was achieved by incubating the protein with the Bir A biotin protein ligase,
D-biotin, and ATP (52). Free biotin was removed through gel filtration using
AKTA FPLC (GE Healthcare). The efficiency of the in vitro biotinylation
was determined through streptavidin-shift assays (3). The proteins displaying high efficiency of biotinylation were used to produce tetramers by
incubating with PE-conjugated streptavidin (Sigma).
The 293T cells were transiently transfected with pDNAM-1 plasmids or
empty vectors. Twenty-four hours after the transfection, cells were treated
with 0.025% trypsin, resuspended in PBS, and stained with the tetramer
(0.05 mg/ml) for 1 h. Subsequently, the cells were analyzed by flow
cytometry (EasyCyte Mini; Guava).
Surface plasmon resonance measurements
5514
STRUCTURE OF CD112 HOMODIMER AND ITS BINDING TO CD226
served between mouse and human.) However, the alignment of
these molecules clearly shows that nectin-2v contains an unusual
extended loop (D-E loop, covering residue 112–117 aligned with
Necl-5) between the D and E strands (Fig. 1B). The D-E loops of
the two molecules of the nectin-2v dimer protrude with the same
orientation (to the upper side of the dimeric interface) and display
a flexible conformation (the flexibility can be derived from the fact
that there is no electronic density observed for molecule 2 of the
nectin-2v dimer). More importantly, the uncommon loop of nectin2v is due to the addition of residues in the protein sequence of
nectin-2 relative to other nectin and Necl family molecules (Fig. 1C),
indicating that this unique loop is an intrinsic feature of nectin-2.
The structures of nectin-1 (three Ig domains) (54), Necl-1 (V
domain alone) (36), and Necl-3 (V domain alone) (38) reveal that
these molecules also form dimers through the interaction of their
V domains. By superimposing nectin-2 on nectin-1, we found that
the overall conformation of the dimer formed by nectin-2v is
similar to nectin-1, with the root mean square deviation (RMSD)
of 1.472 Å (Fig. 2A), much smaller than the RMSD when aligning
with Necl-1 (RMSD = 8.388 Å) or Necl-3 (RMSD = 8.109 Å).
When we superimposed the first molecules in the asymmetric unit
of these four dimeric proteins (Fig. 2B), we found that the second
molecule (molecule 2 in the asymmetric unit) of nectin-2v is located in the same position as nectin-1 molecule 2 but is different
from the position of molecule 2 of Necl-1 and Necl-3.
Detailed analyses of the interface of the two molecules in the
nectin-2v dimer indicated that the formation of the dimer depends
on highly hydrophobic interactions between the two molecules.
The interface area of the two nectin-2v molecules is 861.0 Å2,
which is larger than that of nectin-1 (841.0 Å2), Necl-1 (729.9 Å2),
and Necl-3 (698.0 Å2). Moreover, the interface is mainly formed
by hydrophobic amino acids, including Tyr64, Leu67, Ala84, Met89,
Pro94, Ala143, and Phe145 (Figs. 1C, 2C). These residues within the
interface area of the two molecules have a total exposed hydrophobic area of 470.1 Å2, occupying 55.9% of the total interface
area of the two molecules. The unconventional large hydrophobic
interface of nectin-2v dimer is the consequence of a combination
of nectin-2v characteristic residues (unique residues Tyr64, Leu67,
Ala84, and Pro94 in nectin-2 compared with other nectins/Necls in
these positions). In addition to the hydrophobic interaction, the
two molecules form hydrogen bonds to enhance dimer formation.
The generally symmetric hydrogen bonds are formed between
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
FIGURE 1. Structure of nectin-2v dimer. (A) Overview of the nectin-2v dimer. Two molecules within one asymmetric unit of the nectin-2v structure
clearly display a dimer conformation. Molecule 1 is denoted in green with strands and loops covering with surface. Molecule 2 is represented similarly in
orange. The presumed D-E loop in molecule 2, which has no corresponding electron density, is signified by the orange dotted line. (B) The uncommon D-E
loop within the structure of nectin-2v. Structural superposition of five structure-solved nectin/Necls, including nectin-2 (green), nectin-1 (PDB: 3ALP;
purple), Necl-1 (also known as SynCAM3, PDB: 1Z9M; blue), Necl-3 (also known as SynCAM2, PDB: 3M45; yellow), and Necl-5 (PDB: 3EOW; brown),
demonstrating the unique protrusion of the D-E loop of nectin-2. (C) Structure-based sequence alignment of nectin-2 and other structure-determined nectin
and Necls. Black arrows denote b-strands. Loops above the alignment indicate a-helices. The residues in blue boxes are highly conserved (.80%), and the
residues highlighted in red are completely conserved. The green number “1” denotes residues that form disulfide bonds. Residues that were selected for
mutation in the monomeric FAMP mutant of nectin-2v are highlighted with purple asterisks. The sequence alignment was generated with Clustal X (67) and
ESPript (68).
The Journal of Immunology
5515
Ser66, Asn71, Asn80, His86, Ser92, Glu141, Pro146, Ser149, and
Arg151 in molecule 1 and Ser66, Asn71, Gln81, His86, Ser92, Pro146,
Ser149, and Arg151 in molecule 2 (Fig. 2D).
The role of the hydrophobic interface in nectin-2v dimer
formation
Structural analysis indicated that the dimer conformation of nectin2v mainly depends on the hydrophobic interface between the two
V domains. To further verify the pivotal role of the hydrophobic
interface in the homodimer of nectin-2v and elucidate the key
residues within the interface, we generated a mutant of nectin-2v,
termed the FAMP mutant. Four hydrophobic amino acids (Met89,
Pro94, Ala143, and Phe145) in the dimer interface of nectin-2v were
selected for their nondirect van der Waals interactions with the
corresponding residues in the other dimer molecule and replaced
with the nucleophilic amino acid serine in the FAMP mutant. To
observe whether mutagenesis of the hydrophobic interface can ablate the nectin-2v homodimer, we determined the structure of the
FAMP mutant. Indeed, we found only one FAMP molecule in the
asymmetric unit and no dimer as the one in the structure of nectin-2v
is formed through the crystal packing or symmetric operation, indicating that the FAMP mutant is a monomeric mutant of nectin-2v.
The overall structure of the FAMP mutant is similar to both
molecules 1 and 2 in the homodimer of nectin-2v, with RMSDs
of 0.524 and 0.496 Å, respectively. When we superimposed the
FAMP mutant structure onto molecule 1 of nectin-2v, we found
that the major differences between the two structures involve a
conformational shift of the loops on the borderline of the interface of the nectin-2v homodimer (Fig. 3A). These loops in the
FAMP mutant float away from the direction of the dimer interface
and adopt a loose conformation compared with the molecules in
the homodimer of nectin-2v. Comparison of the electronic density of the mutated residues from the FAMP mutant with WT
nectin-2v clearly demonstrates that Met89, Pro94, Ala143, and
Phe145 of nectin-2v have been replaced by four serines (Fig. 3B).
Further analysis of the hydrophobicity of the altered surface of
the FAMP mutant (where the interface of the homodimer of native nectin-2v is located) indicated that mutation of the four
residues dramatically reduced the area of the hydrophobic interface of nectin-2v (Fig. 3C, 3D), as expected. Transformation of
the nectin-2v dimer into a monomer of the FAMP mutant is
highly correlated with the destruction of the hydrophobic interface of nectin-2v, indicating the pivotal role of the hydrophobic
interface in the formation of nectin-2v homodimers. Thus, the
hydrophobic interface and the ability of nectin-2v to form a tight
dimer may have important roles in its physiological function
(i.e., DNAM-1 binding).
To further verify the monomeric FAMP mutant compared with
the dimeric status of nectin-2v, we performed a series of biochemical assays. First, nectin-2v and the FAMP mutant were subjected to
size-exclusion chromatography (Superdex 200 10/300 GL column).
Nectin-2v was eluted as a single peak at an elution volume of 12 ml,
corresponding to a globular protein with an estimated molecular
mass ∼30 kDa. However, the FAMP mutant was eluted at a volume
of 14 ml, corresponding to a much smaller molecular mass ∼13
kDa (Fig. 4A). Subsequently, we further analyzed the status of the
two proteins by performing native PAGE (Fig. 4B); the results
demonstrated that nectin-2v and the FAMP mutant carried different
negative charges in the native PAGE gel. Both of the proteins
moved from the negative electrode to the positive, but the rate of
nectin-2v migration was much slower than that of the FAMP mutant, indicating the distinct status of the two proteins.
Taken together, both the structural and biochemical results indicate that destruction of the hydrophobic interface of nectin-2v
abolishes its dimerization. Moreover, the successful mutation
assays guided by the structure analyses revealed that the bacterially
produced nectin-2 protein, which was used to determine the
structure, has a naturally folded conformation.
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
FIGURE 2. The dimer formation of nectin-2v.
(A) Alignment of nectin-2 (green) with nectin-1
(purple) shows that the two molecules in these
dimers form perpendicular or even acute angles.
(B) Nectin-2v homodimer formation is similar to
nectin-1 but different from Necl-1 and Necl-3.
When superimposing molecule 1 of these four
structures, molecule 2 of nectin-2 (green) is located in the same position as molecule 2 of
nectin-1 (purple), but molecule 2 of Necl-1
(blue) and Necl-3 (yellow) occupies a different
position. (C) Large hydrophobic interface of the
nectin-2v dimer. The vacuum electrostatic potential surface based on Amber 99 charges of
molecule 1 of nectin-2v shows the hydrophobic
area of the dimer interface. Seven major hydrophobic residues whose side chains point to the
interface are represented by purple sticks. The
four residues in the yellow ellipses were mutated
to generate the monomeric mutant of nectin-2v
(FAMP mutant). (D) Hydrogen bond network
between molecule 1 (green) and molecule 2
(orange) of the nectin-2v dimer.
5516
STRUCTURE OF CD112 HOMODIMER AND ITS BINDING TO CD226
Soluble nectin-2v, but not the monomer mutant, binds to cell
surface-expressed DNAM-1
To test the physiological activity of our recombinant nectin-2v, we
prepared a tetramer of nectin-2v using a biotin-streptavidin system to
FIGURE 4. Biochemical confirmation of
nectin-2v mutant (FAMP mutant) as a monomer. (A) Size-exclusion chromatographs of
nectin-2v (cyan) and the FAMP mutant (purple)
on a Superdex 200 column. Compared with the
early elution of WT nectin-2v at 12 ml, the
peak of the FAMP mutant is at 14 ml, with the
expected molecular mass ∼30 kDa (nectin-2v)
and 13 kDa (FAMP mutant). The approximate
positions of the molecular mass standards of
67.0, 43.0, and 13.7 kDa are marked. (B)
Analysis of the conformational status of nectin2v and FAMP by native PAGE. The results in
(A) and (B) are representative of three independent experiments.
Binding-kinetics analysis of DNAM-1 to nectin-2v and FAMP
mutant measured by BIAcore
Binding affinity of DNAM-1 to nectin-2v was performed using SPR
technology. When DNAM-1 was injected over WT nectin-2v at
25˚C, a fast binding kinetics was observed (Kon and Koff are too
fast and are beyond the BIAcore detection sensitivity) (Fig. 6A).
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
FIGURE 3. Structural comparison of monomer mutant (FAMP mutant)
with WT nectin-2v. (A) Structural alignment of the FAMP mutant (cyan)
with molecule 1 (green) of the nectin-2v dimer. The four residues (Met89,
Pro94, Ala143, and Phe145) in the hydrophobic interface of the nectin-2v
dimer and the corresponding substituted serines in the FAMP mutant are
represented by purple and red sticks, respectively, in the boxes. The conformational shift of the loops on the edge of the dimer interface of nectin-2v
is observed when comparing the two structures. These loops in the FAMP
monomer shift away (as the orientation of black arrows) from the direction
of the dimer interface compared with molecule 1 in the nectin-2v cis-dimer. (B) The electron density at the 1s contour level clearly shows that
residues Met89, Pro94, Ala143, and Phe145 in nectin-2v (purple residues, left
panels) were mutated to four serines in the FAMP mutant (red residues,
right panel). (C and D), Deconstruction of the hydrophobic interface of
nectin-2v cis-dimer. After mutation of the four hydrophobic residues in the
dimer interface, the corresponding vacuum electrostatic potential surface
within the area of the four residues (yellow circles) obviously changed. (C)
WT nectin-2v. (D) FAMP mutant.
assess the binding of nectin-2 to DNAM-1. Cultured 293T cells
were used to generate a cell line transiently expressing DNAM-1.
First, to test the expression status of nectin-2 and its receptor
(DNAM-1) on the surface of the 293T cells, we stained the untransfected cells with anti–nectin-2 and anti–DNAM-1 mAbs (Fig. 5A).
Flow cytometry analysis revealed that 293T cells express nectin-2,
but not DNAM-1, on the cell surface, as previously reported (15).
Nectin-2v tetramers were also used to stain the 293T cells. As expected, no binding of nectin-2v to the 293T cells was observed
(Fig. 5A). These results indicate that 293T cells can act as a model
to study the interaction of nectin-2 and its receptor DNAM-1.
Subsequently, 293T cells were transfected with pDNAM-1
plasmid harboring full-length DNAM-1, and the expression of
DNAM-1 on the transfected cells was verified with flow cytometry
using the anti–DNAM-1 mAb (Fig. 5B), as well as Western blotting (data not shown). The cells were then used to analyze the
binding of nectin-2 to DNAM-1. Notably, tetramers of nectin-2v
bound to the surface of the DNAM-1–transfected cells (Fig. 5B).
Thus, we concluded that recombinant, soluble nectin-2v has the
potential to perform an in vivo biological function, indicating a
natural three-dimensional structural conformation of this protein.
To further elucidate the binding profile of nectin-2v with
DNAM-1, we also prepared tetramers of the monomeric FAMP
mutant of nectin-2v. DNAM-1–expressing 293T cells were then
stained with the tetramers of the FAMP mutant in comparison
with the tetramers of WT nectin-2v (Fig. 5B). The results of flow
cytometry demonstrated that, although the nectin-2v tetramer
binds to the DNAM-1–expressing cells (25% positive cells with
high fluorescence), tetramers of the FAMP mutant cannot stain the
cells (no positive cells with high fluorescence). We also performed
in situ staining of the DNAM-1–expressing 293T cells using
biotinylated nectin-2v and its FAMP mutant (Fig. 5C); the in situ
coloring of the stained cells was developed via HRP. In contrast to
efficient staining of the DNAM-1–expressing cells by nectin-2v,
the cells could not be stained by FAMP mutant. The disruption of
the dimerization interface of nectin-2 leads to a dramatic loss of
the binding of nectin-2 to DNAM-1, indicating that this unusual
hydrophobic interface between the two molecules within the
nectin-2 homodimer plays a role (direct or indirect) in the interaction with DNAM-1.
The Journal of Immunology
5517
The KD value was calculated as 8.97 mM, using a least-squares fit
for the data, assuming a 1:1 interaction in the steady-state affinity
model (Fig. 7), which was confirmed by a Scatchard plot (Fig.
6B). To compare the binding capacity of WT nectin-2v and the
FAMP mutant to DNAM-1, DNAM-1 (200 mM) was injected over
WT nectin-2v and the FAMP mutant at the same time. The results
show that WT nectin-2v binds to DNAM-1 well, but the FAMP
mutant loses its binding capacity to DNAM-1 (Fig. 6C).
Discussion
In this study, a series of biochemical, crystallographic, and cell
biological analyses characterized the homodimer formation and the
binding of nectin-2 to its immune costimulatory receptor DNAM-1.
A perpendicular homodimer structure of nectin-2 V-set was solved.
Furthermore, the pivotal role for the hydrophobic interface, including the key residues, was verified in the dimerization of
nectin-2. The disruption of the hydrophobic interface also abol-
FIGURE 6. BIAcore SPR analysis of
DNAM-1 binding to nectin-2v and its
FAMP monomer mutant. (A) The kinetic
profile of DNAM-1 binding to nectin-2v
is shown. A series of concentrations of
DNAM-1 (ranging from 1.56 to 50 mM)
was used to measure the binding kinetics,
with nectin-2v immobilized on the CM5
chip. (B) Response units were plotted
against protein concentration calculated
from the BCA Kit. KD was calculated as
8.97 mM using a least-squares fit to the
data, assuming a 1:1 interaction, which was
confirmed by a Scatchard plot. (C) A total
of 200 mM DNAM-1 was injected through
flow cells, with nectin-2v and the FAMP
mutant immobilized on the same CM5
chip. Clearly, WT nectin-2v can bind to
DNAM-1 well, but the FAMP mutant loses
its binding capacity.
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
FIGURE 5. Soluble nectin-2v, but not its monomer
mutant (FAMP mutant), binds to cell-expressed
DNAM-1. (A) Cultured 293T cells stained with anti–
DNAM-1, anti–nectin-2 Abs, and tetramers of nectin2v (blue lines). The cyan lines denote the control cells
without any staining. Nectin-2, but not DNAM-1, expression is clearly shown on the 293T cell surface. (B)
Left panel, DNAM-1–transfected 293T cells could
clearly be stained by anti–DNAM-1 Ab (red line). Blue
line represents the control cells stained with anti–
DNAM-1 after the transfection of empty vectors.
DNAM-1–transfected 293T cells are stained with
nectin-2v (middle panel) or FAMP (right panel) tetramers (red lines). The control cells, which were
transfected with empty vector pcDNA3.0, were stained
with these tetramers as negative controls (blue lines).
(C) In situ staining of DNAM-1–transfected 293T cells
with biotinylated nectin-2v or FAMP (2 mg/ml final
concentration). Original magnification 3400. HRPlabeled streptavidin and HRP-DAB substrate reagents
were used to stain the cells. Cells transfected with
pcDNA3.0 were stained by biotinylated nectin-2v as
negative controls. The results shown are representative
of three independent experiments.
5518
STRUCTURE OF CD112 HOMODIMER AND ITS BINDING TO CD226
ishes the binding of nectin-2 to DNAM-1. Thus, these results may
be helpful for the interpretation of the properties of homodimerization of nectin-2 and the interactive mode of nectin-2 with
DNAM-1.
Nectin and Necls are believed to function in both homophilic and
heterophilic dimers, as well as cis- and trans-dimers, in a variety of
cells, including epithelium, immune cells, and neurologic system
cells (24, 27–29). In addition, they interact with a broad spectrum
of ligands/counterreceptors to execute their functions (14–16, 24,
56, 57). More importantly, they can also be used as receptors for
virus entry (31, 58–60). However, with these diverse functions, it
is less well understood how the same molecule functions in different cells, especially with regard to the ability of the same
molecule to bind different ligand/counterreceptors. Structural information about nectin and Necls is very limited. Only three
homodimers (three Ig domains of nectin-1, one Ig domain of Necl-1
and Necl-3) and one monomer (two Ig domains of Necl-5) of
nine members of this family have been described but with different conclusions (36, 38, 54, 55). The structure of nectin-1 is
regarded as a cis-dimer, but Necl-1 is considered a trans-dimer
(36, 38, 54). There is no structural report of any heterodimers for
these family members. Therefore, the true scenario of the structural basis for the diverse binding properties of nectins/Necls
remains a big issue. Recently, our group (61) and Di Giovine
et al. (62) independently discovered that human HSV-1 glycoprotein D (gD) unexpectedly precluded the nectin-1 dimer from
entering host cells (i.e., the binding interface of nectin-1 to gD
is almost the same as the nectin-1 dimer interface), revealing
the monomeric nectin-1 binding mode to gD, rather than the suspected nectin-1 dimeric binding.
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
FIGURE 7. Proposed model of nectin-2 dimer formation and the
binding to DNAM-1. (A and B) Two dimer models based on the structure
of nectin-2v determined in our study. Cell surface (blue double membrane)-expressed nectin-2 forms trans-dimer (A) and/or cis-dimer (B)
through the interaction of V domains. The C1 and C2 domains (purple and
yellow circles) may contribute to cis-dimer formation (33, 35). (C and D)
Two possible binding models of nectin-2 with DNAM-1. (C) DNAM-1
(green ellipses) expressed on cell surfaces (red double membrane) binds to
nectin-2 through the same interface as in the nectin-2v dimer. (D) DNAM-1
interacts with the cis-dimer of nectin-2.
The structure of nectin-2v clearly displays a tight homodimer
formed through a large hydrophobic interface. The recently determined structure of another nectin family member, nectin-1,
confirms the cis-homodimer binding, mainly through a network
of hydrogen bonds (54). Via analysis of the nectin-1 dimer, Narita
et al. (54) proposed that the trans-interaction of nectins may depend on the upper side of the nectin cis-dimer. However, in our
study, tetramers of recombinant nectin-2v did not bind to 293T
cells, which express nectin-2 on their surface (Fig. 5A). One interpretation of this phenomenon is that soluble recombinant
nectin-2v exists as a cis-dimer in solution. The nonbinding of this
cis-dimer to cell surface-expressed nectin-2 indicates that transinteraction of nectin-2 expressed on different cell surfaces may
also use a similar hydrophobic interface as in the cis-dimer. This
implies that both cis- and trans-interaction of nectin-2 may occur
via a similar manner through the dimerization of the first Ig-like
V-set domain (Fig. 7A, 7B). In contrast, on the surface of the same
cell, nectin-2 would have two statuses, monomer and cis-dimer,
that exist in equilibrium. The monomers can trans-bind to each
other through their first V-set domain. This mode of trans-interaction of nectin-2 can also be applied to the heterointeraction of
different nectins. There is also a possibility that nectin-2 on different cells behaves differently (i.e., as either a monomer or cisdimer). All of these questions remain to be answered. In addition
to the contribution of V-set domain in the dimer formation, previous cell biological studies indicated that the C-set domains of
nectin or nectin-like molecules are involved in cis-interaction
(Fig. 7A) (33–35). Related studies from different groups independently demonstrated that the C-set domains of nectin-2 and Necl-2
(SynCAM1) on the surface of the cells are responsible for the
lateral clustering of the cis-interaction, which could complement
or even promote the trans-interaction of these molecules (33, 35).
Nectin-2v, produced through the prokaryotic expression system
in our study, has no posttranslational modified residues. However, it was demonstrated that the posttranslational modification
of adhesion molecules can modulate their homophilic adhesion
and other functions (36). The KD value of nectin-2 binding to
DNAM-1 determined in our study is slightly lower than the KD
value reported by Tahara-Hanaoka and colleagues (23). They produced the nectin-2 proteins through an eukaryotic expression system featured with postmodifications of the protein. This may indicate that postmodified residues of nectin-2 also contribute to the
nectin-2/DNAM-1 interaction. Fogel et al. (38) determined that Nglycosylation within the V domains of Necls, also known as
SynCAMs, differentially affects Necl properties. The N-glycans
at Asn70/Asn104 of Necl-2 (SynCAM1) increase its homophilic
interactions, whereas N-glycosylation of Necl-3 (SynCAM2) at
Asn60 hinders the homodimer formation. It was determined that
Necl-2 V domain accommodates the glycosylation sites Asn70/
Asn104 flanking the homo-binding interface, which may favor
dimer formation by limiting the conformational space available to
the V domain. In contrast, an N-glycan on residue Asn60 of Necl-3
was identified within the adhesive interface of its V domain, which
may interfere with homodimer formation. When we analyzed the
potential N-glycosylation sites within the V domain of nectin-2
through prediction (63), we found that Asn107 may be N-glycosylated. Based on our structure of nectin-2v, Asn107 locates out of
the homodimer-binding interface and near the V/C-set domain
interface, which indicates that N-glycosylation may augment
homodimer formation of nectin-2.
In our study, the abolition of nectin-2 binding to DNAM-1, by
altering the dimer-binding interface of nectin-2, indicates two
possibilities of the binding mode between nectin-2 and DNAM-1
(Fig. 7C, 7D): the binding of DNAM-1 to nectin-2, like HSV-1 gD
The Journal of Immunology
Acknowledgments
We thank Dr. Fuliang Chu, Dr. Beiwen Zheng, Dr. Guangwen Lu, Dr. Zheng
Fan, Shihong Zhang, Dr. Yi Shi, and Qun Yan for excellent assistance and
suggestions throughout the project. Assistance from the staff at the Shanghai
Synchrotron Radiation Facility (SSRF Beamline BL17U) (Shanghai, China)
and the Argonne National Laboratory (Chicago, IL) is acknowledged.
Disclosures
The authors have no financial conflicts of interest.
References
1. Bryceson, Y. T., M. E. March, H. G. Ljunggren, and E. O. Long. 2006. Activation, coactivation, and costimulation of resting human natural killer cells.
Immunol. Rev. 214: 73–91.
2. Cerwenka, A., and L. L. Lanier. 2001. Natural killer cells, viruses and cancer.
Nat. Rev. Immunol. 1: 41–49.
3. Liu, J., S. Zhang, S. Tan, B. Zheng, and G. F. Gao. 2011. Revival of the identification of cytotoxic T-lymphocyte epitopes for immunological diagnosis,
therapy and vaccine development. Exp. Biol. Med. (Maywood) 236: 253–267.
4. Lanier, L. L. 2005. NK cell recognition. Annu. Rev. Immunol. 23: 225–274.
5. Dunn, G. P., C. M. Koebel, and R. D. Schreiber. 2006. Interferons, immunity and
cancer immunoediting. Nat. Rev. Immunol. 6: 836–848.
6. Vivier, E., E. Tomasello, M. Baratin, T. Walzer, and S. Ugolini. 2008. Functions
of natural killer cells. Nat. Immunol. 9: 503–510.
7. Shibuya, A., D. Campbell, C. Hannum, H. Yssel, K. Franz-Bacon,
T. McClanahan, T. Kitamura, J. Nicholl, G. R. Sutherland, L. L. Lanier, and
J. H. Phillips. 1996. DNAM-1, a novel adhesion molecule involved in the cytolytic function of T lymphocytes. Immunity 4: 573–581.
8. Chen, L., X. Xie, X. Zhang, W. Jia, J. Jian, C. Song, and B. Jin. 2003. The
expression, regulation and adhesion function of a novel CD molecule, CD226, on
human endothelial cells. Life Sci. 73: 2373–2382.
9. Kojima, H., H. Kanada, S. Shimizu, E. Kasama, K. Shibuya, H. Nakauchi,
T. Nagasawa, and A. Shibuya. 2003. CD226 mediates platelet and megakaryocytic cell adhesion to vascular endothelial cells. J. Biol. Chem. 278: 36748–
36753.
10. Dardalhon, V., A. S. Schubart, J. Reddy, J. H. Meyers, L. Monney, C. A. Sabatos,
R. Ahuja, K. Nguyen, G. J. Freeman, E. A. Greenfield, et al. 2005. CD226 is
specifically expressed on the surface of Th1 cells and regulates their expansion
and effector functions. J. Immunol. 175: 1558–1565.
11. Shibuya, K., J. Shirakawa, T. Kameyama, S. Honda, S. Tahara-Hanaoka,
A. Miyamoto, M. Onodera, T. Sumida, H. Nakauchi, H. Miyoshi, and
A. Shibuya. 2003. CD226 (DNAM-1) is involved in lymphocyte functionassociated antigen 1 costimulatory signal for naive T cell differentiation and
proliferation. J. Exp. Med. 198: 1829–1839.
12. Iguchi-Manaka, A., H. Kai, Y. Yamashita, K. Shibata, S. Tahara-Hanaoka,
S. Honda, T. Yasui, H. Kikutani, K. Shibuya, and A. Shibuya. 2008. Accelerated
tumor growth in mice deficient in DNAM-1 receptor. J. Exp. Med. 205: 2959–2964.
13. Dieudé, P., M. Guedj, M. E. Truchetet, J. Wipff, L. Revillod, G. Riemekasten,
M. Matucci-Cerinic, I. Melchers, E. Hachulla, P. Airo, et al. 2011. Association of
the CD226 Ser(307) variant with systemic sclerosis: evidence of a contribution
of costimulation pathways in systemic sclerosis pathogenesis. Arthritis Rheum.
63: 1097–1105.
14. Tahara-Hanaoka, S., A. Miyamoto, A. Hara, S. Honda, K. Shibuya, and
A. Shibuya. 2005. Identification and characterization of murine DNAM-1
(CD226) and its poliovirus receptor family ligands. Biochem. Biophys. Res.
Commun. 329: 996–1000.
15. Bottino, C., R. Castriconi, D. Pende, P. Rivera, M. Nanni, B. Carnemolla,
C. Cantoni, J. Grassi, S. Marcenaro, N. Reymond, et al. 2003. Identification of
PVR (CD155) and Nectin-2 (CD112) as cell surface ligands for the human
DNAM-1 (CD226) activating molecule. J. Exp. Med. 198: 557–567.
16. Pende, D., C. Bottino, R. Castriconi, C. Cantoni, S. Marcenaro, P. Rivera,
G. M. Spaggiari, A. Dondero, B. Carnemolla, N. Reymond, et al. 2005. PVR
(CD155) and Nectin-2 (CD112) as ligands of the human DNAM-1 (CD226)
activating receptor: involvement in tumor cell lysis. Mol. Immunol. 42: 463–
469.
17. Gilfillan, S., C. J. Chan, M. Cella, N. M. Haynes, A. S. Rapaport, K. S. Boles,
D. M. Andrews, M. J. Smyth, and M. Colonna. 2008. DNAM-1 promotes activation of cytotoxic lymphocytes by nonprofessional antigen-presenting cells and
tumors. J. Exp. Med. 205: 2965–2973.
18. Pende, D., G. M. Spaggiari, S. Marcenaro, S. Martini, P. Rivera, A. Capobianco,
M. Falco, E. Lanino, I. Pierri, R. Zambello, et al. 2005. Analysis of the receptorligand interactions in the natural killer-mediated lysis of freshly isolated myeloid
or lymphoblastic leukemias: evidence for the involvement of the Poliovirus receptor (CD155) and Nectin-2 (CD112). Blood 105: 2066–2073.
19. Tahara-Hanaoka, S., K. Shibuya, H. Kai, A. Miyamoto, Y. Morikawa,
N. Ohkochi, S. Honda, and A. Shibuya. 2006. Tumor rejection by the poliovirus
receptor family ligands of the DNAM-1 (CD226) receptor. Blood 107: 1491–
1496.
20. Pende, D., R. Castriconi, P. Romagnani, G. M. Spaggiari, S. Marcenaro,
A. Dondero, E. Lazzeri, L. Lasagni, S. Martini, P. Rivera, et al. 2006. Expression
of the DNAM-1 ligands, Nectin-2 (CD112) and poliovirus receptor (CD155), on
dendritic cells: relevance for natural killer-dendritic cell interaction. Blood 107:
2030–2036.
21. Reymond, N., A. M. Imbert, E. Devilard, S. Fabre, C. Chabannon, L. Xerri,
C. Farnarier, C. Cantoni, C. Bottino, A. Moretta, et al. 2004. DNAM-1 and PVR
regulate monocyte migration through endothelial junctions. J. Exp. Med. 199:
1331–1341.
22. Bachelet, I., A. Munitz, D. Mankutad, and F. Levi-Schaffer. 2006. Mast cell
costimulation by CD226/CD112 (DNAM-1/Nectin-2): a novel interface in the
allergic process. J. Biol. Chem. 281: 27190–27196.
23. Tahara-Hanaoka, S., K. Shibuya, Y. Onoda, H. Zhang, S. Yamazaki,
A. Miyamoto, S. Honda, L. L. Lanier, and A. Shibuya. 2004. Functional characterization of DNAM-1 (CD226) interaction with its ligands PVR (CD155) and
nectin-2 (PRR-2/CD112). Int. Immunol. 16: 533–538.
24. Takai, Y., J. Miyoshi, W. Ikeda, and H. Ogita. 2008. Nectins and nectin-like
molecules: roles in contact inhibition of cell movement and proliferation. Nat.
Rev. Mol. Cell Biol. 9: 603–615.
25. Takai, Y., K. Irie, K. Shimizu, T. Sakisaka, and W. Ikeda. 2003. Nectins and
nectin-like molecules: roles in cell adhesion, migration, and polarization. Cancer
Sci. 94: 655–667.
26. Fuchs, A., and M. Colonna. 2006. The role of NK cell recognition of nectin and
nectin-like proteins in tumor immunosurveillance. Semin. Cancer Biol. 16: 359–
366.
27. Kurita, S., H. Ogita, and Y. Takai. 2011. Cooperative role of nectin-nectin and
nectin-afadin interactions in formation of nectin-based cell-cell adhesion. J. Biol.
Chem. 286: 36297–36303.
28. Miyahara, M., H. Nakanishi, K. Takahashi, K. Satoh-Horikawa, K. Tachibana,
and Y. Takai. 2000. Interaction of nectin with afadin is necessary for its clustering at cell-cell contact sites but not for its cis dimerization or trans interaction.
J. Biol. Chem. 275: 613–618.
29. Togashi, H., K. Kominami, M. Waseda, H. Komura, J. Miyoshi, M. Takeichi, and
Y. Takai. 2011. Nectins establish a checkerboard-like cellular pattern in the
auditory epithelium. Science 333: 1144–1147.
30. Ogita, H., and Y. Takai. 2006. Nectins and nectin-like molecules: roles in cell
adhesion, polarization, movement, and proliferation. IUBMB Life 58: 334–343.
31. Mendelsohn, C. L., E. Wimmer, and V. R. Racaniello. 1989. Cellular receptor for
poliovirus: molecular cloning, nucleotide sequence, and expression of a new
member of the immunoglobulin superfamily. Cell 56: 855–865.
32. Xu, Z., and B. Jin. 2010. A novel interface consisting of homologous immunoglobulin superfamily members with multiple functions. Cell. Mol. Immunol. 7:
11–19.
33. Momose, Y., T. Honda, M. Inagaki, K. Shimizu, K. Irie, H. Nakanishi, and
Y. Takai. 2002. Role of the second immunoglobulin-like loop of nectin in cellcell adhesion. Biochem. Biophys. Res. Commun. 293: 45–49.
34. Yasumi, M., K. Shimizu, T. Honda, M. Takeuchi, and Y. Takai. 2003. Role of
each immunoglobulin-like loop of nectin for its cell-cell adhesion activity.
Biochem. Biophys. Res. Commun. 302: 61–66.
35. Fogel, A. I., M. Stagi, K. Perez de Arce, and T. Biederer. 2011. Lateral assembly
of the immunoglobulin protein SynCAM 1 controls its adhesive function and
instructs synapse formation. EMBO J. 30: 4728–4738.
36. Dong, X., F. Xu, Y. Gong, J. Gao, P. Lin, T. Chen, Y. Peng, B. Qiang, J. Yuan,
X. Peng, and Z. Rao. 2006. Crystal structure of the V domain of human Nectinlike molecule-1/Syncam3/Tsll1/Igsf4b, a neural tissue-specific immunoglobulinlike cell-cell adhesion molecule. J. Biol. Chem. 281: 10610–10617.
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
binding to nectin-1, also occupies the dimer interface, disrupting
dimer formation or cis-dimerization of nectin-2 is necessary for its
interaction with the DNAM-1 receptor. Our current work cannot
rule out either of these two models, and future work should focus
on this. However, Tahara-Hanaoka et al. (23) found that DNAM-1
binding to nectin-2–transfected cells was augmented substantially
when the cells are pretreated with the anti–nectin-2 mAb (R2.477.1),
which recognizes an epitope at one of the C set Ig-like domains of
nectin-2 and blocks homophilic binding (64, 65). This suggested
that DNAM-1 binding to nectin-2 on the cell surface may compete
for the nectin-2 homophilic binding site, which supports the first
mode that we proposed (Fig. 7C).
Elucidation of the molecular mechanism of the homodimer
formation of nectin-2 and its interaction with immune costimulator
receptor DNAM-1 will help us to understand cell–cell adhesions
and several other immune processes, such as the cytotoxicity of
NK and CTL cells toward tumor cells (18, 19), the interaction
between NK cells and DCs, the elimination of stimulated T cells
by NK cells (66), monocyte extravasation through the endothelium (21), and the interaction of eosinophils with mast cells (22).
Our findings increase the general understanding of DNAM-1 and
its ligands in immune recognition and activity, as well as pave the
way for immune diagnosis and therapy of tumors or autoimmune
diseases through DNAM-1 and nectin-2/Necl5–dependent intervention.
5519
5520
STRUCTURE OF CD112 HOMODIMER AND ITS BINDING TO CD226
54. Narita, H., Y. Yamamoto, M. Suzuki, N. Miyazaki, A. Yoshida, K. Kawai,
K. Iwasaki, A. Nakagawa, Y. Takai, and T. Sakisaka. 2011. Crystal Structure of
the cis-Dimer of Nectin-1: implications for the architecture of cell-cell junctions.
J. Biol. Chem. 286: 12659–12669.
55. Zhang, P., S. Mueller, M. C. Morais, C. M. Bator, V. D. Bowman, S. Hafenstein,
E. Wimmer, and M. G. Rossmann. 2008. Crystal structure of CD155 and electron
microscopic studies of its complexes with polioviruses. Proc. Natl. Acad. Sci.
USA 105: 18284–18289.
56. Fuchs, A., M. Cella, E. Giurisato, A. S. Shaw, and M. Colonna. 2004. Cutting
edge: CD96 (tactile) promotes NK cell-target cell adhesion by interacting with
the poliovirus receptor (CD155). J. Immunol. 172: 3994–3998.
57. Yu, X., K. Harden, L. C. Gonzalez, M. Francesco, E. Chiang, B. Irving, I. Tom,
S. Ivelja, C. J. Refino, H. Clark, et al. 2009. The surface protein TIGIT suppresses T cell activation by promoting the generation of mature immunoregulatory dendritic cells. Nat. Immunol. 10: 48–57.
58. Warner, M. S., R. J. Geraghty, W. M. Martinez, R. I. Montgomery,
J. C. Whitbeck, R. Xu, R. J. Eisenberg, G. H. Cohen, and P. G. Spear. 1998. A
cell surface protein with herpesvirus entry activity (HveB) confers susceptibility
to infection by mutants of herpes simplex virus type 1, herpes simplex virus type
2, and pseudorabies virus. Virology 246: 179–189.
59. Mühlebach, M. D., M. Mateo, P. L. Sinn, S. Prüfer, K. M. Uhlig, V. H. Leonard,
C. K. Navaratnarajah, M. Frenzke, X. X. Wong, B. Sawatsky, et al. 2011.
Adherens junction protein nectin-4 is the epithelial receptor for measles virus.
Nature 480: 530–533.
60. Noyce, R. S., D. G. Bondre, M. N. Ha, L. T. Lin, G. Sisson, M. S. Tsao, and
C. D. Richardson. 2011. Tumor cell marker PVRL4 (nectin 4) is an epithelial
cell receptor for measles virus. PLoS Pathog. 7: e1002240.
61. Zhang, N., J. Yan, G. Lu, Z. Guo, Z. Fan, J. Wang, Y. Shi, J. Qi, and G. F. Gao.
2011. Binding of herpes simplex virus glycoprotein D to nectin-1 exploits host
cell adhesion. Nat. Commun. 2: 577.
62. Di Giovine, P., E. C. Settembre, A. K. Bhargava, M. A. Luftig, H. Lou,
G. H. Cohen, R. J. Eisenberg, C. Krummenacher, and A. Carfi. 2011. Structure of
herpes simplex virus glycoprotein D bound to the human receptor nectin-1. PLoS
Pathog. 7: e1002277.
63. Blom, N., T. Sicheritz-Pontén, R. Gupta, S. Gammeltoft, and S. Brunak. 2004.
Prediction of post-translational glycosylation and phosphorylation of proteins
from the amino acid sequence. Proteomics 4: 1633–1649.
64. Lopez, M., F. Cocchi, L. Menotti, E. Avitabile, P. Dubreuil, and G. CampadelliFiume. 2000. Nectin2alpha (PRR2alpha or HveB) and nectin2delta are lowefficiency mediators for entry of herpes simplex virus mutants carrying the
Leu25Pro substitution in glycoprotein D. J. Virol. 74: 1267–1274.
65. Lopez, M., M. Aoubala, F. Jordier, D. Isnardon, S. Gomez, and P. Dubreuil.
1998. The human poliovirus receptor related 2 protein is a new hematopoietic/
endothelial homophilic adhesion molecule. Blood 92: 4602–4611.
66. Ardolino, M., A. Zingoni, C. Cerboni, F. Cecere, A. Soriani, M. L. Iannitto, and
A. Santoni. 2011. DNAM-1 ligand expression on Ag-stimulated T lymphocytes
is mediated by ROS-dependent activation of DNA-damage response: relevance
for NK-T cell interaction. Blood 117: 4778–4786.
67. Thompson, J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin, and D. G. Higgins.
1997. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 25: 4876–4882.
68. Gouet, P., X. Robert, and E. Courcelle. 2003. ESPript/ENDscript: Extracting and
rendering sequence and 3D information from atomic structures of proteins.
Nucleic Acids Res. 31: 3320–3323.
Downloaded from http://www.jimmunol.org/ by guest on June 14, 2017
37. Read, R. J. 2001. Pushing the boundaries of molecular replacement with maximum likelihood. Acta Crystallogr. D Biol. Crystallogr. 57: 1373–1382.
38. Fogel, A. I., Y. Li, J. Giza, Q. Wang, T. T. Lam, Y. Modis, and T. Biederer. 2010.
N-glycosylation at the SynCAM (synaptic cell adhesion molecule) immunoglobulin interface modulates synaptic adhesion. J. Biol. Chem. 285: 34864–
34874.
39. Chen, Y., P. Liu, F. Gao, H. Cheng, J. Qi, and G. F. Gao. 2010. A dimeric
structure of PD-L1: functional units or evolutionary relics? Protein Cell. 1: 153–
160.
40. Li, X., J. Liu, J. Qi, F. Gao, Q. Li, X. Li, N. Zhang, C. Xia, and G. F. Gao. 2011.
Two distinct conformations of a rinderpest virus epitope presented by bovine
major histocompatibility complex class I N*01801: a host strategy to present
featured peptides. J. Virol. 85: 6038–6048.
41. Liu, J., L. Dai, J. Qi, F. Gao, Y. Feng, W. Liu, J. Yan, and G. F. Gao. 2011.
Diverse peptide presentation of rhesus macaque major histocompatibility complex class I Mamu-A 02 revealed by two peptide complex structures and insights
into immune escape of simian immunodeficiency virus. J. Virol. 85: 7372–7383.
42. Qian, X., J. Qi, F. Chu, J. Liu, Q. Li, and J. Yan. 2009. Crystallization and
preliminary X-ray analysis of the V domain of human nectin-2. Acta Crystallogr.
Sect. F Struct. Biol. Cryst. Commun. 65: 615–617.
43. Otwinowski, Z., and W. Minor. 1997. Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276: 307–326.
44. Schneider, T. R., and G. M. Sheldrick. 2002. Substructure solution with
SHELXD. Acta Crystallogr. D Biol. Crystallogr. 58: 1772–1779.
45. McCoy, A. J., R. W. Grosse-Kunstleve, P. D. Adams, M. D. Winn, L. C. Storoni,
and R. J. Read. 2007. Phaser crystallographic software. J. Appl. Cryst. 40: 658–
674.
46. Cowtan, K. 1994. Joint CCP4 and ESF-EACBM Newsletter on Protein Crystallography. 31: 34–38.
47. Perrakis, A., M. Harkiolaki, K. S. Wilson, and V. S. Lamzin. 2001. ARP/wARP
and molecular replacement. Acta Crystallogr. D Biol. Crystallogr. 57: 1445–
1450.
48. Emsley, P., and K. Cowtan. 2004. Coot: model-building tools for molecular
graphics. Acta Crystallogr. D Biol. Crystallogr. 60: 2126–2132.
49. Murshudov, G. N., A. A. Vagin, and E. J. Dodson. 1997. Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr. D
Biol. Crystallogr. 53: 240–255.
50. Adams, P. D., P. V. Afonine, G. Bunkóczi, V. B. Chen, I. W. Davis, N. Echols,
J. J. Headd, L. W. Hung, G. J. Kapral, R. W. Grosse-Kunstleve, et al. 2010.
PHENIX: a comprehensive Python-based system for macromolecular structure
solution. Acta Crystallogr. D Biol. Crystallogr. 66: 213–221.
51. Laskowski, R. A., M. W. Macarthur, D. S. Moss, and J. M. Thornton. 1993.
PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl. Cryst. 26: 283–291.
52. Liu, J., Y. Sun, J. Qi, F. Chu, H. Wu, F. Gao, T. Li, J. Yan, and G. F. Gao. 2010.
The membrane protein of severe acute respiratory syndrome coronavirus acts as
a dominant immunogen revealed by a clustering region of novel functionally and
structurally defined cytotoxic T-lymphocyte epitopes. J. Infect. Dis. 202: 1171–
1180.
53. Liu, J., P. Wu, F. Gao, J. Qi, A. Kawana-Tachikawa, J. Xie, C. J. Vavricka,
A. Iwamoto, T. Li, and G. F. Gao. 2010. Novel immunodominant peptide presentation strategy: a featured HLA-A*2402-restricted cytotoxic T-lymphocyte
epitope stabilized by intrachain hydrogen bonds from severe acute respiratory
syndrome coronavirus nucleocapsid protein. J. Virol. 84: 11849–11857.