Phosphinic peptide matrix metalloproteinase-9 inhibitors by solid-phase synthesis using a building block approach

Article abstract of DOI:10.1002/(SICI)1521-3765(19991001)5:10<2877::AID-CHEM2877>3.0.CO;2-Z

The solid-phase synthesis of an array of different pseudopeptides containing a phosphinic glycine-leucine moiety (-GΨ{P(O)OH-CH₂}L-)[¹] is described. The resulting pseudopeptides were shown to act as matrix metalloproteinase-9 (MMP-9

Full text of DOI:10.1002/(SICI)1521-3765(19991001)5:10<2877::AID-CHEM2877>3.0.CO;2-Z

FULL PAPER
Phosphinic Peptide Matrix Metalloproteinase-9 Inhibitors by Solid-Phase
Synthesis Using a Building Block Approach
Jens Buchardt,[a] Mercedes Ferreras,[b] Christian Krog-Jensen,[a] Jean-Marie DelaisseÂ,[b]
Niels Tñkker Foged,[b] and Morten Meldal*[a]
Abstract: The solid-phase synthesis of
an array of different pseudopeptides
containing a phosphinic glycine ± leu-
cine moiety (-GY{P(O)OH-CH2}L-)[1]
is described. The resulting pseudopep-
tides were shown to act as matrix metal-
loproteinase-9 (MMP-9) inhibitors.
Starting from available materials, the
protected amino acid isosters benzyloxy-
carbonyl aminomethylphosphinic acid
(glycine analogue) and ethyl a-isobutyl-
acrylate (leucine analogue) were syn-
thesized and coupled with the bis(tri-
methylsilyl)phosphonite. Protective group
interchange yielded a protected phos-
phinic dipeptide building block 1 ready
for use in solid-phase peptide synthesis.
Solid-phase peptide synthesis was per-
formed with 9-fluorenylmethyloxycar-
bonyl (Fmoc) chemistry on a polyeth-
ylene glycol polyamide (PEGA) support
and the coupling of 1 (1.5 equiv) was
carried out with standard activation. The
P4 ± P2 and P2' ± P4' positions of the
pseudopeptides were designed by anal-
ogy of the cleavage sequences of differ-
ent natural extracellular matrix protein
substrates with a synthetic peptide sub-
strate of MMP-9. The crude peptides
were obtained in high yield and purity as
determined by RP-HPLC, and were
characterized by electrospray mass spec-
trometry and amino acid analysis after
purification. Enzyme kinetic investiga-
tions with MMP-9 of the purified pep-
tide inhibitors showed Ki values in the
range from mm to nm.
Keywords: enzyme inhibitors
´
matrix metalloproteinases ´ peptide
isosters ´ solid-phase synthesis
to bone resorption, the migration of osteoclasts[15, 16] and their
anchoring to the extracellular bone matrix is controlled by
MMPs, although their specific roles in these processes as well
as the resorption process still remain to be elucidated.
The evaluation of substrate specificity of MMP-9 towards
small peptide substrates[17] showed that P, L, G/A, and L/I are
preferred in the P3, P2, P1, and P1' subsites, respectively. This
has been utilized in a quenched fluorogenic substrate Mca-
PLG L-Dpa-AR-NH2,[18±20] which has high kcat/KM values for
MMP-1, MMP-2, and MMP-3, in addition to the high activity
to MMP-9. Truncation of the substrate yielded poor kcat/KM
values;[17] thus a good substrate must cover the P3 ± P2'
subsites. Recently, specific peptide substrates for MMP-9
containing P, Hyp, and M or L in the P3, P1, and P1' subsites,
respectively, have also been found by solid-phase combinato-
rial methods.[21] In the case of natural protein substrates, the
specificity of MMP-12 has been proven to be similar.[6] Thus,
MMP-12 cleaves the bone matrix proteins osteopontin (OPN)
and bone sialo protein (BSP) from different species at
conserved sites with G or A, and L or I in the P1 and P1'
subsites, respectively.
Introduction
Matrix metalloproteinases (MMPs),[2, 3] also known as ma-
trixins, are a family of zinc-containing endoproteases involved
in a wide variety of biological processes. These include
embryo development, reproduction, tissue resorption, blood
vessel formation, and wound healing. MMPs are also believed
to play a major role in pathological conditions such as
arthritis, tumor invasion, and metastasis. In particular, the
resorption of organic bone constituents is controlled by
MMPs produced by both bone-forming osteoblasts,[4] and
bone-degrading osteoclasts.[5, 6] Cysteine proteases, such as
cathepsin K, and MMPs,[6, 8±13] in particular MMP-9, MMP-12,
and MMP-14,[14] are produced by osteoclasts.[5, 7, 8] In addition
[a] Prof. Dr. M. Meldal, J. Buchardt, MSc., Dr. C. Krog-Jensen
Carlsberg Laboratory, Department of Chemistry
Gamle Carlsbergvej 10, DK-2500 Valby (Denmark)
Fax : (‡ 45)3327-4708
E-mail: mpm@crc.dk
Â
[b] Dr. N. T. Foged, Dr. J.-M. Delaisse, Dr. M. Ferreras
Department of Basic Research, Center for Clinical
and Basic Research
Ballerup Byvej 222, DK-2750 Ballerup (Denmark)
Fax : (‡ 45)4468-4220
The observed preference of MMP-9 to cleave between the
amino acids G/A and L/I suggests that good inhibitors of
MMP-9 could be obtained by replacing, for example, the G L
cleavage site in a good substrate with a phosphinic -GY-
E-mail: ntf@ccbr.dk
Chem. Eur. J. 1999, 5, No. 10
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0510-2877 $ 17.50+.50/0
2877

M. Meldal et al.
FULL PAPER
{P(O)OH-CH2}L- segment. Phosphinic peptides have previ-
ously been used to prepare protease inhibitors;[22±24] in some
cases the actual G ± L phosphinic segment -GY{P(O)OH-
CH2}L- has been used.[25±27] In these reports the phosphinic part
was placed at the N-terminus of a relatively short sequence of
about two to four amino acids including the phosphinic part.
However, as only relatively long peptides seem to act as good
substrates of MMP-9, longer sequences might be necessary in
order to prepare potent inhibitors. This phenomenon was
reported for MMP-3,[28] where the elongation of the P3 subsite
of a phosphinic peptide increased the affinity.
protected phosphinic dipeptides that meet the existing pep-
tide synthesis protocol requirements with the use of Fmoc-
amino acids.
The specific roles of MMPs during bone resorption,
osteoclast migration, and anchoring still remain to be solved.
Therefore, specific and strong competitive inhibitors of
MMPs are desired as a tool to investigate these processes,
not only under physiological conditions, but also under
pathological conditions such as osteoporosis and bone meta-
stasis. Adapting the strategy of Yiotakis et al.,[52] we set out to
prepare the protected G L pseudodipeptide
1
(see
The inhibitory effect of phosphorus acids on esterases has
been known for a long time.[29] This type of compounds has
been used extensively for the construction of peptide isosters
as competitive protease inhibitors, especially by Bartlett and
co-workers,[27, 30±34] with traditionally three different types of
isosters being used, namely the phosphonamidic, phosphonic,
and phosphinic group. Phosphinic compounds have proven to
be as potent as the other compounds,[25, 27, 35] and in addition
are particularly stable towards hydrolysis.[36] By coordinating
to zinc, the oxy-anion of phosphinic acids mimics the unstable
tetrahedral transition state of the peptide bond cleavage site, a
feature which is crucial to their function as inhibitors.[27, 36] The
phosphinic moiety has been utilized for the construction of
many inhibitors of a variety of enzymes such as MMP-3,[28, 35]
collagenases,[25, 37, 38] other zinc metalloproteases,[22, 26, 27, 39±41]
HIV-protease,[42, 43] other aspartic proteases,[32, 44, 45] ligas-
es,[23, 24, 46, 47] and yet other enzymes.[48±50] Phosphinic peptides
have been prepared both in solution[27, 28, 37, 44] and more
recently on the solid phase.[22, 39, 40, 51±53] The solid-phase syn-
thesis technique[54] has a number of advantages in the
synthesis of sequential oligomeric molecules. Recently, Yio-
takis et al.[52] have developed a strategy for preparing suitably
Scheme 1), and subsequently incorporate the phosphinic
segment into peptides on the solid phase. The sequences of
these pseudopeptides were based on the cleavage sites of
MMP-12 in the bone matrix proteins OPN and BSP,[6] and the
frequently encountered MMP cleavage site in the synthetic
peptide substrate H-Abz-GPLG LY(NO2)AR-NH2.[55] Thus,
replacement of G and L at the cleavage sites in these
substrates by a -GY{P(O)OH-CH2}L- phosphinic segment
would produce putative inhibitors (Figure 1).
O
sequence 2
sequence 1
P
*
N
OH
O
H
Figure 1. Putative phosphinic peptide MMP inhibitor containing the
phosphinic -GY{P(O)OH-CH2}L- segment.
Results and Discussion
Building block synthesis: The key step in the strategy for the
preparation of 1 is the formation of the pseudopeptide bond
by coupling of the amino acid isosters, benzyloxycarbonyl
aminomethyl phosphinic acid (2b) (protected glycine analogue)
and ethyl a-isobutylacrylate (protected leucine analogue).
This is followed by exchange of protective groups (Scheme 1).
Ethyl a-isobutylacrylate was prepared analogous to the
previously described procedure for ethyl acrylates;[56] how-
ever, preparation of the intermediate compound 2a was not
immediately achieved. Neither the method by Baylis et al.[57]
nor the most recently described preparation[58] were successful
in our hands. Therefore, a third procedure[59, 60] was attempted
and proved to be successful. Thus, reaction of ethyl (dieth-
oxymethyl)phosphinate[61, 62] with 1,3,5-(diphenylmethyl)-
hexahydro-s-triazine[57] followed by complete deprotection
with concentrated aqueous HBr gave the desired compound
2a. Since large amounts of the final building block were
needed for combinatorial investigations, 2a was prepared on a
multigram scale, while the synthesis was still efficient (73%
yield). Benzyloxycarbonyl (Cbz) protection of 2a was ach-
ieved by a standard procedure[63] and the product 2b was
obtained in pure form by crystallization, although at this point
the product was difficult to analyze by NMR spectroscopy.
Abstract in Danish: Fastfase syntese af et antal forskellige
pseudopeptider indeholdende en glycin-leucin phosphinsyredel
(-GY{P(O)OH-CH2}L-[1]) er beskrevet. Pseudopeptiderne
viste sig at vñre effektive inhibitorer af matrix metalloprotei-
nase 9. De beskyttede aminosyreanaloger, benzyloxycarbonyl
aminomethyl phosphinsyre (glycin-analog) og ethyl a-isobu-
tylakrylat (leucin-analog), blev syntetiseret udfra kommercielt
tilgñngelige stoffer og koblet via den til phosphinsyren
svarende bis(trimethylsilyl)phosphonit. Efter udskiftning af
Ê
beskyttelsesgrupper opnaedes en phosphinsyre dipeptid bygge-
blok 1 designet til brug ved fastfase peptidsyntese. Fastfase
Ê
peptidsyntesen blev udfùrt vha. Fmoc-kemi pa en PEGA-resin
og koblingen af byggeblokken 1 (1.5 equiv) blev foretaget med
en gñngs aktiveringsmetode. Aminosyrerne i P4 ± P2 og P2' ± P4'
Ê
positionerne i pseudopeptiderne var udvalgt saledes, at de
Ê
opnaede pseudopeptider lignede MMP-9 substrater, heriblandt
Ê
bade matrixprotein substrater og et syntetisk peptidsubstrat.
Ê
Raprodukterne fra peptidsyntesen havde ifùlge RP-HPLC hùj
renhed og blev isoleret i hùjt udbytte. Efter oprensning vha.
preparativ HPLC blev peptiderne karakteriseret ved elektro-
spray massespektrometri og aminosyre analyse. Enzymkine-
tiske undersùgelser af de oprensede peptid-inhibitorer med
1
Both H and 13C NMR spectra recorded in organic solvents
were impossible to assign; the proton spectrum contained
only broad signals with few visible couplings. However, when
acquired in ꢀ0.3% NaOD in D2O, a useful spectrum with
Ê
MMP-9 gav Ki vñrdier i omradet fra mm til nm.
2878
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0510-2878 $ 17.50+.50/0
Chem. Eur. J. 1999, 5, No. 10

MMP-9 Inhibitors
2877±2884
more problematic. Work-up of
the complex product mixture
was difficult and the yield de-
creased to 35%. Thus, this re-
action did not seem suitable for
large-scale synthesis particular-
ly considering the amount of
expensive silver reagent re-
quired. Alternative procedures
for esterification were there-
fore considered (Scheme 2).
Esterification of the phosphinic
O
O
P
EtO
EtO
i)
H
P H
H
OH
OEt
O
P
R1
H
iii), iv)
N
H
OH
Ph
Ph
Ph
Ph
2a: R1=H
ii)
N
NH2
v)
2b: R1=Cbz
Ph
N
N
Ph
Ph
Ph
vi), vii)
acid
with
1-adamantanol
(AdOH) mediated by ethyl-
(dimethylamino)-propylcarbo-
diimide (EDC)[66] was not suc-
cessful. However, reaction of
the corresponding phosphinic
acid chloride (3') (generated
cleanly in situ with (COCl)2/
DMF[67] as confirmed by 31P
NMR spectroscopy) with so-
dium 1-adamantanolate in
THF at 608C gave 4 in 61%
yield. The moderate yield is
O
O
P
x)
O
R1
N
P
R3
OH
*
*
Fmoc
*
N
H
O
R2
O
H
OAd
O
3: R1=Cbz, R2=H, R3=Et
4: R1=Cbz, R2=Ad, R3=Et
R1=Cbz, R2=Ad, R3=OH
1
viii)
ix)
Scheme 1. Preparation of building block 1, suitable for the solid-phase synthesis of phosphinic peptides
containing the phosphinic -GY{P(O)OH-CH2}L- moiety: i) CH(OEt)3, RT, 55% (distilled); ii) aqueous CH2 O,
cat. KOH, 908C, 87%; iii) toluene, 1008C; iv) 48% aqueous HBr, D, 73%; v) Cbz-Cl, K2CO3, H2O, RT, 72%;
vi) HMDS, 1108C; vii) CH2 CH(iBu)-CO2Et, 908C, 96%; viii) AdBr, Ag2O, CHCl3, D, 89%; ix) NaOH (EtOH),
ˆ
ˆ
RT; x) H2/Pd(C), Fmoc-OSu, NaHCO3, MeOH/EtOAc/H2O, RT, 65%.
most probably a result of the fact that both the phosphinic
acid and the nucleophile (AdO ) are very sterically hindered.
Thus, an SN1-type reaction must be considered as the most
effective reaction type for the esterification of phosphinic
acids with a tertiary alcohol. Therefore, the reaction of
sharp peaks was obtained and the significant double triplet
from the phosphorus-bonded proton was clearly visible. This
phenomenon has not previously been reported for other
similar phosphinic acids. The broad peaks persisted until the
phosphinic acid had been protected, suggesting that aggrega-
tion or micelle formation occurred as a result of hydrogen
bonding in the amphiphilic compound (phosphinic acids are
known to form hydrogen-bonded dimers in nonpolar sol-
vents).[64]
O
OEt
*
Cbz
P
Conversion of 2b into 3 was performed by Michael addition
of the phosphinate to ethyl a-isobutylacrylate.[52, 65] Reaction
of 2b with five equivalents of hexamethyldisilazane (HMDS)
produced the corresponding bis(trimethylsilyl)phosphonite in
situ, and subsequent addition to the acrylate yielded the
pseudodipeptide 3. The NMR spectral analysis of 3 was
rendered difficult for the same reason as for 2b. The problems
were overcome by recording the spectra in aqueous base. The
basic conditions gave rise to a partial hydrolysis of the
carboxylic ester as seen by slow disappearance of a broad
signal at d ˆ 4.12 (ethyl ester) and formation of a sharp
quartet at d ˆ 3.64 (EtOD). Similarly, in the 31P NMR
spectrum, a peak at d ˆ 36.5 slowly replaced the peak at d ˆ
34.5, when the sample was left at room temperature. The
crude product was collected in 96% yield and was essentially
pure according to analysis by thin-layer chromatography
(TLC) and NMR spectroscopy. However, the analysis of the
product by electrospray mass spectrometry (ES-MS) was
difficult because of the low intensity of the molecular ion.
A previously suggested procedure for the subsequent AgI-
mediated protection of the phosphinic acid with 1-adamantyl
bromide (AdBr)[52] was not reproducible in our hands.
However, rearranging the order of reagent addition gave a
satisfactory reaction (yield of 4 89%) on a medium scale.
Performing this reaction on a larger scale (74.5 mmol) was
*
N
OAd
O
H
4
AdOH/
EDC,
0%
AdBr/
Ag2O
89%
or
or
AdO-C(NH)-CCl3/
Lewis acid, 0%
O
P
OEt
*
Cbz
N
H
OH
O
3
(COCl)2/DMF
100%
AdO-Na+
61%
O
P
OEt
4
*
Cbz
*
N
Cl
O
H
3'
Scheme 2. Reaction conditions investigated to improve the phosphinic
acid esterification in dipeptide analogue 3.
Chem. Eur. J. 1999, 5, No. 10
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0510-2879 $ 17.50+.50/0
2879

M. Meldal et al.
FULL PAPER
1-adamantyl trichloroacetimidate with 3 catalyzed by differ-
ent acids and Lewis acids was investigated. However, only
trace amounts of 4 were formed in this reaction. 1-Adamantyl
trichloroacetimidate was prepared in 96% yield by a standard
procedure.[68] These observations indicate that the AgI-
mediated SN1-type reaction with 1-adamantyl bromide is a
superior method for esterification of the phosphinic di-
peptide 3.
Both mass spectrometric and NMR spectral analysis of the
compound were possible after the esterification of the
phosphinic acid. However, the esterification produced a chiral
center at the phosphinic acid[27, 52] resulting in a mixture of two
diastereomeric sets of enantiomers. This complicated the
analysis of 1H and 13C NMR spectra, and gave rise to two sets
of resonances in the 13C and 31P NMR spectra.
was found that the procedure[52] with 1-adamantyl bromide
and Ag2O was superior to other methods.
Phosphinic peptide synthesis: Phosphinic peptides 5a/b ± 11a/
b (Table 1) were prepared in a multiple-column peptide
synthesizer[70] by using a standard peptide synthesis[71] proto-
col on the PEGA[72] (polyethylene glycol polyamide)-resin.
The couplings of building block 1 performed well in the Fmoc-
peptide synthesis approach. The acid lability of the phosphinic
adamantyl ester is comparable to that of normal tBu-
protected amino acids, and the protected derivatives can
therefore be converted into the phosphinic acid by the
trifluoroacetic acid (TFA) treatment, which simultaneously
cleaves the peptide side chain protective groups. Coupling of 1
was performed with 1.5 equiv after activation with O-(benzo-
triazol-1-yl)-N,N,N',N'-tetramethyluronium tetrafluoroborate
(TBTU)[73] and N-ethylmorpholine.
The peptides contained the racemic building block 1, and
each peptide was therefore a mixture of two diastereomers
(indicated by a and b). This was characteristically observed in
analytical HPLC as two closely eluting components of equal
peak intensity. The impurities, which were minor, also
frequently appeared as two close peaks. The representative
analytical HPLC chromatogram of the crude peptides 6a and
b in Figure 2 shows the high purity of the crude pseudopep-
tides. The individual diastereomers were separated on a
semipreparative column. Not all diasteromers could be
completely separated to the baseline. However, from a
synthetic point of view a troublesome resolution of the
valuable building block prior to peptide assembly would not
be sensible, since the diastereomers of the final peptide
analogues could all be separated and isolated in good yield.
The properties and inhibitory activities of the phosphinic
peptides 5a/b ± 11a/b are presented in Table 2. The sequences
of the inhibitors 5a/b ± 10a/b were selected from cleavage
sites in proteins known to be natural substrates of MMP-12
and MMP-14.[6] These proteins, bone sialo protein (BSP) and
osteopontin (OPN), are noncollagenous proteins present in
the extracellular matrix of the bone. Both MMP-12
and MMP-14 cleave bovine OPN selectively at
Carboxylic ester hydrolysis of 4 was followed by the
exchange of Cbz at the a-amino group for Fmoc (9-
fluorenylmethyloxycarbonyl), which has previously been
performed on other phosphinic dipeptides by a two-step
procedure.[52] However, the desired product 1 was not
observed when following the reported method. Although it
has previously been indicated that the phosphinic adamantyl
ester is unstable to standard hydrogenolytic conditions,[52]
a
one-pot procedure could be developed in which 9-fluorenyl-
methyl succinimidyl carbonate (Fmoc-OSu) was added di-
rectly to the hydrogenation mixture containing NaHCO3 as a
basic buffer. The Fmoc group is known to be rather stable
towards hydrogenolytic conditions.[69] However, prolonged
exposure of the mixture to hydrogen at atmospheric pressure
resulted in decomposition of the formed Fmoc-protected
dipeptide analogue 1. Therefore, simultaneous hydrogenolysis
and Fmoc protection were performed for 80 min, leading to
almost complete removal of the Cbz group and a reasonable
yield (65%) of 1 on a millimolar scale. A tenfold scale-up of
this procedure resulted in lower yield (41%), plus the
formation of other by-products.
In summary, a straight-forward procedure for preparation
of a protected glycine ± leucine phosphinic dipeptide building
block 1 was developed. For a difficult step in this synthesis, the
protection of the phosphinic acid as its 1-adamantyl ester, it
Table 1. Structures of the phosphinic peptide inhibitors 5a/b ± 11a/b containing the -GY{P(O)OH-CH2}L- phosphinic moiety.
Sequence[a]
P1 ± P1'
Amino acid analysis[b]
P5
P4
P3
P2
P2'
P3'
P4'
P5'
A
G
K
L
P
Q[c]
R
S
V
Y
5a
5b
6a
6b
7a
7b
8a
V
V
A
A
A
A
V
V
L
L
L
L
L
L
P
Y
Y
Y
Y
Y
Y
A
A
A
A
A
A
L
GY{PO2H-CH2}L
GY{PO2H-CH2}L
GY{PO2H-CH2}L
GY{PO2H-CH2}L
GY{PO2H-CH2}L
GY{PO2H-CH2}L
GY{PO2H-CH2}L
GY{PO2H-CH2}L
GY{PO2H-CH2}L
GY{PO2H-CH2}L
GY{PO2H-CH2}L
GY{PO2H-CH2}L
GY{PO2H-CH2}L
GY{PO2H-CH2}L
K
K
K
K
R
R
W
W
W
W
Q
Q
Y
Y
S
S
S
S
S
S
L
L
L
L
L
L
A
A
R
R
G
G
G
G
P
G
G
1.01
1.01
1.00
1.01
±
1.02
1.02
1.03
1.02
1.02
1.03
1.99
2.00
1.03
1.02
1.01
1.01
2.00
2.00
1.02
1.00
1.02
1.02
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
2.04
2.06
2.00
1.99
1.99
1.99
1.06
1.08
±
±
±
±
±
±
0.98
0.95
±
±
±
±
0.95
0.94
±
±
±
±
±
±
±
±
±
±
1.01
1.02
±
1.04
1.04
±
0.94
0.94
0.94
0.94
0.94
0.96
±
±
±
±
±
0.99
1.00
±
±
0.99
0.97
±
±
±
±
±
±
±
0.99
0.97
1.01
1.01
0.98
0.98
-
-
±
±
±
±
1.06
1.05
±
±
±
±
±
±
1.05
1.03
±
G
G
G
G
0.98
0.98
0.97
0.99
0.99
0.99
1.97
1.99
8b
9a
P
G
G
G
G
R
R
9b
10a
10b
11a
11b
±
±
±
±
0.97
0.97
A
A
G
G
G
G
P
L
±
±
±
[a] All peptides are of the type H-phosphinic peptide-OH (as depicted in Figure 1); the absolute configuration of the -GY{PO2H-CH2}L- moiety was not
determined. [b] W-residues were not detected because the samples were analyzed with aqueous 6m HCl which results in partial decomposition of the indole
ring. [c] Q-residues were detected as E-residues.
2880
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0510-2880 $ 17.50+.50/0
Chem. Eur. J. 1999, 5, No. 10

MMP-9 Inhibitors
2877±2884
The individual diastereomers displayed up to two orders of
magnitude difference in inhibitory activities towards MMP-9.
This is expected as a result of the strict stereochemical
requirements in the active site of the enzyme. Noteworthy, for
all peptides a tendency was observed for the later eluting
compound to give the highest Ki values; this indicates the
same order of elution of the seven diastereomeric pairs.
However, this conclusion would require independent proof by
the structures. It does, however, suggest that there is a
correlation between the overall polarity of the compound and
its potency as an inhibitor. When comparing similar sequences
(such as 5a/b and 6a/b, or 8a/b and 9a/b), there is no clear
tendency for the longer sequence to constitute the most
effective inhibitor: 5a/b is one order of magnitude more
potent than 6a/b, while there is no significant difference in Ki
value between the pairs 8a/b and 9a/b.
Figure 2. Analytical HPLC chromatogram of the crude phosphinic peptide
inhibitor 6.
The substrate specificity of MMPs is determined not only
by the amino acid sequence in the vicinity of the cleavage site.
There is evidence that MMP-2, a smaller gelatinase, contains
two substrate recognition sites, one at the active site, control-
ling the specificity of small peptide substrates, and one further
away from the active site, which is responsible for the
specificity of large protein substrates such as gelatin and
collagen. This is also true for MMP-9, known to bind to
extracellular bone matrix proteins with the hemopexin
domain or the fibronectin-like insert, and this may account
for the lower activity of the inhibitors based on the MMP
cleavage site of protein substrates 5a/b ± 10a/b compared with
the peptide substrates 11a/b.
Table 2. Characteristics of the phosphinic peptide inhibitors 5a/b ± 11a/b
containing the -GY{P(O)OH-CH2}L- phosphinic moiety.
Compound MW
Found ES-MS Rf value Yield in SPPS[b]
‡
[M ‡ H]
[min][a]
[mg] ![%] Ki [mm]
5a
5b
6a
6b
7a
7b
8a
8b
9a
985.1
985.6
985.5
730.6
730.4
786.1
786.3
918.5
918.5
764.4
764.5
706.4
706.4
20.0
21.5
19.0
20.0
20.0
21.0
23.0
25.0
29.0
31.0
23.0
25.5
25.0
25.5
8 !24
8 !24
6 !24
7 !28
5 !19
7 !26
9 !29
7 !22
6 !23
5 !19
6 !25
5 !21
5 !15
7 !20
0.29
7.8
3.5
70
0.68
16
17
390
11
150
29
8.8
0.0006
0.0034
985.1
729.8
729.8
785.8
785.8
918.1
918.1
763.9
763.9
705.8
705.8
9b
10a
10b
11a
11b
1009.1 1009.5
1009.1 1009.4
Conclusion
[a] Each HPLC analysis was carried out by going from 0 to 100% B-buffer
in 50 min, i. e., a gradient of 2% Bmin (see Experimental Section).
[b] Yields are based on the loading of the resin measured after the first
coupling.
The synthesis of phosphinic pseudopeptides containing the
-GY{P(O)OH-CH2}L- phosphinic part was described. These
were prepared by standard solid-phase peptide synthesis using
a protected phosphinic dipeptide building block 1 to introduce
the phosphinic moiety. The building block was prepared by a
useful, but somewhat lengthy procedure starting from avail-
able materials. Although analysis of the intermediate phos-
phinic acids was difficult, the procedure gave access to the
final building block 1 well designed for the solid-phase
peptide synthesis by the Fmoc-based strategy. Thus, coupling
of a small excess of the building block in a standard Fmoc-
based peptide synthesis with conventional coupling proce-
dures in a multiple-column synthesizer provided 14 inhibitors
as seven diastereomeric pairs, which could easily be sepa-
rated. The crude pseudopeptides obtained from the solid-
phase synthesis were of high purity according to HPLC
analysis. They proved to be moderate to excellent inhibitors of
MMP-9. In fact, one of the inhibitors (11a) has a Ki value
which is in the range of the most potent existing inhibitors for
MMP-9. Finally, this work is a further example, in which the
replacement of the amino acids in the P1 and P1' subsites of a
good protease substrate with a phosphinic dipeptide may
result in an efficient inhibitor of the same protease.
1
VAYG159 L160 KSR which is analogous to the human OPN
sequence, VVYG159 L160RSK. Similarly, both proteases
cleave bovine BSP selectively at GLAA134 I135WLP and
human BSP at GLAA129 I130QLP.[6] Since MMPs generally
share similar substrate specificities, it was assumed that these
substrate sequences for MMP-12 and MMP-14 could be used
to construct potential inhibitors of MMP-9. Thus, the G ± L
and A ± I cleavage sites of OPN and BSP, respectively, were
replaced with the -GY{P(O)OH-CH2}L- phosphinic part to
obtain inhibitors. As seen in Table 2, the Ki values are
generally in the mm range, and thus considerably higher than
the values for the inhibitors derived from the small peptide
substrates 11a/b. These inhibitors were based on a sequence
derived from a quenched fluorogenic MMP substrate,[18]
H-Abz-GPLG LY(NO2)AR-NH2, which has previously been
1
used in our investigations (kcat/KM ˆ 1.5  105 m 1 s
for
MMP-9). Replacement of the G- and L-residues at the
cleavage site by a -GY{P(O)OH-CH2}L- phosphinic part gave
the very effective inhibitors 11a/b. In fact, 11a (Ki ˆ 0.6nm for
MMP-9) is almost as effective as the well known highly potent
hydroxamate/sulfide-based general MMP inhibitor BB-94[74]
(Ki ˆ 0.05nm for MMP-9).
Current work in our laboratory is aiming at the extension of
this method to the synthesis of pseudopeptide split-and-
combine solid-phase inhibitor libraries with various phos-
Chem. Eur. J. 1999, 5, No. 10
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0510-2881 $ 17.50+.50/0
2881

M. Meldal et al.
FULL PAPER
P-(Benzyloxycarbonylaminomethyl)-P-(ethyl 2-isobutylpropionate-3-yl)-
phosphinic acid (3): A mixture of 2b (5.00 g, 21.8 mmol) and HMDS
(23 mL, 109 mmol, 5 equiv) was stirred under Ar in a dried 250 mL round-
bottom flask. The flask was heated to 1158C, where 2b melted and then
reprecipitated forming a suspension in HMDS. The reaction was accom-
panied by the liberation of NH3. The temperature was maintained at 1158C
for 2 h after which time the temperature was lowered to 958C within
35 min. Ethyl a-isobutylacrylate (4.76 mL, 28.3 mmol, 1.3 equiv) was added
dropwise to the opaque mixture within 35 min. After stirring for 3.5 h, the
mixture became clear and colorless. The temperature was lowered to 708C
and EtOH (65 mL) was carefully added. Cooling to room temperature and
subsequent concentration in vacuo yielded a white solid, which was
redissolved in ethyl acetate (570 mL) and washed with HCl (50 mL, 1m).
The organic phase was separated and the aqueous phase was extracted with
ethyl acetate (3 Â 25 mL). The combined organic extracts were washed with
water (2 Â 50 mL), saturated brine (50 mL), dried with Na2SO4, and
concentrated to give a white solid, which was dried at high vacuum for 4 h.
Yield 8.03 g (96%). The product could also be precipitated in chloroform
by addition of light petroleum. However, no significant increase in purity
was observed. The crude product, which was pure according to TLC (Rf ˆ
0.21, chloroform/methanol 5:1 ‡ 5% acetic acid) and NMR spectroscopy,
was characterized by 1D and 2D NMR (1H and 1H COSY NMR). 1H NMR
(500 MHz, 0.3% NaOD in MeOD/C5D5N 1:1, 258C): d ˆ 0.81 (m, 3H), 0.86
(m, 3H), 1.15 (t, 3J(H,H) ˆ 7.5 Hz, 3H), 1.42 (m, 1H), 1.49 (m, 1H), 1.56 (m,
1H), 1.79 (m, 1H), 2.10 (m, 1H), 3.05 (m, 1H), 3.40 (m, 1H), 3.50 (m, 1H),
4.12 (m, 1H), 5.80 (m, 1H); NH was not visible due to H/D exchange. The
aromatic signals were superimposed with residual C5H5N in the deuterated
pyridine, but were visible in another solvent mixture (250 MHz, 0.3%
NaOD in D2O/MeOD 1:1, 258C) at d ˆ 7.50 (s, 5H); 31P NMR (250 MHz,
1% NaOD in D2O/MeOD 1:1, 258C): d ˆ 34.5; 13C NMR spectral data
could not be obtained due to low solubility in aqueous NaOD; ES-MS m/z
phinic acid isosteric dipeptides as building blocks in combi-
nation with substrate indicators.[76]
Experimental Section
Abbreviations: AMPA: 4-aminophenylmercuric acetate, DCM: dichloro-
methane, DMF: N,N-dimethyl formamide, DIPEA: diisopropylethyl-
amine, HMBA: hydroxymethylbenzoic acid, MSNT: 2,4,6-mesitylenesul-
fonyl-3-nitro-1,2,4-triazole, Dhbt-OH: (3-hydroxy-4-oxo-3,4-dihydro-1,2,3-
benzotriazine), VLC: vacuum liquid chromatography. The IUPAC recom-
mendations for amino acid notation with one letter codes has been
employed, see http://www.chem.qmw.ac.uk/iupac.
General procedures: Anhydrous solvents were obtained by storing
analytical quality solvents over 3 or 4 Š activated molecular sieves; the
water content was then verified to be below 30 ppm by Karl Fischer
titration, except for DMF which was fractionally distilled and stored over
4 Š molecular sieves. PEGA-resin was purchased from Polymer Labora-
tories, England. All other synthetic starting materials were purchased from
Fluka, Aldrich, NovaBiochem, or Bachem and used without further
purification. NMR data were acquired on a Bruker Avance DRX 250 or
Varian 500 MHz Unity Inova spectrometer and were referenced to CHCl3
(d ˆ 7.24, 1H and d ˆ 77.0, 13C), HDO (d ˆ 4.72, 1H), MeOD (d ˆ 3.30, 1H),
or 85% aqueous H3PO4 (d ˆ 0, 31P). Electrospray mass spectra were
obtained on a Fisons VG Quattro 5098 mass spectrometer (mobile phase
50% aq. MeCN, 8 mLmin 1, sample: 10 mL ꢀ20 pmolmL 1). Melting points
were measured on a Büchi B-540 and were uncorrected. VLC[77] was
performed with a tightly packed column of Merck silica gel 60H; TLC
plates used were Merck silica gel 60F254 on aluminum. TLC plates were
visualized with a UV lamp, and developed with the AMC reagent (21 g
(NH4)6Mo7O24, 1 g Ce(SO4)2, 31 mL H2SO4, 500 mL water) or ninhydrin
spray (5 g ninhydrin in 100 mL EtOH). Analytical HPLC was performed
on a Waters system (490E detector, two 510 pumps with gradient controller
and Æ 8 mm RCM C18 column); the semipreparative purification of peptide
analogues 5a/b ± 11a/b was carried out on a Merck-Hitachi system (L-4250
UV/Vis detector and L-6250 pump) connected to a Waters Æ 25 mm RCM
C18-column. All RP-HPLC procedures were carried out with a linear
gradient system consisting of two buffers: A (0.1% TFA in water) and B
(0.1% TFA in MeCN/H2O 9:1). A Kontron SFM-25 spectrofluorometer
used for inhibitor assays.
Aminomethylphosphinic acid (2a):[78] Phosphinic acid 2a was efficiently
prepared analogous to a literature procedure[59] on a 300 mmol scale,
although the pure amino acid could not be successfully precipitated with
propylene oxide as previously reported.[57±60] Therefore, the aqueous
solution (ꢀ300 mL) of the crude hydrobromide was concentrated to
dryness with silica gel (ꢀ200 g), split in two portions, each portion placed
on a short VLC column (Æ 10 cm, L 10 cm, 250 mL fractions) and purified
with acetone (2 to 4 fractions) and acetic acid/water/acetone/isopropyl
alcohol 2:2:8:5 (15 to 20 fractions). The product was obtained in 73% yield
with this procedure, which was pure according to the 1H and 31P NMR
spectra,[57] but still contained 16% hydrobromide as determined by
elemental analysis.
‡
(%): 386.1 [M ‡ H] ; calcd for C18H28NO6P: 385.2.
O-Adamantyl P-(benzyloxycarbonyl-aminomethyl)-P-(ethyl 2-isobutyl-
propionate-3-yl)phosphinate (4):
Procedure A [from P-(benzyloxycarbonylaminomethyl)-P-(ethyl 2-isobu-
tylpropionate-3-yl)phosphinic acid (3)]: A mixture of phosphinic acid 3
(820 mg, 2.12 mmol) and Ag2O (986 mg, 4.25 mmol, 2 equiv) was refluxed
under Ar in dry chloroform (3 mL) for 15 min, after which a solution of
AdBr (503 mg, 2.34 mmol, 1.1 equiv) was added dropwise to the refluxing
suspension over a period of 30 min. Reflux was continued for 1 h and the
mixture was stirred overnight. The crude mixture was filtered through
celite, concentrated, and purified by VLC with toluene/ethyl acetate 3:1 as
eluent to afford 4 as a highly viscous syrup (977 mg, 89%). When pure, the
compound crystallized as a white solid. TLC: Rf ˆ 0.40 (toluene/ethyl
acetate 1:3).[79] The 1H NMR spectrum was assigned with 1H COSY and
13C NMR with 1H ± 13C COSY experiments. 1H NMR (250 MHz, CDCl3,
3
3
258C): d ˆ 0.86 (d, J(H,H) ˆ 7.0 Hz, 3H), 0.92 (d, J(H,H) ˆ 7.0 Hz, 3H),
1.22 (t, 3J(H,H) ˆ 7.0 Hz, 3H), 1.22 (t, 3J(H,H) ˆ 7.0 Hz, 3H), 1.33 (m, 1H),
1.49 (m, 1H), 1.56 (m, 1H), 1.58 (brs, 6H), 1.72 (m, 1H), 2.01 (brs, 6H),
2.23 (brs, 3H), 2.18 (m, 1H), 2.80 (m, 1H), 3.50 (brm, 2H), 4.10 (m, 2H),
5.10 (s, 2H), 5.30 (m, 1H), 7.31 (s, 5H); 13C NMR (250 MHz, CDCl3, 258C):
d ˆ 14.1, 22.3 (2d), 25.8, 31.1, 31.4 (2d), 35.6, 37.4 (d), 41.4 (2d), 43.4 (2d),
44.7 (d), 60.6 (d), 67.6, 83.3, 128.1 ± 130.0; 31P NMR (250 MHz, CDCl3,
‡
258C): d ˆ 44.1 and 44.9; ES-MS m/z (%): 520.5 (90) [M ‡ H] , 542.4 (100)
Benzyloxycarbonyl aminomethylphosphinic acid (2b): Phosphinic acid 2b
was prepared from 2a analogous to a standard procedure.[63] Compound 2a
(5.00 g, 52.6 mmol) was dissolved in water with K2CO3 (17.0 g, 123 mmol,
2.34 equiv), and Cbz-Cl (95%, 10.3 mL, 68.4 mmol, 1.3 equiv) was added
dropwise at 08C. After the solution was stirred at RT for 19 h, further Cbz-
Cl (2 Â 0.5 equiv) was added and the solution was stirred for 1 h until
complete consumption of 2a was observed. The aqueous phase was diluted
with water (50 mL) and washed with ethyl acetate (3 Â 50 mL) and toluene
(30 mL). It was then mixed with ice (150 g) and acidified with concentrated
aqueous HCl (21.8 mL, 5 equiv). The product was dissolved in ethyl acetate
(300 mL) and the aqueous phase was extracted with ethyl acetate (2 Â
150 mL). The organic phase was dried with MgSO4 and concentrated to
yield a white solid (8.68 g, 72%). The crude product was crystallized from
ethyl acetate/light petroleum to afford white crystals in 63% yield. M.p.
‡
‡
[M ‡ Na] , 557.8 (5) [M ‡ K] ; C28H42NO6P (519.6): calcd C 64.72, H 8.15,
N 2.70; found C 64.40, H 7.92, N 2.74.
Procedure B [from P-(benzyloxycarbonyl-aminomethyl)-P-(ethyl 2-isobu-
tylpropionate-3-yl)phosphinic acid chloride (3')]: A suspension of phos-
phinic acid 3 (60 mg, 0.16 mmol) and DMF (24 mL, 0.32 mmol, 2 equiv) in
anhydrous THF (1.5 mL) was cooled to 08C under Ar and (COCl)2 (15 mL,
0.17 mmol, 1.1 equiv) was added dropwise. The reaction was accompanied
by evolution of gas. After 20 min, the formation of the acid chloride 3' was
complete as verified by 31P NMR spectroscopy (250 MHz, CDCl3, 258C):
d ˆ 62.8 and 63.9. NaH (60% in oil, 6 mg, 0.16 mmol, 1 equiv) was added to
a solution of 1-adamantanol (26 mg, 0.17 mmol, 1.1 equiv) in anhydrous
THF (1 mL) and heated to 658C. The mixture was stirred for 20 min after
which the solution of 3' in THF was added dropwise over a period of 5 min.
Precipitation of NaCl was observed shortly after addition of 3'. After
stirring for 1 h, the mixture was concentrated and purified by preparative
TLC (2 mm silica-coated plates on glass, toluene/ethyl acetate 1:1) to yield
49 mg (61%).
1
95 ± 968C; H NMR (250 MHz, 0.3% NaOD in D2O, 258C): d ˆ 3.18 (dd,
3
1
2J(H,P) ˆ 10.5 Hz, J(H,H) ˆ 1.6 Hz, 2H), 5.07 (s, 2H), 6.86 (dt, J(H,P) ˆ
523.3 Hz, 3J(H,H) ˆ 2.0 Hz, 1H), 7.36 (s, 5H); 31P NMR (250 MHz, 1%
NaOD in D2O, 258C): d ˆ 21.6.
2882
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0510-2882 $ 17.50+.50/0
Chem. Eur. J. 1999, 5, No. 10

MMP-9 Inhibitors
2877±2884
O-Adamantyl P-(9-fluorenylmethyloxycarbonyl-aminomethyl)-P-(2-iso-
butylpropionic acid-3-yl)phosphinate (1): Aqueous NaOH (11.8 mL, 4m,
11.8 mmol, 4 equiv) was added dropwise to a solution of phosphinate 4
(5.55 g, 11.8 mmol) in EtOH (100 mL). The resulting opaque solution was
stirred for 24 h, after which 4 was consumed. The solution was concen-
trated, partitioned between ethyl acetate (100 mL) and water (50 mL),
cooled to 08C, and aqueous HCl (20 mL 2.4m, then 8 mL 1m) was slowly
added to adjust the pH to 2. The aqueous phase was extracted twice with
ethyl acetate (50 mL), the combined organic phases were dried with MgSO4
and concentrated to dryness to afford a white solid carboxylic acid (6.30 g;
100% yield corresponds to 5.82 g) that consisted of one major product
according to TLC (Rf ˆ 0.40, toluene/ethyl acetate 1:3 ‡ 5% acetic acid).
The crude carboxylic acid (200 mg, 407 mmol) was dissolved in ethyl
acetate/MeOH/water (11 ‡ 6 ‡ 0.5 mL) and hydrogenated at atmospheric
pressure in the prescence of Pd (5% on activated carbon, 86 mg, 0.1 equiv),
NaHCO3 (171 mg, 2.0 mmol, 5 equiv) and Fmoc-OSu (206 mg, 610 mmol,
1.5 equiv) for 80 min. Vacuum (10 mmHg) was then applied for 30 min and
the mixture was stirred overnight. Filtration through celite and concen-
tration gave the crude product as a sticky solid (411 mg), which was purified
by VLC with chloroform/methanol 30:1 as eluent to afford pure 1 as a solid
foam. Yield: 154 mg (65%). TLC: Rf ˆ 0.49 (toluene/ethyl acetate 1:3 ‡
5% acetic acid). The 1H NMR-spectrum was assigned with 1H ± 1H COSY
and 1H ± 13C COSY spectra. 1H NMR (250 MHz, CDCl3, 258C): d ˆ 0.89
(m, 6H), 1.40 (m, 1H), 1.47 (s, 6H), 1.65 (m, 1H), 1.66 (m, 1H), 1.70 (m,
1H), 1.95 (s, 6H), 1.99 (s, 6H), 2.39 (m, 1H), 2.84 (m, 1H), 3.53 (m, 1H),
3.78 (m, 1H), 4.22 (m, 1H), 4.22 (m, 1H), 4.38 (m, 1H), 6.89 (d, 1H), 7.14 ±
7.74 (m); 13C NMR (250 MHz, CDCl3, 258C): 22.3 (2d), 25.5 (d), 30.7 (2d),
31.1, 35.5, 37.8 (t), 42.0 (2d), 42.5 (t), 44.2 (d), 47.0 (s), 67.5 (s), 84.0 (2d),
119.7 (s), 125.3 (d), 125.6, 127.0, 127.5, 141.1, 143.7 (d), 144.1, 156.9 (t), 177.8;
31P NMR (250 MHz, CDCl3, 25%): d ˆ 46.1 and 48.1; ES-MS m/z (%):
(0.1m, 2 Â 1.5 mL each, 1 h) and each portion of the resin was washed with
H2O (4 Â 0.5 mL) and MeOH (2 Â 0.5 mL). The resulting solutions of 5a/
b ± 11a/b (ꢀ5mm) were neutralized with HCl (0.1m, 2.5 mL) and charac-
terized by analytical HPLC, electrospray mass spectrometry, and amino
acid analysis as presented in Tables 1 and 2.
The individual diastereomers of the peptides 5a/b ± 11a/b were isolated by
preparative HPLC: The crude peptide solutions were lyophilized and
redissolved in a suitable mixture of H2O, MeOH and TFA. After filtration,
the solutions were loaded to a semipreparative reversed-phase HPLC
1
column and purified using the standard buffers A and B (2% Bmin
,
10 mLmin 1) and UV detection at 215 nm. Fractions containing the correct
compounds according to matrix-assisted laser desorption/ionisation time-
of-flight mass spectrometry (MALDI-TOF) were collected and lyophi-
lized.
Enzyme kinetic assays: The Ki values of inhibitors 5a/b ± 11a/b against
MMP-9 were determined with increasing concentrations of inhibitors,
which were preincubated with a constant amount (0.2nm final concen-
tration) of MMP-9 for one hour at 378C. APMA-activated mouse
recombinant MMP-9 was obtained and purified from the cell culture
medium of a baby hamster kidney (BHK) cell line overexpressing MMP-
9.[81] The remaining free enzyme was then analyzed by fivefold dilution of
the preincubation mixture into an assay buffer (50mm Tris pH 7.5, 150mm
NaCl, 10mm CaCl2, 50mm ZnSO4, 0.05% (v/v) Brij-35) containing 6.5mm of
[18]
the quenched fluorogenic MMP-substrate Mca-PLG ± L-Dpa-AR-NH2
and incubated at 378C. The time course of the reaction was followed by
recording the increase in fluorescence (lex ˆ 320 nm, lem ˆ 387 nm). The
initial velocities were calculated and plotted against the concentration of
inhibitor. The Ki value was determined by nonlinear regression fit.
‡
580.3 (100) [M ‡ H] ; C33H42NO6P (579.7): calcd C 68.38 H 7.30, N 2.42;
found C 67.68, H 7.50, N 2.30.
Phosphinic peptides 5a/b ± 11a/b: Synthesis protocol: Amino acid cou-
plings were allowed to run for at least 3 h and complete coupling was
monitored by a negative Kaiser test.[71] After coupling of one amino acid or
building block 1, the resin was washed with DMF, followed by deprotection
of a-amino group with piperidine (20% in DMF) for 2 and 10 min. After
the resin was washed with DMF, the cycle was repeated with another amino
acid or building block 1. Volumes of washing solvent were 1 to 2 times the
volume necessary to swell the resin and washings were 6 Â 1 min unless
stated otherwise. The peptides were synthesized by multiple-column
peptide synthesis[70] (MCPS) in a 20 well teflon block with sintered teflon
Acknowledgment
This work has been supported by VñkstFonden and by the SPOCC
program under the Danish National Research Foundation.
[1] This nomenclature means that the peptide bond -C(O)-NH- between
glycine and leucine has been replaced by the peptide isosteric bond
-P(O)OH-CH2-.
[2] H. Nagase, ªMatrix Metalloproteinasesº in Zinc Metalloproteases in
Health and Disease (Ed.: N. M. Hooper), Taylor and Francis, London,
1996, pp. 153 ± 204.
filters. For each diastereomeric pair of peptides
a PEGA-resin
(0.45 mmolmg 1, 75 mg, 34 mmol) derivatized with Fmoc-Gly-HMBA was
used. This was prepared as follows:
A preactivated (10 min) mixture of HMBA (657 mg, 4.32 mmol, 3 equiv),
TBTU (1.33 g, 4.15 mmol, 2.88 equiv) and N-ethylmorpholine (727 mL,
662 mg, 5.76 mmol, 4 equiv) in anhydrous DMF (25 mL) was added to the
resin (3.01 g swelled in DMF). After 2 h, the resin was washed with DMF,
DCM, and then lyophilized overnight. The resin was swollen in dry DCM
(40 mL) and Fmoc-G-OH was coupled with MSNT[80] (two couplings,
65 min and 75 min): Fmoc-G-OH (1.28 g, 4.32 mmol, 3 equiv) was dis-
solved in dry DCM (20 mL) together with N-methyl imidazole (258 mL,
266 mg, 3.24 mmol, 2.25 equiv), and when dissolved, MSNT (1.28 g,
4.32 mmol, 3 equiv) was added. The mixture was immediately added to
the resin. After the second reaction, the resin was washed with DCM,
lyophilized, and a fraction (525 mg, 238 mmol) of the resin was weighed out
for the synthesis. It was then swelled in DMF and equally distributed
between seven wells, the remaining 13 wells being sealed off. Couplings of
normal amino acids were performed with Na-Fmoc-protected amino acid
pentafluorophenyl esters (3 equiv) in dry DMF and Dhbt-OH (1 equiv) as
catalyst, by the protocol described above. Coupling of the building block 1
(1.5 equiv) was accomplished with TBTU activation (1.44 equiv) as above,
and was completed in 4.5 h. After the last amino acid was attached to the
resin, Fmoc was removed and the resins were washed with DMF and 12 Â
DCM, dried by lyophilization for 1.5 h and then treated with a cocktail
[3] H. Birkedal-Hansen, W. G. I. Moore, M. K. Bodden, L. J. Windsor, B.
Birkedal-Hansen, A. DeCarlo, J. A. Engler, Crit. Rev. Oral Bio. Med.
1993, 4, 197 ± 250.
[4] J. A. Lorenzo, C. C. Pilbeam, J. F. Kalinowski, M. S. Hibbs, Matrix
1992, 12, 282 ± 290.
Â
[5] J.-M. Delaisse, G. Vaes, Mechanism of Mineral Solubilization and
Matrix Degradation in Osteoclastic Bone Resorption, in Biology and
Physiology of the Osteoclasts (Eds.: B. R. Rifkin, C. V. Gay), CRC
Press, 1992, pp. 289 ± 314.
Â
[6] H. Peng, I. Byrjalsen, M. Ferreras, N. T. Foged, J.-M. Delaisse, Bone
1998, 23, S552.
Â
[7] J.-M. Delaisse, Y. Eeckhout, G. Vaes, Biochem. Biophys. Res.
Commun. 1984, 125, 441 ± 447.
Â
[8] V. Everts, J.-M. Delaisse, W. Korper, W. Beertsen, J. Bone Miner. Res.
1998, 13, 1420 ± 1430.
[9] T. Sato, M. C. Ovejero, P. Hou, A.-M. Heegaard, M. Kumegawa, N. T.
Â
Foged, J.-M. Delaisse, J. Cell Sci. 1997, 110, 589 ± 596.
Â
[10] J.-M. Delaisse, Y. Eeckhout, C. Francois-Gillet, P. Henriet, Y. Su, G.
Vaes, R. Baron, J. Cell Sci. 1993, 106, 1071 ± 1082.
[11] K. Tezuka, K. Nemoto, Y. Tezuka, T. Sato, Y. Ikeda, M. Kobori, H.
Kawashima, H. Egushi, Y. Hakeda, M. Kumegawa, J. Biol. Chem.
1994, 269, 15006 ± 15009.
composed
of
TFA/DCM/H2O/MeSPh/(CH2SH)2/triisopropylsilane
66.5:20:5:5:2.5:1 for 2.5 h to remove peptide side chain protective groups.
The resins were washed with 4 Â AcOH (95% in H2O, 5 min), 2 Â DMF,
2 Â DIPEA (5% in DMF), 2 Â DMF, 15 Â DCM, and lyophilized (1 h). The
deprotected peptides were cleaved off the resin by treatment with NaOH
[12] P. Reponen, C. Sahlberg, C. Munaut, I. Thesleff, K. Tryggvason, J. Cell
Biol. 1994, 124, 1091 ± 1102.
[13] Y. Okada, K. Naka, K. Kawamura, T. Matsumoto, I. Nakanishi, N.
Fujimoto, H. Sato, M. Seiki, Lab. Invest. 1995, 72, 1 ± 12.
Chem. Eur. J. 1999, 5, No. 10
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0510-2883 $ 17.50+.50/0
2883

M. Meldal et al.
FULL PAPER
[14] Also referred to membrane-type 1 MMP (MT1-MMP)
[50] C.-H. Lin, S. Chen, D. S. Kwon, J. K. Coward, C. T. Walsh, Chem. Biol.
1997, 4, 859 ± 866.
Â
[15] L. Blavier, J.-M. Delaisse, J. Cell Sci. 1995, 108, 3649 ± 3659.
Â
[16] T. Sato, N. T. Foged, J.-M. Delaisse, J. Bone Miner. Res. 1998, 13, 59 ±
[51] J.-M. Campagne, J. Coste, L. Guillou, A. Heitz, P. Jouin, Tetrahedron
Lett. 1993, 34, 4181 ± 4184.
66.
Â
[17] S. Netzel-Arnett, Q.-X. Sang, W. G. I. Moore, M. Navre, H. Birkedal-
Hansen, H. E. Van Wart, Biochemistry 1993, 32, 6427 ± 6432.
[18] C. G. Knight, F. Willenbrock, G. Murphy, FEBS Lett. 1992, 296, 263 ±
266.
[52] A. Yiotakis, S. Vassiliou, J. Jiracek, V. Dive, J. Org. Chem. 1996, 61,
6601 ± 6605.
[53] P. H. Dorff, G. Chiu, S. W. Goldstein, B. P. Morgan, Tetrahedron Lett.
1998, 39, 3375 ± 3378.
[19] Bachem catalogue 1999, D11, product No. M-1895; Mca: (7-methox-
ycoumarin-4-yl)acetyl, Dpa: N-3-(2,4-dinitrophenyl)-L-2,3-diamino-
propionyl.
[20] ± indicates the preferred cleavage site for MMPs throughout the
article.
[54] J. S. Früchtel, G. Jung, Angew. Chem. 1996, 108, 19 ± 46; Angew. Chem.
Int. Ed. Engl. 1996, 35, 17 ± 42.
[55] Abz: 2-aminobenzoyl, Y(NO2): 3-nitrotyrosine.
[56] H. Stetter, H. Kuhlmann, Synthesis 1979, 29 ± 30.
[57] E. K. Baylis, C. D. Campbell, J. G. Dingwall, J. Chem. Soc. Perkin
Trans. 1 1984, 2845 ± 2853.
Â
[21] M. Renil, M. Ferreras, J.-M. Delaisse, N. T. Foged, M. Meldal, J. Pept.
Sci. 1998, 4, 195 ± 210.
[58] X.-Y. Jiao, C. Verbruggen, M. Borloo, W. Bollaert, A. De Groot, R.
Dommisse, A. Haemers, Synthesis 1994, 23 ± 24.
[59] J. G. Dingwall, J. Ehrenfreund, R. G. Hall, Tetrahedron 1989, 45,
3787 ± 3808.
[60] P. P. McCleery, B. Tuck, J. Chem. Soc. Perkin Trans. 1 1989, 1319 ±
1329.
[22] J. Lloyd, J. B. Schmidt, J. T. Hunt, J. C. Barrish, D. K. Little, A. A.
Tymiak, Bioorg. Med. Chem. Lett. 1996, 6, 1323 ± 1326.
[23] W. H. Parsons, A. A. Patchett, H. G. Bull, W. R. Schoen, D. Taub, J.
Davidson, P. L. Combs, J. P. Springer, H. Gadebusch, B. Weissberger,
M. E. Valiant, T. N. Mellin, R. D. Busch, J. Med. Chem. 1988, 31,
1772 ± 1778.
[24] B. A. Ellsworth, N. J. Tom, P. A. Bartlett, Chem. Biol. 1996, 3, 37 ± 44.
[25] R. E. Galardy, D. Grobelny, Z. P. Kortylewicz, L. Poncz, Matrix 1992,
259 ± 262.
[26] D. Grobelny, U. B. Goli, R. E. Galardy, Biochemistry 1989, 28, 4948 ±
4951.
[27] B. P. Morgan, J. M. Scholtz, M. D. Ballinger, I. D. Zipkin, P. A.
Bartlett, J. Am. Chem. Soc. 1991, 113, 297 ± 307.
[28] J. L. Goulet, J. F. Kinneary, P. L. Durette, R. L. Stein, R. K. Harrison,
M. Izquierdo-Martin, D. W. Kuo, T.-Y. Lin, W. K. Hagmann, Bioorg.
Med. Chem. Lett. 1994, 4, 1221 ± 1224.
[29] S. A. Bernhard, L. E. Orgel, Science 1959, 130, 625 ± 626.
[30] N. E. Jacobsen, P. A. Bartlett, J. Am. Chem. Soc. 1981, 103, 654 ± 657.
[31] P. A. Bartlett, C. K. Marlowe, Biochemistry 1983, 22, 4618 ± 4624.
[32] P. A. Bartlett, W. B. Kezer, J. Am. Chem. Soc. 1984, 106, 4282 ± 4283.
[33] K. A. Mookhtiar, C. K. Marlowe, P. A. Bartlett, H. E. Van Wart,
Biochemistry 1987, 26, 1962 ± 1965.
[61] M. J. Gallagher, H. Honegger, Tetrahedron Lett. 1977, 34, 2987 ± 2990.
[62] M. J. Gallagher, H. Honegger, Aust. J. Chem. 1980, 33, 287 ± 294.
[63] P. E. Maligres, I. Houpis, K. Rossen, A. Molina, J. Sager, V. Upadhyay,
K. M. Wells, R. A. Reamer, J. E. Lynch, D. Askin, R. P. Volante, P. J.
Reider, Tetrahedron 1997, 53, 10983 ± 10992.
[64] D. E. C. Corbridge, Phosphorus: An Outline of Its Chemistry, Bio-
chemistry, and Technology, 4th ed., Elsevier, Amsterdam, 1990.
[65] E. A. Boyd, A. C. Regan, K. James, Tetrahedron Lett. 1992, 33, 813 ±
816.
[66] W. P. Malachowski, J. K. Coward, J. Org. Chem. 1994, 59, 7625 ± 7634.
[67] R. S. Rogers, Tetrahedron Lett. 1992, 33, 7473 ± 7474.
[68] F. Cramer, K. Pawelzik, H. J. Baldauf, Chem. Ber. 1958, 91, 1049 ±
1054.
[69] T. W. Greene, P. G. M. Wuts, Protective Groups in Organic Synthesis,
2nd ed., John Wiley and Sons, New York, 1991.
[70] M. Meldal, I. Svendsen, K. Breddam, F.-I. Auzanneau, Proc. Natl.
Acad. Sci. USA 1994, 91, 3314 ± 3318.
[34] P. P. Giannousis, P. A. Bartlett, J. Med. Chem. 1987, 30, 1603 ± 1609.
[35] C. G. Caldwell, S. P. Sahoo, S. A. Polo, R. R. Eversole, T. J. Lanza,
S. G. Mills, L. M. Niedzwiecki, M. Izquierdo-Martin, B. C. Chang,
R. K. Harrison, D. W. Kuo, T.-Y. Lin, R. L. Stein, P. L. Durette, W. K.
Hagmann, Bioorg. Med. Chem. Lett. 1996, 6, 323 ± 328.
[36] P. Jouin, J. Coste, N. Galeotti, J.-M. Campagne, P. Bedos, L. Guillou,
Lett. Pept. Sci. 1995, 2, 233 ± 238.
[37] A. Yiotakis, A. Lecoq, S. Vassiliou, I. Raynal, P. Cuniasse, V. Dive, J.
Med. Chem. 1994, 37, 2713 ± 2720.
[38] A. Yiotakis, A. Lecoq, A. Nicolaou, J. Labadie, V. Dive, Biochem. J.
1994, 303, 323 ± 327.
[71] E. Atherton, R. C. Sheppard, Solid-Phase Peptide Synthesis:
Practical Approach, IRL Press, Oxford, 1989.
[72] F.-I. Auzanneau, M. Meldal, K. Bock, J. Pept. Sci. 1995, 1, 31 ± 44.
[73] R. Knorr, A. Trzeciak, W. Bannwarth, D. Gillesen, Tetrahedron Lett.
1989, 30, 1927 ± 1930.
A
[74] S. A. Watson, Cancer Res. 1995, 55, 3629 ± 3633.
[75] J. L. Seltzer, H. Weingarten, K. T. Akers, M. L. Eschbach, G. A.
Grant, A. Z. Eisen, J. Biol. Chem. 1989, 264, 19583 ± 19586.
[76] M. Meldal, I. Svendsen, J. Chem. Soc. Perkin Trans. 1 1995, 1591 ±
1596.
[77] N. M. Target, J. P. Kilcoyne, B. Green, J. Org. Chem. 1979, 44, 4962 ±
4964.
Â
[39] J. Jiracek, A. Yiotakis, B. Vincent, A. Lecoq, A. Nicolaou, F. Checler,
V. Dive, J. Biol. Chem. 1995, 270, 21701 ± 21706.
[78] The synthesis of 2a involves the use of the pentavalent hypophos-
phorus acid derivative, ethyl (diethoxymethyl)phosphinate. Anhy-
drous hypophosphorus acid is reported to be required for the
preparation of this compound, but we observed that explosive
substances were formed in this reaction when hypophosphorus acid
containing less than 5% water was used (rigorously dried by
coevaporation several times with 1,4-dioxane below RT). Instead,
95% aqueous hypophosphorus acid obtained by simple rotary
evaporation of the commercial 50% acid at high vacuum was found
to be sufficient for this purpose.
Â
[40] J. Jiracek, A. Yiotakis, B. Vincent, F. Checler, V. Dive, J. Biol. Chem.
1996, 271, 19606 ± 19611.
[41] B. A. McKittrick, A. W. Stamford, X. Weng, K. Ma, S. Chackalaman-
nil, M. Czarniecki, R. M. Cleven, A. B. Fawzi, Bioorg. Med. Chem.
Lett. 1996, 6, 1629 ± 1634.
[42] A. Peyman, K.-H. Budt, J. Spanig, D. Ruppert, Angew. Chem. 1993,
105, 1852 ± 1854; Angew. Chem. Int. Ed. Engl. 1993, 32, 1720 ± 1722.
[43] D. Grobelny, E. M. Wondrak, R. E. Galardy, S. Oroszlan, Biochem.
Biophys. Res. Commun. 1990, 169, 1111 ± 1116.
[44] M. C. Allen, W. Fuhrer, B. Tuck, R. Wade, J. M. Wood, J. Med. Chem.
1989, 32, 1652 ± 1661.
[45] P. A. Bartlett, J. E. Hanson, P. P. Giannousis, J. Org. Chem. 1990, 55,
6268 ± 6274.
[46] M. E. Tanner, S. Vaganay, J. van Heijenoort, D. Blanot, J. Org. Chem.
1996, 61, 1756 ± 1760.
[47] D. J. Miller, S. M. Hammond, D. Anderluzzi, T. D. H. Bugg, J. Chem.
Soc. Perkin Trans. 1 1998, 131 ± 142.
[79] The acid-protected phosphinic dipeptide building blocks are present
as two diastereomeric sets of enantiomers. These are indicated by two
closely running spots on TLC, but also by the multiple signals
observed for some resonances in both 1H, 13C, and 31P NMR. Rf values
given here are mean values for the two spots.
[80] B. Blankemeyer-Menge, M. Nimtz, R. Frank, Tetrahedron Lett. 1990,
31, 1701 ± 1704.
[81] The BHK cell line was purchased from University Technologies
International Inc., Calgary University, Calgary, Canada.
[48] J. Hiratake, H. Kato, J. Oda, J. Am. Chem. Soc. 1994, 116, 12059 ±
12060.
[49] C. Verbruggen, S. De Craecker, P. Rajan, X.-Y. Jiao, M. Borloo, K.
Smith, A. H. Fairlamb, A. Haemers, Bioorg. Med. Chem. Lett. 1996, 6,
253 ± 258.
Received: February 25, 1999 [F 1637]
2884
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0510-2884 $ 17.50+.50/0
Chem. Eur. J. 1999, 5, No. 10