Small multivalent architectures mimicking homotrimers of the TNF superfamily member CD40L: Delineating the relationship between structure and effector function

Article abstract of DOI:10.1021/ja073169m

Synthetic multivalent ligands, owing to the presence of multiple copies of a recognition motif attached to a central scaffold, can mediate clustering of cell surface receptors and thereby function as effector molecules. This paper dissects the relationship between structure and effector function of synthetic multivalent ligands targeting CD40, a cell surface receptor of the tumor necrosis factor receptor (TNF-R) superfamily. Triggering CD40 signaling in vivo can be used to enhance immunity against intracellular pathogens or tumors. A series of multimeric molecules has been prepared by systematically varying the shape and the valency of the central scaffold, the nature and the length of the linker as well as the sequence of the receptor binding motif. The data reported here (i) suggest that radial distribution of CD40-binding units and C ₃-symmetry are preferred for optimal binding to CD40 and signaling, (ii) underscore the importance of choosing an appropriate linker to connect the receptor binding motif to the central scaffold, and (iii) show the versatility of planar cyclic α- and β-peptides as templates for the design of CD40L mimetics. In particular, the (Ahx)₃-B trimeric scaffold-linker combination equally accommodated binding elements derived from distinct CD40L hot-spot regions including AA″ loop and β-strand E. The use of miniCD40Ls such as those reported here is complementary to other approaches (recombinant ligands, agonistic anti-receptor antibodies) and may find interesting therapeutic applications. Furthermore, the results disclosed in this paper provide the basis for future design of other TNF family member mimetics.

Full text of DOI:10.1021/ja073169m

Published on Web 10/13/2007
Small Multivalent Architectures Mimicking Homotrimers of the
TNF Superfamily Member CD40L: Delineating the
Relationship between Structure and Effector Function
Nathalie Trouche,† Se´bastien Wieckowski,† Weimin Sun,†,‡ Olivier Chaloin,†
Johan Hoebeke,† Sylvie Fournel,*,† and Gilles Guichard*,†
Contribution from the CNRS, Institut de Biologie Mole´culaire et Cellulaire, laboratoire
d’Immunologie et Chimie The´rapeutiques, 15 rue Rene´ Descartes, 67084 Strasbourg, France
Received May 4, 2007; E-mail: g.guichard@ibmc.u-strasbg.fr; s.fournel@ibmc.u-strasbg.fr
Abstract: Synthetic multivalent ligands, owing to the presence of multiple copies of a recognition motif
attached to a central scaffold, can mediate clustering of cell surface receptors and thereby function as
effector molecules. This paper dissects the relationship between structure and effector function of synthetic
multivalent ligands targeting CD40, a cell surface receptor of the tumor necrosis factor receptor (TNF-R)
superfamily. Triggering CD40 signaling in vivo can be used to enhance immunity against intracellular
pathogens or tumors. A series of multimeric molecules has been prepared by systematically varying the
shape and the valency of the central scaffold, the nature and the length of the linker as well as the sequence
of the receptor binding motif. The data reported here (i) suggest that radial distribution of CD40-binding
units and C3-symmetry are preferred for optimal binding to CD40 and signaling, (ii) underscore the importance
of choosing an appropriate linker to connect the receptor binding motif to the central scaffold, and (iii)
show the versatility of planar cyclic R- and â-peptides as templates for the design of CD40L mimetics. In
particular, the (Ahx)3-B trimeric scaffold-linker combination equally accommodated binding elements derived
from distinct CD40L hot-spot regions including AA′′ loop and â-strand E. The use of miniCD40Ls such as
those reported here is complementary to other approaches (recombinant ligands, agonistic anti-receptor
antibodies) and may find interesting therapeutic applications. Furthermore, the results disclosed in this
paper provide the basis for future design of other TNF family member mimetics.
Introduction
decade the number of reports of synthetic molecules capable
of blocking protein-protein interactions in a specific manner
Protein-protein interactions are central to most biological
processes from cellular communication to programmed cell
death and provide the basis for interesting therapeutic applica-
tions. Small molecules that can bind at protein-protein inter-
faces are thus of interest to inhibit undesired biological processes
or to promote desired ones.1 The design of such molecules,
however, is hampered by the large protein surfaces at interplay
(1000-2000 Å2 is generally considered as a “standard size”
for the area buried in the interaction2), as well as by the difficulty
to identify defined “hot spot” residues, and to handle confor-
mational, nonsequential epitopes. Nevertheless, in the past
has grown steadily.1 On the other hand, few protein mimetics
with effector functions, that can both bind to a cell surface
receptor and activate downstream signaling pathways have been
described.
Ligand-induced oligomerization or aggregation of cell surface
receptors (e.g., cytokine receptors, receptor tyrosine kinases,
tumor necrosis factor (TNF) receptors, T-cell receptors (TCRs),
Toll-like receptors (TLRs), and G-protein coupled receptors
(GPCRs) for naming a few) is an important mechanism to
initiate and control signaling in a number of biological
processes.3-8 In this context, (small) synthetic molecules binding
avidly to multiple copies of a receptor may be used to activate
signaling pathways similar to the cognate ligand. Early studies
† Immunologie et Chimie The´rapeutiques (ICT).
‡ Present Address: Institute of Biological Science and Technology,
Beijing Jiaotong University, 3 Shang Yuan Cun, Haidian District, 100044
Beijing, China.
(3) (a) Austin, D. J.; Crabtree, G. R.; Schreiber, S. L. Chem. Biol. 1994, 1,
131-6. (b) Klemm, J. D.; Crabtree, G. R.; Schreiber, S. L. Annu. ReV.
Immunol. 1998, 16, 569-592.
(1) For review, see: (a) Davis, J. M.; Tsou, L. K.; Hamilton, A. D. Chem.
Soc. ReV. 2007, 36, 326-334. (b) Zhao, L.; Chmielewski, J. Curr. Opin.
Struct. Biol. 2005, 15, 31-34. (c) Yin, H.; Hamilton, A. D. Angew. Chem.,
Int. Ed. 2005, 44, 4130-4163. (d) Arkin, M. R.; Wells, J. A. Nat. ReV.
Drug DiscoVery 2004, 3, 301-317. (e) Pagliaro, P.; Felding, J.; Audouze,
K.; Nielsen, S. J.; Terry, R. B.; Krog-Jensen, C.; Butcher, S. Curr. Opin.
Chem. Biol. 2004, 8, 442-449. (f) Berg, T. Angew. Chem., Int. Ed. 2003,
42, 2462-2481; Angew. Chem. 2003, 115, 2566-2586. (g) Toogood, P.
L. J. Med. Chem. 2002, 45, 1543-1558. (h) Peczuh, M. W.; Hamilton, A.
D. Chem. ReV. 2000, 100, 2749-2494. (i) Cochran, A. G. Chem. Biol.
2000, 7, R85-R94.
(4) (a) Heldin, K.-H. Cell 1995, 80, 214-223. (b) Heldin, C. H.; Ostman, A.
Growth Factor ReV. 1996, 7, 3-10.
(5) (a) Schlessinger, J. Cell 2000, 103, 211-225. (b) Lemmon, M. A.;
Schlessinger, J. Methods Mol. Biol. 1998, 84, 49-71.
(6) (a) Germain, R. N. Curr. Biol. 1997, 7, R640-644. (b) Alarco´n, B.; Swamy,
M.; van Santen, H. M.; Schamel, W. W. A. EMBO J. 2006, 7, 490-495.
(7) (a) Ozinsky, A.; Underhill, D. M.; Fontenot, J. D.; Hajjar, A. M.; Smith,
K. D.; Wilson, C. B.; Schroeder, L.; Aderem, A. Proc. Natl. Acad. Sci.
U.S.A. 2000, 97, 13766-13771. (b) Choe, J.; Kelker M. S.; Wilson, I. A.
Science 2005, 309, 581-585. (c) Kirk, P.; Bazan, J. Immunity, 2005, 23,
347-350.
(2) (a) Lo Conte, L.; Chothia, C.; Janin, J. J. Mol. Biol. 1999, 285, 2177-
2198. (b) Chakrabarti, P.; Janin, J. Proteins 2002, 47, 334-343.
9
13480
J. AM. CHEM. SOC. 2007, 129, 13480-13492
10.1021/ja073169m CCC: $37.00 © 2007 American Chemical Society

Relationship between Structure and Effector Function
A R T I C L E S
on synthetic multivalent systems (mainly for antagonizing
unwanted biological processes) have clearly demonstrated that
it is possible through polyvalency to achieve tight binding with
ligands of low or modest surface areas.9-13 The possibility to
create molecules with effector functions through multivalency
has been recognized more recently.11-15 In the field of cytokine
and growth factor mimetics, dimeric peptides, C2-symmetric
organic molecules as well as dendrimers capable of dimerizing
and activating erythropoietin receptor (EPO-R), thrombopoietin
receptor (TPO-R), granulocyte-colony stimulating factor (G-
CSF) receptor, nerve growth factor (NGF) receptor TrkA, or
neurotrophin-3 (NT-3) receptor TrkC are worth mentioning.16,17
While protein-ligand induced dimerization is the mechanism
of activation of receptors for cytokines and growth factors,
signaling by members of the tumor necrosis factor receptor
(TNF-R) family involves receptor trimerization.18,19 The geom-
etry of the resulting ligand-receptor complex is favorable to
the formation of an intracellular signaling complex.4a,18 TNF-R
family members can transduce a variety of intracellular signals
culminating in cell proliferation, differentiation, survival, and
death. Most of the 19 ligands and 29 receptors of the TNF/
TNF-R superfamilies play an active role in the development,
maintenance, and function of the immune system.18 A few
members are also implicated in other physiological processes
such as bone remodeling or development of skin appendages.
CD40 is expressed on antigen-presenting cells (APCs) such as
dendritic cells (DCs) and B cells and, by interacting with its
ligand CD40L (CD154), leads to their activation and differentia-
tion.20 Blockade of the CD40/CD40L pathway is a potential
immunomodulatory strategy for B- and T-cell-mediated diseases.
Administration of anti-CD40L antibodies has given encouraging
results in the treatment of experimental autoimmune diseases
as well as in the treatment of allograft rejection.21,22 Alterna-
tively, triggering CD40 signaling in ViVo can be used to enhance
immunity against intracellular pathogens or tumor cells. The
use of agonistic anti-CD40 antibodies for example has been
shown to increase the efficacy of peptide-based antitumor
vaccines.23 These findings support a clinical use of CD40-
stimulating agents as components of anticancer immunotherapies
or anti-infectious vaccines. Like other members of the TNF
family, CD40L monomers self-assemble around a three-fold
symmetry axis to form noncovalent homotrimers that can each
bind three receptor molecules. The crystal structure of CD40L
extracellular domain homotrimers24 and a homology model of
the CD40L/CD40 complex25 have provided the basis for the
(17) (a) Wrighton, N. C.; Farrell, F. X.; Chang, R.; Kashyap, A. K.; Barbone,
F. P.; Mulcahy, L. S.; Johnson, D. L.; Barrett, R. W.; Jolliffe, L. K.; Dower,
W. J. Science 1996, 273, 458-463. (b) Livnah, O.; Stura, E. A.; Johnson,
D. L.; Middleton, S. A.; Mulcahy, L. S.; Wrighton, N. C.; Dower, W. J.;
Jolliffe, L. S. Wilson, I. A. Science 1996, 273, 464-471. (c) Johnson, D.
L.; Farrell, F. X.; Barbone, F. P.; McMahon, F. J.; Tullai, J.; Kroon, D.;
Freedy, J.; Zivin, R. A.; Mulcahy, L. S.; Jolliffe, L. K. Chem. Biol. 1997,
4, 939-950. (d) Cwirla, S. E.; Balasubramanian, P.; Duffin, D. J.;
Wagstrom, C. R.; Gates, C. M.; Singer, S. C.; Davis, A. M.; Tansik, R. L.;
Mattheakis, L. C.; Boytos, C. M.; Schatz, P. J.; Baccanari, D. P.; Wrighton,
N. C.; Barrett, R. W.; Dower, W. J. Science 1997, 276, 1696-1699. (e)
Qureshi, S. A.; Kim, A. M.; Konteatis, Z.; Biazzo, D. E.; Motamedi, H.;
Rodrigues, R.; Boice, J. A.; Calaycay, J. R.; Bednarek, M. A.; Griffin, P.;
Gao, Y.-D.; Chapman, K.; Mark, D. F. Proc. Natl. Acad. Sci. U.S.A. 1999,
12, 12156-12161. (f) Goldberg, J.; Jin, Q.; Ambroise, Y.; Satoh, S.;
Desharnais, J.; Capps, K.; Boger, D. L. J. Am. Chem. Soc. 2002, 124, 544-
555. (g) Tian, S.-S.; Lamb, P.; King, A. G.; Miller, S. G.; Kessler, L.;
Luengo, J. I.; Averill, L.; Johnson, R. K.; Gleason, J. G.; Pelus, L. M.;
Dillon, S. B.; Rosen, J. Science 1998, 281, 257-259. (h) Pattarawarapan,
M.; Reyes, S.; Xia, Z.; Zaccaro, M. C.; Saragovi, H. U.; Burgess, K. J.
Med. Chem. 2003, 46, 3565-3567.
(8) (a) Bouvier, M. Nat. ReV. Neurosci. 2001, 2, 274-286. (b) George, S. R.;
O’Dowd, B. F.; Lee, S. P. Nat. ReV. Drug DiscoVery 2002, 1, 808-820.
(c) Park, P. S.; Filipek, S.; Wells, J. W.; Palczewski, K. Biochemistry 2004,
43, 15643-15656. (d) Prinster, S. C.; Hague, C.; Hall, R. A. Pharmacol.
ReV. 2005, 57, 289-298. (e) Bulenger, S.; Marullo, S.; Bouvier, M. Trends
Pharmacol. Sci. 2005, 26, 131-137.
(9) For early examples of carbohydrate-based multivalent inhibitors (glyco-
polymers and glycodendrimers) of influenza virus hemagglutinin binding
to host cells, see: (a) Spaltenstein, A.; Whitesides, G. M. J. Am. Chem.
Soc. 1991, 113, 686-687. (b) Sabesan, S.; Duus, J. O.; Domaille, P.; Kelm,
S.; Paulson, J. C. J. Am. Chem. Soc. 1991, 113, 5865-5866. (c) Gamian,
A.; Chomik, M.; Laferriere, C. A.; Roy, R. Can. J. Microbiol. 1991, 37,
233-237. (d) Kingery-Wood, J. E.; Williams, K. W.; Sigal, G. B.;
Whitesides, G. M. J. Am. Chem. Soc. 1992, 114, 7303-7305. (e) Sparks,
M. A.; Williams, K. W.; Whitesides, G. M. J. Med. Chem. 1993, 36, 778-
783. (f) Roy, R.; Pon, R. A.; Tropper, F. D.; Andersson, F. O. J. Chem.
Soc., Chem. Commun. 1993, 264-265. (g) Roy, R.; Zanini, D.; Meunier,
S. J.; Romanowska, A. J. Chem. Soc., Chem. Commun. 1993, 1869-1872.
(10) Selected recent examples of multivalent inhibitors: (a) Dimick, S. M.;
Powell, S. C.; McMahon, S. A.; Moothoo, D. N.; Naismith, J. H.; Toone,
E. J. J. Am. Chem. Soc. 1999, 121, 10286-10296. (b) Kitov, P. I.;
Sadowska, J. M.; Mulvey, G.; Armstrong, G. D.; Ling, H.; Pannu, N. S.;
Read, R. J.; Bundle, D. R. Nature 2000, 403, 669-672. (c) Merritt, E. A.;
Zhang, Z.; Pickens, J. C.; Ahn, M.; Hol, W. G.; Fan, E. J. Am. Chem. Soc.
2002, 124, 8818-8824. (d) Gestwicki, J. E.; Cairo, C. W.; Strong, L. E.;
Oetjen, K. A.; Kiessling, L. L. J. Am. Chem. Soc. 2002, 124, 14922-
14933; Gradl, S. N.; Felix, J. P.; Isacoff, E. Y.; Garcia, M. L.; Trauner, D.
J. Am. Chem. Soc. 2003, 125, 12668-12669. (e) Thumshirn, G.; Hersel,
U.; Goodman, S. L.; Kessler, H. Chem. Eur. J. 2003, 9, 2717-2725. (f)
Boturyn, D.; Coll, J.-L.; Garanger, E.; Favrot, M.-C.; Dumy, P. J. Am.
Chem. Soc. 2004, 126, 5730-5739. (g) Wolfenden, M. L.; Cloninger, M.
J. J. Am. Chem. Soc. 2005, 127, 12168-12169. (h) Arosio, D.; Fontanella,
M.; Baldini, L.; Mauri, L.; Bernardi, A.; Casnati, A.; Sansone, F.; Ungaro,
R. J. Am. Chem. Soc. 2005, 127, 3660-3661.
(11) For early examples of bivalent opioid ligands, see: (a) Portoghese, P. S. J.
Med. Chem. 2001; 44, 2259-2269. (b) Erez; M.; Takemori, A. E.;
Portoghese, P. S. J. Med. Chem. 1982, 25, 847-849; Portoghese, P. S.;
Larson, D. L.; Sayre, L. M.; Yim, C. B. (c) Ronsisvalle, G.; Tam, S. W.;
Takemori, A. E. J. Med. Chem. 1986, 29, 1855-1861.
(12) Choi, S. K. Synthetic MultiValent Molecules: Concepts and Biochemical
Applications; Wiley-Interscience: Hoboken, NJ, 2004.
(13) For reviews, see: (a) Mammen, M.; Choi, S.-K.; Whitesides, G. M. Angew.
Chem., Int. Ed. 1998, 37, 2755-2794. (b) Kiessling, L. L.; Gestwicki, J.
E.; Strong, L. E. Curr. Opin. Chem. Biol. 2000, 4, 696-703. (c) Kiessling,
L. L.; Pohl, N. L. Chem. Biol. 1996, 3, 71-77.
(14) For review, see: Kiessling, L. L.; Gestwicki, J. E.; Strong, L. E.; Angew.
Chem., Int. Ed. 2006, 45, 2348-2368.
(18) (a) Bodmer, J. L.; Schneider, P.; Tschopp, J. Trends Biochem. Sci. 2002,
27, 19-26. (b) Locksley, R. M.; Killen, N.; Lenardo, M. J. Cell 2001,
104, 487-501.
(19) Although the conventional view is that members with the TNF-R family
exist as independently diffusing monomeric species and trimerize upon
ligand binding, some evidence suggests however, that several of these
receptors may exist as pre-assembled oligomers prior to ligand binding. It
has been shown that a conserved domain in the extracellular region of TNF-
R, Trail receptor 1, and CD40, distal from the domain that interacts with
the ligand, serves as a preligand-binding assembly domain (PLAD). Chan,
F. K.; Chun, H. J.; Zheng, L.; Siegel, R. M.; Bui, K. L.; Lenardo, M. J.
Science 2000, 288, 2351-2354.
(20) For reviews, see: (a) van Kooten, C.; Banchereau, J. J. Leuk. Biol. 2000,
67, 1-17. (b) Diehl, L.; Den, Boer, A. T.; van der Voort, E. I.; Melief, C.
J.; Offringa, R.; Toes, R. E. J. Mol. Med. 2000, 78, 363-371. (c) Grewal,
I. S.; Flavell, R. A. Annu. ReV. Immunol. 1998, 16, 111-135. (d) Van
Kooten, C.; Banchereau, J. AdV. Immunol. 1996, 61, 1-77. (e) Grewal, I.
S.; Flavell, R.A. Immunol. Today 1996, 17, 410-414. (f) Noelle, R. J.
Immunity 1996, 4, 415-419.
(21) (a) Howard, L. M.; Miga, A. J.; Vanderlugt, C. L.; Dal Canto, M. C.; Laman,
J. D.; Noelle, R. J.; Miller, S. D. J. Clin. InVest. 1999, 103, 281-290. (b)
Kirk, A. D.; et al. Nat. Med. 1999, 5, 686-693.
(22) Phase I/II studies in humans with lupus nephritis or multiple sclerosis
revealed significant immunomodulatory effects of anti-CD40L monoclonal
antibody treatment. However, because of some thromboembolic complica-
tions in early clinical trials, further human studies with anti-CD40L antibody
have been temporary halted, and future clinical use is uncertain. See
comment in: Couzin, J. Science 2005, 307, 1712-1715.
(15) Selected examples: (a) Gestwicki, J. E.; Strong, L. E.; Kiessling, L. L.
Chem. Biol. 2000, 7, 583-591. (b) Gestwicki, J. E.; Kiessling, L. L. Nature
2002, 415, 81-84. (c) Kramer, R. H.; Karpen, J. W. Nature 1998, 395,
710-713. (d) Pattarawarapan, M.; Reyes, S.; Xia, Z.; Zaccaro, M. C.;
Saragovi, H. U.; Burgess, K. J. Med. Chem. 2003, 46, 3565-3567. (e)
Puffer, E. B.; Pontrello, J. K.; Hollenbeck, J. J.; Kink, J. A.; Kiessling, L.
L. ACS Chem. Biol. 2007, 2, 252-262.
(23) (a) Diehl, L.; den Boer, A. T.; Schoenberger, S. P.; van der Voort, E. I.;
Schumacher, T. N.; Melief, C. J.; Offringa, R.; Toes, R. E. Nat. Med. 1999,
5, 774-779. (b) van Mierlo, G. J.; den Boer, A. T.; Medema, J. P.; van
der Voort, E. I.; Fransen, M. F.; Offringa, R.; Melief, C. J.; Toes, R. E.
Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5561-5566.
(24) Karpusas, M.; Lucci, J.; Ferrant, J.; Benjamin, C.; Taylor, F. R.; Strauch,
K.; Garber, E.; Hsu, Y. M. Structure 2001, 9, 421-429.
(16) Reviews on small molecule cytokine mimetics: (a) Whitty, A.; Borysenko,
C. W. Chem. Biol. 1999, 6, R107-R118.
(25) Singh, J.; Garber, E.; Van Vlijmen, H.; Karpusas, M.; Hsu, Y. M.; Zheng,
Z.; Naismith, J.; Thomas, D. Protein Sci. 1998, 7, 1124-1135.
9
J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007 13481

A R T I C L E S
Trouche et al.
CD40-binding peptides with their linkers were prepared on solid
support starting from a 2-chlorotrityl chloride resin as illustrated
in Scheme 2 for the Lys-Gly-Tyr-Tyr-Ahx sequence. Peptide
assembly on core structures was generally performed with
BOP29 as the coupling agent in DMF for 24 h, after which the
crude fully protected multimeric construct recovered by filtration
after precipitation with a saturated NaHCO3 solution was washed
extensively with ethyl acetate and dried under high vacuum.
All protecting groups were removed by treatment with trifluo-
roacetic acid (TFA) to afford crude ligands. Finally, ligands
were purified by C18 RP-HPLC and recovered in overall yields
ranging from 6 to 37%. It is noteworthy that the synthesis of
A-1 was routinely achieved on >100 mg scale with an overall
yield exceeding 25%. In an optimized process, the linear peptide
precursor of scaffold A, H-(Lys(Boc)-D-Ala)3-OH was as-
sembled by solid-phase synthesis on a 2-chlorotrityl chloride
resin. All ligands were identified by matrix-assisted laser
desorption/ionization mass spectrometry (MALDI-MS) and their
homogeneity was assessed by C18 RP-HPLC with purity of all
peptides determined to be >99%, (see Supporting Information
for detailed experimental procedures and characterization of all
investigated ligands).
Variations in Shape and Size of the Central Core. The
nature of a scaffold for polyvalent display can significantly
impact on the biological activity of synthetic multivalent ligands.
Features such as geometry, valency, and rigidity must be
carefully tailored to ensure effective distribution of appended
recognition patterns. This has been recognized in a number of
multivalent systems with a variety of poly(oligo-)valent scaffolds
(e.g., small synthetic scaffolds, dendrimers, polymers, liposomes,
proteins, and viral particles).9,10,12-14 Our previous finding that
CD40L mimetics based on trimeric architectures A and B but
not branched E exhibit effector functions in several cellular
assays (maturation of D1 dendritic cell line,30 apopotosis of B
lymphoma cells31) and promote control of parasitemia during
experimental T. cruzi infections in mice tend to suggest that
the rigidity of the central core is one determinant of the activity
of CD40L mimetics.26a,27 This can be rationalized in terms of
the minimization of the conformational entropic cost.13a It also
suggests that effective trivalent display of a given CD40-binding
motif with geometry and distances that match those in the
cognate protein trimer imposes constraints on the overall
geometry and dimensions of the core + linker element. To
ascertain this trend, we have now considered an enlarged set of
trimeric architectures falling into three categories (Chart 1):
macrocyclic (A-D), branched (E,F), and heterocyclic (G-I),
varying in size (A ≈ C > B > I > G > E > H > F, based on
fully extended structures) and rigidity. Partially N-methylated
cyclo-D,L-R-peptide C which displays a smaller number of
H-bond donors than A was intended to modulate the possible
self-assembling properties of the resulting ligands.32 A macro-
cyclic ligand with receptor-binding motifs attached to the upper
rim of a calix[6]triamine core structure (type D)33 was designed
Figure 1. X-ray structure of CD40L homotrimer (PDB 1ALY)22 viewed
down the C3 axis and schematic representation of trimeric architecture used
as CD40L mimetics.
rational design and synthesis of a first-generation of CD40L
mimetics.26,27 We have shown that rigid trimeric scaffolds can
serve to distribute a CD40-binding motif with geometry and
distances that could match that of the natural CD40L protein
(Figure 1). A short loop sequence from the surface of CD40L
encompassing three “hot-spot” residues (Lys143, Tyr145, and
Tyr146 in the AA′′ loop) critical for binding to CD40, was
initially chosen as a CD40-binding motif.
The gain in avidity that could result from a multivalent
interaction with CD40 molecules was expected to overcome the
relatively low binding energy of the short tetrapeptide sequence.
Two trimeric molecules A-1 and B-1 built on the rigid cyclo-
(Lys-D-Ala)3 and cyclo-(â3-HLys) peptide rings A and B,
respectively (see Chart 1), with amino hexanoic acid (Ahx)
residues as additional spacer arms were found to bind in a
specific manner to CD40 and to compete with the binding of
CD40L homotrimers.26,27 In various cell-based assays, A-1 and
B-1 displayed some effector functions of the much larger
recombinant CD40L homotrimers, i.e., induction of apoptosis
of B lymphomas and maturation of DCs. Both CD40L mimetics
were shown recently to be effective in ViVo and to promote the
control of Trypanosoma cruzi (i.e., the etiological agent of the
Chagas’disease) infection in an experimental mouse model by
overcoming the immunosuppression usually induced by the
parasite.28 However, the molecular determinants of the design
of CD40L mimetics (hereby referred to as “miniCD40Ls”) with
optimal CD40 binding properties and effector functions are not
yet fully delineated. To assess the contribution of each individual
component of the trivalent architecture, we have now undertaken
a detailed study by systematically varying the shape and the
valency of the core structure, the geometry and the length of
the spacer arm as well as the nature of the CD40 binding motif.
Results and Discussion
Ligand Synthesis. All CD40L mimetics described in this
study (except E-1, see Supporting Information) were synthesized
according to Scheme 1 by fragment coupling in solution of a
fully protected peptide fragment encompassing the receptor
binding motif and a spacer arm, to the corresponding amino-
functionalized core structure A-0 to L-0 (R ) H). The protected
(26) (a) Fournel, S.; Wieckowski, S.; Sun, W.; Trouche, N.; Dumortier, H.;
Bianco, A.; Chaloin, O.; Habib, M.; Peter, J.-C.; Schneider, P.; Vray, B.;
Toes, R. E.; Offringa, R.; Melief, C. J. M.; Hoebeke, J.; Guichard, G. Nat.
Chem. Biol. 2005, 1, 377-382. (b) Bianco, A.; Fournel, S.; Wieckowski,
S.; Hoebeke, J.; Guichard, G. Org. Biomol. Chem. 2006, 4, 1461-1463.
(27) (a) Hymowitz, S. G.; Ashkenazi, A. Nat. Chem. Biol. 2005, 1, 353-354.
(b) Gibson, S. E.; Castaldi, M. P. Angew. Chem., Int. Ed. 2006, 45, 4718-
4720.
(28) Habib, M.; Chamekh, M.; Noval, Rivas, M.; Wieckowski, S;, Sun, W.;
Bianco, A.; Trouche, N.; Chaloin, O.; Dumortier, H.; Goldman, M.;
Guichard, G.; Fournel, S.; Vray, B. J. Immunol. 2007, 178, 6700-6704.
(29) Castro, B.; Dormoy, J. R.; Evin, G.; Selve, C. Tetrahedron Lett. 1975,
1219-1222.
(30) Schuurhuis, D. H.; Laban, S.; Toes, R. E. M.; Ricciardi-Castagnoli, P.;
Kleijmeer, M. J.; van der Voort, E. I. H.; Rea, D.; Offringa, O.; Geuze, H.
J.; Melief, C. J. M.; Ossendorp, F. J. Exp. Med. 2000, 192, 145-150.
(31) B lymphoma cells undergo apoptosis upon engagement of CD40 by
recombinant CD40L. See: Frisan, T.; Donati, D.; Cervenak, L.; Wilson,
J.; Masucci, M. G.; Bejarano, M. T. Int. J. Cancer 1999, 83, 772-779.
(32) Bong, D. T.; Clark, T. D.; Granja, J. R.; Ghadiri, M. R. Angew. Chem.,
Int. Ed. 2001, 40, 988-1011 and refs therein.
9
13482 J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007

Relationship between Structure and Effector Function
A R T I C L E S
Chart 1. Trimeric Architectures and Ligands Based on Scaffolds A-I
Scheme 1. General Synthetic Approach to Trimeric CD40L
Mimetics
Screening of molecules A-1 to I-1 at 250 nM in a SPR
binding assay with immobilized rshCD40:mIg confirmed mac-
rocyclic A-1 and B-1 and newly identified heterocyclic G-1 as
the best CD40 binders in the series in terms of binding capacity
and interaction stability (Figures 2a, S2, and S3). A similar trend
was observed when analyzing kinetic binding parameters
calculated from SPR sensorgrams acquired for various concen-
trations of miniCD40Ls (ka versus kd plots with isoaffinity lines
are shown in Figure 2b). Kinetic values were all derived from
the 1:1 interaction Langmuir model (see Experimental Section
for description of equations used and Figure S3 for raw SPR
data).35,36 Whereas the majority of CD40L/CD40 interactions
showed binding kinetics suitable for determination of the kinetic
to investigate the effect of a more axial (assuming a cone
conformation) distribution of the CD40-binding motif.34
The corresponding ligands A-1 to I-1 obtained by appending
the pentapeptide sequence Lys-Gly-Tyr-Tyr-Ahx (1) have been
evaluated for binding to rshCD40:mIg [a bivalent form of CD40
consisting of recombinant human CD40 fused to a mouse IgG
Fc domain (see Figure S1 in Supporting Information)] using
surface plasmon resonance (SPR) and for their capacity to
induce apoptosis of the human B lymphoma cell line BL41
(Figures 2 and 3).
(34) For use of calixarenes as multivalent templates and for the recognition of
protein surfaces, see: (a) Baldini, L.; Casnati, A.; Sansone, F.; Ungaro, R.
Chem. Soc. ReV. 2007, 36, 254-266. (b) Reference 9g. (c) Park, H. S.;
Lin Q.; Hamilton, A. D. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5105-
5109. (d) Blaskovich, M. A.; Lin, Q.; Delarue, F. L.; Sun, J.; Park, H. S.;
Coppola, D.; Hamilton, A. D.; Sebti, S. M. Nat. Biotechnol. 2000, 18,
1065-1070. (e) Park, H. S.; Lin, Q.; Hamilton, A. D. J. Am. Chem. Soc.
1999; 121, 8-13.
(33) Gac, S.; Zeng, X.; Reinaud, O.; Jabin, I. J. Org. Chem. 2005, 70, 1204-
1210.
9
J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007 13483

A R T I C L E S
Trouche et al.
Scheme 2. Synthesis of the Protected Pentapeptide
Boc-Lys(Boc)-Gly-Tyr(OtBu)-Tyr(OtBu)-Ahx-OH (1) (See
Supporting Information for Details)a
aConditions: (a) Fmoc-6-aminocaproic acid, DIEA, CH2Cl2; (b) 25%
piperidine/DMF; (c) SPPS: Fmoc-Xaa-OH, BOP, HOBt, DIEA, DMF; (d)
HFIP/CH2Cl2 (60:40 v/v).
Figure 3. Comparison of compounds A-1 to I-1 for their ability to inhibit
binding of soluble human CD40L (hCD40L:mCD8) to CD40 and to induce
apoptosis of the human B lymphoma cell BL41. (a) Schematic representation
of the SPR inhibition experiment. (b) Gray bars: values of the inhibition
experiments. Percentages of inhibition at 1 µM of mimetics are represented.
They were derived from the slope of the CD40Ls mixed to the mimetics
response signal substracted from the response signal of the mimetics alone.
Results are representative of two independent experiments. ND: not
determinable because of aggregation; black line: induction of apoptosis of
human lymphoma cell line BL41 after 16 h of treatment. Values compared
are from percentages of specific apoptosis (see Experimental Section)
obtained at 10 µM of the different mimetics. Results are expressed as means
( SEM of three independent experiments.
3a. At 1 µM, only A-1 and B-1 were found to significantly
inhibit the binding of soluble recombinant hCD40L:mCD8 to
rshCD40:mIg fusion protein (Figure 3b). Molecule G-1 was only
a weak inhibitor. In parallel, all ligands were screened in a cell-
based assay (Figure 3b). Their capacity to induce apoptosis of
CD40-positive human Burkitt B lymphoma cells was measured
by the decrease of the mitochondrial membrane potential (∆ψm)
using DiOC6(3) dye as previously described.26a,37 A-1 and B-1
induced a high level of apoptosis at 10 µM, whereas all
remaining ligands were only weakly active (C-1, D-1, G-1) or
ineffective (E-1, F-1, H-1, I-1).
Ligands A-1 and B-1 differ from other tested ligands by the
shape of their central core. The general structural features of
Figure 2. Comparison of compounds A-1 to I-1 for binding to CD40 by
surface plasmon resonance (SPR). (a) Screening of the molecules in a
comparative assay at 250 nM. Binding and stability report points for each
miniCD40L were determined from the association and dissociation phases
on the sensorgrams at 235 and 250 s, respectively (see Figures S2 and S3).
Responses were baseline substracted and adjusted for the refractive index
(RI) and the molecular weight of each molecule. Molecules are thus
compared in a plot reflecting the extent of binding and stability of the
interaction with CD40. The best ligands are encircled. (b) Graphical
summary of the data generated from different concentrations of the mimetics.
Kinetics values were all derived from the 1:1 Langmuir model. All results
are representative of at least two independent experiments.
(35) In the case of A-1, SPR data fit better with a trivalent model than with the
classical “Langmuir 1:1” monovalent one [Chi square (ø2) values which
indicate that the accuracy of the fits are 0.43 and 0.85, respectively],
suggesting that this interaction is of one miniCD40L for three CD40
receptors stoichiometry. Wieckowski, S.; Trouche, N.; Chaloin, O.;
Guichard, G.; Fournel, S.; Hoebeke, J. Biochemistry, 2007, 46, 3482-
3493.
(36) Because of substantial but unknown contribution of avidity effects to the
strength of binding that results from CD40 being immobilized to high
density on the CM5 sensor chip, the affinities measured in our SPR binding
assay represent apparent binding affinities rather than true KD values. For
an interesting discussion on the determination of true KD values for TNF
family ligands binding to their cognate receptors and introduction of a
“solution-phase Biacore binding assay”, see : Day, E. S.; Cachero, T. G.;
Qian, F.; Sun, Y.; Wen, D.; Pelletier, M.; Hsu, Y. M.; Whitty, A.
Biochemistry 2005, 44, 1919-1931.
rate constants, SPR analysis was hampered for ligands C-1 and
D-1 (partially N-methylated cyclo-D,L-hexapeptide and calix-
[6]arene cores, respectively) because of the formation of
aggregates in the conditions of the assay. Trimeric ligands were
also compared in an inhibition SPR assay as shown in Figure
(37) Quantification of A-1-mediated apoptosis in Burkitt lymphoma cells by
an annexin V-propidium iodide colabeling technique (which enables the
recognition of early and late apoptotic cells) and by terminal deoxynucle-
otide transferase dUTP nick-end labeling (TUNEL) gave comparable results
(ref 26a and data not shown).
9
13484 J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007

Relationship between Structure and Effector Function
A R T I C L E S
Chart 2
cyclo-(Xaa-D-Xbb)n and cyclo-(â3-HXaa)n peptides have been
extensively investigated both in solution by NMR spectroscopy
and in the solid state by XRD and PXRD.32,38-41 In both series,
the peptide backbone generally adopts a flat-ring conformation
with side chains occupying equatorial positions along the ring’s
edge.42 The results obtained by SPR and in cell-based assays
with compounds A-1 to I-1 suggest in first approximation that
a radial distribution of the CD40-binding motif is preferred.
Among all ligands tested, A-1 and B-1 are those potentially
providing the longest distance between the center of the trivalent
system and the NH connecting the Ahx linker. Thus, our data
also support the view that the distance between the center of
the trivalent system and the NH of the flexible linker connecting
the CD40 binding motif should be above a minimum length
(vide infra, section on modulation of the linker length).
Comparison to Mono-, Di-, and Tetravalent Systems. In
the next series of experiments, the importance of a C3-symmetric
trivalent display was assessed by systematically varying the
number of CD40 recognition elements appended to the cyclic
D,L-peptide core (Chart 2).
We found that removing one CD40-binding motif while
maintaining a radial distribution (A-1 f J-1) resulted in a strong
decrease of the CD40-binding capacity (Figure 4) and biological
effects, i.e., apoptosis of human B lymphoma cells (Figure 5a)
and maturation of mouse dendritic cells D1 (Figure 5b).30 Not
surprisingly, ligand K-1 with only one CD40-binding motif
appended to a cyclic D,L-hexapeptide core as well as the
corresponding H-Lys-Gly-Tyr-Tyr-Ahx-OH pentapeptide26a was
inactive.
Figure 4. Influence of the valency of the ligands on their capacity to bind
CD40. Direct binding of A-1, J-1, and K-1 to CD40 as shown by SPR. All
data were obtained for mimetics concentration from 100 to 800 nM.
D-Ala)4,40 L-1, was synthesized and tested for CD40 binding
and B lymphoma apoptosis (Figure 6). However, L-1 turned
out to be less soluble than A-1. Although SPR binding data
obtained at low concentration between 6 nM and 100 nM may
be indicative of specific binding, the C4-symmetric ligand L-1
was found to extensively aggregate above 200 nM under SPR
conditions (see Supporting Information, Figure S4). Moreover,
L-1 was not able to induce significant apoptosis of B lymphomas
at 10 µM. Overall these results together with those obtained
with J-1 and K-1 support the view that C3-symmetry of the
core structure and radial distribution of appended peptides are
both determinants of the activity of CD40L mimetics.
Modulation of the Linker Portion. As discussed previously,
the choice of an appropriate linker to connect the binding
moieties in a multivalent system can be a determinant. Rather
than a rigid linker difficult to optimize, we opted for rather
flexible oligomethylene or oligo(ethylene glycol) type spacers
that we felt to be more appropriate to sample conformational
space. In the crystal structure of the CD40L homotrimer, the
distance between the R carbon (RC) of each hot-spot residue,
Lys143 to Tyr146, and the centroid of the equilateral triangle
formed by the RC of the three cognate residues in the trimer is
between 17 and 22 Å. The R-amino group of the next residue
in the sequence (Thr147) is positioned ca. 16 Å apart from the
center of the trimer. When designing ligands based on scaffold
A, the Ahx linker was originally selected to approach this
distance. If one assumes that the Lys(Ahx) side chain on the
D,L-cyclohexapeptide core A adopts a fully extended conforma-
tion (∼14.9 Å) as in the crystal structure of a related Ahx-Ahx
peptide segment43 and that the diameter of the core cyclopeptide
is ∼7.5-8 Å [based on the crystal structures38,39 of related cyclo-
We also investigated higher-order multimeric structures. A
C4-symmetric tetrameric analogue of A-1 based on cyclo-(Lys-
(38) For studies on partially N-alkylated C3-symmetric cyclic D,L-hexapeptide
structures, see: (a) Tomasi, L.; Lorenzi, G. P. HelV. Chim. Acta 1987, 70,
1012-1016. (b) Saviano, M.; Lombardi, A.; Pedone, C.; Di, Blasio, B.;
Sun, X. C.; Lorenzi, G. P. J. Incl. Phenom. 1994, 18, 27-36. (c) Sun, X.
C.; Lorenzi, G. P. HelV. Chim. Acta 1994, 77, 1520-1526.
(39) For crystal structures of cyclo-(L-Val-D-Val)3 and cyclo-(L-Phe-D-Phe)3
hexapeptides, see: Pavone, V.; Benedetti, E.; Di Blasio, B.; Lombardi,
A.; Pedone, C.; Tomasich, L.; Lorenzi, G. Biopolymers 1989, 28, 215-
223.
(40) For studies on cyclic D,L-octapeptides, see: (a) Ghadiri, M. R.; Granja, J.
R.; Milligan, R. A.; McRee, D. E.; Khazanovich, N. Nature 1993, 366,
324-327. (b) Khazanovich, N.; Granja, J. R.; McRee, D. E.; Milligan, R.
A.; Ghadiri, M. R. J. Am. Chem. Soc. 1994, 116, 6011-6012. (c)
Hartgerink, J. D.; Granja, J. R.; Milligan, R. A.; Ghadiri, M. R. J. Am.
Chem. Soc. 1996, 118, 43-50.
(41) (a) For the NMR structure determination of cyclo-(â3-HGlu)3 in D2O, see:
Gademann, K.; Seebach, D. HelV. Chim. Acta 1999, 82, 957-962. (b) For
XRD studies of three stereoisomers of cyclo-(â3-HAla)4, see: Seebach,
D.; Matthews, J. L.; Meden, A.; Wessels, T.; Baerlocher, C.; McCusker,
L. B. HelV. Chim. Acta 1997, 80, 173-182.
(42) Although cyclo-L,D-peptides show propensity to self-assemble into H-bonded
tubular stacks in the solid state and in nonpolar solvents, there is no
indication that A-1 aggregates in aqueous solution.
9
J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007 13485

A R T I C L E S
Trouche et al.
Chart 3
pared a series of B-1 variants (Chart 3) with oligomethylene
[-(CH2)n-] tethers of length varying from n ) 3 to 7 (Chart
3).
Compound B-2, in which the oligomethylene chain is reduced
by two carbons (n ) 3), is no longer able to induce apoptosis
of human B lymphoma BL41 cells (Figure 7a) and exhibits
drastically reduced binding to CD40 (Figure 7b). Significant
binding to CD40 is retained when the length of the spacer is
decreased by only one carbon (B-3), although its association
rate is lower than that for B-1 (Figure 7b), but the capacity to
induce apoptosis is not restored (Figures 7a and S5 for an
annexin V-propidium iodide colabeling analysis). In contrast,
analogues of B-1 with one and two additional methylene groups
(B-4, n ) 6, and B-5, n ) 7, respectively) induced significantly
higher levels of apoptosis compared to B-1 (Figures 7a and S5).
The finding that the level of apoptosis is considerably enhanced
as the spacer length increased from a three- to an optimal six-
carbon chain, and remains constant upon an additional increment
in chain length from n ) 6 to n ) 7, is consistent with other
studies.45,46 It has been shown previously, that in the case of
flexible tethers, the effective molarity and potency are the
highest for an optimal length and are only moderately affected
by increments in chain length beyond this value.41 Unfortunately,
aggregation of B-4 and B-5 under SPR conditions precluded
their evaluation for binding to CD40. Further increments in chain
length from the 7-carbon chain to the 10-carbon chain led to a
compound (B-6) with even lower solubility that could not be
evaluated in the two assays. To increase the water solubility of
the ligands, we also explored modification of the spacer by
Figure 5. Influence of the valency on effector functions. (a) Induction of
apoptosis of human lymphoma cell line BL41. Percentage of specific
apoptosis measured by reduction of DiOC6(3) dye uptake (see Experimental
Section) induced by the different mimetics at the indicated concentrations
after 16 h of incubation. Results are expressed as means ( SD of two
independent experiments. (b) Expression of the maturation markers (CD80,
CD54, CD86, and CD40) by the mouse dendritic cell line D1 as measured
by flow cytometry after 24 h of incubation with the different mimetics at
the indicated concentrations; white histogram: medium control.
Figure 6. Tetrameric versus trimeric ligands. Percentage of specific
apoptosis induced by L-1 and A-1 at the indicated concentrations after 16
h of incubation detected by reduction of DiOC6(3) dye uptake (see
Experimental Section). Cells used in the assay are human lymphoma cells
BL41 (black) which express CD40, Jurkat cells (white) which do not express
CD40, and B lymphoma Raji cells (gray bars) which express high level of
CD40. Results are expressed as means ( SEM of three independent
experiments. Molecular structure of the tetravalent mimetic is shown in
the inset.
(L-Val-D-Val)3, cyclo-(L-Phe-D-Phe)3, and cyclo-(N-MeLeu-D-
Leu)3], one can approximate the distance between the center of
the core and the NH of the Ahx residue in A-1 to a maximum
value of ca. 18-19 Å. The corresponding value in B-1 should
be smaller by ca. 1.5 Å because of the smaller diameter of the
12-membered (â3-HLys)3 core structure.41,44
(44) The crystal structures of a number of isosteric cyclic tri(â-hydroxy
alkanoates) have been solved. In crystal, these compounds generally adopt
a flat, nearly equilateral triangular shape with all carbonyl groups pointing
in one direction, very similar to the model calculated for cyclo-(â3-HGlu)3
(ref 36a) with a distance between the center of gravity of the triangle and
the âC ranging from 2.62 to 2.75 Å. See: (a) Seebach, D.; Muller, H.-M.;
Burger, H. M.; Plattner, D. A. Angew. Chem., Int. Ed. Engl. 1992, 31,
434. (b) Seebach, D.; Hoffmann, T.; Kuhnle, F. N. M.; Lengweiler, U. D.
HelV. Chim. Acta 1994, 77, 2007-2034.
To investigate how the length and the nature of the linker
can modulate the activity of CD40L mimetics, we have pre-
(45) Sridhar, J.; Wei, Z.-L.; Nowak, I.; Lewin, N. E.; Ayres, J. A.; Pearce, L.
V.; Blumberg, P. M.; Kozikowski, A. P. J. Med. Chem. 2003, 46, 4196-
4204.
(43) Kasai, N.; Yamanaka, T.; Miki, K.; Tanaka, N.; Post, B.; Morawetz, H.
(46) Krishnamurthy, V. M.; Semetey, V.; Bracher, P. J.; Shen, N.; Whitesides,
G. M. J. Am. Chem. Soc. 2007, 129, 1312-1320.
Acta Crystallogr., Sect. B 1981, 37, 1628-1630.
9
13486 J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007

Relationship between Structure and Effector Function
A R T I C L E S
Figure 8. Reverse â-turn type conformation of the Lys143-Gly-Tyr-Tyr146
CD40-binding motif in the crystal structure of CD40L.24 The values of φ
and ψ angles for residues i+1 (Gly144) and i+2 (Tyr145) are indicated.
Chart 4
Figure 7. Influence of the length and nature of the linker. (a) Induction of
apoptosis of human B lymphoma BL41 cells measured by reduction of
DiOC6(3) dye uptake (see Experimental Section). Percentages of specific
apoptosis induced by the different mimetics at the indicated concentrations
after 16 h of incubation. Results are expressed as means ( SD of two
independent experiments. (b) Direct binding to CD40 of B-1 analogues at
200 nM as shown by SPR. Sensorgrams for B-4 and B-5 are not shown
because of aggregation in the SPR conditions. All results are representative
of at least two independent experiments.
substituting the saturated chain in B-5 with a oligo(ethylene
glycol) chain (EGn) of similar length. This replacement resulted
in a ligand (B-7) with dramatically reduced binding to CD40
and no apoptotic activity. No improvement was observed upon
increasing further the number of oligo(ethylene glycol) units
(e.g., B-8).
Overall the data obtained with the â-peptide core structure
B suggest that n-amino-alkanoic acids as tethers are superior
to corresponding oligoethylene glycol amino acids to distribute
the CD40-binding motif and that a minimum carbon chain length
(e.g., n ) 5-7) is required to ensure binding to CD40 and
effector functions. This result may partially explain why ligands
such as F-1 and H-1 based on TREN and isocyanurate fail both
in binding significantly to CD40 and in inducing apoptosis of
B lymphoma cells. Even by assuming a fully extended confor-
mation for the central core and the Ahx residue allowing a radial
distribution of the CD40 binding motif, the distance between
the center of the trimeric scaffold and the NH of the Ahx residue
in F-1 and H-1 does not exceed 13 Å which is below the optimal
16 Å value.
The strong dichotomy observed between oligomethylene and
oligo(ethylene glycol) tethers was not expected, and its origin
remains unclear. Such a differential behavior has been docu-
mented previously in other systems including bivalent protein
kinase C ligands45 and might result from significant conforma-
tional differences and/or from a different propensity to interact
with the surface of proteins.
binding of each amino acid in the Lys-Gly-Tyr-Tyr sequence,
we have previously prepared analogues of B-1 singly substituted
at each position by an alanyl (or glycyl) residue.26a We found
that Lys and both Tyr residues are critical to binding and effector
functions, consistent with data obtained from site-directed
mutagenesis studies using recombinant CD40L molecules.47
However, the glycyl residue is quite permissive and can be
substituted by Ala without major decrease in binding to CD40
and with no strong reduction in the apoptotic activity on
lymphoma cells.26a Examination of the geometry of the AA′′
loop in the crystal structure of CD40L reveals that the four
residues of the Lys-Gly-Tyr-Tyr CD40 binding motif located
(47) (a) Bajorath, J.; Chalupny, N. J.; Marken, J. S.; Siadak, A. W.; Skonier, J.;
Gordon, M.; Hollenbaugh, D.; Noelle, R. J.; Ochs, H. D.; Aruffo, A.
Biochemistry 1995, 34, 1833-1844. (b) Bajorath, J.; Marken, J. S.;
Chalupny, N. J.; Spoon, T. L.; Siadak, A. W.; Gordon, M.; Noelle, R. J.;
Hollenbaugh, D.; Aruffo, A. Biochemistry 1995, 34, 9884-9892.
MiniCD40Ls Based on Conformationally Restrained Re-
ceptor Binding Motifs. To delineate the contribution to CD40
9
J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007 13487

A R T I C L E S
Trouche et al.
Figure 10. Part of the CD40 binding surface showing the relative positions
of the putative CD40-binding sequences Lys143-Gly-Tyr-Tyr146 (AA′′ loop)
and Arg203-Leu-Leu-Ile-Arg207 (â-strand E) selected for the construction
of first- and second-generation CD40L mimetics.
Figure 9. Induction of apoptosis of human B lymphoma BL41 cells by
analogues of B-1 containing modifications in the CD40-binding motif.
Percentages of specific apoptosis are shown at the indicated concentrations
after 16 h of incubation. Results are expressed as means ( SEM of three
independent experiments.
Scheme 3 a
at the top of the loop adopt a â-turn type conformation centered
on Gly144 (i+1) and Tyr145 (i+2) residues. The corresponding
backbone torsion angle values φi+1, ψi+1, and φi+2 indicated in
Figure 8 are close to values of an ideal reverse type-II′ â-turn
[φi+1 ) +60° ((20°), ψi+1 ) -120° ((20°), and φi+2 ) -80°
((20°)]. However, the precise conformation of the loop upon
CD40 binding to CD40L remains to be elucidated.
To reduce the conformational space accessible to the glycyl
residue and to approach the geometry of a reverse turn segment,
we decided to introduce conformationally restricted amino acid
residues such as D-Pro and Aib (R-amino isobutyric acid) which
are known to facilitate â-turn conformations (Chart 4).48
Whereas favored (φ,ψ) angles for D-Pro (+60°,+30° and
+60°,-120°, respectively) are compatible with the requirements
of the i+1 residue in both type-I′ and type-II′ â-turns, the (φ,ψ)
space of Aib restricted to (+60°,+30°) and (-60°,-30°) values
is compatible with type-I and type-I′ â-turns. Interestingly,
substituting D-Pro for Gly (B-1 f B-9) does not compromise
binding to CD40 in SPR experiments. Analysis of kinetic data
using the classical Langmuir monovalent model and local fittings
revealed similar dissociation constants with KD of 120 and 195
nM for B-1 and B-9, respectively (see Supporting Information,
Figure S6).49 It is worth mentioning that the induction of
apoptosis on human lymphoma cells was increased by this muta-
tion, B-9 being significantly more potent than B-1 (Figure 9).
Preventing the possible formation of a reverse turn by
substituting L-Pro for Gly (f B-10) was detrimental for the
apoptotic properties. Similarly, the insertion of an Aib residue
at the i+1 position gave B-11 which was not active on CD40
positive human lymphoma cells. Overall, these data give some
hints about the bioactive conformation(s) of the Lys-Gly-Tyr-
Tyr CD40-binding segment in miniCD40Ls and suggest
a Reagents and conditions: a: Grubbs catalyst II; CH2Cl2, 50 °C, 7 days;
b: A-0 (0.3 equiv), BOP, DIEA, DMF, 24 h; c: TFA/H2O (95:5).
possibilities to further modulate/optimize effector functions
of miniCD40Ls using appropriate modifications centered on
the Gly residue. The insertion of pseudopeptide bonds between
i and i+1 residues to create local conformational perturba-
tions that could eventually modulate the effector functions of
miniCD40Ls B-1 and B-9 was also evaluated (Chart 4). Whereas
the introduction of a reduced peptide (methylene amino) bond
[denoted ψ(CH2-N) according to Spatola’s nomenclature
for pseudopeptide bonds]50 was found to be detrimental for
biological activity, analogues B-13 and B-14 with ethylene
(48) For reviews, see: (a) Venkatraman, J.; Shankaramma, S. C.; Balaram, P.
Chem. ReV. 2001, 101, 3131-3152. (b) Rose, G. D.; Gierasch, L. M.; Smith,
J. A. AdV. Protein Chem. 1985, 37, 1-110.
(49) It is noteworthy that similar results were obtained for analogous ligands
A-1 and A-9 (ref 35). In this case, KD calculated using the Langmuir 1:1
model and local fitting models were 121 and 183 nM, respectively. Analysis
of cooperativity using Hill plots suggested that the A-1/CD40 interaction
is cooperative (Hill coefficient h ) 2.9) and that this cooperativity is lost
upon replacing Gly by D-Pro (h ) 1.0).
9
13488 J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007

Relationship between Structure and Effector Function
A R T I C L E S
Figure 11. Comparison of miniCD40Ls with binding motifs derived from distinct CD40L hot-spot regions [e.g., Lys143-Gly-Tyr-Tyr146 from the AA′′ loop
(B-1), Arg203-Leu-Leu-Ile-Arg207 from â-strand E (B-17), and Gly199-Arg-Phe-Ile-Arg203 from DE loop (B-18)]. (a) Direct binding to CD40 at different
concentrations of mimetics as shown by SPR. Dashed lines represent experimental curves, straight lines represent fitting curves. All results are representative
of at least two independent experiments. (b) Induction of apoptosis of human B lymphoma cells BL41 and Jurkat cells (human CD40-negative cells) after
16 h of incubation with 10 µM of the different mimetics (+) or in culture medium alone (-). Results are representative of three independent experiments.
Percentage of apoptotic cells is shown above the apoptotic left-peak on each histogram.
ψ(CH2-CH2) and methylene ψ(CH2) bonds were significantly
more potent than B-1 in inducing apoptosis of CD40 positive
B lymphoma cells (Figure 9 and Figure S5). Noteworthy, using
the trivalent model, B-14 was found to bind CD40 with an
overall KD of 2.8 nM calculated using the trivalent model with
local fitting (the Langmuir 1:1 model gave poor fitting, see
Supporting Information, Figure S7). The substitution of a ψ-
(CH2) methylene linkage for the amide bond between Lys and
Gly consists in the replacement of the Lys-Gly segment by a
γ4-Lys residue. Similarly, substitution of γ-amino acid for
R-amino acid residues in the central type-II′ â-turn forming
tetrapeptide fragment of octreotide, a potent analogue of
somatostatin, led to the discovery of open-chain peptidomimetics
with nanomolar affinities for certain human somatostatin
receptor subtypes.51
Alternatively, the conformational space accessible to the linear
receptor binding peptide can be dramatically reduced through
cyclization. Ring-closing methatesis of fully protected linear
hexapeptides 17 and 18 containing Lys-Gly-Tyr-Tyr and Lys-
D-Pro-Tyr-Tyr epitopes and a free carboxylic acid tail, using
Grubbs catalyst,52 afforded cyclopeptides 15 and 16 ready for
coupling onto triamine cores. The corresponding ligands A-15
and A-16 with cyclic CD40 binding segments were synthesized
and purified according to our general procedure and evaluated
for binding to recombinant human CD40 by SPR and for effector
functions (see Scheme 3). Although binding to CD40 was
observed in direct SPR experiments (see Supporting Information,
Figure S8),53 A-15 and A-16 turned out to be inactive in the
apoptotic assay (see Supporting Information, Figure S9). One
can speculate that the loss of activity observed is associated
with unwanted conformational changes and low population of
a type-II′ â-turn conformation. Alternatively, the readily ac-
cessible and recently reinvestigated DL4 cyclopentapeptide
template could be used to enforce type-II′ â-turn conformation
with the D-amino acid residue at the i+1 position.54
Second-Generation Mimetics Based on a New CD40-
Binding Motif. Molecular modeling experiments and X-ray
crystal data indicate that the surface area buried upon complex-
ation of CD40L and CD40 is ca. 850 Å2 and is highly polar.25
Residues of CD40L in contact with CD40 are distributed over
two CD40L molecules. Although the Lys-Gly-Tyr-Tyr loop
(52) Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Org. Lett. 1999, 1, 953-
956.
(53) In the case of B-15, fitting of SPR data was better achieved with divalent
and trivalent models (ø2 ) 0.271 and 0.269, respectively) than with the
“Langmuir 1:1” monovalent model (ø2 ) 0.843).
(54) (a) Mierke, D. F.; Kurz, M.; Kessler, H. J. Am. Chem. Soc. 1994, 116,
1042-1049. (b) Heller, M.; Sukopp, M.; Tsomaia, N.; John, M.; Mierke,
D. F.; Reif, B.; Kessler, H. J. Am. Chem. Soc. 2006, 128, 13806-13814.
(50) Spatola, A. F. In Chemistry and Biochemistry of Amino Acids, Peptides
and Proteins; Weinstein, B., Ed.; Marcel Dekker Inc.: New York, 1983;
Vol. 7, pp 267-357.
(51) Seebach, D; Schaeffer, L.; Brenner, M.; Hoyer, D. Angew. Chem., Int. Ed.
2003, 42, 776-778.
9
J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007 13489

A R T I C L E S
Trouche et al.
motif from one subunit is critical for CD40L binding to CD40,
site-directed mutagenesis studies have identified another hot-
spot region located on â-strand E and encompassing residues
Arg203 and Arg207 (Figure 10).
It has been suggested from site-directed mutagenesis experi-
ments and calculations that these positively charged side-chains
(Lys143, Arg203, and Arg207) can make contacts with oppositely
charged residues on CD40, namely Glu66, Glu74, Asp84, and
Glu117 and that the resulting CD40-CD40L charged pairs
Arg203-Glu74, Arg207-Asp84, Arg207-Glu117 have a net stabilizing
effect on the complex. On the basis of distance considerations,
calculation of desolvation energy and total electrostatic stabiliza-
tion, the Arg203-Glu74 pair (3.2 Å) was proposed to be the
strongest salt bridge in the model. These data thus suggested to
us that other sequences, derived from the second symmetry
related subunit or combinations of sequences could in principle
be used to design second generation CD40L mimetics. Two
linear pentapeptide sequences mapping residues Gly199 to Arg207
on loop DE and â-strand E and containing Arg203 (N- or
C-terminal) were considered as possible CD40 binding units.
The corresponding trimeric architectures were assembled on
â-tripeptide core B with an Ahx linker and compared directly
to miniCD40L B-1 in SPR and cellular assays. The results are
reported in Figure 11.
Figure 12. Schematic representation of the two equilateral triangles whose
vertices are formed by R-amino groups of residues immediately following
Tyr146 (green) and Arg207 (pink) according to the X-ray crystal coordinates
of homotrimeric CD40L.24 Distances from centroid to vertex are indicated.
peptides of 5-8 Å diameter (e.g., A and B) with side arms
linked to flexible oligomethylene spacers are very versatile
systems for trivalent display of CD40-binding motifs.55 For
example, miniCD40Ls B-1 and B-17 which utilize distinct
CD40-binding elements, namely Lys143-Tyr146 on loop AA′′ and
Arg203-Arg207 on â-strand E have the same trimeric scaffold-
linker combination (Ahx)3-B. Examination of the crystal
structure of homotrimeric CD40L24 shows, however, that the
two equilateral triangles whose vertices are formed by the three
R-amino groups of the residue immediately following Tyr146
and Arg207 (Thr147 and Ala208, respectively) largely differ in size,
with centroid-to-vertex distances of ca. 16 and 10 Å, respectively
(Figure 12).
SPR measurements at different concentrations (Figure 11a)
showed that compound B-17 (R ) H-Arg-Leu-Leu-Ile-Arg),
but not B-18 (R ) H-Gly-Arg-Phe-Ile-Arg), binds significantly
to CD40. Analysis of kinetic data using the Langmuir 1:1 model
and local fitting gave an mean KD of 4.4 nM for B-17 (vs 120
nM for B-1). Like for B-1, local fitting using a trivalent model
gave lower ø2 values (0.54 versus 7.5 for the langmuir 1:1 model
and 2:1 for the divalent model), suggesting binding in a trivalent
manner. Interestingly, B-17 was also found to be as potent as
B-1 in inducing apoptosis of human lymphoma B cells (42.5%
apoptosis versus 40.9% at 10 µM after 16 h of treatment, Figure
11b). Preliminary data also indicate that B-17 like B-1 is
effective in inducing the maturation of the D1 dendritic cell
line (data not shown). Overall, these biochemical and biological
data underscore further the versatility of rigid and planar R-
and â-cyclopeptide based scaffolds A and B for the design of
miniCD40Ls.
Although this observation underscores the versatility of the
(Ahx)3-B system, it also suggests that a scaffold-linker
combination consisting of a shorter linker or a smaller scaffold
(e.g., heterocyclic) could be considered in principle to tether
the Arg203-Arg207 binding moiety.
All TNF superfamily members share the same overall
topology and are active as homotrimers. Thus, one can conceive
that a general approach based on the use of peptide templates
A and B could be transposable to other TNF family members.
While this work was in progress, researchers at Genentech and
Affymax, independently reported peptide sequences that specif-
ically bind to the human DR5 (hDR5), one of the signaling
receptors for TNF-related apoptosis-inducing ligand (TRAIL).56
Interestingly, oligomeric forms were generated either by fusion
of these peptides to leucine zipper variants57 that form either
dimers, trimers, or tetramers56a or simply by dimerization.56b
Avidity to hDR5 was significantly improved through oligomer-
ization of DR5-binding peptides and some of these multimeric
Conclusion
Molecules with multiple copies of a recognition motif
appended to a central scaffold have unique properties that differ
from those of monovalent compounds. In particular, designed
multivalent systems can mediate cell-surface receptor oligo-
merization and display effector functions.14 MiniCD40Ls are
relatively small (<3 kDa), rationally designed synthetic mul-
timeric architectures that mimic, both in vitro26 and in vivo28
the effects of CD40L, a noncovalent homotrimeric protein
belonging to the TNF superfamily. Because they are chemically
defined and tunable, synthetic multivalent ligands are very
attractive molecules to delineate molecular determinants of
effector functions and to control receptor assembly in signal
transduction.14 Our data are consistent with C3-symmetry and
radial distribution of the CD40-binding motif being two essential
features leading to mimetics with effector functions. The
distance between the center of the scaffold and the binding
moiety is also a critical element of the design. Flat cyclic
(55) For selected examples of polyvalent systems based on cyclic peptide
templates of different topology, see: (a) Singh, Y.; Dolphin, G. T.; Razkin,
J.; Dumy, P. ChemBioChem 2006, 7, 1298-1314. (b) Reference 10e. (c)
Zhang, Z.; Liu, J.; Verlinde, C. L. M. J.; Hol, W. G. J.; Fan, E. J. Org.
Chem. 2004, 69, 7737-7740.
(56) (a) Li, B.; Russell, S. J.; Compaan, D. M.; Totpal, K.; Marsters, S. A.;
Ashkenazi, A.; Cochran, A. G.; Hymowitz, S. G.; Sidhu, S. S. J. Mol.
Biol. 2006, 361, 522-36. (b) Angell, Y. M.; et al. In Understanding Biology
Using Peptides; Blondelle, S. E., Ed.; Proceedings of the Nineteenth
American Peptide Symposium, Series: American Peptide Symposia, Vol.
9; Springer: New York, 2006; pp 405-406.
(57) Harbury, P. B.; Zhang, T.; Kim, P. S.; Alber, T. Science, 1993, 262, 1401-
1407.
9
13490 J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007

Relationship between Structure and Effector Function
A R T I C L E S
was purchased from Biacore AB (Uppsala, Sweden). The 3,3′-
dihexyloxacarbocyanine iodide DiOC6(3) dye was purchased from
Interchim (Montluc¸on, France). Fluorescein isothiocyanate (FITC)-
conjugated anti-mouse CD40 (3/23 clone), FITC anti-human CD40
(5C3 clone), phycoerythrin (PE) anti-mouse CD54 (3E2 clone), PE anti-
mouse CD80 (16-10A1 clone), and PE anti-mouse CD86 (GL1 clone)
mAbs were purchased from Pharmingen (San Jose, CA). Low endotoxin
recombinant mouse granulocyte/macrophage colony stimulating factor
(rmGM-CSF) was purchased from USBiological (Swampscott, MA).
constructions were found to display significant cytotoxic activity.
These results which confirm that peptide scaffolding on mul-
tivalent core structures is a viable approach to generate
miniTNFs, also expand the set of eligible multimeric architec-
tures. It is worth mentioning that self-organization into nonco-
valent homotrimers with C3-symmetry axis is not limited to TNF
family members and is also a structural feature of proteins such
as TNF-receptor-associated factor (TRAF),58 collectins [e.g.,
mannan binding lectins (MBLs), surfactant proteins SP-A and
SP-D],59 members of the C1q family,60 adenovirus fiber head
(knob domain) emanating from the vertices of the icosahedral
capsid,61 or envelope (Env) glycoproteins of HIV type-1 (HIV-
1).62 It remains to be seen whether these protein trimers could
also be mimicked (or targeted) with trimeric architectures such
as those reported here.63
By signaling through CD40, CD40L plays an active role in
a variety of immune responses, allowing B cells to differentiate,
proliferate, and isotype class switch, activating antigen present-
ing cells (e.g., dendritic cells) and mediating T-cell help for
cytotoxic T lymphocytes. Therefore, small molecule CD40L
mimetics could be potentially useful as immunopotentiators.
TNF family ligands are type-II transmembrane proteins with
an extracellular, homotrimeric C-terminal TNF homology
domain that is frequently released as a soluble cytokine upon
proteolytic processing. However, receptor trimerization of
TNF-R superfamily members by their soluble ligands is not
always sufficient to reach the activation threshold. In particular,
full activation of CD40 which leads to B-cell proliferation and
isotype switching requires higher-order receptor oligomerization.
This is typically achieved not only by membrane-bound CD40L
but also by cross-linking with antibodies or by using artificial
fusion proteins consisting of multiple copies (two or four) of
CD40L trimers.64 By analogy, it is tempting to speculate that
similar synergistic effects could be generated through multiva-
lent display of miniCD40Ls. CD40L mimetics disclosed in this
work together with their higher-order multimeric versions would
certainly constitute a unique set of probes to investigate the
relationship between receptor oligomerization and cell signal
transduction. This possibility will be addressed in future
extension of this work.
Surface Plasmon Resonance Analysis. BIAcore 3000 (Biacore AB)
was used to evaluate the binding of CD40L mimetics to CD40. Flow
cells of a CM5 sensor chip (Research Grade, Biacore AB) were
precoated with rabbit polyclonal antibodies directed against mouse
immunoglobulin (Ig) domain (RAM-Ig, Biacore AB) using amine
coupling at 30 µg/mL in 10 mM acetate buffer, pH 5.5 according to
the manufacturer’s instructions. The chip was then flushed with 1 M
ethanolamine hydrochloride pH 8.5 (Biacore AB) and 50 mM HCl to
eliminate unbound antibody. Generally ca. 10,000 response units (RU),
corresponding to 10 ng/mm2, of RAM-Ig were immobilized. Biosensor
assays were performed at 25 °C with HBS-EP buffer [10 mM HEPES
(pH 7.4), containing 0.15 M NaCl, 3.4 mM EDTA, and 0.005% v/v
surfactant P20] as running buffer. Capture of rshCD40:mIg and of
LG11-2, a mouse IgG2a mAb directed against H2B histone used as
irrelevant control (purified by protein G affinity chromatography from
hybridoma supernatant), was performed on individual flow cells at a
flow rate of 5 µL/min and at a concentration that would achieve
equivalent protein mass binding. A small amount of protein was
captured, ca. 450 RU, sufficient for the detection of the binding of the
CD40 ligands, and limiting the mass transport and rebinding artefacts.
CD40L mimetics and rshCD40L:mCD8 were then injected (kinetic
mode) at a flow rate of 30 µL/min over the control and CD40 channels
for 4 min and allowed to dissociate for an additional 3 min. The flow
cells were regenerated for 30 s with 50 mM HCl. A 1-Hz acquisition
mode was sufficient for subsequent good assessment of kinetic
constants. Control sensorgrams were subtracted from the CD40 ones
and analyzed by BIAevaluation 4.1 with various models.
The simplest “Langmuir 1:1” model was first used to assess the
binding kinetics of the interaction. A complete description of equations
used for kinetics data determination can be found in the Biacore AB
Handbook and is summarized in Supporting Information. The trivalent
analyte model was also used to evaluate the stoichiometry and to get
some details on the mechanism of the binding, in particular the distinct
rates for all binding and conformational change events, but due to
complexity of the equations and the limited integrative possibilities
offered by the BIAevaluation software, cooperativity was assessed with
the Req values (from a trivalent model with freedom in the determination
of the association and dissociation rate constants of the first interaction)
and the Rmax value as well as all other rate constants globally processed
for the sets of concentrations of the CD40-ligands.
Experimental Section
Synthesis of Ligands. See Supporting Information.
Biological Reagents. Recombinant soluble human CD40-mouse Ig
fusion protein (rshCD40:mIg) and recombinant soluble human CD40L
coupled to the mouse CD8-R extracellular domain (rshCD40L:mCD8)
were purchased from Ancell corporation (Bayport, MN). Surfactant P20
Accuracy of the fits was measured by the Chi square (ø2) value,
which evaluates the deviation of the fitted model from the experimental
points. Residuals were also shown to give better insight into whether
a model is accurate, as the bandwidth and its shape reveal the differences
between the fitted curve and the experimental data.
(58) McWhirter, S. M.; Pullen, S. S.; Holton, J. M.; Crute, J. J.; Kehry, M. R.;
Alber, T. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 8408-8413.
(59) Hakansson, K.; Reid, K. B. Protein Sci. 2000, 9, 1607-1617.
(60) Kishore, U.; Gaboriaud, C.; Waters, P.; Shrive, A. K.; Greenhough, T. J.;
Reid, K. B.; Sim, R. B.; Arlaud, G. J. Trends Immunol. 2004, 25, 551-
561.
Surface Plasmon Resonance Analysis: Summary of the Equa-
tions Used for the Determination of the Kinetics Data. The simplest
model, the Langmuir 1:1 one, was first used to assess the binding
kinetics of the interaction. In the following equations, L denotes the
ligand captured on the immobilized rabbit anti-mouse Ig, in our case
the receptor CD40, and A denotes the analyte injected on the chip,
that is either of the CD40L or miniCD40Ls. When A interacts with L
forming the LA complex due to diffusion, collision, and binding, in
function of time with the association rate constant ka (M-1 s-1), but
can also disappear with dissociation rate constant kd (1/s) to release A
from L (eq 1).
(61) Bewley, M. C.; Springer, K.; Zhang, Y. B.; Freimuth, P.; Flanagan, J. M.
Science 1999, 286, 1579-1583.
(62) Zhu, P.; Liu, J.; Bess, J., Jr.; Chertova, E.; Lifson, J. D.; Grise, H.; Ofek,
G. A.; Taylor, K. A.; Roux, K. H. Nature 2006, 441, 847-852.
(63) Synthetic trivalent miniproteins designed to target the CD4-binding sites
on the HIV-1 Env glycoprotein gp120 trimers have been reported recently.
Li, H.; Guan, Y.; Szczepanska, A.; Moreno-Vargas, A. J.; Carmona, A.
T.; Robina, I.; Lewis, G. K.; Wang, L. X. Bioorg. Med. Chem. 2007, 15,
4220-4228.
(64) Holler, N.; Tardivel, A.; Kovacsovics-Bankowski, M.; Hertig, S.; Gaide,
O.; Martinon, F.; Tinel, A.; Deperthes, D.; Calderara, S.; Schulthess, T.;
Engel, J.; Schneider, P.; Tschopp, J. Mol. Cell. Biol. 2003, 23, 1428-
1440.
9
J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007 13491

A R T I C L E S
Trouche et al.
Results are expressed as the percentage of specific apoptosis
according to the following formula:
kaf
rkd
L + A
LA
(1)
% of specific apoptosis )
Equilibrium is reached when the rate of association (eq 2) equals
the rate of dissociation, that is when the number of dissociation events
compensates the number of association events.
100 (% apoptotic treated cells -
% spontaneous apoptotic control cells)
(100 - % spontaneous apoptotic control cells)
d[LA]
dt
) ka[L][A] - kd[LA]
(2)
Culture and Maturation of D1 Mouse Dendritic Cell Line. D1
cells have been described as a MHC class II-positive growth factor-
dependent immature dendritic cells, derived from adult mouse spleen
maintained in lineage without being transformed.30 Cells were cultured
in nontreated plastic dishes IMDM with HEPES and L-glutamine
(Cambrex), supplemented with 10% of heat-decomplemented FBS,
gentamicin (10 µg/mL), 10 µM â-mercaptoethanol, and 10 ng/mL of
rmGM-CSF. For cellular assays, 2 × 105 cells/mL were cultured for
24 h in 24-well untreated polystyrene microplates (Evergreen Scientific,
Los Angeles, CA). Then, fresh medium containing the various
miniCD40Ls was added. After an additional 24 h of incubation, cells
were washed with cold PBS and harvested with 2 mL of PBS containing
2 mM EDTA. After centrifugation, cells were resuspended in cold PBS
and analyzed for cell surface phenotyping by flow cytometry.
Flow Cytometry Analysis. For measurement of apoptosis, lym-
phoma cells were analyzed directly after the staining procedure. For
phenotyping, D1 cells were stained in PBS containing 2% FBS at 4
°C for 20 min with the various antibodies used at concentrations
recommended by the manufacturer. After two washes in PBS, cells
were analyzed by flow cytometry with a FACSCalibur. At least 1 ×
105 events were acquired for each experiment using CellQuest version
3.3 (Becton Dickinson, Pont de Claix, France), and the data were
processed with WinMDI, version 2.8 freeware (Joseph Trotter, Scripps
Research Institute, http://facs.scripps.edu/software.html).
Under our conditions, equilibrium cannot be reached at the end of
the injection phase, but values at equilibrium could be extrapolated,
especially the dissociation constant KD (M), using the kinetic values
obtained from the individual experimental curves of the association
and dissociation phases (eq 3).
k
[L][A]
[LA]
)
d ) KD
(3)
ka
Rewriting eq 2 in terms of concentration and responses, we obtain
eq 4.
dR
dt
) kaC(Rmax - Rt) - kdRt
(4)
In eq 4, Rt corresponds to the response signal of the formed complex
at t time, C represents the concentration of CD40-ligands, and Rmax is
the maximal feasible response signal, depending on the total CD40-
binding capacity (or activity) on the sensor chip. In this case we assume
that free ligand equals maximal capacity minus bound receptor.
The data extracted from the experimental curves as well as the known
input parameters are used by the algorithm and assessed by numerical
integration to calculate the different parameters needed. One can then
determine the values for signal response at equilibrium (Req), that is
for t value tending to infinite, from the kinetic constants and the Rmax
calculated parameter (eq 5).
Acknowledgment. We thank J. P. Briand and S. Muller for
their continuous support. This work was supported by the CNRS,
the Ministe`re de la Recherche (ACI “Jeunes Chercheurs”), “La
Ligue contre le Cancer” and INCA (Cancer: de´tection
d’innovations 2006, #50). W.S. and S.W. were supported by
grants from the Ministe`re de la Recherche and from the
Fondation pour la Recherche Me´dicale (FRM). N.T. was
supported by “La Ligue contre le Cancer”. We thank N.
Bonnefoy-Berard and M. Flacher for providing the lymphoma
cells, C. J. M. Melief for providing the dendritic cell line D1,
M. Monestier for providing the LG11-2 hybridoma. We
acknowledge O. Reinaud for a generous gift of calix[6]triamine
D-0.
kaConc × Rmax
Req
)
(5)
kaConc + kd
Culture of Lymphoma Cells and Measurement of Apoptosis.
BL41 Burkitt lymphoma and Jurkat human T leukemia cells were
cultured in RPMI 1640 (Cambrex Bioscience, Verviers, Belgium)
supplemented with 10% of heat-decomplemented fetal bovine serum
(FBS; Dominique Dutscher, Brumath, France) and gentamicin (10 µg/
mL, Cambrex). For apoptosis assays, cells (5 × 105 cells/mL) were
incubated at 37 °C in flat-bottom 96-well plates at the indicated times
and concentrations in the presence of various miniCD40Ls in 200 µL
of complete medium. After incubation, cells were washed with
phosphate-buffered saline (PBS), and apoptosis was evaluated by
measurement of a decrease in mitochondrial transmembrane potential
(∆ψm) associated with a reduction of the cationic dye DiOC6(3) uptake,
as detected by flow cytometry. For this, cells were resuspended in 300
µL of PBS containing 20 nM of DiOC6(3), incubated at 37 °C for 20
min, and then directly analyzed by flow cytometry.
Supporting Information Available: Detailed procedures and
characterization data of all ligands as well as additional surface
plasmon resonance (SPR) sensorgrams and analysis; complete
refs 21b and 56b. This material is available free of charge via
the Internet at http://pubs.acs.org.
JA073169M
9
13492 J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007