Structure, Vol. 13, 825–832, May, 2005, ©2005 Elsevier Ltd All rights reserved. DOI 10.1016/j.str.2005.03.009
Structure of the Angiopoietin-2 Receptor
Binding Domain and Identification of Surfaces
Involved in Tie2 Recognition
William A. Barton,1,2Dorothea Tzvetkova,1
and Dimitar B. Nikolov1,*
1Structural Biology Program
Memorial Sloan-Kettering Cancer Center
1275 York Avenue
New York, New York 10021
for receptor recognition and binding, the coiled-coil
motif mediates homo- or heterodimerization (Davis et
al., 2003; Fiedler et al., 2003; Ward and Dumont, 2002).
While Ang2 can exist as a dimer, both Ang2 and Ang1
are observed primarily as tetramers, hexamers, and
higher-order aggregates in solution (Davis et al., 1996,
2003; Fiedler et al., 2003). Thus, it is thought that the
superclustering region assembles the preformed angio-
poietin coiled-coil dimers into higher-order oligomers.
While Ang1 is a constitutive agonist, Ang2 is a
context-dependent one. It, for example, is incapable of
activating endogenous Tie2 in an endothelial cell cul-
ture system (Maisonpierre et al., 1997). Within this sys-
tem, activation of Tie2 by Ang1 can be competitively
inhibited by Ang2, indicating that they function by act-
ing through the same binding site (Fiedler et al., 2003;
Maisonpierre et al., 1997). This has led to the hypothe-
sis that a general physiological role of Ang2 is to serve
as a competitive antagonist, competing with Ang1 for
In order to gain a more accurate molecular under-
standing of angiopoietin/Tie2 interactions, we deter-
mined the three-dimensional structure of the Ang2 fi-
brinogen-like receptor binding domain (Ang2-RBD) at
2.4 Å resolution. The structure reveals a patch of sur-
face residues conserved within the angiopoietin family
that is likely involved in receptor recognition. Structure-
based mutagenesis and in vitro binding studies corro-
borate the role of this molecular surface in Tie2 binding.
The angiopoietins comprise a small class of secreted
glycoproteins that play crucial roles in the maturation
and maintenance of the mammalian vascular and lym-
phatic systems. They exert their effects through a
member of the tyrosine kinase receptor family, Tie2.
Angiopoietin/Tie2 signaling is unique among tyrosine
kinase receptor-ligand systems in that distinct angio-
poietin ligands, although highly homologous, can
function as agonists or antagonists in a context-
dependent manner. In an effort to understand this mo-
lecular dichotomy, we have crystallized and deter-
mined the 2.4 Å crystal structure of the Angiopoietin-2
(Ang2) receptor binding region. The structure reveals
a fibrinogen fold with a unique C-terminal P domain.
Conservation analysis and structure-based mutagen-
esis identify a groove on the Ang2 molecular surface
that mediates receptor recognition.
The angiopoietins are the secreted ligands of the Tie
family of protein receptor tyrosine kinases (RTK) (Davis
et al., 1996, 2003; Maisonpierre et al., 1997; Ramsauer
and D’Amore, 2002; Valenzuela et al., 1999). Tie2 is one
of a handful of endothelial-specific RTKs that also in-
clude VEGFR. Several groups have shown Tie2 to in-
teract with all four members of the angiopoietin family
(Davis et al., 1996, Valenzuela et al., 1999). Binding
studies have identified the first Ig domain and the three
EGF repeats of Tie2 as essential for angiopoietin bind-
ing (Fiedler et al., 2003).
The four known angiopoietins share significant amino
acid sequence identity and contain a small amino-ter-
minal region that modulates angiopoietin clustering
(superclustering region), followed by a rather large
coiled-coil motif (Davis et al., 1996; Maisonpierre et al.,
1997; Ramsauer and D’Amore, 2002; Valenzuela et al.,
1999). A short linker separates the coiled-coil motif
from a fibrinogen domain at the carboxy terminus.
Binding experiments with Ang1 truncation mutants
have shown that the coiled-coil region is dispensable
for Tie2 binding, although it is responsible for a rather
modest increase in Kd(Davis et al., 2003). It is proposed
that, while the fibrinogen domain is solely responsible
Functional Characterization of the Angiopoietin-2
Receptor Binding Domain
The Ang2 receptor binding domain and the Tie2 ligand
binding domain (Tie2-LBD) were expressed as secreted
Fc-fusion proteins from stably transfected HEK293 cell
lines. To assess the biological activity of our angiopoie-
tin and Tie constructs, we employed several binding
assays. First, gel filtration chromatography was used to
confirm the high-affinity Ang2/Tie2 interaction, in which
we monitored the retention of either Ang2 or Tie2, or a
mixture of the two on an analytical size-exclusion col-
umn. When injected alone, Ang2-RBD elutes with a
retention volume of 17.9 ml corresponding to an ap-
proximate molecular weight of 25 kDa (Figure 1A). Al-
ternatively, Tie2-LBD, when chromatographed alone,
elutes with a retention volume of 15.2 ml and an ap-
proximate molecular weight of 55 kDa (Figure 1B). How-
ever, when Tie2-LBD and Ang2-RBD are injected to-
gether (with Ang2-RBD in slight molar excess), two
peaks are observed with retention volumes of 14.4 ml
(or approximately 70 kDa) and 17.9 ml (or w25 kDa)
(Figure 1C). The first peak at w70 kDa corresponds to
a 1:1 Tie2-LBD/Ang2-RBD complex, while the second
peak corresponds to Ang2-RBD alone. Samples (frac-
tions 12–20) from each run were resolved on SDS-PAGE
and are shown in Figures 1A–1C. Panel (C) clearly il-
lustrates the comigration of Tie2 and Ang2.
2Present address: Institute for Structural Biology and Drug Discov-
ery, Virginia Commonwealth University, 800 East Leigh Street, Suite
212, Richmond, Virginia 23219.
Crystals grown in the presence of zinc were harvested and flash
frozen in the cold stream of an X-Stream cooling system in the
mother liquor with 30% PEG-550 as a cryoprotectant, while crys-
tals grown in the presence of PEG-4000 were rapidly transferred to
a cryo-buffer consisting of the mother liquor with an additional 20%
glycerol. The data were collected at NSLS Brookhaven beamline
X9A and CHESS beamline F2. Oscillation photographs were integ-
rated, merged, and scaled by using DENZO and SCALEPACK,
respectively (Otwinowski and Minor, 1997). Subsequent calcula-
tions were done with autoSHARP and the CCP4 program suite
(CCP4, 1994; Evans and Bricogne, 2002). The structure was deter-
mined with a combination of MAD phasing and molecular replace-
ment. Initially, the program AMORE was used to identify the loca-
tion of the fibrinogen domain by using the horseshoe crab
Tachylectin 5A (PDB ID 1JC9) as a search model (Fiedler et al.,
2003; Jones et al., 1991). Simultaneously, a zinc anomalous data
set was collected by utilizing a single zinc atom involved in mediat-
ing an intermolecular crystal contact. The program autoSHARP
was used to identify the location of the zinc atom and to refine its
position and occupancy for phase calculations. Density modifica-
tion with DM improved the zinc MAD maps that proved to be of
sufficient quality to confirm the placement of the fibrinogen domain
(CCP4, 1994). The model was subjected to 100 rounds of improve-
ment in the program ARP/wARP by using the phases calculated
by autoSHARP and phase constraints with a blurring factor of 1.0
(Perrakis et al., 2001). Refinement proceeded with iterative rounds
of model adjustments (by using the molecular graphics program
O), molecular dynamics, and energy minimization in CNS (Brunger
et al., 1998; Jones et al., 1991). The first model was refined at 3.2 Å
resolution to R and Rfreevalues of 26.9 and 32.8, respectively, for
1725 atoms and 215 amino acids.
At this point, crystals grown from MgCl2and PEG that diffracted
to a higher resolution were identified. The high-resolution structure
was determined by using molecular replacement with the zinc
Ang2-RBD structure as a search model. The model was subjected
to adjustments with O and refined in CNS. Stereochemical analysis
of the refined models with PROCHECK (CCP4, 1994) revealed main
chain and side chain parameters better than or within the typical
range of values for protein structures.
The F468A/Y474A/Y475A mutant was crystallized from 0.2 M am-
monium sulfate, 0.1 M sodium acetate (pH 5.2), 22% PEG-4000.
Crystals were transferred to mother liquor with 20% glycerol for
flash freezing. A 2.25 Å native data set was collected at CHESS
beamline F2. The CCP4 program AMORE was used to locate four
monomers in the asymmetric unit. In an attempt to keep the model
unbiased within the loop containing the mutations, residues 467–
476 were omitted in the initial model. The model was subsequently
subjected to additional rounds of refinement within the CNS suite
prior to manually building the missing region by using the graphical
CCP4 suite: programs for X-ray crystallography. Acta Crystallogr. D
Biol. Crystallogr. 50, 760–763.
Davis, S., Aldrich, T.H., Jones, P.F., Acheson, A., Compton, D.L.,
Jain, V., Ryan, T.E., Bruno, J., Radziejewski, C., Maisonpierre, P.C.,
and Yancopoulos, G.D. (1996). Isolation of Angiopoietin-1, a ligand
for the Tie2 receptor by secretion-trap expression cloning. Cell 87,
Davis, S., Papadopoulos, N., Aldrich, T.H., Maisonpierre, P.C., Hu-
ang, T., Kovac, L., Xu, A., Leidich, R., Radziejewska, E., Rafique, A.,
et al. (2003). Angiopoietins have distinct modular domains essential
for receptor binding, dimerization, and superclustering. Nat. Struct.
Biol. 10, 38–44.
Evans, G., and Bricogne, G. (2002). Triiodide derivatization and
combinatorial counter-ion replacement: two methods for enhanc-
ing phasing signal using laboratory Cu Kalpha X-ray equipment.
Acta Crystallogr. D Biol. Crystallogr. 58, 976–991.
Fiedler, U., Krissl, T., Koidl, S., Weiss, C., Koblizek, T., Deutsch, U.,
Martiny-Baron, G., Marme, D., and Augustin, H.G. (2003). Angio-
poietin-1 and Angiopoietin-2 share the same binding domains in
the Tie2 receptor involving the first Ig-like loop and the epidermal
growth factor-like repeats. J. Biol. Chem. 278, 1721–1727.
Jones, T.A., Zou, J.Y., Cowan, S.W., and Kjeldgaard, M. (1991). Im-
proved methods for building protein models in electron density
maps at the location of errors in these models. Acta Crystallogr. A
Kairies, N., Beisel, H.G., Fuentes-Prior, P., Tsuda, R., Muta, T., Iwa-
naga, S., Bode, W., Huber, R., and Kawabata, S. (2001). The 2.0-Å
crystal structure of tachylectin 5A provides evidence for the com-
mon origin of the innate immunity and the blood coagulation sys-
tems. Proc. Natl. Acad. Sci. USA 98, 13519–13524.
Kawabata, S., and Tsuda, R. (2002). Molecular basis of non-self
recognition by the horseshoe crab tachylectins. Biochim. Biophys.
Acta 1572, 414–421.
Koopman, J., Haverkate, F., Briet, E., and Lord, S.T. (1991). A con-
genitally abnormal fibrinogen (Vlissingen) with a 6-base deletion in
the γ-chain gene, causing defective calcium binding and impaired
fibrin polymerization. J. Biol. Chem. 266, 13456–13461.
Maisonpierre, P.C., Suri, C., Jones, P.F., Bartunkova, S., Weigand,
S.J., Radziejewski, C., Compton, D., McClain, J., Aldrich, T.H., Pa-
padopoulos, N., et al. (1997). Angiopoietin-2, a natural antagonist
for Tie2 that disrupts in vivo angiogenesis. Science 277, 55–60.
Otwinowski, Z., and Minor, W. (1997). Processing of X-ray diffrac-
tion data collected in oscillation mode. Methods Enzymol. 276,
Perrakis, A., Harkiolaki, M., Wilson, K.S., and Lamzin, V.S. (2001).
ARP/wARP and molecular replacement. Acta Crystallogr. D Biol.
Crystallogr. 57, 1445–1450.
Pratt, K.P., Cote, H.C.F., Chung, D.W., Stenkamp, R.E., and Davie,
E.W. (1997). The primary fibrin polymerization pocket: three-dimen-
sional structure of a 30 kDa C-terminal γ chain fragment complexed
with the peptide Gly-Pro-Arg-Pro. Proc. Natl. Acad. Sci. USA 94,
Ramsauer, M., and D’Amore, P.A. (2002). Getting Tie(2)d up in angi-
ogenesis. J. Clin. Invest. 110, 1615–1617.
Valenzuela, D.M., Griffiths, J.A., Rojas, J., Aldrich, T.H., Jones, P.F.,
Zhou, H., McClain, J., Copeland, N.G., Gilbert, D.J., Jenkins, N.A.,
et al. (1999). Angiopoietin 3 and 4: diverging gene counterparts in
mice and humans. Proc. Natl. Acad. Sci. USA 96, 1904–1909.
Ward, N.L., and Dumont, D.J. (2002). The angiopoietins and Tie2/
Tek: adding to the complexity of cardiovascular development.
Semin. Cell Dev. Biol. 13, 19–27.
Yee, V.C., Pratt, K.P., Cote, H.C., Trong, I.L., Chung, D.W., Davie,
E.W., Stenkamp, R.E., and Teller, D.C. (1997). Crystal structure of
a 30 kDa C-terminal fragment from the gamma chain of human
fibrinogen. Structure 5, 125–138.
The authors would like to thank Dr. Min Lu (Biochemistry Depart-
ment of Cornell University) for performing the analytical ultracentrif-
ugation experiments. This work was supported by the National In-
stitute of Health (R01-HL077429) and the New York State Spinal
Cord Injury Research Program.
Received: October 27, 2004
Revised: March 3, 2005
Accepted: March 6, 2005
Published: May 10, 2005
Brunger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P.,
Grosse-Kunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M.,
Pannu, N.S., et al. (1998). Crystallography and NMR system: a new
software suite for macromolecular structure determination. Acta
Crystallogr. D Biol. Crystallogr. 54, 905–921.
CCP4 (Collaborative Computational Project, Number 4) (1994). The
Coordinates have been deposited in the Protein Data Bank (PDB
accession numbers 1Z3S and 1Z3U).