Cyclic RGD β-Lactam Peptidomimetics Induce Differential Gene Expression in Human Endothelial Cells

Article abstract of DOI:10.1002/cbic.201000572

RGD peptidomimetics touch DNA: A few RGD cyclic peptidomimetics, each incorporating a densely substituted β-lactam moiety, were prepared, and their effect on extracellular adhesion and intracellular gene regulation through α_Vβ₃ integ

Full text of DOI:10.1002/cbic.201000572

DOI: 10.1002/cbic.201000572
Cyclic RGD b-Lactam Peptidomimetics Induce Differential Gene Expression
in Human Endothelial Cells
Jesus M. Aizpurua,*[a] Josꢀ Ignacio Ganboa,[a] Claudio Palomo,*[a] Iraida Loinaz,[a] Joseba Oyarbide,[a]
Xabier Fernandez,[a] Eva Balentovꢁ,[a] Raluca M. Fratila,[a] Azucena Jimꢀnez,[a] Josꢀ Ignacio Miranda,[a]
Antonio Laso,[b] Silvia ꢂvila,[b] and Josꢀ Luis Castrillo[b]
Angiogenesis is a fundamental step in the transition of solid
tumors from a dormant state to a malign one. Many of the
low-molecular-weight anti-angiogenic drug candidates mimic
the short peptide epitope Arg-Gly-Asp (RGD),[1] disrupting the
extracellular matrix/integrin adhesion and, ultimately, leading
to tumor cell apoptosis. In contrast to the detailed structural
information available for the extracellular adhesion inhibition
of endothelial cells through the recognition of integrins (typi-
cally aVb3) by RGD peptidomimetics,[2] most aspects of possible
intracellular angiogenic gene regulation caused by peptidomi-
metics remain unexplored.[3] Importantly, dysfunction of this
signaling system is suspected to be behind the resistance phe-
nomena developed in anti-angiogenic therapies.[4]
ulation. In this context, we set out to design alternative low-
molecular-weight RGD probes for interaction with aVb3 integrin
and gene regulation in HUVECs. Ideally, these peptidomimetics
should: 1) contain a minimal scaffold to prevent undesired
scaffold/integrin interactions, 2) have a uniform and predicta-
ble scaffold conformation, 3) bear a maximum of recognition
groups, including hydrophobic or hydrophilic ones, and
4) permit the deletion of selected residues from the RGD triad
without global shape change.
We selected five-membered and four-membered small cyclic
peptides for development. The residual flexibility of these
cyclopeptides can be further constrained by incorporating
lactam bridges between neighboring amino acids to stabilize
protein secondary structure motifs characterized by combina-
tions of b-turns and/or g-turns.[8] Several lactam pseudopep-
tides resulting from such an extension of Freidinger’s con-
cept[9] (Scheme 1; 1!2) have been explored with the goals of
mimicking receptor-bound conformations of bioactive pep-
tides and of providing pharmacophore information for non-
peptide drug design.[10] However, despite its apparent simplici-
ty, this design is not always reliable for the precise positioning
of a maximum number of recognition groups around the pseu-
dopeptide cyclic core. Because the interresidual lactam bridge
created by modification of the side chain (R1) shares recogni-
tion and constraint functions, the design of mimetic libraries
becomes difficult and non-general, when synthetically achieva-
ble. An alternative way to constrain cyclic peptides is based on
the incorporation of a d-amino acid and an N-alkyl-amino acid
into the macrocycle, as illustrated by the remarkable aVb3
antagonist cilengitide (3, cyclo-[Arg-Gly-Asp-d-Phe-N(Me)Val])
developed by Kessler et al.[11]
Inside the endothelial cell, dozens of proteins mediate or
control the signaling pathways of angiogenesis after integrin
activation, but only a couple of kinases (JNK, ERK) and tran-
scription factors (NFkB, FoxO) are able to promote gene regu-
lation[5] (Figure 1). In addition, large environmental ligands,
such as vascular endothelial growth factors (VEGFs)[6] or pep-
tide hormones,[7] are required to elicit proangiogenic gene reg-
Here we report an alternative, more versatile solution to the
above design problem, by introducing a-amino-b-lactam scaf-
folds originating from aCH/NH proton mimicry (Scheme 2; 1!
4). This “b-lactam scaffold-assisted design” (b-LSAD) approach
is based on the separation of recognition and constraint
groups, and has previously proven efficient for promotion of
unusually stable type-II (II’) b-turn peptidomimetics from ex-
tended open native peptides.[12] Accordingly, the straight appli-
cation of the b-LSAD principle to RGD cyclic mimetics led us to
the b-lactam pseudopeptides 4a–d, which fulfilled several
structural requirements to trigger quite different signaling in-
teractions with aVb3 integrin. The hydrophobic benzyl group in
the d-Phe residue of cilengitide (3), for instance, was replaced
by the strongly hydrophilic 1,2-dihydroxyethyl moiety in the
mimetic 4a or by the poorly hydrophobic methyl group in
compound 4b. Conversely, the hydrophobic R1–R3 substituents
Figure 1. Signaling pathways initiated by integrins affecting the gene regula-
tion of angiogenesis. ERK=extracellular-signal-regulated kinase, FAK=focal
adhesion kinase, FoxO=forkhead box O, JNK=c-Jun N-terminal kinase,
NFkB=nuclear factor k of B-cells.
[a] Prof. Dr. J. M. Aizpurua, Prof. Dr. J. I. Ganboa, Prof. Dr. C. Palomo,
Dr. I. Loinaz, Dr. J. Oyarbide, X. Fernandez, Dr. E. Balentovꢀ,
Dr. R. M. Fratila, Dr. A. Jimꢁnez, Dr. J. I. Miranda
Departamento de Quꢂmica Orgꢀnica I
University of the Basque Country UPV-EHU
Centro Joxe Mari Korta, Avenida de Tolosa 72, 20018 San Sebastian (Spain)
Fax: (+34)943-015959
E-mail: jesusmaria.aizpurua@ehu.es
[b] Dr. A. Laso, Dr. S. ꢃvila, Dr. J. L. Castrillo
Genetadi Biotech AG
Edificio 502, Parque Tecnolꢄgico de Bizkaia, 48160 Derio (Spain)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201000572.
ChemBioChem 2011, 12, 401 – 405
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
401

Scheme 2. Reagents and conditions: a) Cbz-Asp(OtBu)-OH, EDC, HOBt, Et3N,
CH2Cl2. b) HF·Pyr, then BAIB, TEMPO, MeCN/H2O. c) H-Arg(Pbf)R4, EDC, HOBt,
CH2Cl2. d) PhSH, K2CO3, MeCN, RT, 2 h. e) Cbz-Asp(OtBu)-OH, HATU, KHCO3,
DMF. f) H2/Pd-C, MeOH. g) HATU, HOAt, KHCO3, DMF. h) F3CO2H. BAIB=[bis-
(acetoxy)iodo]benzene, EDC=3-[(dimethylamino)propyl]-3-ethylcarbodiimide
hydrochloride, HATU=O-(7-azabenzotriazol-1-yl)-N,N,N’,N’-tetramethyluroni-
um hexafluorophosphate, HOAt=1-hydroxy-7-azabenzotriazole, HOBt=hy-
droxybenzotriazole, Ns=2-nosyl, Pbf=2,2,4,6,7-pentamethyl-2,3-dihydroben-
zofuran-5-sulfonyl, TEMPO=2,2,6,6-tetramethyl-1-piperidinyloxyl.
Scheme 1. Lactam pseudopeptide design in cyclic peptides. Left: Freiding-
er’s concept: the R1 group in 2 requires modification to fix a b-turn confor-
mation and acts simultaneously as a recognition (functional) element. Right:
b-lactam scaffold-assisted design (b-LSAD): formal insertion of a single
carbon atom into 1 (aCꢀH+HꢀN!CH2) mimics two hydrogen atoms and
constrains the cyclic pentapeptide 4, preserving the original R1 group exclu-
sively for molecular recognition. Bottom) Some b-LSAD cyclic RGD mimet-
ics—4a–d—prepared in this work.
because tetrapeptides are very reluctant to afford 12-mem-
bered macrocycles.[14] A prearranged b-turned conformation of
8d with the b-lactam ring at the (i+1) position would account
for the easy cyclization observed in this case.[15]
were preserved in compounds 4c–d, but in 4d the central Gly
residue was removed from the RGD sequence.
The cyclic integrin probes 4a–d could easily be prepared
from the known a-amino-b-1,3-dioxolanyl-b-lactam 5 and the
a-amino-a-alkyl-b-lactams 6[13] (Scheme 2). Compound 5 was
coupled to Cbz-Asp(OtBu)-OH and the N-tert-butyldimethylsilyl-
oxymethyl moiety was then subjected to desilylation and oxi-
dation with the BAIB/TEMPO system to provide 7a in 54%
overall yield. Conversely, coupling of the carboxylic b-lactams 6
with the protected arginines H-Arg(Pbf)GlyOBn and H-Arg-
(Pbf)OBn gave the b-lactam pseudopeptides 7b, 7c, and 7d in
65, 78, and 70% yields, respectively. Next, peptide coupling at
the carboxylic group of 7a and the N termini of 7b–d afforded
the b-lactam macrocyclic precursors 8a–d in 70–84% overall
yields. These compounds were deprotected under catalytic hy-
drogenation conditions and cyclized under high-dilution condi-
tions, and the resulting b-lactam cyclic peptides were depro-
tected again with trifluoroacetic acid to provide the target
compounds 4a–d. The cyclization of 8d to 4d is noteworthy
NMR/MD conformation analysis of the RGD b-lactam cyclic
peptides 4a–d in water solution revealed the greatly prevalent
formation of highly populated single clusters lacking inner hy-
drogen-bonded b-turns or g-turns. This was evident from the
large thermal coefficients (jDd/DTj>ꢀ3 ppbKꢀ1) recorded for
the amide NH protons and from the ROESY interproton dis-
tance data sets (Figure 2; see pages S18–S24 in the Supporting
Information for details). These observations were in line with
the reported conformation of cilengitide (3), which also fea-
tures a peptide backbone without stable hydrogen bonds (Dd/
DT=d-Phe: ꢀ4.1; Asp: ꢀ6.1; Gly: ꢀ4.9; Arg: ꢀ9.3 ppbKꢀ1).[11]
Only the mimetic 4b, devoid of bulky substituents in the b-
lactam ring, displayed a significantly low thermal coefficient
(ꢀ1.2 ppbKꢀ1) for the a-amido-b-lactam NH group, consistent
with the formation of a strong inverse g-turn around the Asp
residue and stabilized by the (b-lactam)NꢀH···O=C(Gly) hydrogen
bond. Surprisingly, though, the strong betagenicity displayed
402
www.chembiochem.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 401 – 405

mimetics tested, except for the a-methyl-substituted com-
pound 4b, displayed IC50 values comparable to that of 3
(IC50 =5 mm under the test conditions). The lower activity of 4b
was not unexpected, because the generally accepted RGD/
aVb3 integrin interaction model assumes that hydrophobic
groups, such as the Bn residue of d-Phe in cilengitide, increase
the affinity of the mimetic for the integrin receptor.[17] More
surprisingly, the polyhydroxylic mimetic 4a showed a slightly
higher activity (IC50 =1.5 mm) than cilengitide (pages S27–30 in
the Supporting Information).[18] Although we cannot provide
an explanation for this observation, the structural information
provided by compound 4a might be useful for the exploration
of a novel, as yet unknown, hydrophilic pocket in aVb3 integrin
and its exploitation in nonpeptide mimetic design. Finally, the
tetrapeptide mimetic 4d, the product of formal deletion of the
Gly residue from pentapeptide 4c, was also a strong antago-
nist of aVb3 integrin. This finding is in good agreement with
the accepted model correlating the selectivity of RGD mimetics
for aVb3 integrin to a distance shorter than ꢁ6.0 ꢄ between
the Cb atoms of the Asp and Arg residues.[19]
In order to study the intracellular effect of the RGD mimetics
with higher affinity for integrin binding, we conducted a gene
expression microarray assay. HUVECs were treated separately
with a 10ꢀ5 m concentration of cilengitide (3) or of compounds
4a, 4c, or 4d for 48 h. After analysis of the whole 20500
human genes, the microarray assay provided a two-color out-
put image of normalized gene expression data (Figure 4 and
pages S32–35 in the Supporting Information). Genes that were
at least 1.5-fold differentially expressed on three of four arrays
were scored as significant. From this transcriptomic analysis we
could identify up to 225 activated genes and 64 inhibited
genes for compound 3, 221 activated and 66 inhibited for 4a,
227 activated and 56 inhibited for 4c, and up to 198 activated
and 87 inhibited for 4d after the RGD mimetic cell treatment.
Over 90% of these genes showed a similar transcriptional pat-
tern in cells treated with compounds 3, 4a, and 4c. Less than
10% of these genes were conserved in cells treated with com-
pound 4d. Out of all the up- or down-regulated genes, we
selected 17 genes known to be related to angiogenesis[20] for a
more detailed analysis (Figure 4). After treatment with cilengi-
tide (3), ten of these genes were activated and seven were
blocked. The pattern of gene expression was mainly conserved
(15/17 genes) for compounds 4a and 4c (14 of 17 genes). On
the other hand, the lack of correlation in the expression of
three genes (CDC7, ADAM6, and GCKR) demonstrated the
highly specific “in vivo” cell activity after ligand–receptor bind-
ing.
Figure 2. Five-structure overlays for compounds 4a–d calculated from NMR
interproton distance restrictions (ROESY, tmix =200 ms) determined in H2O/
D2O (9:1). Amide thermal coefficient values are in ppbKꢀ1. All hydrogen
atoms except amide NH have been omitted for clarity.
by a-amino-b-lactams in open peptides vanished in the cyclic
pentapeptide and tetrapeptide counterparts, as evidenced by
the absence of any (Asp)C=O···HꢀN(Arg) hydrogen bond in the
mimetics 4a–d. Finally, the peptide backbone strain in the b-
lactam macrocycles varied with the azetidinone ring substitu-
tion position (the b-substituted compound 4c was more rigid
than the a-substituted 4a or 4b) and also with the number of
residues (the tetrapeptide 4d was extremely rigid). These ob-
servations confirmed the suitability of the b-LSAD design as a
general scaffolding tool for positioning of different recognition
groups in predefined spatial dispositions.
We next compared the antagonist affinities of the RGD b-
lactam mimetics 4a–d and of the highly active cilengitide (3)
against aVb3 integrin on human endothelial cells (HUVECs) by
means of an adhesion inhibition assay (Figure 3).[16] All the
The microarray results were further confirmed by gene-spe-
cific quantitative mRNA assays (qRT-PCR). The pattern expres-
sion of, for example, TGFBR2, TRIP12, FOXC1, and ITGA9 genes,
with use of specific amplification primers and hybridization
probes, was identical to that previously shown by microarray
technology (pages S34–S36 in the Supporting Information). In
addition, no significant changes were observed in the gene ex-
pression of the apoptosis-associated tyrosine kinase (AATK)[22]
and the apoptotic regulator BAX (BL2-associated X protein)[23]
genes, suggesting that apoptosis was not induced after 72 h
Figure 3. Adhesion inhibition of HUVECs on vitronectin-coated cell-culture
plates promoted by the RGD b-lactam ligands 4a–d and cilengitide (3). Ad-
herent cells were counted after 1 h and their adhesion (%) relative to un-
treated control experiment is shown. HUVEC=human umbilical vein endo-
thelial cell.
ChemBioChem 2011, 12, 401 – 405
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
403

the Gobierno Vasco (ETORTEK-inanoGUNE IE-08/225). We thank
SGIker (UPV/EHU) for NMR facilities.
Keywords: antitumor agents
·
cell adhesion
·
gene
expression · peptides · peptidomimetics
[1] a) R. Haubner, W. Schmitt, G. Hçlzemann, S. L. Goodman, A. Jonczyk, H.
Kessler, J. Am. Chem. Soc. 1996, 118, 7881–7891; b) L. Belvisi, A. Bernar-
di, A. Checchia, L. Manzoni, D. Potenza, C. Scolastico, M. Castorina, A.
Cupelli, G. Giannini, P. Carminati, C. Pisano, Org. Lett. 2001, 3, 1001–
1004; c) S. Urman, K. Gaus, Y. Yang, U. Strijowski, N. Sewald, S. De Pol, O.
Reiser, Angew. Chem. 2007, 119, 4050–4053; Angew. Chem. Int. Ed.
2007, 46, 3976–3978; d) A. S. M. da Ressurreicao, A. Vidu, M. Civera, L.
Belvisi, D. Potenza, L. Manzoni, S. Ongeri, C. Gennari, U. Piarulli, Chem.
Eur. J. 2009, 15, 12184–12188.
[2] a) J.-P. Xiong, T. Stehle, R. Zhang, A. Joachimiak, M. Frech, S. L. Good-
man, M. A. Arnaout, Science 2002, 296, 151–155; b) L. Marinelli, A. Lave-
chia, K. E. Gottschalk, E. Novellino, H. Kessler, J. Med. Chem. 2003, 46,
4393–4404; c) D. Craig, M. Gao, K. Schulten, V. Vogel, Structure 2004,
12, 2049–2058; d) T. Cupido, J. Spengler, J. Ruiz-Rodriguez, J. Adan, F.
Mitjans, J. Piulats, F. Albericio, Angew. Chem. 2010, 122, 2792–2797;
Angew. Chem. Int. Ed. 2010, 49, 2732–2737.
Figure 4. Angiogenesis-related gene regulation of DNA samples extracted
from HUVECs after treatment with the RGD b-lactam ligands 4a–d or with
cilengitide (3). Numbers refer to a binary logarithmic scale, and colors refer
to activation (red) or inhibition (blue). For gene hierarchical clustering and
gene function, see page S32 in the Supporting Information.
[3] For a genetic and cell biological analysis of integrin outside-in signaling
see: K. R. Legate, S. A. Wickstrçm, R. Fꢅssler, Genes Dev. 2009, 23, 397–
418.
[4] For a recent review on the resistance to anti-angiogenic therapies, see:
Y. Crawford, N. Ferrara, Trends Pharmacol. Sci. 2009, 30, 624–630.
[5] T. Kume, Trends Cardiovasc. Med. 2008, 18, 224–228.
[6] S. P. Ivy, J. Y. Wick, B. M. Kaufman, Nat. Rev. Clin. Oncol. 2009, 6, 569–
579.
[7] C. Clapp, S. Thebault, M. C. Jeziorski, G. Martinez De La Escalera, Physiol.
Rev. 2009, 89, 1177–1215.
of treatment. Surprisingly, the gene expression pattern induced
after cell treatment with the cyclic tetrapeptide 4d was almost
the opposite (activation vs. inhibition) of that of cilengitide (3)
in 15 out of 17 genes. These data suggest an “in vivo” pro-
angiogenic prevalent effect of compound 4d, which might act
as an agonist ligand of the RGD receptor.[21]
In conclusion, the incorporation of a-amino-b-lactam moiet-
ies into cyclic pentapeptides and tetrapeptides prevents b-turn
formation around the b-lactam residue, affording highly rigidi-
fied b-lactam cyclopeptides. This design (b-LSAD) permits the
precise spatial positioning of a variety of recognition groups,
as illustrated in the preparation of the cyclic RGD b-lactam
mimetics 4a–d. With the exception of the a-methyl-substitut-
ed mimetic 4b, most of the RGD compounds prepared exhibit-
ed a cilengitide-like antagonist activity against aVb3 integrin in
an “in-vitro” cell assay, but they induced compound-specific in-
tracellular angiogenesis-related gene expression. Cilengitide (3)
and the cyclic RGD b-lactam pentapeptides 4a–c show a very
similar pattern of gene expression, whereas the tetrapeptide
mimetic 4d (with a deleted Gly residue) has the opposite
effect in intracellular gene regulation. This example suggests
that a very specific signaling modification of intracellular an-
giogenesis pathways can be achieved by using suitably de-
signed low-molecular-weight peptidomimetics. It also outlines
the potential usefulness of automated genetic tests, in combi-
nation with current adhesion assays, to establish more accu-
rately the anti- or proangiogenic profiles of novel integrin li-
gands. Further studies to establish the gene-regulation effects
of other glycine-deleted RGD peptidomimetics are underway
in our laboratory.
[8] a) C. Gilon, C. Mang, E. Lohof, A. Friedler, H. Kessler in Houben–Weyl
Methods of Organic Chemistry, Vol. E22b: Synthesis of Peptides and Pepti-
domimetics (Eds.: M. Goodman, A. Felix, L. Moroder, C. Toniolo) Thieme,
Stuttgart, 2003, pp. 461–542; b) M. Gurrath, G. Mꢆller, H. Kessler, M. Au-
mailley, R. Timpl, Eur. J. Biochem. 1992, 210, 911–921.
[9] a) R. M. Freidinger, D. F. Veber, D. S. Perlow, J. R. Brooks, R. Saperstein,
Science 1980, 210, 656; b) R. M. Freidinger, J. Med. Chem. 2003, 46,
5553–5566.
[10] G. V. Nikiforovich, K. E. Kçvꢀr, W.-J. Zhang, G. R. Marshall, J. Am. Chem.
Soc. 2000, 122, 3262–3273.
[11] M. A. Dechantsreiter, E. Planker, B. Mathꢅ, E. Lohof, G. Hçlzemann, A.
Jonczyk, S. L. Goodman, H. Kessler, J. Med. Chem. 1999, 42, 3033–3040.
[12] C. Palomo, J. M. Aizpurua, A. Benito, J. I. Miranda, R. M. Fratila, C.
Matute, M. Domercq, F. Gago, S. Martin-Santamaria, A. Linden, J. Am.
Chem. Soc. 2003, 125, 16243–16260.
[13] C. Palomo, J. M. Aizpurua, E. Balentova, A. Jimenez, J. Oyarbide, R. M.
Fratila, J. I. Miranda, Org. Lett. 2008, 10, 101–104.
[14] J. M. Beierle, W. S. Horne, J. H. van Maarseveen, B. Waser, J. C. Reubi,
M. R. Ghadiri, Angew. Chem. 2009, 121, 4819–4823; Angew. Chem. Int.
Ed. 2009, 48, 4725–4729.
[15] For a review on conformation-directed macrocyclizations, see: J. Blan-
kenstein, J. Zhu, Eur. J. Org. Chem. 2005, 1949–1964.
[16] M. Kawaguchi, R. Hosotani, S. Ohishi, N. Fujii, S. S. Tulachan, M. Koizumi,
E. Toyoda, T. Masui, S. Nakajima, S. Tsuji, J. Ida, K. Fujimoto, M. Wada, R.
Doi, M. Imamura, Biochem. Biophys. Res. Commun. 2001, 288, 711–717.
[17] R. Haubner, D. Finsinger, H. Kessler, Angew. Chem. 1997, 109, 1440–
1456; Angew. Chem. Int. Ed. Engl. 1997, 36, 1374–1389.
[18] Three additional diastereomers of compound 4a were prepared, their
conformations in water were studied, and their binding abilities were
tested: see pages S20–S21 in the Supporting Information.
[19] a) S. Kostidis, A. Stavrakoudis, N. Biris, D. Tsoukatos, C. Sakarellos, V. Tsi-
karis, J. Pept. Sci. 2004, 10, 494–509; b) M. Salvati, F. M. Cordero, F. Pisa-
neschi, F. Melani, P. Gratteri, N. Cini, A. Bottoncetti, A. Brandi, Bioorg.
Med. Chem. 2008, 16, 4262–4271. See also ref. [2d].
Acknowledgements
[20] N. G. Costouros, S. K. Libutti, Expert Opin. Biol. Ther. 2002, 2, 545–556.
[21] For an example of pro-angiogenic behavior of RGD-mimetic integrin in-
hibitors, see: A. R. Reynolds, I. R. Hart, A. R. Watson, J. C. Weiti, R. G.
This work was supported by the Ministerio de Educaciꢄn y Cien-
cia (MEC, Spain, Project: CTQ2006-13891/BQU), UPV/EHU, and
404
www.chembiochem.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 401 – 405

Silva, S. D. Robinson, G. Da Violante, M. Gourlaouen, M. Salih, M. C.
Jones, D. T. Jones, G. Saunders, V. Kostourou, F. Perron-Sierra, J. C.
Norman, G. C. Tucker, K. M. Hodivala-Dilke, Nat. Med. 2009, 15, 392–
400.
[23] J. Ding, Z. Zhang, G. J. Roberts, M. Falcone, Y. Miao, Y. Shao, X. C. Zhang,
D. W. Andrews, J. Lin, J. Biol. Chem. 2010, 285, 28749–28763.
Received: September 22, 2010
[22] M. Tomomura, N. Morita, F. Yoshikawa, A. Konishi, H. Akiyama, T. Furui-
chi, H. Kamiguchi, Neuroscience 2007, 148, 510–521.
Published online on January 10, 2011
ChemBioChem 2011, 12, 401 – 405
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
405