Identification of a novel class of small-molecule antiangiogenic agents through the screening of combinatorial libraries which function by inhibiting the binding and localization of proteinase MMP2 to integrin αvβ3

Article abstract of DOI:10.1021/ja003579+

The process of new blood vessel growth from existing vasculature, known as angiogenesis, is critical to several pathological conditions, most notably cancer. Both MMP2, which degrades the extracellular matrix (ECM), and integrin α_vβ₃

Full text of DOI:10.1021/ja003579+

1280
J. Am. Chem. Soc. 2001, 123, 1280-1288
Identification of a Novel Class of Small-Molecule Antiangiogenic
Agents through the Screening of Combinatorial Libraries Which
Function by Inhibiting the Binding and Localization of Proteinase
MMP2 to Integrin RVâ3
Dale L. Boger,*,† Joel Goldberg,† Steve Silletti,‡,§ Torsten Kessler,‡ and David A. Cheresh‡
Contribution from the Departments of Chemistry, Immunology, and Vascular Biology and The Skaggs
Institute for Chemical Biology, The Scripps Research Institute, 10550 North Torrey Pines Road,
La Jolla, California 92037
ReceiVed October 3, 2000
Abstract: The process of new blood vessel growth from existing vasculature, known as angiogenesis, is
critical to several pathological conditions, most notably cancer. Both MMP2, which degrades the extracellular
matrix (ECM), and integrin RVâ3, which contributes to endothelial cell attachment to the ECM, are critically
involved in this process. Recent findings have shown that MMP2 is localized in an active form on the surface
of invasive endothelial cells based on its ability to directly bind integrin RVâ3, suggesting that disrupting this
protein-protein interaction may represent a new target for the development of angiogenesis inhibitors. The
screening of small molecule libraries led to the identification of compounds which disrupt the MMP2-RVâ3
interaction in an in vitro binding assay. A prototypical inhibitor was further found to prevent the degradation
of the protein matrix without directly inhibiting MMP2 activity or disrupting the binding of RVâ3 to its classical
ECM ligand, vitronectin. The synthesis and screening of analogues and substructures of this lead compound
allowed the identification of requisite structural features for inhibition of MMP2 binding to RVâ3. This led to
the synthesis of a more water-soluble derivative which maintains the in vitro biological properties and has
potent antiangiogenic and antitumor activity in vivo, validating the target as one useful for therapeutic
intervention.
Introduction
a crucial role in degrading the extracellular matrix (ECM) during
tumor-induced angiogenesis.6 Peptide and peptidomimetic MMP2
inhibitors have been developed which diminish tumor growth
and metastasis.7 Another key participant in angiogenesis is the
integrin RVâ3, which mediates cellular interactions with the
ECM.8 This heterodimeric cell-surface protein recognizes certain
matrix proteins containing the RGD (Arg-Gly-Asp) peptide
sequence, leading to integrin clustering and ultimately supporting
endothelial cell migration and survival. Soluble RGD peptides,
nonpeptide RGD mimetics, and antibodies that disrupt RVâ3
ligation have been shown to inhibit angiogenesis.9 In addition,
a number of investigators have developed antagonists of integrin
RIIbâ3, which plays a key role in thrombosis and platelet
aggregation,10 some of which are currently approved for
thrombotic disease. Because of the similarity of the ligands
recognized by the different integrins, an inherent obstacle in
the design of effective RVâ3 antagonists for use in antiangiogenic
therapy is integrin selectivity (e.g., RVâ3 vs RIIbâ3).
The development of new blood vessels from existing vascu-
lature, a process known as angiogenesis, has been the subject
of considerable research.1 This process is normally involved in
wound repair, inflammation, and embryonic development,
whereas aberrant angiogenesis is characteristic of several
pathological conditions,2 including arthritis, ocular retinopathy,
and tumor growth and metastasis.3 Progress in recent years has
led to a better understanding of angiogenesis, revealing that it
is a multistep process involving vascular cell activation, matrix
degradation, and cell migration, proliferation, and differentiation.
A number of proteolytic enzymes and cell-surface receptors are
involved, and the process can be promoted or inhibited by the
action of growth factors and other endogenous regulators.4
Several approaches, targeting angiogenesis at different stages,
are being pursued as potential cancer therapies. Recent estimates
indicate that at least 40 pharmaceutical firms are developing
angiogenesis inhibitors with at least 27 drugs currently in clinical
trials.5
Recently, a novel interaction between MMP2 and integrin
RVâ3 was observed when these proteins were found to colocalize
on the surface of angiogenic blood vessels in vivo.11 Integrin
MMP2 (gelatinase A), a member of the matrix metallopro-
teinase family, is secreted by vascular endothelial cells and plays
(5) Brower, V. Nature Biotechnol. 1999, 17, 963. Klohs W. D.; Hamby
J. M. Curr. Opin. Biotechnol. 1999, 10, 544.
(6) Chen, J.-M.; Aimes, R. T.; Ward, G. R.; Youngleib, G. L.; Quigley,
J. P. J. Biol. Chem. 1991, 266, 5113.
(7) Kahari, V.-M.; Saarialho-Kere, U. Ann. Med. 1999, 31, 34.
(8) Brooks, P. C.; Clark, R. A. F.; Cheresh, D. A. Science 1994, 264,
569. For a review of endogenous ligand binding to integrins, see: Plow, E.
F.; Haas, T. A.; Zhang, L.; Loftus, J.; Smith, J. W. J. Biol. Chem. 2000,
275, 21785.
(9) Giannis, A.; Rubsam, F. Angew. Chem., Int. Ed. Engl. 1997, 36, 588.
(10) Eldred, C. D.; Judkins, B. D. Prog. Med. Chem. 1999, 36, 29.
† Department of Chemistry and The Skaggs Institute for Chemical
Biology.
‡ Departments of Immunology and Vascular Biology.
§ Current address: Department of Pediatrics-Whitier Institute, University
of California, San Diego, 9894 Genesee Ave., La Jolla, CA 92037.
(1) Klagsbrun, M.; Moses, M. A. Chem. Biol. 1999, 6, R217.
(2) Folkman, J. Nature Med. 1995, 1, 27.
(3) Folkman, J. New Engl. J. Med. 1971, 285, 1182.
(4) Fan, T.-P. D.; Jaggar, R.; Bicknell, R. Trends Pharmacol. Sci. 1995,
16, 57.
10.1021/ja003579+ CCC: $20.00 © 2001 American Chemical Society
Published on Web 01/26/2001

NoVel Class of Small-Molecule Antiangiogenic Agents
J. Am. Chem. Soc., Vol. 123, No. 7, 2001 1281
Figure 1. Top: Depiction of the MMP2-RVâ3 interaction and its role
in angiogenesis. Bottom: A small-molecule antagonist can disrupt
angiogenesis by inhibiting the localization of MMP2 to the cell surface
through binding to RVâ3.
RVâ3 binds MMP2, and this interaction is a requisite step in
the cellular utilization of the enzyme on the surface of invasive
endothelial cells.12 A noncatalytic, 193 residue C-terminal
fragment of MMP2, termed PEX, was found to inhibit MMP2
binding to RVâ3 and to indirectly block its cell surface
proteolytic activity.13 PEX was further shown to disrupt angio-
genesis and tumor growth in the chick chorioallantoic mem-
brane. A naturally occurring form of PEX can be detected in
vivo in association with cells expressing RVâ3, suggesting its
role as an endogenous regulator of angiogenesis. These observa-
tions indicate that disrupting the binding of MMP2 to integrin
RVâ3 may be a promising approach to controlling angiogenesis
and could lead to novel therapeutic agents for angiogenesis-
dependent diseases including cancer (Figure 1).
The exact nature of the MMP2-RVâ3 interaction is unknown,
although it has been established that the integrin binding site is
distinct from that which recognizes the RGD sequence in
traditional high-affinity ligands (e.g., vitronectin).12 Without
further knowledge of the target structure available, a combina-
torial chemistry14 approach was undertaken in search of small
molecule MMP2-RVâ3 antagonists. Herein we report the
discovery of the first members of a new class of antiangiogenic
compounds that derive their activity by inhibiting this protein-
protein interaction and validate this target as a useful one for
therapeutic intervention.
Figure 2.
studying protein-protein interactions.15,16 This approach allows
for the rapid generation of multimilligram quantities of com-
binatorial mixtures and individual compounds with two or more
binding groups separated by variable linkers that can interact
with one or more protein targets. The screening of such
compound libraries as antagonists of the MMP2-RVâ3 interac-
tion was undertaken to identify lead compounds as potential
angiogenesis inhibitors. As described in the following sections,
this led to the initial identification of 1 (A6B10C4), its
refinement to 19, and its subsequent simplification to provide
34 which also possesses improved solubility and in vivo activity
(Figure 2).
Library Screening Results. Mixture library 1 was found to
have antagonist activity in an assay which measures the in vitro
binding of MMP2 to RVâ3. This library consists of 600
symmetrical dimeric structures which contain three diversity
subunits (Figure 3). The library was synthesized by combining
We have previously described solution phase methodology
designed for the generation of chemical libraries suitable for
(15) Boger, D. L.; Goldberg, J.; Jiang, W.; Chai, W.; Ducray, P.; Lee, J.
K.; Ozer, R. S.; Andersson, C.-M. Bioorg. Med. Chem. 1998, 6, 1347.
(16) Cheng, S.; Comer, D. D.; Williams, J. P.; Boger, D. L. J. Am. Chem.
Soc. 1996, 118, 2567. Boger, D. L.; Tarby, C. M.; Caporale, L. H. J. Am.
Chem. Soc. 1996, 118, 2109. Cheng, S.; Tarby, C. M.; Comer, D. D.;
Williams, J. P.; Caporale, L. H.; Boger, D. L. Bioorg. Med. Chem. 1996,
4, 727. Boger, D. L.; Ducray, P.; Chai, W.; Jiang, W.; Goldberg, J. Bioorg.
Med. Chem. Lett. 1998, 8, 2339. Boger, D. L.; Ozer, R. S.; Andersson,
C.-M. Bioorg. Med. Chem. Lett. 1997, 7, 1903. Boger, D. L.; Chai, W.
Tetrahedron 1998, 54, 3955. Boger, D. L.; Chai, W.; Ozer, R. S.; Andersson,
C.-M. Bioorg. Med. Chem. Lett. 1997, 7, 463.
(11) Brooks, P. C.; Stro¨mblad, S.; Sanders, L. C, von Schalscha, T. L.;
Aimes, R. T.; Stetler-Stevenson, W. G.; Quigley, J. P.; Cheresh, D. A. Cell
1996, 85, 683.
(12) Silletti, S.; Cheresh, D. A. Fibrinolysis Proteolysis 1999, 13, 226.
(13) Brooks, P. C.; Silletti, S.; von Schalscha, T. L.; Friedlander, M.;
Cheresh, D. A. Cell 1998, 92, 391.
(14) Floyd, C. D.; Leblanc, C.; Whittaker, M. Prog. Med. Chem. 1999,
36, 91.

1282 J. Am. Chem. Soc., Vol. 123, No. 7, 2001
Boger et al.
Figure 3. The structure of mixture library 1. The library was
synthesized and screened as mixtures containing single R1 (A1-A6)
and R2 (B1-B10) substitutions, linked by an equimolar mixture of
subunit R3 (C1-C10): 60 mixtures of 10 compounds, see ref 15.
six different R1 amines with 10 different R2 amines through an
iminodiacetic acid linker. Each of these combinations was
dimerized by coupling to a mixture of 10 dicarboxylic acids
(R3) providing 60 mixtures of 10 compounds.17 The screen was
set up to measure the amount of labeled (biotinylated) MMP2
which binds to RVâ3 fixed to a 96-well plate in the presence of
the library mixtures, using high-throughput ELISA methods.18
The mixtures were tested at a concentration of 50 µM (5 µM
per component) in 5% DMSO-aqueous buffer, and the data
are reported as % binding versus control (DMSO) in Figure
4.19 Mixtures containing the N-R-CBZ-lysine methyl ester (B10)
subunit were the most effective at inhibiting MMP2 binding,
especially in combination with the A6 (4-fluorophenethylamine)
group. The six mixtures containing the B10 subunit were re-
screened at lower concentrations (3 and 10 µM), and again the
Figure 4. Binding (% of control) of labeled MMP2 to RVâ3 in the
presence of library 1. Purified RVâ3 was coated onto microtiter plate
wells, which were subsequently blocked with casein and incubated (1.5
h, 37 °C) with biotinylated MMP2 (5 nM) and the library components
(50 µM mixture concentration, 5% DMSO in the testing media). After
removal of unbound enzyme, specific protein binding was quantified
by measuring the activity of an HRP-linked antibiotin antibody as
described in ref 13. Each A and B combination is linked by a mixture
of 10 dicarboxylic acids, C1-C10, see Figure 3. Shorter bars (lower
MMP2 binding) represent higher inhibitory activity. Observed binding
>100% is attributed to the insolubility of some library members.
A6B10 combination was effective. The A1B10 and A2B10
mixtures also showed good activity in this concentration range
(Figure 5). The 30 individual components making up these three
mixtures were prepared and screened in the binding assay at
3 µM. The data for the A6B10 series are presented in Figure 5.
The results revealed that the specific substitution on the benzene
ring in the R1 binding groups (A1, A2, or A6) and the exact
(17) The synthesis of this library is described in detail in ref 15.
(18) This assay is described in ref 13.
(19) Instances of observed binding >100% were attributed to precipitation
due to the poor solubility of some agents at this concentration, and these
results were not further investigated.

NoVel Class of Small-Molecule Antiangiogenic Agents
J. Am. Chem. Soc., Vol. 123, No. 7, 2001 1283
Figure 5. Inhibition of binding by mixtures from library 1 containing
the B10 subunits, and the inhibitory activity of individual compounds
from the deconvolution of the A6B10 mixture. Assays were performed
in triplicate as described for Figure 4; bars are the mean ( standard
error. Compound 1, A6B10C4 was identified as a lead structure for
further investigation.
nature of the linking dicarboxylic acid (R3) were not paramount
to the observed reduction in MMP2 binding, implying that the
B10 substitution as the R2 subunit was key to the antagonist
activity. Despite the greater tolerance for variation in the R1
and R3 substituents, it was observed that the A6 substitution
(4-fluorophenethylamine) led to the best activity and that the
C4 linker (isophthalic acid) was a good choice for future studies
since it was one of the best R3 substitutions in the A6B10 series
and was active in combination with the other R1 groups (A1
and A2, data not shown).
A series of 1, A6B10C4 analogues were prepared (77
individual compounds) on the basis of the library 1 template,
which contain variations in the N-R-CBZ-lysine methyl ester
(B10) subunit.20 All were screened in the binding assay, and
25 key analogues were identified as having interesting activity
and examined more closely (Figures 6 and 7). Of these, the
4-(trifluoromethyl)benzyl carbamate analogue 19 was observed
to be the most active, providing concentration-dependent
inhibition of MMP2 binding. Notably 19 at 3 µM was more
effective than cold MMP2 (200×, 1 µM) at disrupting the
labeled MMP2-RVâ3 binding and roughly 50% as effective at
the equimolar concentration of 1 µM, indicating that 19 is quite
effective as disrupting this protein-protein interaction (Figure
7). In addition, analysis of the B10 analogues in this series
provides insight into the functionality required and further
modifications that could be expected to be tolerated. Significant
changes to the methyl ester group are generally acceptable, and
the primary carboxamide 6 is just as active in the binding assay
as 1, A6B10C4. Interestingly, analogue 25 which lacks the ester
substitution altogether, maintains moderate activity, but com-
Figure 6. Structures of key 1, A6B10C4 analogues screened to
improve binding activity. Each analogue contains the 4-fluorophen-
ethylamine (A6) subunit in position R1 and the isophthalic acid (C4)
linker in position R3, with modifications to B10 in position R2. The
structures of the full set (77) of compounds may be found in Supporting
Information.
pound 23, the enantiomer of 1, A6B10C4, loses all activity.
These results suggested that certain ester or other substitutions
at the lysine carboxylate may influence binding but are not key
to the observed activity and that this site could be modified in
subsequent analogues. Comparisons of 9-22 provide informa-
tion on the impact of modifications to the lysine R-amino group.
Benzoyl amide 9 was inactive in the binding assay and was
selected as a negative control in subsequent studies (see below).
Urea and sulfonamide analogues (12 and 14) were less effective
than 1, A6B10C4. Analogues 15, 16, and 22 have diminished
activity, but the other carbamates display equivalent or improved
activity, most notably the electron poor benzyl carbamate
derivative 19. The importance of the chain length which
(20) The structures of the complete set of library members are provided
in Supporting Information.

1284 J. Am. Chem. Soc., Vol. 123, No. 7, 2001
Boger et al.
Scheme 1
carbonate and subsequent addition to L-Lys(NHBOC)-OCH3,
providing 27. N-BOC deprotection (HCl) and coupling (PyBrOP,
74%) to 2815 provided 29. N-BOC deptrotection (HCl) and
dimerization upon reaction with isophthaloyl dichloride com-
pleted the synthesis and provided 19. A radiolabel was
incorporated by saponification of the two methyl esters (LiOH,
96%) followed by esterification of the resulting dicarboxylic
acid 30 with [14C]-methanol mediated by 1-(3-(dimethylamino)-
propyl)-3-ethylcarbodiimide hydrochloride (EDCI) and catalytic
4-(dimethylamino)pyridine (DMAP) in 35% yield.
A series of compounds was prepared which lacked the R1
subunit present in the library compounds (Scheme 2), and these
proved key to defining the structural features of 19 contributing
to its activity. This was achieved by coupling 27 to N-BOC-
Gly (EDCI, 96%), providing 31. N-BOC deprotection (HCl)
provided the free amine (32, 99%), which was dimerized by
reaction with isophthaloyl dichloride, providing 33 in 68% yield.
The corresponding dicarboxylic acid (34) was prepared by
saponification of the two methyl esters (LiOH, 93%). A
radiolabel was incorporated into compound 34 by repeating its
synthesis utilizing N-BOC-[1-14C]-Gly. Smaller substructure
analogues of 19 were prepared by reacting intermediate 32 with
benzoyl chloride (90%), providing 35. Carboxylic acid 36 was
prepared by treatment of 35 with LiOH in 88% yield. An
additional substructure analogue with functionality designed to
increase aqueous solubility was prepared by coupling 32 with
diglycolic acid, resulting in carboxylic acid 37 in 77% yield.
Similar modifications to the inactive control compound 9 were
made for study where the 4-fluorophenethylamine subunit was
omitted and the methyl esters on the lysine portion were
hydrolyzed to improve aqueous solubility (Scheme 3). The
synthesis was achieved by reacting the free amine of N-ꢀ-BOC-
lysine methyl ester with benzoyl chloride, providing benzoyl
amide 38 in 78% yield. The N-BOC protecting group was
Figure 7. (a) Inhibition of MMP2 binding in the presence of 1,
A6B10C4 and analogues 2-26. Assays were performed as described
for Figure 4. (b) Compound 19 was identified as the most potent agent
which displays concentration-dependent inhibition of binding. Com-
pound 9 was selected as a control compound since it displays no
significant inhibitory activity despite its overall structural similarity.
separates the carbamate from the molecule core can be inferred
from analogues 24-26, where a spacing of 5 methylene units
(25), as with CBZ-Lys, provides greater activity than the
shortened (24) or extended (26) analogues.
Synthesis of 19 and Derivatives. Having identified 19 as
the most potent agent derived from the library screening, the
compound was resynthesized to allow more detailed evaluation
of its biological properties. A series of additional analogues and
partial structures based on 19 were also prepared which defined
the functionality necessary for antagonist activity.
Compound 19 was prepared in three steps starting with
commercially available N-ꢀ-BOC-L-lysine methyl ester (Scheme
1). The carbamate was installed in 99% yield by reaction of
4-(trifluoromethyl)benzyl alcohol with N,N′-disuccinimidyl

NoVel Class of Small-Molecule Antiangiogenic Agents
J. Am. Chem. Soc., Vol. 123, No. 7, 2001 1285
Scheme 2
Figure 8. 14C-labeled 19 binds specifically to RVâ3. Purified RVâ3 (10
µg/mL) was plated onto microtiter plate wells overnight at 4 °C prior
to blocking with casein and incubation with [14C]-19 (3 µM). Bound
agent was competed off the integrin by a 25-fold molar excess of cold
19, but not 9 or the RGD-peptide c(RGDfV). See text for details.
Scheme 3
Figure 9. [14C]-19 binds integrin RVâ3 but has no significant binding
to MMP2 in the same concentration range. Purified proteins were coated
onto microtiter plates prior to blocking with casein and incubation with
[14C]-19.
phobicity). After incubation (at 3 µM) with fixed RVâ3 and
removal of unbound compound, [14C]-19 was found to adhere
to the integrin, as measured by the radioactivity (CPMs) retained
on the protein (Figure 8). Incubation in the presence of a 25-
fold molar excess of cold 19 significantly reduced the amount
of bound agent, whereas incubation in the presence of a 25-
fold molar excess of (cold) control compound 9 did not affect
the binding of [14C]-19. In a similar experiment measuring the
interaction of [14C]-19 to fixed MMP2 (Figure 9), no significant
binding was observed. These results indicate that the antagonist
activity observed in the MMP2-RVâ3 binding assay is derived
from the binding of 19 to RVâ3. The nature of the 19-RVâ3
interaction was shown to be independent of the integrin site
that recognizes the Arg-Gly-Asp sequence, as the high-affinity
cyclic RGD peptide cyclo(Arg-Gly-Asp-D-Phe-Val)21 did not
compete for [14C]-19 binding (Figure 8). Moreover 19 did not
inhibit the binding of vitronectin to RVâ3, further indicating that
the binding site for 19 is distinct from that which binds RGD-
ligands (Figure 10). In an assay which measures the ability of
purified MMP2 to degrade a [3H]-collagen-IV matrix, compound
cleaved (HCl) and the resulting free amine coupled to N-BOC-
Gly (EDCI, 90%). After deprotection, this diamide (39) was
dimerized by reaction with isophthaloyl dichloride, providing
dimethyl ester 40 (62%). Dicarboxylic acid 41 was prepared
by saponification of the two lysine methyl esters (LiOH, 73%).
Biological Properties of 19 and Derivatives. Compound 19
was examined in detail to determine its target protein and to
define its biological properties. Benzoyl amide 9 was selected
as a negative control compound since it was found to lack
antagonist activity in the binding assay, despite its similar overall
structure and physical properties (e.g., solubility and hydro-
(21) Haubner, R.; Gratias, R.; Diefenbach, B.; Goodman, S. L.; Jonczyk,
A.; Kessler, H. J. Am. Chem. Soc. 1996, 118, 7461.

1286 J. Am. Chem. Soc., Vol. 123, No. 7, 2001
Boger et al.
Figure 11. Inhibition of MMP2 binding to RVâ3 by partial structures
27, 29, 33, and 35 (see Schemes 1 and 2 for structures). The assay was
performed in triplicate as described for Figure 4; bars are the mean (
standard error.
Figure 10. Binding of biotinylated MMP2 (left) and vitronectin (VN)
(right) to integrin RVâ3. Compound 19 disrupts the binding of MMP2
but does not affect the binding of the integrin’s endogenous ligand,
vitronectin, which recognizes RVâ3 through its RGD sequence. Com-
pound 9 has no significant inhibitory effect for either protein-protein
interaction. Purified RVâ3 (10 µg/mL, 50 µL/well) was coated onto
microtiter plates and incubated with biotinylated MMP2 and VN in
the presence of 9 and 19. Protein binding was assessed by measuring
the activity of an HRP-conjugated antibiotin antibody, followed by color
development using TMB peroxidase substrate.
19 (at 3 µM) did not have an effect on the enzyme’s proteolytic
activity. This indicates that, at concentrations where binding
inhibition was observed, 19 does not function as a direct enzyme
inhibitor.
These experiments support the expectation that compound
19 may disrupt the ability of tumor cells to utilize MMP2 to
degrade ECM proteins in a manner analogous to that of PEX.
This agent does not interfere with the binding of RVâ3 to its
classical RGD ligands nor does it function as a direct protease
inhibitor. Since it is believed that MMP2 binding to RVâ3 is
necessary for angiogenesis as demonstrated with PEX, the in
vivo effects of 19 on bFGF-induced angiogenesis were examined
in chick embryo and mouse models. Although preliminary
results provided indications of the predicted in vivo activity (data
not shown), the experiments were hampered by the poor physical
properties of the agent, most notably its low aqueous solubility.
Since the primary screen of our libraries indicated that not all
of the functionality of 19 was necessary and that there were
sites on the molecule that could be modified to improve its
properties, we examined a series of substructure analogues and
derivatives of 19 (see Schemes 1-3) that had better solubility
characteristics in search of an agent more suited for in vivo
testing.
Lysine derivative 27, the simplest substructure of 19, was
found to display no measurable inhibition in the MMP2-RVâ3
binding assay, nor did the monomeric iminodiacetic diamide
29 (Figure 11). Some inhibition was observed with 33 at low
concentrations; however at higher concentrations it appeared
to suffer from poor solubility. Since 33 comprises dimerized
product 19 without the R1 (4-fluorophenethylamine) subunit,
its activity further demonstrates that this portion of the molecule
may not be required for biological activity. Amide 35 was almost
as effective as 19, despite consisting of only a portion of the
dimerized structure (Figure 11). These results indicated that the
R1 subunit or a dimerized product displaying two identical
binding groups may not be necessary.
Figure 12. Inhibition of MMP2 binding to RVâ3 by water-soluble
derivatives 30, 32, 34, 36, 37, and 41 (see Schemes 1-3 for structures).
The assay was performed in triplicate as described for Figure 4; bars
are the mean ( standard error. Compound 34 was identified as having
comparable inhibitory activity to that of compound 19.
are shown in Figure 12. Compounds 32 and 37, which lack the
benzoyl (or isophthaloyl) amides present in all previously active
agents, displayed no inhibition. Dicarboxylic acid 30, however,
had some activity, consistent with previously studied analogues
which indicated that the ester was not necessary for binding.
Substructure 34, the dicarboxylic acid derivative of 33, was
found to effectively inhibit MMP2 binding at a level similar to
19, despite lacking the R1 (4-fluorphenethylamine, A6) and
methyl ester substituents. Monocarboxylic acid 36 also appeared
to have moderate activity in the binding assay, although its
activity was weak and variable in subsequent studies. Compound
41, a water-soluble derivative of control compound 9, was
similarly found to lack antagonist activity in the binding assay
and was thus selected to serve as a control for subsequent
experiments.
Compound 34 was selected for detailed study and was found
to have activity comparable to 19 in all the additional in vitro
assays described above. Thus, it was shown to bind RVâ3, but
not MMP2, in a dose-dependent manner and it did not inhibit
the binding of vitronectin to RVâ3. It also displayed minimal
interaction with integrin R5â1, which does not recognize MMP2.
The RVâ3 binding of [14C]-34 is inhibited with cold 34 but not
control compound 41. Compound 34 was further found to block
the ability of RVâ3-transfected melanoma cells to utilize MMP2
and degrade a collagen IV matrix in a â3-dependent manner,
and as was observed with 19, it had no direct inhibitory effect
on purified MMP2 catalytic activity. Compound 34 also
Results for an important set of substructures which contain
further modifications designed to increase aqueous solubility

NoVel Class of Small-Molecule Antiangiogenic Agents
J. Am. Chem. Soc., Vol. 123, No. 7, 2001 1287
the crude hydrochloride salt was suspended in CH2Cl2 (30 mL), treated
with isophthaloyl dichloride (305 mg, 1.5 mmol) and i-Pr2NEt (1.04
mL, 6.0 mmol), and stirred for 4 h at 25 °C. The reaction mixture was
diluted with EtOAc (500 mL) and washed with 10% aqueous HCl (2
× 300 mL) and 5% aqueous Na2CO3 (300 mL), dried (Na2SO4), and
evaporated. Flash chromatography (SiO2, 1:4.5:4.5 MeOH/CH2Cl2/
EtOAc) provided 1.54 g (77%) of 19 as a white powder: [R]25D -5.5
possesses a number of features which made it more suited
for in vivo testing, including better aqueous solubility, a lower
molecular weight, and the lack of labile functionality (i.e., the
methyl ester). Full details of the in vitro properties as well as
in vivo biological activity of 34 and [14C]-34 are described
elsewhere.22 Importantly, it was found to be a potent inhibitor
of angiogenesis in a chick CAM model and led to the near
complete reduction of the growth of solid tumors in vivo,
validating this new target and representing a rare example of a
small molecule therapeutic intervention through disruption of
a specific protein-protein interaction.23,24
1
(c 2.8, CH3OH); H NMR (CD3OD, 400 MHz) δ 7.62 (m, 4H), 7.52
(m, 4H), 7.42 (m, 4H), 7.19 (m, 4H), 6.96 (m, 4H), 5.14 (m, 4H), 4.16
(m, 2H), 4.13 (m, 2H), 4.08 (m, 2H), 3.99 (m, 4H), 3.68 (s, 6H), 3.45-
3.35 (m, 4H), 3.25-3.11 (m, 4H), 2.82-2.70 (m, 4H), 1.82 (m, 2H),
1.69 (m, 2H) 1.60-1.34 (m, 8H); 13C NMR (CD3OD, 100 MHz) δ
174.5, 173.4, 171.1, 170.8, 169.9 (d, J ) 241.1 Hz), 158.2, 142.7, 136.8,
136.2, 131.5, 130.8 (q, J ) 31.8 Hz), 130.0, 129.5, 128.9, 128.8, 126.3,
125.6 (q, J ) 268.7 Hz), 116.0 (d, J ) 20.2 Hz), 66.5, 55.4, 55.3,
53.0, 52.7, 42.0, 40.1, 35.5, 32.0, 29.7, 24.2; IR (film) νmax 3291, 2936,
1725, 1651, 1326 cm-1; FABHRMS (NBA-CsI) m/z 1459.4015 (M +
Cs+, C64H70F8N8O14 requires 1459.3938).
Experimental Section
(S)-Methyl 6-(((tert-Butyloxy)carbonyl)amnio)-2-((4-trifluoro-
methyl)benzyloxycarbonyl)hexanoate (27). A solution of N,N′-
disuccinimidyl carbonate (5.38 g, 21 mmol) in acetonitrile (150 mL)
was treated with 4-(trifluoromethyl)benzyl alcohol (2.87 mL, 21 mmol)
and Et3N (5.8 mL, 42 mmol) and stirred at 25 °C. After 3 h, the reaction
mixture was added to a flask containing N-ꢀ-BOC-lysine methyl ester
(4.2 g, 14 mmol) in acetonitrile and stirred for an additional 3 h. The
solvent was evaporated and the residue dissolved in CH2Cl2 (250 mL)
and washed with 10% aqueous HCl (2 × 200 mL) and saturated
aqueous NaHCO3 (200 mL). Flash chromatography (SiO2, 3:1 CH2Cl2/
EtOAc) provided 6.4 g (99%) of 27 as a pale yellow oil: [R]25D -8.9
N,N′-Bis-[(2-[4-fluorophenyl]ethyl)carboxamidomethyl]-N,N′-bis-
[N-(5-(S)-carboxy-5-[((4-trifluoromethyl)benzyloxycarbonyl)amino]-
pentyl)carboxamidomethyl]benzene-1,3-dicarboxamide (30). A so-
lution of 19 (100 mg, 0.075 mmol) in THF-MeOH (0.8 mL, 1:1) was
treated with LiOH‚H2O (12.6 mg, 0.30 mmol) dissolved in H2O (0.2
mL) and stirred for 2 h at 0 °C. The reaction mixture was quenched by
the addition of 10% aqueous HCl (10 mL) and extracted with EtOAc
(3 × 10 mL). The combined organic layers were washed with saturated
1
(c 5.6, CH3OH); H NMR (CDCl3, 400 MHz) δ 7.57 (d, J ) 8.1 Hz,
aqueous NaCl (10 mL), dried (Na2SO4), and evaporated to afford 95
1
mg (96%) of 30 as a white powder: [R]25 +0.8 (c 5.0, CH3OH); H
2H), 7.39 (d, J ) 8.1 Hz, 2H), 5.70 (d, J ) 7.9 Hz, 1H), 5.13 (m, 2H),
4.71 (m, 1H), 4.28 (m, 1H), 3.67 (s, 3H), 3.03 (m, 2H), 1.78 (m, 1H),
1.64, (m, 1H) 1.46-1.32 (m, 4H) 1.35 (s, 9H); 13C NMR (CDCl3, 100
MHz) δ 172.9, 156.2, 155.8, 140.4, 130.1 (q, J ) 32.0 Hz), 127.8,
125.3, 122.9 (q, J ) 270.0 Hz), 79.1, 65.8, 53.7, 52.3, 39.8, 31.7, 29.5,
D
NMR (DMSO-d6, 400 MHz) δ 12.54 (br s, 2H), 8.63 (m, 1H), 8.43
(m, 2H), 8.30 (m, 1H), 7.74 (m, 4H), 7.69 (m, 2H), 7.57 (m, 4H), 7.40
(m, 4H), 7.24 (m, 4H), 7.09 (m, 4H), 5.15 (m, 4H), 4.14 (m, 2H),
4.02-3.87 (m, 8H), 3.31 (m, 4H), 3.208 (m, 4H), 2.74 (m, 4H), 1.71
(m, 2H), 1.62 (m, 2H) 1.50-1.34 (m, 8H); 13C NMR (CD3OD, 100
MHz) δ 175.8, 173.5, 171.1, 170.8, 162.9 (d, J ) 241.2 Hz), 158.3,
142.8, 136.8, 136.2, 131.5, 130.8 (q, J ) 32.0 Hz), 130.1, 129.6, 128.8,
126.3, 125.6 (q, J ) 270.1 Hz), 116.1 (d, J ) 21.7 Hz), 66.5, 55.3,
55.2, 53.0, 42.0, 40.1, 35.3, 32.1, 29.7, 24.3; IR (film) νmax 3287, 2928,
1705, 1659, 1320 cm-1; MALDIFTMS m/z 1321.4493 (M + Na+,
C62H66F8N8O14 requires 1321.4468).
28.4, 22.2; IR (film) νmax 3357, 2952, 1790, 1745, 1524 cm-1
;
FABHRMS (NBA-NaI) m/z 463.2044 (M + H+, C21H29F3N2O6 requires
463.2056).
(S)-N-((tert-Butyloxy)carbonyl)-N′-((4-fluorophenyl)ethyl)-N′′-(5-
methoxycarbonyl)-5-(4-(trifluoromethyl)benzyloxycarbonyl)amino)-
pentyl)iminodiacetic Acid Diamide (29). A solution of 27 (2.2 g, 4.8
mmol) in CH2Cl2 (3 mL) was treated with 4 N HCl-dioxane (10 mL)
and stirred for 20 min at 25 °C. Solvent and excess acid were removed
under reduced pressure, and the crude hydrochloride salt was dissolved
in DMF (40 mL), treated with N-((tert-butyloxy)carbonyl)-N′-(2-(4-
fluorophenyl)ethyl)iminodiacetic acid monoamide15 (28, 1.68 g, 4.8
mmol), PyBrOP (3.3 g, 7.1 mmol), and i-Pr2NEt (5.0 mL, 29 mmol),
and stirred for 1 h at 25 °C. The reaction mixture was diluted with
EtOAc (400 mL) and washed with 10% aqueous HCl (2 × 300 mL)
and saturated aqueous NaHCO3 (300 mL). Flash chromatography (SiO2,
[14C]-19. A suspension of 30 (1.7 mg, 1.3 µmol) and EDCI (2.0
mg, 10.3 µmol) in DMF (20 µL) was treated with a solution of 14CH3-
OH (American Radiolabeled Chemicals, 0.3 mCi, 57 mCi/mmol, 5.2
µmol) in CH2Cl2 (0.3 mL) and a solution of DMAP (73 µg, 0.6 µmol)
in CH2Cl2 (35 µL) and stirred for 4 h at 0 °C. The reaction mixture
was diluted with EtOAc (3 mL), washed with 10% aqueous HCl (3 ×
3 mL) and saturated aqueous NaHCO3 (3 mL), and dried (Na2SO4).
PTLC (SiO2, 2:3:3 EtOH/CHCl3/EtOAc) provided 0.6 mg (35%) of
[14C]-19 as a white film. This material was identical to the correspond-
1:1 CH2Cl2/EtOAc) provided 2.47 g (74%) of 29 as a white foamy
1
solid: [R]25 -7.1 (c 4.5, CH3OH); H NMR (CDCl3, 400 MHz) δ
1
D
ing unlabeled dimethyl ester (19) by H NMR and TLC. The relative
8.23 and 7.59 (two m, 1H), 7.58 (d, J ) 8.1 Hz, 2H), 7.43 (m, 2H),
7.13 (m, 2H), 7.06 and 6.78 (two m, 1H), 6.94 (m, 2H), 5.70 (dd, J )
12.9 and 8.2 Hz, 1H), 5.11 (m, 2H), 4.31 (m, 1H), 3.85-3.72 (m, 4H),
3.71 (s, 3H), 3.49 (m, 2H), 3.22 (m, 2H), 2.79 (m, 2H), 1.81-1.39 (m,
6H) 1.38 (s, 9H); 13C NMR (CDCl3, 100 MHz) δ 172.8, 170.0, 169.9,
155.8, 154.8, 161.4 (d, J ) 242.7 Hz), 140.2, 134.4, 130.1 (q, J )
33.4 Hz), 130.0, 127.7, 125.3 (q, J ) 3.0 Hz), 123.8 (q, J ) 299.9
Hz), 115.0, 81.2, 65.7, 53.9, 53.8, 53.3, 52.2, 40.8, 38.6, 34.4, 31.5,
activity was approximately 104 mCi/mmol.
(S)-Methyl 6-[2-(((tert-Butyloxy)carbonyl)amino)acetamido]-2-
[(4-trifluoromethyl)benzyloxycarbonyl]hexanoate (31). A solution
of 27 (2.7 g, 5.8 mmol) in CH2Cl2 (3 mL) was treated with 4 N HCl-
dioxane (10 mL) and stirred for 20 min at 25 °C. Solvent and excess
acid were removed under reduced pressure, and the crude hydrochloride
salt was dissolved in DMF (50 mL), treated with N-BOC-Gly (1.0 g,
5.8 mmol), EDCI (1.2 g, 6.4 mmol), and i-Pr2NEt (2.0 mL, 11.6 mmol),
and stirred for 12 h at 25 °C. The reaction mixture was diluted with
EtOAc (400 mL) and washed with 10% aqueous HCl (3 × 250 mL)
and saturated aqueous NaHCO3 (250 mL), dried (Na2SO4), and
28.4, 28.0, 22.4; IR (film) νmax 3267, 2935, 1708, 1657, 1511 cm-1
;
FABHRMS (NBA-CsI) m/z 831.2026 (M + Cs+, C33H42F4N4O8 requires
831.1993).
evaporated to provide 2.89 g (96%) of 31 as a white foamy solid: [R]25
N,N′-Bis-[(2-[4-fluorophenyl]ethyl)carboxamidomethyl]-N,N′-bis-
[N-(5-(S)-(methoxycarbonyl)-5-[((4-trifluoromethyl)benzyloxycar-
bonyl)amino]pentyl)carboxamidomethyl]benzene-1,3-dicarboxam-
ide (19). A solution of 29 (2.10 g, 3.0 mmol) in CH2Cl2 (5 mL) was
treated with 4 N HCl-dioxane (10 mL) and stirred for 0.5 h at 25 °C.
Solvent and excess acid were removed under reduced pressure, and
D
1
-10.4 (c 2.5, CH3OH); H NMR (CDCl3, 400 MHz) δ 7.59 (m, 2H),
7.45 (m, 2H), 6.19 (m, 1H), 5.51 (m, 1H), 5.13 (m, 2H), 4.32 (m, 2H),
3.74 (s, 3H), 3.73 (m, 2H), 3.26 (m, 2H), 1.81-1.39 (m, 6H) 1.44 (s,
9H); 13C NMR (CD3OD, 100 MHz) δ 174.5, 172.2, 158.2, 158.1, 142.7,
130.8 (q, J ) 31.8 Hz), 128.8, 126.3, 125.5 (q, J ) 269.7 Hz), 80.5,
66.5, 55.3, 52.7, 44.6, 39.8, 32.1, 29.8, 28.6, 24.0; IR (film) νmax 3320,
2932, 1721, 1692, 1326 cm-1; FABHRMS (NBA-CsI) m/z 652.1234
(M + Cs+, C23H32F3N3O7 requires 652.1247).
(22) Silletti, S.; Kessler, T.; Goldberg, J.; Boger, D. L.; Cheresh, D. A.
Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 119.
(23) Cochran, A. G. Chem. Biol. 2000, 7, R85.
(24) In vitro cytotoxic activity (L1210 IC50) of key structures: 19, 6 µM;
34, 45 µM; 9, 70 µM; 41, >100 µM (34% inhibition at 100 µM).
(S)-Methyl 6-[2-(Amino)acetamido]-2-[(4-trifluoromethyl)ben-
zyloxycarbonyl]hexanoate Hydrochloride (32). A solution of 31 (350

1288 J. Am. Chem. Soc., Vol. 123, No. 7, 2001
Boger et al.
mg, 0.67 mmol) in CH2Cl2 (2 mL) was treated with 4 N HCl-dioxane
(5.0 mL) and stirred at 25 °C. After 0.5 h, solvent and excess acid
mL) was added to a 4 mL vial containing [1-14C]-N-BOC-Gly (1.5
mg, 0.0085 mmol), treated with i-Pr2NEt (2 µL) and a solution of EDCI
(5.0 mg, 0.026 mmol) in CH2Cl2 (0.1 mL), and stirred for 3 h at 25
°C. The reaction mixture was diluted with EtOAc (2.0 mL), washed
with 10% aqueous HCl (3 × 1.0 mL), saturated aqueous Na2CO3 (1.0
mL), and saturated aqueous NaCl (1.0 mL), and evaporated. PTLC
(SiO2, EtOAc/CHCl3 1:1) provided 1.8 mg (41%) of [14C]-31 as a white
film.
A 4 mL vial containing [14C]-31 (1.8 mg, 3.5 µmol) was treated
with 4 N HCl-dioxane (0.25 mL) and the reaction stirred for 0.5 h at
25 °C. Solvent and excess acid were evaporated under a stream of N2,
and the resulting crude hydrochloride salt was treated with isophthaloyl
dichloride (355 µg, 0.0018 mmol) in CH2Cl2 (0.1 mL) and i-Pr2NEt
(2.4 µL, 0.014 mmol) in CH2Cl2 (0.05 mL). After 3 h at 25 °C, the
reaction mixture was directly purified by PTLC (SiO2, MeOH/CHCl3/
EtOAc 1:9:9) to provide 1.2 mg (71%) of [14C]-33.
A solution of [14C]-33 (1.2 mg, 1.2 µmol) in THF-MeOH (0.2 mL,
1:1) was treated with a solution of LiOH‚H2O (0.4 mg) in H2O (0.05
mL) at 0 °C and stirred for 1 h. The reaction mixture was diluted with
MeOH (1.0 mL), treated with Dowex 50WX-8 acidic cation-exchange
resin (200 mg), and stirred for 1 min. The mixture was then filtered
through cotton wool and the filtrate evaporated to provide 1.1 mg (94%)
of [14C]-34, relative activity approximately 110 mCi/mmol. This
compound and all its synthetic intermediates were identical to the
corresponding unlabeled material.
Solid Phase Integrin Binding Assays. Integrin binding assays were
performed as described previously.13 Briefly, purified integrins were
adsorbed onto microtiter wells (1-5 µg/mL, 50 µL/well) overnight at
4 °C. Plates were washed to remove unbound integrin and blocked
with 1% Caseinblocker (Pierce, Rockford, IL) for 1 h at 37 °C. Purified
biotinylated MMP2 (bMMP2) in binding buffer (50 mM Tris pH 8,
150 mM NaCl, 1 mM MgCl2, 1 mM CaCl2, 0.5 mM MnCl2) was added
to the wells at a concentration of 3-5 nM in the presence or absence
of compounds, or buffer vehicle alone. As a control, blocked wells
without integrin were examined for binding as well. Biotinylated
vitronectin (bVN; 1 µg/mL) was used as a reference for binding assays.
After 90 min of binding at 37 °C, wells were washed extensively with
binding buffer, and bound bMMP2 was detected with an HRP-labeled
anti-biotin mAb. Antibody complexes were visualized with the single
reagent peroxidase substrate TMB (BioRad, Hercules, CA), and the
reaction was stopped with 0.18 M H2SO4. Binding was quantitated on
a Kinetic Microplate Reader (Molecular Devices, Sunnyvale, CA).
For the assessment of direct integrin binding by 19, purified integrin
RVâ3 (10 µg/mL, 50 µL/well) was coated onto Immulon-4 microtiter
wells (Dynatech Laboratories, Inc., Chantilly, VA), which were
subsequently blocked as above and incubated with a titration of [14C]-
labled 19 prior to the addition of a 150 µL binding buffer containing
0.1% Tween-20 and aspiration of all liquid. Dried wells were separated
and immersed in BetaMax liquid scintillation cocktail (ICN, Costa
Mesa, CA) for quantitation. From this binding curve a subsaturating
concentration (3 µM) of [14C]-labeled 19 was chosen for inhibitor
studies in the presence or absence of a 25-fold molar excess of unlabeled
19 or 9, or 1 equiv of cyclic RGD peptide c(RGDfV).21 Biotinylated
vitronectin was used as a control and detected colorimetrically with
HRP-conjugated anti-biotin mAb as described above.
were removed under reduced pressure, providing 300 mg (99%) of 31
1
as a pale yellow oil: [R]25 -10.4 (c 3.0, CH3OH); H NMR (CD3-
D
OD, 400 MHz) δ 7.65 (d, J ) 8.2 Hz, 2H), 7.48 (d, J ) 8.2 Hz, 2H),
5.18 (d, J ) 13.3 Hz, 1H), 5.16 (d, J ) 13.3 Hz, 1H), 4.16 (m, 1H),
3.70 (s, 3H), 3.65 (s, 2H), 3.21 (t, J ) 7.1 Hz, 2H), 1.84 (m, 1H), 1.68
(m, 1H), 1.54 (m, 2H), 1.42 (m, 2H); 13C NMR (CD3OD, 100 MHz) δ
174.5, 167.1, 158.2, 142.8, 130.8 (q, J ) 31.8 Hz), 128.8, 126.3, 125.5
(q, J ) 270.1 Hz), 66.5, 55.4, 52.7, 41.5, 40.2, 32.1, 29.6, 24.1; IR
(film) νmax 3317, 2954, 1718, 1684, 1530, 1327 cm-1; MALDIFTMS
(DHB) m/z 442.1586 (M + Na+, C18H24F3N3O5 requires 442.1566).
N,N′-Bis-[(5-(S)-(methoxycarbonyl)-5-[((4-trifluoromethyl)ben-
zyloxycarbonyl)amino]pentyl)carboxamidomethyl]benzene-1,3-
dicarboxamide (33). A solution of 31 (2.05 g, 4.0 mmol) in CH2Cl2
(3.0 mL) was treated with 4 N HCl-dioxane (10.0 mL) and stirred for
20 min at 25 °C. Solvent and excess acid were removed under reduced
pressure, and the crude hydrochloride salt was suspended in DMF
(40 mL), treated with isophthaloyl dichloride (400 mg, 2.0 mmol) and
i-Pr2NEt (1.4 mL, 8.0 mmol), and stirred for 1 h at 25 °C. The reaction
mixture was diluted with EtOAc (400 mL) and washed with 10%
aqueous HCl (3 × 200 mL) and 5% aqueous Na2CO3 (200 mL), dried
(Na2SO4), and evaporated. Flash chromatography (SiO2, 1:4.5:4.5
MeOH/CH2Cl2/EtOAc) provided 1.30 g (68%) of 33 as a yellow
powder: [R]25D -6.4 (c 2.1, CH3OH); 1H NMR (CDCl3, 400 MHz) δ
8.29 (m, 2H), 8.11 (m, 2H), 7.87 (m, 4H), 7.53 (m, 2H), 7.39 (m, 2H),
6.88 (m, 2H), 5.94 (m, 2H), 5.08 (m, 4H), 4.30 (m, 2H), 4.01 (m, 4H),
3.70 (s, 6H), 3.23 (m, 4H), 1.77 (m, 2H), 1.67 (m, 2H), 1.51 (m, 4H),
1.37 (m, 4H); 13C NMR (CD3OD, 100 MHz) δ 174.6, 171.5,
169.3, 158.3, 142.7, 135.3, 131.6, 130.8 (q, J ) 32.2), 129.8, 128.8,
127.6, 126.3, 125.5 (q, J ) 269.2), 66.5, 55.4, 52.7, 44.1, 40.0, 32.1,
29.8, 24.1; IR (film) νmax 3305, 2951, 1716, 1651, 1538 cm-1
;
FABHRMS (NBA-CsI) m/z 1101.2398 (M + Cs+, C44H50F6N6O12
requires 1101.2445).
N,N′-Bis[(5-(S)-carboxy-5-[((4-trifluoromethyl)benzyloxycarbonyl)-
amino]pentyl)carboxamidomethyl]benzene-1,3-dicarboxamide (34).
A solution of 33 (0.95 g. 0.98 mmol) in THF-MeOH (8.0 mL, 3:1)
was treated with LiOH‚H2O (165 mg, 3.9 mmol) in H2O (2.0 mL) and
stirred at 0 °C. After 2 h, the reaction was quenched by the addition of
10% aqueous HCl (20 mL) and the mixture was extracted with EtOAc
(3 × 50 mL). The organic layers were combined and washed with
saturated aqueous NaCl (50 mL), dried (Na2SO4), and evaporated to
provide 0.86 g (93%) of 34 as a white powder: [R]25 -0.6 (c 3.2,
D
CH3OH); 1H NMR (CD3OD, 400 MHz) δ 8.40 (m, 1H), 8.03 (dd, J )
1.8, 7.8 Hz, 2H), 7.62 (d, J ) 8.1 Hz, 4H), 7.53 (t, J ) 6.1 Hz, 1H),
7.51 (d, J ) 8.1 Hz, 4H), 5.16 (d, J ) 13.3 Hz, 2H), 5.13 (d, J ) 13.3
Hz, 2H), 4.12 (m, 2H), 4.00 (s, 4H), 3.22 (t, J ) 6.6 Hz, 4H), 1.85 (m,
2H), 1.70 (m, 2H), 1.55 (m, 4H), 1.37 (m, 4H); 13C NMR (CD3OD,
100 MHz) δ 175.9, 171.6, 169.4, 158.4, 142.8, 135.4, 131.6, 130.5 (q,
J ) 31.8), 129.8, 128.8, 127.7, 126.3, 125.6 (q, J ) 268.7), 66.5, 53.3,
44.2, 40.1, 32.2, 29.8, 24.2; IR (film) νmax 3334, 2933, 1718, 1646,
1631, 1528 cm-1; MALDIFTMS (DHB) m/z 963.2953 (M + Na+,
C42H46F6N6O12 requires 963.2976).
[14C]-34. A solution of [1-14C]-glycine (American Radiolabeled
Chemicals, 1.0 mCi, 55 mCi/mmol, 0.018 mmol) in 0.1 N HCl was
transferred to a 4 mL vial, and the solvent was removed under a stream
of N2. The resulting residue was treated with a solution of NaHCO3
(4.6 mg, 0.054 mmol) in H2O (0.25 mL) and a solution of di-tert-butyl
dicarbonate (10.5 µL, 0.045 mmol) in THF (0.25 mL) and stirred at
25 °C. After 12 h the solution had become homogeneous and was
diluted with H2O (1.0 mL) and washed with Et2O (2 × 1.0 mL). The
aqueous solution was acidified by the addition of 10% aqueous HCl
(0.5 mL) and extracted with EtOAc (4 × 1.0 mL). The combined
extracts were dried (Na2SO4) and evaporated under a stream of N2 to
provide 2.9 mg (92%) of [1-14C]-N-BOC-glycine as a white film.
A solution of 27 (50 mg, 0.11 mmol) in CH2Cl2 (1 mL) was treated
with 4 N HCl-dioxane (1 mL) and stirred for 1 h at 25 °C. The reaction
mixture was diluted with EtOAc (25 mL) and washed with 10% aqueous
Na2CO3 (25 mL) and saturated aqueous NaCl (25 mL), dried (Na2-
SO4), and evaporated to provide the deprotected lysine as a colorless
film. A portion of this free amine (4.5 mg, 0.013 mmol) in DMF (0.2
Acknowledgment. This work was supported by the National
Institutes of Health (D.L.B.; CA78045 and D.A.C.; CA45726
and CA50286) and the award of predoctoral (J.G.; NDSEG,
1995-1998) and postdoctoral (S.S.; NCI NRSA 1F32 CA72192-
01; T.K.; Mildred-Scheel-Stipedium by Deutsche Krebshilfe)
fellowships.
Supporting Information Available: Synthesis and charac-
terization of compounds 35-41 and the structures of the
complete set (77) of 1, A6B10C4 analogues containing modi-
fications to R2 (1, A6B11C4-A6B87C4). This material is
available free of charge via the Internet at http://pubs.acs.org.
JA003579+