Advanced glycation end product recognition by the receptor for AGEs

Article abstract of DOI:10.1016/j.str.2011.02.013

Nonenzymatic protein glycation results in the formation of advanced glycation end products (AGEs) that are implicated in the pathology of diabetes, chronic inflammation, Alzheimer's disease, and cancer. AGEs mediate their effects primarily through a receptor-dependent pathway in which AGEs bind to a specific cell surface associated receptor, the Receptor for AGEs (RAGE). N _ε-carboxy-methyl-lysine (CML) and N_ε-carboxy- ethyl-lysine (CEL), constitute two of the major AGE structures found in tissue and blood plasma, and are physiological ligands of RAGE. The solution structure of a CEL-containing peptide-RAGE V domain complex reveals that the carboxyethyl moiety fits inside a positively charged cavity of the V domain. Peptide backbone atoms make specific contacts with the V domain. The geometry of the bound CEL peptide is compatible with many CML (CEL)-modified sites found in plasma proteins. The structure explains how such patterned ligands as CML (CEL)-proteins bind to RAGE and contribute to RAGE signaling.

Full text of DOI:10.1016/j.str.2011.02.013

Structure
Article
Advanced Glycation End Product Recognition
by the Receptor for AGEs
Jing Xue,1,5 Vivek Rai,2,5 David Singer,3 Stefan Chabierski,3 Jingjing Xie,4 Sergey Reverdatto,1 David S. Burz,1
Ann Marie Schmidt,2 Ralf Hoffmann,3 and Alexander Shekhtman1,
*
1Department of Chemistry, State University of New York at Albany, Albany, NY 12222, USA
2Division of Endocrinology, Department of Medicine, New York University Langone Medical Center, 550 1st Avenue, NY, NY 10016, USA
3Institute of Bioanalytical Chemistry and Center for Biotechnology and Biomedicine, Universita¨ t Leipzig, Leipzig 04103, Germany
4College of Biotechnology and Pharmaceutical Engineering, Nanjing, University of Technology, Nanjing 210009, PR China
5Theses authors contributed equally to this work
*Correspondence: ashekhta@albany.edu
DOI 10.1016/j.str.2011.02.013
SUMMARY
consists of three extracellular immunoglobulin domains, V, C1,
and C2, a transmembrane helix and a short cytosolic tail (Hori
Nonenzymatic protein glycation results in the forma- et al., 1995; Neeper et al., 1992). In spite of their structural diver-
sity, AGEs bind only to the V domain of RAGE (Kislinger et al.,
1999; Xie et al., 2008). This binding does not accelerate clear-
tion of advanced glycation end products (AGEs) that
are implicated in the pathology of diabetes, chronic
ance or degradation but rather begins a sustained period of
cellular activation mediated by receptor-dependent signaling,
leading to inflammation. It is proposed that RAGE activation is
inflammation, Alzheimer’s disease, and cancer.
AGEs mediate their effects primarily through
a receptor-dependent pathway in which AGEs bind
to a specific cell surface associated receptor, the
Receptor for AGEs (RAGE). N3-carboxy-methyl-
lysine (CML) and N3-carboxy-ethyl-lysine (CEL),
largely responsible for the pathogenicity associated with AGEs
(Hofmann et al., 1999; Schmidt et al., 1995). Soluble RAGE
(sRAGE), a variant containing extracellular V, C1, and C2
domains serves as a scavenger and is a potent inhibitor of
constitute two of the major AGE structures found in
tissue and blood plasma, and are physiological
RAGE signaling (Hori et al., 1995).
Other endogenous ligands are implicated in amplifying RAGE-
ligands of RAGE. The solution structure of a CEL- dependent proinflammatory signaling: cytokine-like mediators
of the S100 family (Hofmann et al., 1999) and amphoterin (Tagu-
chi et al., 2000), a nuclear protein released by necrotic cells.
Unlike AGEs and possibly amphoterin that bind to a single
containing peptide-RAGE V domain complex reveals
that the carboxyethyl moiety fits inside a positively
charged cavity of the V domain. Peptide backbone
atoms make specific contacts with the V domain.
The geometry of the bound CEL peptide is compat-
ible with many CML (CEL)-modified sites found in
plasma proteins. The structure explains how such
patterned ligands as CML (CEL)-proteins bind to
domain, all three extracellular domains of RAGE are involved in
binding S100 proteins: S100A12 binds to the C1 domain (Xie
et al., 2007), S100B binds to V and C1 (Dattilo et al., 2007; Osten-
dorp et al., 2007), and S100A6 binds to V and C2 (Leclerc et al.,
2007). The differences between RAGE binding sites may
account for the diverse cellular responses caused by S100
RAGE and contribute to RAGE signaling.
proteins. Another major class of AGEs, imidazolones have also
been implicated as possible ligands of RAGE (Thornalley,
1998). The structural diversity of RAGE ligands and the fact
that RAGE recognizes a class of ligands, such as AGEs, led to
the hypothesis that RAGE is a pattern recognition receptor (Cha-
INTRODUCTION
The products of nonenzymatic glycation and oxidation of vakis et al., 2003).
proteins, the early and the advanced glycation end products
Despite the fact that AGE-RAGE biology has been studied for
(AGEs), are a heterogeneous class of compounds that form more than 20 years, very little is known about the structural
under diverse circumstances in response to cellular stress biology of AGE-RAGE complexes. This is mostly due to the
(Brownlee et al., 1984). Specific AGE compounds, namely extensive heterogeneity of AGEs created by glycation reactions:
N3-carboxy-methyl-lysine (CML) and N3-carboxy-ethyl-lysine glycation reactions are not largely dependent on sequence
(CEL) (Figure 1A), due to their interactions with the Receptor specificity, and lysine and arginine residues, which are particu-
for AGEs (RAGE), have been linked to complications of diabetes larly susceptible to glycation, are very common in proteins. The
and chronic inflammation, the severity of Alzheimer’s disease, binding of AGE-modified proteins to the V domain of RAGE
and cancer (Ishiguro et al., 2005; Ramasamy et al., 2005a, depends weakly on either the primary or tertiary structure of
2005b; Thornalley, 1999; Valente et al., 2010). RAGE is located the AGE modified sites. And while individual CML (CEL) struc-
in the major histocompatibility complex class III (MHC III) region tures bind to the V domain of RAGE weakly, constitutive oligo-
suggesting its involvement in immune responses (Schmidt and merization of RAGE provides a mechanism for increasing the
Stern, 2001; Sugaya et al., 1994). RAGE is a member of the number of binding sites and subsequently, the binding affinity
immunoglobulin superfamily of cell surface molecules and of AGEs (Thornalley, 1998; Xie et al., 2008).
722 Structure 19, 722–732, May 11, 2011 ª2011 Elsevier Ltd All rights reserved

Structure
AGE Recognition by RAGE
Figure 1. CEL-Containing Peptides Can Be
Produced by Both Glycation and Chemical
Synthesis
(A) Fructosyllysine, CML, and CEL are the prod-
ucts of early and advanced glycoxidation of
sugars.
(B) Synthetic CEL-peptide, DEF(CEL)ADE,
contains
a
correct chemical structure of
N3-carboxy-ethyl-lysine. To characterize CEL-
peptide, we collected two 2D homonuclear NMR
experiments, 1H,1H TOCSY and 1H,1H ROESY.
The 1H,1H TOCSY strip shows through bond
correlation between the amide proton (8 ppm) and
side-chain protons Ha (4.1 ppm), Hb (1.7 ppm), Hg
(1.2 ppm), Hd (1.5 ppm), and H3 (2.85 ppm) of CEL.
The 1H,1H ROESY strip shows through space
correlations between H3 and carboxyethyl protons
CH(CH3)-COOH (3.8 ppm) and CH(CH3)-COOH
(1.7 ppm) of CEL, confirming the presence of
a proper chemical structure of N3-carboxy-ethyl-
lysine.
See also Figure S1.
homologous surface in the variable
domain of an antibody is used for hetero-
dimerization. We showed that both the
CEL moiety and the peptide backbone
make specific contacts with the V domain
of RAGE. The structure suggests how
RAGE can bind a large variety of CEL
(CML)-modified
proteins.
Specific
binding of CML (CEL)-modified proteins
to RAGE suggests that the receptor func-
tions as a sensor of cellular stress.
RESULTS
A variety of extracellular proteins contain-
ing CML or CEL and possessing unre-
lated primary sequences bind to RAGE.
Surprisingly, free CML and CEL do not
bind to RAGE (Xie et al., 2008). Only
CML or CEL embedded in a peptidic
structure are capable of binding (Xie
et al., 2008). Thus, to structurally charac-
terize AGE-RAGE interactions, we
synthesized short peptides containing
CML and CEL (Figure 1B; see Figure S1
Many AGE structures are presentin plasmaat verylow concen- available online). The amino acid sequences of the peptides
trations and thus, are unlikely to contribute significantly to RAGE represent the primary glycation sites found in human serum
signaling (Thornalley et al., 2003). At the same time, CML and CEL albumin, HSA, a major plasma protein.
The affinities of CML (CEL) peptides for the V domain were
are major AGEs found in tissue and blood plasma and have been
established as ligands of RAGE (Schmidt et al., 1996; Thornalley estimated by monitoring changes in the native tryptophan fluo-
et al., 2003). Structural information is critical to understand the rescence of the V domain upon ligand binding. Unmodified
mechanisms of RAGE signaling and to design and optimize effec- peptides, F(K)DLGEE and DEF(K)ADE, bind to the V domain
tive therapeutic agents against RAGE-caused pathologies.
with low affinities, whereas the binding of modified peptides is
Here, we solved a solution structure of a CEL-containing ꢀ6- to 7-fold greater (Table 1; Figure S2). As expected, the fluo-
peptide V domain complex. Structure of the V domain is closely rescence titration indicated that the binding depends only
homologous to the variable domain of an antibody. CEL peptide slightly on the primary sequence of the peptide and even less
binds to the surface formed by the C, F, and G strands. The so on the type of AGE modification. For example, the modified
Structure 19, 722–732, May 11, 2011 ª2011 Elsevier Ltd All rights reserved 723

Structure
AGE Recognition by RAGE
RAGE V domain is different from that of the immunoglobulin-
like domains found in cell adhesion receptor molecules, such
as ICAM and JAM, suggesting a different mode of action. The
closest structural analog is a variable domain of the IgG1
P20.1 FAB light chain (PDB code 2ZPK; Nogi et al., 2008) Figures
2C and 4A; Figure S4).
Table 1. Binding Affinities of the CML Peptides for the Wild-Type
and Mutants of V Domain
R98A
K52A
K52A, R98A
V Domain
Kd, mMa
Peptide
V Domain
Kd, mM1
V Domain
Kd, mM1
V Domain
Kd, mM1
Sequence
DEF(CML)ADE
DEF(CEL)ADE
DEFKADE
97
3
5
725 17
528 19
830 27
The V domain structure consists of seven strands connected
by six loops, to form two beta-sheets. The disulfide bond
between Cys38 and Cys99 links these sheets into a beta-sand-
wich structure. Two antiparallel beta strands, B and D, are
shorter than in the conventional, IgG1 P20.1 FAB V-type domain
(Bork et al., 1994) (Figure 2C). This arrangement allows loops BC
and C’D to be packed against the core of the protein rather than
protruding into the solvent. Loops BC and C’D contain the hyper-
variable CDR1 and CDR2 regions of the antibody that are
involved in antigen binding (Figure 2C; Figure S4). Two other
104
685 26
560 32
912 42
617 24
—
—
—
—
—
—
—
—
—
—
—
—
F(CML)DLGEE
F(CEL)DLGEE
FKDLGEE
87
93
5
6
673 38
a Dissociation constant was obtained by fitting fluorescence titration data
with a single site binding isotherm.
peptides F(CML)DLGEE and DEF(CML)ADE bind with 97 and regions distinct from conventional immunoglobulin structures
87 mM affinity, respectively, and F(CEL)DLGEE and DEF(CEL) are a helix between strands C’ and D and the apparent lack of
ADE bind with 104 and 93 mM affinity, respectively. Overall, the the hydrogen bonds necessary to form the C’’ strand conven-
binding affinities for all of the modified peptides examined tionally found in variable-type domains.
were very similar to one another.
The electrostatic potential mapped onto the molecular surface
Since CML and CEL differ only by a methyl group (Figure 1; of the V domain is shown in Figure 4B. There is only one obvious
Figure S1), we hypothesized that the CML and CEL peptides hydrophobic cavity close to the C terminus formed by Ile 30, Pro
bind to the same molecular surface of the V domain. NMR chem- 87, Ala 88, Ile 91, Tyr 118 around loop A’B and loop EF. The elec-
ical shifts of backbone amide protons and nitrogens are exqui- trostatic surface potential of the V domain reveals that the
sitely sensitive to the changes in chemical environment induced molecular surface is covered by positive charges. In particular,
by ligand binding. These changes allowed us to identify amino there are two areas where positive charges are densely localized
acid residues of the V domain that are affected by the binding to form a cationic center. The first one consists of Lys52, Arg98,
of either DEF(CML)ADE (CML-PEP) or DEF(CEL)ADE (CEL- and Lys110 and is positioned on a relatively flat molecular
PEP) (Figures 2B and 2C; Figure S3). Titrating the peptides into surface. IgG1 utilizes this area to form heterodimers between
the V domain led to gradual changes in the chemical shift of light and heavy chain variable domains (Figure 4A; Figure S4).
the residues Asn25, Lys52, Glu97, Arg98, Cys99, Ala101, and The second area of positive charge consists of Lys43, Lys44
Lys110. This set of residues was virtually the same for CML- on loop BC, and Arg104 on loop FG. Conventional antibodies
PEP and CEL-PEP suggesting that the V domain does not utilize the latter surface for specific ligand binding (Bork et al.,
discriminate between these species and the same interaction 1994; Nogi et al., 2008).
surface is involved in CML and CEL binding.
The binding conformation of the CEL-PEP peptide backbone
To elucidate the chemical nature of the interaction between is bent at CEL and forms a loop possibly due to steric hindrance
the V domain of RAGE and CEL (CML)-containing proteins, we caused by the bulky CEL group (Figures 3 and 4). The CEL-PEP
2
˚
studied the solution structure of the V domain (residues 23– binding site is small, spanning 180 A and sits on a positively
123 of the 356 residue RAGE) in complex with the 7 residue charged groove formed by the C, F, and G strands. Conventional
CEL-PEP. The structured regions comprise residues 23–92 of antibodies use this surface for heterodimerization (Figure 4A;
the V domain and 3–5 of CEL-PEP. Based on the constraints ob- Figure S4). The binding site for CEL-PEP is relatively flat, which
tained from NMR experiments, 25 structures with the lowest probably reflects the lack of specificity for primary structure
target function values were superimposed (Figure 3; Table S1). observed in binding experiments.
The solution structure of the V domain-CEL-PEP complex is
The CEL moiety of CEL-PEP makes the majority of intermolec-
similar to that of the free V domain (Matsumoto et al., 2008) ular contacts (Figure 4C; Table S2). CEL-PEP also contacts only
and V domain within a VC1 construct (Koch et al., 2010). The a limited number of residues in the V domain of RAGE (Figure 4C;
root mean square (rms) deviations between the solution struc- Table S2). Ionic interactions predominate CEL-PEP-V domain
ture of the complexed and free V domain (PDB code 2E5E) or interactions since an increase in ionic strength from 200 to
V domain within VC1 construct (PDB code 3CJJ) are 1.8 and 400 mM abolished CEL-PEP binding. We detected an extended
˚
˚
1.4 A for the backbone and 2.6 and 1.9 A for all heavy atoms in network of nOes confidently placing the negatively charged car-
of the ordered regions. The large loop between strands C’ and boxyethyl head group of CEL into the groove formed by posi-
D was poorly defined in the free V domain (Matsumoto et al., tively charged Lys52, Lys110, and Arg98 (Figure 4C; Table S2
2008) and excluded from structural comparison.
and Figure S4C). The rest of CEL fits snugly into a groove formed
The solution structure of the V domain (Figure 4A) closely by the F and G strands, which may provide additional hydro-
resembles the structure of conventional immunoglobulin phobic contacts. To further demonstrate that Lys52 and Arg98
V-type domains. Secondary structure elements were labeled play a major part in the binding of CML (CEL)-containing
following the immunoglobulin convention (Bork et al., 1994; Cho- peptides to the V domain, we made single R98A and K52A,
thia and Jones, 1997). The network of hydrogen bonds in the and double R98A, K52A mutants of the V domain. The binding
724 Structure 19, 722–732, May 11, 2011 ª2011 Elsevier Ltd All rights reserved

Structure
AGE Recognition by RAGE
Figure 2. RAGE V Domain Does Not Discriminate between CML- and CEL-Containing Peptides
(A) Overlay of 15N-HSQC of free (black) and CML-PEP bound (red) V domain.
(B) Overlay of 15N-HSQC of free (black) and CEL-PEP bound (red) V domain.
(C) Sequence alignment of V domain and a V-type domain from a heavy chain antibody IgG1 P20.1. Amino acid residues involved in CML (CEL) binding are in red.
CDR1 and CDR2, the hypervariable regions of IgG1 P20.1, are in blue. Secondary structure elements are shown above the sequences.
See also Figure S3.
affinities of CML-PEP and CEL-PEP for the mutant V domains for CEL will lead to loss of binding. Importantly, the methyl group
are 5- to 10-fold weaker than for the wild-type V domain, and of CEL is oriented away from the binding site suggesting that
comparable to that of unmodified PEP (Table 1; Figure S2).
this group does not contribute to the binding free energy. This
The CEL moiety binds to the V domain in an extended confor- configuration of CEL would explain why CML, which lacks the
mation that is longer than any of the amino acids side chains. As methyl group, has a very similar binding affinity and fits into
a result, substituting a negatively charged side chain such as Glu a very similar interaction surface. Consistent with our previous
Structure 19, 722–732, May 11, 2011 ª2011 Elsevier Ltd All rights reserved 725

Structure
AGE Recognition by RAGE
Figure 3. Stereo View of the Overlay of 25
Lowest Energy CEL-PEP V Domain Back-
bone Traces (PDB code 2L7U)
N and C termini of the V domain and CEL-PEP are
indicated. Figure is prepared by using Molmol
(Koradi et al., 1996). See also Table S1.
extremely flexible (Povey et al., 2008).
This suggests that even CML (CEL)-modi-
fied sites, which do not initially possess
proper binding geometry, may be able to
bind to the V domain by using induced fit.
Soluble RAGE (sRAGE) works as
a scavenger to remove ligands capable
binding experiments, an early glycation end product, fructosylly- of activating RAGE expressing cells (Hofmann et al., 1999). We
sine (Figure 1A), which possesses a bulky head group, will not fit used this function of sRAGE to test whether our structure of
into the CEL binding site on the V domain (Xie et al., 2008).
the CEL-PEP-V domain reflects physiological interactions
Our earlier observations suggest that interactions between the between RAGE and CML modified proteins. Instead of the
CEL moiety and V domain are not sufficient to provide stable V domain that binds to CML peptides only weakly, we used
binding since neither free CEL nor CML stably bind to the V the VC1 domain of RAGE that dimerizes (Xie et al., 2007; Koch
domain (Xie et al., 2008). We hypothesized that the peptide back- et al., 2010; Zong et al., 2010) (Figure 5A; Figure S5A) and
bone conformation is important for CEL-PEP binding to the V thus, is capable of binding multiple CMLs with increased binding
domain. Indeed the backbone amide proton of CEL-PEP Ala5 affinity. CML-modified BSA was used as a RAGE activating
˚
is located within 2.5 A of the V domain Asn112 side-chain ligand. To interrogate the functional implications of these find-
carbonyl group and possesses geometry consistent with ings, we used two RAGE-expressing cell types, C6 glioma cells
a hydrogen bond (Figure 4C). The chemical shift of the Ala5 and primary murine aortic vascular smooth muscle cells (VSMC)
amide proton exhibits a downfield shift from 7.6 to 8.2 ppm (Brett et al., 1993; Taguchi et al., 2000). Cellular activation was
upon CEL-PEP binding to the V domain, which is also indicative monitored by observing the increase in phosphorylation levels
of hydrogen bond formation. We detected intermolecular of two key signal transduction proteins previously shown to be
contacts between Hb of the Ala5 side chain and Hd of Arg114, downstream of RAGE: ERK and P38 (MAPK)(Kislinger et al.,
which may strengthen the CEL-PEP peptide backbone-V 1999).
domain interaction (Table S2 and Figure S4C). Additional hydro-
Wild-type VC1 inhibited CML-BSA induced RAGE signaling in
phobic contacts are supplied by V domain residues Trp61, Ile96, both C6 glioma and VSMC cells, as evidenced by no increase in
and Phe97, which together extend the hydrophobic surface the levels of phosphorylated ERK (pERK) and phosphorylated
along face of Phe3, since we detected an intermolecular nOe MAPK (pP38) relative to the control lane (Figures 5B and 5C;
between Hb of Phe3 and Hd of Ile96 (Figure S4C). At the same Figures S5B and S5C). Total ERK and P38 MAPK levels re-
time, the interaction between the benzene ring of Phe3 and the mained constant for all experiments. The two single mutants,
V domain is not specific since we did not identify an extensive R98A and K52A, and one double mutant of VC1, R98A, K52A
network of nOes from the V domain to Phe3 benzene ring (Fig- (KRA), failed to inhibit RAGE signaling induced by CML-BSA,
ure 4C; Table S2). The side chain of Phe3 is not well defined in strongly suggesting that the identified CEL (CML) binding site
the V domain-CEL-PEP complex (Figure S4D). In addition, during is critical for the CML-BSA induced activation of RAGE.
the CEL-PEP-V domain titration residues Trp61, Ile96, and
Phe97 did not undergo the large chemical shift changes ex- DISCUSSION
pected for ring current shifts (Perkins and Wuthrich, 1979). This
hydrophobic surface is large enough to accommodate any The structure of the V domain with its physiological ligand, CEL-
side chain, possibly providing an increase in binding affinity PEP, reveals how RAGE is able to recognize CML (CEL)-modi-
without a subsequent increase in specificity.
fied proteins when the modification site can possess very
To rationalize the ability of RAGE to bind various CML- and different primary, secondary, and tertiary structures. The
CEL-modified proteins, we structurally aligned CEL-PEP with V domain of RAGE makes molecular contacts with both the
three-dimensional structures of loops from BSA reported to CEL (CML) moiety and the peptide backbone in the immediate
contain major glycation sites (Wa et al., 2007; Figure 4D). Three vicinity of CEL (CML). The V domain does not discriminate
out of six loops possessed conformations that are compatible between CEL and CML. The negative charge of CEL (CML) is
with binding to the V domain. Importantly, the conformation of a critical determinant for binding to the positively charged
the backbone amide group that follows the AGE-modified lysines V domain surface. The extended geometry of CEL (CML) and
in these loops are almost identical to that of CEL-PEP. Moreover, the distance from the negative carboxyl group to the peptide
it is known that glycation disrupts local secondary structure backbone is also critical for molecular recognition. These two
rendering the area immediately around the modified site conditions will preclude binding of either negatively charged
726 Structure 19, 722–732, May 11, 2011 ª2011 Elsevier Ltd All rights reserved

Structure
AGE Recognition by RAGE
Figure 4. Solution Structure of the CEL-PEP-V Domain Complex
(A) Structure of CEL-PEP bound to V domain. V domain is shown in ribbon representation. Elements of secondary structure are labeled following the immu-
noglobulin convention (Bork et al., 1994).
(B) Electrostatic potential is mapped onto the molecular surface of the V domain. Positively and negatively charged surfaces are indicated in blue and red,
respectively.
˚
(C) V domain amino acid residues located within 5 A from the CEL moiety of CEL-PEP. Putative hydrogen bond between the backbone amide proton of Ala5 and
the side-chain carbonyl group of Asn112 is indicated. Carbon atoms of CEL-PEP and V domain are in yellow and cyan, respectively.
(D) Structural alignment of CEL-PEP with three short segments from human serum albumin (HSA)-containing lysines (Wa et al., 2007), 11-FKD, 56-AKT, and 261-
AKY. The lysines in these sequences were shown to be glycated under elevated concentrations of D-glucose.
See also Figure S4 and Table S2.
amino acids or bulky early glycation products, such as fructosyl- most only a single CML (CEL)-modified lysine. Since binding
lysine, to the CEL-PEP binding site. The torsional angles of the of a single CML peptide to the V domain is weak this event
CEL-PEP peptide backbone when bound to the V domain, allow will not trigger RAGE signaling (Xie et al., 2008). Under cellular
different sequences within the AGE-modified protein to fit into stress, caused by different pathologies including diabetes, Alz-
the binding site. The latter explains why RAGE can bind to heimer’s disease, or cancer, the concentration of CML (CEL)
various CML (CEL)-modified proteins found in tissue and in proteins increases, thus increasing the amount of multiple-
plasma.
modified proteins (Brownlee, 1995). Since RAGE is a constitu-
Binding of CEL-PEP to the V domain is clearly specific. Is the tive oligomer (Xie et al., 2007; Koch et al., 2010; Zong et al.,
binding of CML (CEL) proteins to RAGE physiologically impor- 2010), polyvalent engagement of CML proteins and RAGE
tant and how does it relate to the immune response caused by results in tight binding, thus, triggering the RAGE-dependent
RAGE signaling? CML and CEL-modified proteins are found immune response. This mechanism suggests a physiological
under normal physiology (Thornalley et al., 2003). However, role for RAGE as a sensor of cellular stress caused by various
their concentration is low and individual proteins have at pathologies (Schmidt et al., 2000). It also highlights an
Structure 19, 722–732, May 11, 2011 ª2011 Elsevier Ltd All rights reserved 727

Structure
AGE Recognition by RAGE
Figure 5. Mutants of the VC1 Domains of
RAGE Fail to Suppress CML-BSA-Induced
RAGE Signaling
(A) Cartoon model of how RAGE dimerization
promotes
V domain binding to multiple CML
moieties on CML-BSA (ribbon). The molecular
surface of the V domain involved in CML(CEL)
binding is in red. Only two V domains are shown.
Lysines of BSA, which may undergo glycation are
in yellow. The cartoon model was prepared by
using SWISS-PDB Viewer (Guex and Peitsch,
1997).
(B and C) Single K52A and R98A, and double
K52A, R98A (KRA) mutants of the VC1 domains do
not interfere with CML-BSA induced RAGE
signaling in both C6 rat glioma (B) and mouse
VSMC cells (C).
See also Figure S5.
chains of Asp and Glu were protected with tBu
and the side chain of Lys with 4-methyltrityl
(Mtt). Fmoc-amino acid derivatives were acti-
vated with N,N’-diisopropylcarbodiimide (DIC)
and N-hydroxy-benzotriazole (HOBt) at equimolar
ratios, added to the resin with eight molar
excess, and reacted at room temperature for
70 min. After completion of the synthesis, the
Fmoc group was cleaved with piperidine and
the free N terminus was protected again with
a
mixture of di-tert-butyl dicarbonate (Boc2O,
20 eq) and N,N-diisopropylethylamine (DIPEA,
10 eq) in N,N-dimethylformamide (DMF) at room
temperature overnight. The Mtt group was
cleaved with 2% (v/v) trifluoroacetic acid (TFA)
and 1% (v/v) triisopropylsilane (TIS) in dichloro-
methane (DCM) two times for 30 min. The resin
was washed three times with DCM and dried
before the unprotected 3-amino groups of the
lysine residues were alkylated with 2-chlorotrityl
3-bromopropanoate (or 2-chlorotrityl bromoace-
tate) (20 eq) in the presence of DIPEA (10 eq) in
DCM to obtain the protected CEL (or CML)
peptides, respectively. Peptides were cleaved
with TFA containing 12.5% (v:v) of a scavenger
mixture (ethanedithiole, m-cresole, thioanisole
and water, 1:2:2:2) at room temperature for
2 hr. The peptides were precipitated with cold di-
ethyl ether and purified by RP-HPLC by using
a linear aqueous acetonitrile gradient (5 mmol
ammonium formate [pH 3.2]) as ion pair reagent
on a Jupiter C18-column (21.2 mm internal diameter, 250 mm length,
opportunity to interfere with RAGE signaling by obstructing low
affinity ligand binding. The structure of the CML-PEP-V domain
will rationalize these efforts.
15 mm particle size, 30 nm pore size) (Phenomenex Inc., Torrance, CA).
The purities of the peptides were confirmed by RP-HPLC and the molecular
weight by matrix-assisted laser desorption/ionization time-of-flight mass
spectrometry (MALDI-TOF-MS; 4700 proteomic analyzer; Applied Biosys-
tems GmbH, Darmstadt, Germany) using an a-cyano-4-hydroxycinnamic
acid matrix.
EXPERIMENTAL PROCEDURES
Reagents and Chemicals
Restriction enzymes and Taq polymerase were from NEB. All other chemicals
used were reagent grade or better.
Preparation of 2-Chloro-Trityl 3-Bromopropanoate
and 2-Chloro-Trityl 3-Bromoacetate
3-Bromopropionic acid (9.6 mmol, 0.86 ml) or bromoacetic acid (9.6 mmol,
3.0 g) were mixed with equimolar amounts of 2-chlorotrityl chloride (1.34 g)
and DIPEA (1.65 ml) in dichloromethane (25 ml). This mixture was stirred at
room temperature for 1 hr before the solvent was removed on a rotary evapo-
rator. The dry products were used without further purification to alkylate the
side chains of lysine residues.
Solid-Phase CML Peptide Synthesis
CML- and CEL-containing peptides were synthesized on a multiple synthe-
sizer SYRO2000 (MultiSynTech GmbH, Witten, Germany) by using 9-fluore-
nylmethoxycarbonyl/tert-butyl (Fmoc/tBu)-chemistry on Wang resin (Atherton
et al., 1978) utilizing a global postsynthetic alkylation approach. The side
728 Structure 19, 722–732, May 11, 2011 ª2011 Elsevier Ltd All rights reserved

Structure
AGE Recognition by RAGE
Plasmid Construction
The extent of chemical modification was determined colorimetrically by using
2,4,6-trinitrobenzenesulfonic acid to measure the difference spectrum
between lysine residues of modified and unmodified protein preparations
(Habeeb, 1966). The extent of lysine modification of CML-BSA preparations
was 21%.
Human RAGE cDNA library clone BC020669 was obtained from Open Biosys-
tems and used as a template for PCR amplifications. DNA coding for the
V domain (amino acids 24–125) was PCR-amplified using Taq polymerase
and oligonucleotides 50-TTTCATATGGCTCAAAACATCACAGCCCGGATTGG
and 30-TTTGTCGACTCATTCTGGCTTCCCAGGAATCTG containing flanking
50-NdeI and 30-SalI restriction sites. The restriction-digested PCR products
were ligated into expression vector pET28a (Novagen), which confers kana-
mycin resistance. The resulting plasmid, pET28-V, expresses a C-terminal
His-tagged V domain of RAGE. DNA coding for VC1 fragments (amino acids
23–243) were PCR amplified using oligonucleotides 50-TTTCATATGGCT
CAAAACATCACAGCCCGGATTGG and 30-TTTGAGCTCCACCACCAATTG
GACCTCCTCCAG containing 50-NdeI and 30-XhoI restriction sites. DNA frag-
ments were subcloned into the NdeI and XhoI sites of pET15b vector (Nova-
gen), which confers ampicillin resistance. The resulting plasmid, pET15b-
VC1, expresses an N-terminal His-tagged VC1 domain.
Site-Directed Mutagenesis of the V and VC1 Domains
To singly or doubly mutate the V domain of RAGE, the QuikChange II XL Site-
Directed Mutagenesis Kit (Strategene) was used. Following mutagenic PCR,
pET28-V was restriction digested with DpnI for 1 hr and transformed into
E. coli strain DH10B. Mutated plasmids were isolated and purified using
Mini-Prep Kit (QIAGEN). DNA sequencing identified plasmids pET28-R98A-
V, pET28-K52A-V, or pET28-K52A-R98A-V, which code for the appropriate
mutant V domain.
To singly or doubly mutate the VC1 domain of RAGE, High-Fidelity DNA
polymerase (New England BioLabs) was used. Following mutagenic PCR,
pET15-VC1 was transformed into E. coli strain XL-1 Blue. Mutated plasmids
were isolated and purified using a Mini-Prep Kit (QIAGEN). DNA sequencing
identified plasmids pET15-R98A-VC1, pET15-K52A-VC1, and pET15-K52A-
R98A-VC1, which code for the appropriate mutant VC1 domain.
Labeling, Expression, and Purification of Wild-Type and Mutant
V Domains
To uniformly label the V domain of RAGE, pET28-V, pET28-R98A-V, pET28-
K52A-V, or pET28-K52A-R98A-V were transformed into Escherichia coli strain
BL21(DE3) Codon+ (Novagen). For U-15N labeling, cells were grown at 37ꢁC in
minimal medium (M9) containing 35 mg/liter kanamycin and 1 g/liter [15N]
ammonium chloride as the sole nitrogen source. For U-13C,15N labeling, cells
were grown at 37ꢁC in M9 medium containing 35 mg/liter kanamycin, 1 g/liter
NMR Experiments
CEL- and CML-containing peptides were assigned using 2D 1H,1H TOCSY
and 1H,1H ROESY experiments (Cavanagh et al., 1996), which provide through
bond and through space proton connectivities. Protein samples of the
uniformly labeled, [U-13C,15N] and [U-15N], V domain, with concentrations
ranging from 60 to 300 mM were dissolved in NMR buffer (10 mM potassium
phosphate [pH 6.5], 100 mM NaCl, 0.02% (w/v) NaN3, in 90%H2O/10%D2O)
and unlabeled peptide, CEL-PEP or CML-PEP, was added to a 1.2 molar
excess. To obtain backbone resonance assignments of the [U-13C,15N]
[
15N]ammonium chloride, and 2 g/liter [13C]glucose instead of unlabeled
glucose as the sole carbon source. Cells were grown to 0.7 OD600 at 37ꢁC,
induced with 0.5 mM isopropyl 1-thio-b-d-galactopyranoside (IPTG), and
grown overnight. Cells were harvested and resuspended in 20 mM HEPES-
Na [pH 7.0] buffer, containing 8 M urea and heat lysed at 100ꢁC for 10 min.
The lysate was centrifuged, and the supernatant was loaded onto a nickel-
nitrilotri-acetic acid-agarose (Ni-NTA) column (QIAGEN). The column was
washed with 20 mM HEPES-Na buffer (pH 7.0) and the protein was allowed
to renature on the column before eluting with 20 mM HEPES-Na (pH 7.0), con-
taining 500 mM imidazole. Fractions containing the eluted protein were pooled
and dialyzed into NMR buffer (10 mm sodium phosphate [pH 6.5], 100 mM
NaCl, 0.02% (w/v) NaN3). The C-terminal His tag of the V domain was cleaved
by thrombin (Novagen) at room temperature for 1 hr before gel filtration chro-
matography on a SE-75 column (Amersham Biosciences). The fractions con-
taining the eluted protein were concentrated by using Ultra-Centricones (Milli-
pore). Purity was estimated to be >95% by Coomassie-stained SDS-PAGE.
V
domain-CEL-PEP complex, standard triple resonance spectra 1H,15
N
HSQC, HN(CA)CO, HNCO, HN(CO)CA, HNCA, CBCA(CO)NH, and HNCACB
(Cavanagh et al., 1996) were acquired at 298 K using an Avance Bruker spec-
trometer operating at a 1H frequency of 700 MHz equipped with a single Z-axis
gradient cryoprobe. To obtain the side-chain resonance assignments of
V domain bound to CEL-PEP, 1H, 13C HSQC, 1H, 13C 3D NOESY-HSQC,
and 3D HCCH-TOCSY experiments (Cavanagh et al., 1996) were performed.
To assign intramolecular nOes in CEL-PEP,
a
15N-filtered 13C filtered
NOESY-HSQC (Cavanagh et al., 1996; Iwahara et al., 2001; Zwahlen et al.,
1997) was acquired on an NMR sample containing 300 mM [U-13C,15N]
V
domain and 200 mM unlabeled CEL-PEP, and a
13C-double-filtered
NOESY-HSQC (Cavanagh et al., 1996; Iwahara et al., 2001; Zwahlen et al.,
1997) was acquired on a NMR sample with 350 mM [U-13C,15N] V domain
and 200 mM unlabeled CEL-PEP. To assign intermolecular nOes, a 15N-edited
13C filtered NOESY-HSQC (Iwahara et al., 2001; Zwahlen et al., 1997) was
acquired on an NMR sample containing 320 mM [U-13C,15N] V domain and
1 mM unlabeled CEL-PEP, and a 13C-edited, 13C filtered NOESY-HSQC
(Iwahara et al., 2001; Zwahlen et al., 1997) was acquired on the NMR sample
with 300 mM [U-13C,15N] V domain and 1 mM unlabeled CEL-PEP. To identify
the peaks from intramolecular 13C-bound protons, the NOESY spectra were
acquired both with and without heteronuclear 13C decoupling during the
indirect proton acquisition period. Peaks that were split in the absence of
decoupling were assigned to intramolecular V domain nOes. All spectra
were processed using TOPSPIN 2.1 (Bruker, Inc), and assignments were
made using CARA (Masse and Keller, 2005).
Expression and Purification of Wild-Type and Mutant VC1 Domains
RAGE fragments were overexpressed in E. coli strain OrigamiB-(DE3) (Nova-
gen), grown at 37ꢁC to OD600 ꢀ0.8, adjusted to 20ꢁC for 30 min, induced
with 1 mM IPTG, and allowed to express for 4–6 hr. Cells were lysed at 4ꢁC
in lysis buffer (20 mM Tris-HCl [pH 8.0], 20 mM imidazole, 300 mM NaCl) in
the presence of lysozyme (5 mg/ml), followed by sonication (5 min with
a 50% duty cycle). Clarified lysate was initially purified on a Ni-NTA (QIAGEN)
column equilibrated with lysis buffer and eluted with 20 mM Tris-HCl [pH 8.0],
100 mM imidazole, 300 mM NaCl. The C-terminal His tag of the VC1 domain
was cleaved by thrombin (Novagen) at room temperature for 1 hr before gel
filtration chromatography on a SE-75 column (Amersham Biosciences). The
fractions containing the eluted protein were concentrated by using Amicon-
Ultra-Centricones (Millipore). Residual endotoxin was removed from the
sample by repetitive use of EndoTrap Red (Lonza). The endotoxin level of
the final protein solution was determined by using Gel Clot LAL Assay (Lonza)
to be less than 0.1 EU/ml. Purity was estimated to be >95% by Coomassie-
stained SDS-PAGE.
To obtain translational diffusion coefficients, D, gradient diffusion experi-
ments were performed using the pulse sequence described by Ferrage et al.
(2003). A 100 mM protein sample of the uniformly labeled [U-15N] VC1 construct
was dissolved in NMR buffer (10 mM potassium phosphate [pH 7.2], 100 mM
NaCl, 0.02% (w/v) NaN3, in 90%H2O/10%D2O). The attenuation of the NMR
signal due to the increase in the strength of the gradient field was used to
calculate translational diffusion coefficients that, for a spherical Brownian
particle, is inversely proportional to the Stokes’ radius (Schimmel and Cantor,
1980),
Preparation of CML-Bovine Serum Albumin
We followed the protocol (Kislinger et al., 1999) that leads exclusively to CML
modifications of lysine. In brief, CML-BSA was prepared by incubating 5 mM
BSA (fraction V, fatty acid free, endotoxin free bovine serum albumin, EMD) in
150 mM sodium phosphate buffer [pH 7.4], containing 25 mM glyoxylic acid
and 75 mM NaBH3CN, at 50ꢁC for 48 hr. The reaction mixture was dialyzed
against 10 mM sodium phosphate buffer [pH 7.0] and 100 mM NaCl to remove
unreacted glyoxylic acid and NaBH3CN and stored at ꢂ20ꢁC in 20% glycerol.
D = KBT=ð6phraÞ;
(1)
with KBT being the thermal energy, h the viscosity of the solvent and ra the
Stokes’ radius. In our experiment, the diffusion delay was D + 6t = 1 s; each
Structure 19, 722–732, May 11, 2011 ª2011 Elsevier Ltd All rights reserved 729

Structure
AGE Recognition by RAGE
sine-shaped encoding gradient lasted d = 1.3 ms. Translational diffusion coef-
ficients were calculated by fitting integrated amide signal intensities using the
equation:
buffer (Cell Signaling Technology) containing 1 mM phenylmethylsulfonyl fluo-
ride and Complete Protease Inhibitors (Roche Applied Science).
À
Á
S=S0 = exp ꢂ Dk2ðD + 6tÞ ;
(2)
Western Blot Analysis
Total cell lysate was immunoblotted and probed with ERK, p-ERK, P38, and
pP38 specific antibodies (Cell Signaling Technology), HRP-conjugated donkey
anti-rabbit IgG (Amersham Pharmacia) or HRP-conjugated sheep anti-mouse
IgG (Amersham Pharmacia) were used to visualize the bands on the gel. After
probing with the p-ERK and pP38 antibodies, membranes were stripped of
boundimmunoglobulins andreprobed withERKorP38antibody for relative total
protein. Blots were scanned with an AlfaImager TM 2200 scanner with AlfaEase
(AlfaImager) FC 2200 software. Results are reported as a relative of test antigen
to relative total proteins. In all western blot studies, at least three cell lysates per
group were used; results of representative experiments are shown.
where
k = gsGmaxd:
(3)
g is the proton gyromagnetic ratio, s the shape of the encoding and decoding
gradient pulses, and G their peak amplitude.
Structure Calculation
Structural calculations were carried out with Cyana 2.1(Guntert, 2004)
using 986 distance restraints derived from 13C-edited NOESY and 15N-edited
NOESY spectra, 150 backbone torsion angle restraints derived from TALOS
(Cornilescu et al., 1999), 38 restraints for hydrogen bonds, and the restraints
from one disululfide bond between Cys38 and Cys99. nOes were converted
to upper limit distances using the CALIBA module in CYANA (Guntert, 2004).
The reference volume determined by CALIBA was increased two times before
conversion in order to loosen the distance restraints. All upper limit distances
ACCESSION NUMBERS
Coordinates and chemical shift assignments of the CEL-PEP V domain
complex have been deposited in the Protein Data Bank with accession number
2L7U.
˚
for intermolecular nOes were set to 6 A. The geometry of unnatural amino acids
CML and CEL were added to the CYANA library. Backbone torsion angle
restraints for CEL-PEP were estimated by using TALOS (Cornilescu et al.,
1999). These experimental restraints are summarized in Table S1. To perform
CYANA calculations, a single polypeptide chain was constructed for the V
domain and CEL-PEP molecules.
SUPPLEMENTAL INFORMATION
Supplemental Information includes five figures and two tables and can be
found with this article online at doi:10.1016/j.str.2011.02.013.
Refinement
ACKNOWLEDGMENTS
The CYANA-generated distance and angle restraints were converted into CNS
format in CCPN (Fogh et al., 2002). The structure for the nonstandard amino
acids CEL and CML was generated and energy-minimized in PRODRG2
(Schuttelkopf and van Aalten, 2004). A total of 1000 structures were calcu-
lated, and the 200 lowest energy structures were subjected to water refine-
ment and further analysis by PROCHECK_NMR (Laskowski et al., 1996).
80.5% of the V domain residues were in the most favorable regions of Rama-
chandran plot, 18.2% were in the additional allowed regions and 1.3% were in
generously allowed regions. There were no residues in the disallowed regions
of the Ramachandran plot. The structural statistics of the 25 best structures
are reported in Table S1.
This work was supported by funding from the American Diabetes Association
1-06-CD-23 and the National Institutes of Health R01 GM085006-01A2. We
thank Gregory McLaughlin (University at Albany) for his help in preparation
of the manuscript.
Received: October 15, 2010
Revised: February 9, 2011
Accepted: February 10, 2011
Published: May 10, 2011
REFERENCES
Fluorescence Titration
Measurements were performed on a Fluorolog-3 fluorescence spectropho-
tometer (HORIBA Jobin Yvon) at 25ꢁC in a 1 ml stirred cuvette. For fluores-
cence titration experiments, 100 nM of V domain dissolved in 10 mM phos-
phate buffer [pH 6.5] and 100 mM NaCl was used, and 1 mM solution of
CML- and CEL-containing peptides was used to increase the concentration
in 10 nM steps. Titrations in the absence of V domain and in the absence
CML (CEL) peptides were performed as references. Tryptophan fluorescence
was measured using an excitation wavelength of 280 nm. The fluorescence
emission signal was subtracted from the signal of the reference titrations,
and the differences adjusted by the dilution factor were plotted against the final
concentration of added CML (CEL) peptide. Curve fitting (OriginLab) was per-
formed to find the best values for Kd using a single site binding isotherm
approximation (Eftink, 1997).
Atherton, E., Fox, H., Harkiss, D., and Sheppard, R.C. (1978). J. Chem. Soc.
Chem. Commun., 537–539.
Bork, P., Holm, L., and Sander, C. (1994). The immunoglobulin fold. Structural
classification, sequence patterns and common core. J. Mol. Biol. 242,
309–320.
Brett, J., Schmidt, A.M., Zou, Y.S., Yan, S.D., Weidman, E., Pinsky, D., Neeper,
M., Przysiecki, M., shaw, A., Migheli, A., and Stern, D. (1993). Tissue distribu-
tion of the receptor for advanced glycation end products (RAGE): expression in
smooth muscle, cardiac myocyte, and neural tissue in addition to vasculature.
Am. J. Pathol. 143, 1699–1712.
Brownlee, M. (1995). Advanced protein glycosylation in diabetes and aging.
Annu. Rev. Med. 46, 223–234.
Brownlee, M., Vlassara, H., and Cerami, A. (1984). Nonenzymatic glycosylation
and the pathogenesis of diabetic complications. Ann. Intern. Med. 101,
527–537.
Cell Lines and Materials
Wild-type mice primary vascular smooth muscle cells were isolated and
cultured from aortas and employed through passage 5 to 7. Rat C6 glioma
cells were obtained from ATCC (CCL-107) and maintained in Dulbecco’s
modified Eagle’s medium supplemented with 10% fetal bovine serum (Invitro-
gen). Primary vascular smooth muscle cells and C6 glioma cells were seeded
at 1 3 106 cells/100 mm dish in complete medium and grown for 24 hr before
starvation overnight in serum-free medium. After overnight starvation cells
were stimulated with 10 mg/ml CML-BSA or preincubated (2 hr at 4ꢁC)
CML-BSA with at least equimolar amounts of single mutants K52A, R98A,
double mutant K52A-R98A, or wild-type VC1 domain for the indicated time
point, rinsed with ice-cold phosphate-buffered saline, and lysed using lysis
Cavanagh, J., Fairbrother, W.J., Palmer, A.G., and Skelton, N.J. (1996). Protein
NMR Spectroscopy: Principles and Practice (San Diego: Academic Press).
Chavakis, T., Bierhaus, A., Al-Fakhri, N., Schneider, D., Witte, S., Linn, T.,
Nagashima, M., Morser, J., Arnold, B., Preissner, K.T., and Nawroth, P.P.
(2003). The pattern recognition receptor (RAGE) is a counterreceptor for leuko-
cyte integrins: a novel pathway for inflammatory cell recruitment. J. Exp. Med.
198, 1507–1515.
Chothia, C., and Jones, E.Y. (1997). The molecular structure of cell adhesion
molecules. Annu. Rev. Biochem. 66, 823–862.
730 Structure 19, 722–732, May 11, 2011 ª2011 Elsevier Ltd All rights reserved

Structure
AGE Recognition by RAGE
Cornilescu, G., Delaglio, F., and Bax, A. (1999). Protein backbone angle
glycation end products: new insight into AGE-RAGE interaction.
restraints from searching
a
database for chemical shift and sequence
Biochemistry 47, 12299–12311.
homology. J. Biomol. NMR 13, 289–302.
Neeper, M., Schmidt, A.M., Brett, J., Yan, S.D., Wang, F., Pan, Y.C., Elliston,
K., Stern, D., and Shaw, A. (1992). Cloning and expression of a cell surface
receptor for advanced glycosylation end products of proteins. J. Biol. Chem.
267, 14998–15004.
Dattilo, B.M., Fritz, G., Leclerc, E., Kooi, C.W., Heizmann, C.W., and Chazin,
W.J. (2007). The extracellular region of the receptor for advanced glycation
end products is composed of two independent structural units.
Biochemistry 46, 6957–6970.
Nogi, T., Sangawa, T., Tabata, S., Nagae, M., Tamura-Kawakami, K., Beppu,
A., Hattori, M., Yasui, N., and Takagi, J. (2008). Novel affinity tag system using
structurally defined antibody-tag interaction: application to single-step protein
purification. Protein Sci. 17, 2120–2126.
Eftink, M.R. (1997). Fluorescence methods for studying equilibrium macromol-
ecule-ligand interactions. Methods Enzymol. 278, 221–257.
Ferrage, F., Zoonens, M., Warschawski, D.E., Popot, J.L., and Bodenhausen,
G. (2003). Slow diffusion of macromolecular assemblies by a new pulsed field
gradient NMR method. J. Am. Chem. Soc. 125, 2541–2545.
Ostendorp, T., Leclerc, E., Galichet, A., Koch, M., Demling, N., Weigle, B.,
Heizmann, C.W., Kroneck, P.M., and Fritz, G. (2007). Structural and func-
tional insights into RAGE activation by multimeric S100B. EMBO J. 26,
3868–3878.
Fogh, R., Ionides, J., Ulrich, E., Boucher, W., Vranken, W., Linge, J.P., Habeck,
M., Rieping, W., Bhat, T.N., Westbrook, J., et al. (2002). The CCPN project: an
interim report on a data model for the NMR community. Nat. Struct. Biol. 9,
416–418.
Perkins, S.J., and Wuthrich, K. (1979). Ring current effects in the conformation-
dependent NMR chemical shifts of aliphatic protons in the basic pancreatic
trypsin inhibitor. Biochim. Biophys. Acta 576, 409–423.
Guex, N., and Peitsch, M.C. (1997). SWISS-MODEL and the Swiss-
PdbViewer: an environment for comparative protein modeling.
Electrophoresis 18, 2714–2723.
Povey, J.F., Howard, M.J., Williamson, R.A., and Smales, C.M. (2008). The
effect of peptide glycation on local secondary structure. J. Struct. Biol. 161,
151–161.
Guntert, P. (2004). Automated NMR structure calculation with CYANA.
Methods Mol. Biol. 278, 353–378.
Ramasamy, R., Vannucci, S.J., Yan, S.S., Herold, K., Yan, S.F., and Schmidt,
A.M. (2005a). Advanced glycation end products and RAGE: a common thread
in aging, diabetes, neurodegeneration, and inflammation. Glycobiology 15,
16R–28R.
Habeeb, A.F. (1966). Determination of free amino groups in proteins by trinitro-
benzenesulfonic acid. Anal. Biochem. 14, 328–336.
Hofmann, M.A., Drury, S., Fu, C., Qu, W., Taguchi, A., Lu, Y., Avila, C.,
Kambham, N., Bierhaus, A., Nawroth, P., et al. (1999). RAGE mediates a novel
proinflammatory axis: a central cell surface receptor for S100/calgranulin poly-
peptides. Cell 97, 889–901.
Ramasamy, R., Yan, S.F., and Schmidt, A.M. (2005b). The RAGE axis and
endothelial dysfunction: maladaptive roles in the diabetic vasculature and
beyond. Trends Cardiovasc. Med. 15, 237–243.
Hori, O., Brett, J., Slattery, T., Cao, R., Zhang, J., Chen, J.X., Nagashima, M.,
Lundh, E.R., Vijay, S., Nitecki, D., et al. (1995). The receptor for advanced gly-
cation end products (RAGE) is a cellular binding site for amphoterin. Mediation
of neurite outgrowth and co-expression of rage and amphoterin in the devel-
oping nervous system. J. Biol. Chem. 270, 25752–25761.
Schimmel, P.R., and Cantor, C.R. (1980). Biophysical Chemistry: PartII; tech-
niques for the Study of Biologycal Structure and Function (New York: W.H.
Freeman).
Schmidt, A.M., and Stern, D.M. (2001). Receptor for age (RAGE) is a gene
within the major histocompatibility class III region: implications for host
response mechanisms in homeostasis and chronic disease. Front. Biosci. 6,
D1151–D1160.
Ishiguro, H., Nakaigawa, N., Miyoshi, Y., Fujinami, K., Kubota, Y., and Uemura,
H. (2005). Receptor for advanced glycation end products (RAGE) and its
ligand, amphoterin are overexpressed and associated with prostate cancer
development. Prostate 64, 92–100.
Schmidt, A.M., Hori, O., Chen, J.X., Li, J.F., Crandall, J., Zhang, J., Cao, R.,
Yan, S.D., Brett, J., and Stern, D. (1995). Advanced glycation endproducts in-
teracting with their endothelial receptor induce expression of vascular cell
adhesion molecule-1 (VCAM-1) in cultured human endothelial cells and in
mice. A potential mechanism for the accelerated vasculopathy of diabetes.
J. Clin. Invest. 96, 1395–1403.
Iwahara, J., Wojciak, J.M., and Clubb, R.T. (2001). Improved NMR spectra of
a protein-DNA complex through rational mutagenesis and the application of
a sensitivity optimized isotope-filtered NOESY experiment. J. Biomol. NMR
19, 231–241.
Kislinger, T., Fu, C., Huber, B., Qu, W., Taguchi, A., Du Yan, S., Hofmann, M.,
Yan, S.F., Pischetsrieder, M., Stern, D., and Schmidt, A.M. (1999). N(epsilon)-
(carboxymethyl)lysine adducts of proteins are ligands for receptor for
advanced glycation end products that activate cell signaling pathways and
modulate gene expression. J. Biol. Chem. 274, 31740–31749.
Schmidt, A.M., Hori, O., Cao, R., Yan, S.D., Brett, J., Wautier, J.L., Ogawa, S.,
Kuwabara, K., Matsumoto, M., and Stern, D. (1996). RAGE: a novel cellular
receptor for advanced glycation end products. Diabetes 45 (Suppl 3),
S77–S80.
Schmidt, A.M., Hofmann, M., Taguchi, A., Yan, S.D., and Stern, D.M.
(2000). RAGE: a multiligand receptor contributing to the cellular response
in diabetic vasculopathy and inflammation. Semin. Thromb. Hemost. 26,
485–493.
Koch, M., Chitayat, S., Datillo, B.M., Schiefner, A., Diez, J., Chazin, W., and
Fritz, G. (2010). Structural Basis for Ligand recognition and activation of
RAGE. Structure 18, 1342–1352.
Koradi, R., Billeter, M., and Wuthrich, K. (1996). MOLMOL: a program for
Schuttelkopf, A.W., and van Aalten, D.M. (2004). PRODRG: a tool for high-
throughput crystallography of protein-ligand complexes. Acta Crystallogr. D
Biol. Crystallogr. 60, 1355–1363.
display and analysis of macromolecular structures. J. Mol. Graph. 14, 51–55.
Laskowski, R.A., Rullmannn, J.A., MacArthur, M.W., Kaptein, R., and
Thornton, J.M. (1996). AQUA and PROCHECK-NMR: programs for checking
the quality of protein structures solved by NMR. J. Biomol. NMR 8, 477–486.
Sugaya, K., Fukagawa, T., Matsumoto, K., Mita, K., Takahashi, E., Ando, A.,
Inoko, H., and Ikemura, T. (1994). Three genes in the human MHC class III
region near the junction with the class II: gene for receptor of advanced
glycosylation end products, PBX2 homeobox gene and a notch homolog,
human counterpart of mouse mammary tumor gene int-3. Genomics 23,
408–419.
Leclerc, E., Fritz, G., Weibel, M., Heizmann, C.W., and Galichet, A. (2007).
S100B and S100A6 differentially modulate cell survival by interacting with
distinct RAGE (receptor for advanced glycation end products) immunoglobulin
domains. J. Biol. Chem. 282, 31317–31331.
Taguchi, A., Blood, D.C., del Toro, G., Canet, A., Lee, D.C., Qu, W., Tanji, N.,
Lu, Y., Lalla, E., Fu, C., et al. (2000). Blockade of RAGE-amphoterin signalling
suppresses tumour growth and metastases. Nature 405, 354–360.
Masse, J.E., and Keller, R. (2005). AutoLink: automated sequential resonance
assignment of biopolymers from NMR data by relative-hypothesis-prioritiza-
tion-based simulated logic. J. Magn. Reson. 174, 133–151.
Matsumoto, S., Yoshida, T., Murata, H., Harada, S., Fujita, N., Nakamura, S.,
Yamamoto, Y., Watanabe, T., Yonekura, H., Yamamoto, H., et al. (2008).
Solution structure of the variable-type domain of the receptor for advanced
Thornalley, P.J. (1998). Cell activation by glycated proteins. AGE receptors,
receptor recognition factors and functional classification of AGEs. Cell. Mol.
Biol. (Noisy-le-grand) 44, 1013–1023.
Structure 19, 722–732, May 11, 2011 ª2011 Elsevier Ltd All rights reserved 731

Structure
AGE Recognition by RAGE
Thornalley, P.J. (1999). Clinical significance of glycation. Clin. Lab. 45,
glycated end products (RAGE) using symmetric hydrophobic target-binding
263–273.
patches. J. Biol. Chem. 282, 4218–4231.
Thornalley, P.J., Battah, S., Ahmed, N., Karachalias, N., Agalou, S., Babaei-
Jadidi, R., and Dawnay, A. (2003). Quantitative screening of advanced glyca-
tion endproducts in cellular and extracellular proteins by tandem mass spec-
trometry. Biochem. J. 375, 581–592.
Xie, J., Reverdatto, S., Frolov, A., Hoffmann, R., Burz, D.S., and Shekhtman, A.
(2008). Structural basis for pattern recognition by the receptor for advanced
glycation end products (RAGE). J. Biol. Chem. 283, 27255–27269.
Zong, H., Madden, A., Ward, M., Mooney, M.H., Elliott, C.T., and Stitt, A.W.
(2010). Homodimerization is essential for the receptor for advanced glycation
end products (RAGE)-mediated signal transduction. J. Biol. Chem. 285,
23137–23146.
Valente, T., Gella, A., Fernandez-Busquets, X., Unzeta, M., and Durany, N. (2010).
Immunohistochemical analysis of human brain suggests pathological synergism
of Alzheimer’s disease and diabetes mellitus. Neurobiol. Dis. 37, 67–76.
Wa, C., Cerny, R.L., Clarke, W.A., and Hage, D.S. (2007). Characterization of
glycation adducts on human serum albumin by matrix-assisted laser desorp-
tion/ionization time-of-flight mass spectrometry. Clin. Chim. Acta 385, 48–60.
Zwahlen, C., Legault, P., Vincent, S.J.F., Greenblatt, J., Konrat, R., and Kay,
L.E. (1997). Methods for measurement of intermolecular NOEs by multinuclear
NMR spectroscopy: application to a bacteriophage l N-peptide/boxB RNA
complex. J. Am. Chem. Soc. 119, 6711–6721.
Xie, J., Burz, D.S., He, W., Bronstein, I.B., Lednev, I., and Shekhtman, A.
(2007). Hexameric calgranulin C (S100A12) binds to the receptor for advanced
732 Structure 19, 722–732, May 11, 2011 ª2011 Elsevier Ltd All rights reserved