A synthetic method for diversification of the P1′ substituent in phosphinic dipeptides as a tool for exploration of the specificity of the S1′ binding pockets of leucine aminopeptidases

Article abstract of DOI:10.1016/j.bmc.2007.02.042

A novel, general, and versatile method of diversification of the P1′ position in phosphinic pseudodipeptides, presumable inhibitors of proteolytic enzymes, was elaborated. The procedure was based on parallel derivatization of the amino group in the suitably protected phosphinate building blocks with appropriate alkyl and aryl halides. This synthetic strategy represents an original approach to phosphinic dipeptide chemistry. Its usefulness was confirmed by obtaining a series of P1′ modified phosphinic dipeptides, inhibitors of cytosolic leucine aminopeptidase, through computer-aided design basing on the structure of homophenylalanyl-phenylalanine analogue (hPheP[CH₂]Phe) bound in the enzyme active site as a lead structure. In this approach novel interactions between inhibitor P1′ fragment and the S1′ region of the enzyme, particularly hydrogen bonding involving Asn330 and Asp332 enzyme residues, were predicted. The details of the design, synthesis, and activity evaluation toward cytosolic leucine aminopeptidase and aminopeptidase N are discussed. Although the potency of the lead compound has not been improved, marked selectivity of the synthesized inhibitors toward both studied enzymes was observed.

Full text of DOI:10.1016/j.bmc.2007.02.042

Bioorganic & Medicinal Chemistry 15 (2007) 3187–3200
A synthetic method for diversification of the P10 substituent
in phosphinic dipeptides as a tool for exploration of the specificity
of the S10 binding pockets of leucine aminopeptidases
Stamatia Vassiliou,a Metaxia Xeilari,a Athanasios Yiotakis,a Jolanta Grembecka,b
Małgorzata Pawełczak,c Paweł Kafarskib,c and Artur Muchab,*
aLaboratory of Organic Chemistry, Department of Chemistry, University of Athens, Panepistimiopolis Zografou, 15771 Athens, Greece
bDepartment of Bioorganic Chemistry, Faculty of Chemistry, Wrocław University of Technology, Wybrzez_e Wyspian´skiego 27,
50-370 Wrocław, Poland
cInstitute of Chemistry, University of Opole, Oleska 48, 45-052 Opole, Poland
Received 4 May 2006; revised 13 February 2007; accepted 20 February 2007
Available online 22 February 2007
Abstract—A novel, general, and versatile method of diversification of the P10 position in phosphinic pseudodipeptides, presumable
inhibitors of proteolytic enzymes, was elaborated. The procedure was based on parallel derivatization of the amino group in the
suitably protected phosphinate building blocks with appropriate alkyl and aryl halides. This synthetic strategy represents an original
approach to phosphinic dipeptide chemistry. Its usefulness was confirmed by obtaining a series of P10 modified phosphinic dipep-
tides, inhibitors of cytosolic leucine aminopeptidase, through computer-aided design basing on the structure of homophenylalanyl-
phenylalanine analogue (hPheP[CH2]Phe) bound in the enzyme active site as a lead structure. In this approach novel interactions
between inhibitor P10 fragment and the S10 region of the enzyme, particularly hydrogen bonding involving Asn330 and Asp332
enzyme residues, were predicted. The details of the design, synthesis, and activity evaluation toward cytosolic leucine aminopepti-
dase and aminopeptidase N are discussed. Although the potency of the lead compound has not been improved, marked selectivity of
the synthesized inhibitors toward both studied enzymes was observed.
ꢀ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
hydrolytic action, which is not completely clarified yet.
Basing on numerous three-dimensional LAP structures
obtained for both enzyme in native form and complexed
with various inhibitors,6–10 a range of mechanistic
suggestions have been coined.1,5,9,11–13 Despite this, the
mode of LAP action still simulates studies on theoretical
evaluation of zinc environmental effects, functional role
of the active-site residues,14 and role of the water chan-
nels. They are carried out in order to identify the most
likely candidates for nucleophiles active in the hydro-
lytic process.15 Related situation occurs considering
physiological and pathological implications of LAP
activity. Biological role of the enzyme has not been fully
elucidated yet, however, similarly to other aminopepti-
dases, LAP is assumed to be involved in protein modifi-
cation, activation, and degradation as well as in the
metabolism of biologically active peptides and regula-
tion of activity of hormonal and non-hormonal
peptides.2,4 In mammals, LAP processes antigenic pep-
tides for presentation by the major histocompatibility
complex class I molecules.16 Recently, it has been also
Aminopeptidases form a family of metallo-dependent
hydrolases that are widespread in plants, animals, and
bacteria. Being responsible for cleavage of the N-terminal
amino acids from polypeptide chains, these exo-
peptidases play a key role in physiological processing
and degradation of peptides and proteins, and altera-
tions in their activity have been associated with various
pathological disorders, including cancer.1–3 Among
them, bizinc cytosolic leucine aminopeptidase (LAP,
E.C.3.4.11.1), exhibiting the substrate specificity toward
hydrophobic amino acid residues, remains one of the
most extensively studied enzymes.4,5 This interest is
mainly directed toward evaluation of its mechanism of
Keywords: Phosphinic pseudodipeptides; LAP inhibitors; P10 diversi-
fication; Cross-coupling; Alkylation.
*
Corresponding author. Tel.: +48 71 320 3354; fax: +48 71 328
4064; e-mail: artur.mucha@pwr.wroc.pl
0968-0896/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2007.02.042

3188
S. Vassiliou et al. / Bioorg. Med. Chem. 15 (2007) 3187–3200
identified as the key enzyme responsible for glutathione
turnover in liver.17,18 Significantly, over-expression of
LAP has been implicated in certain pathological stress
states including HIV infection, inflammation, cataracts,
and cancer.19–23
bottom of the S10 pocket of APM, as envisaged
after building the model of APM active site based on
homology to human leukotriene A4 hydrolase/
aminopeptidase.27
The above-described results stimulated further studies
directed toward design of novel P10 substituents in pre-
sumable inhibitors of aminopeptidases. In this paper, we
present our attempts to modify the structure of
hPheP[CH2]Phe in order to obtain new phosphinic
dipeptide inhibitors of LAP and APM. Two kinds of
modifications were planned (Fig. 1). The first one relied
on attachment of a heteroatom-containing group
(hydroxyl or amino) to the para position of the phenyl
ring via methylene bridge ensuring its increased rota-
tional freedom, which in turn should enable more flexi-
ble binding and thus facilitate hydrogen bond formation
with Asn330. The second one was introduction of an
additional amino group within the P10 side chain of
the dipeptide mimetic. The predicted binding mode
revealed the proximity of this amino group to the car-
bonyl portion of Asn332 and a potential for hydrogen
bond formation between them.
Biological significance and potential medical importance
of leucine aminopeptidase stimulate continuous interest
in agents regulating its activity.24 Development of orga-
nophosphorus compounds mimicking the high energy
transition state of the peptide bond hydrolysis has been
shown to be one of the most fruitful strategies in this
field.25–28 Recently, we have reported27 rationally
designed phosphinic dipeptide analogues that rank
amongst the most potent low-molecular weight inhibi-
tors of LAP reported so far. Among them, homophenyl-
alanyl-phenylalanine (compound 1, hPheP[CH2]Phe,
Fig. 1) and homophenylalanyl-tyrosine analogues (com-
pound 2, hPheP[CH2]Tyr) appeared to be highly active
(Ki = 66 and 67 nM, respectively, for mixture of four
diastereoisomers). The observation of virtually identical
binding affinities obtained for both 1 and 2 indicates
that the hydroxyl group of the side chain of Tyr at P10
is not involved in any additional interaction with the
enzyme.27 This is a surprising finding since the forma-
tion of the hydrogen bond between this hydroxyl group
and the side-chain amidate of Asn330 at the bottom of
S10 has been predicted by molecular modeling. How-
ever, this interaction apparently has not been created.
The reason of that could derive from unfavorable geom-
etry in terms of the hydrogen bond distance and angle.
To obtain the designed compounds, complex synthetic
problems were solved. As corresponding a-aminoacry-
lates—key electrophilic substrates for Michael addition
methodology appeared synthetically not available, new
general parallel arylation/alkylation procedure for the
phosphinic dipeptide amino derivatives was developed.
The details of this procedure leading to the target struc-
tures are described and discussed. Finally, the binding
affinities of the set of target inhibitors toward both
LAP and APM have been evaluated.
Interestingly, such a structural change (Phe ! Tyr in
P10 position) resulted with much more visible difference
in binding affinities of both compounds to microsomal
leucine aminopeptidase (aminopeptidase N, APM, E.C.
3.4.11.2).27 This monozinc exopeptidase29 was selected
intentionally in order to evaluate binding preferences
toward studied phosphinic pseudodipeptides in compar-
ison with LAP. hPheP[CH2]Tyr (2) appeared much
more effective APM inhibitor (Ki = 36 nM) than any
phosphorus containing dipeptide analogue described in
the literature30 (for recent, extensive review of APM
inhibitors, see also Ref. 31). This finding was explained
by formation of specific hydrogen bond between Tyr
hydroxyl with carboxyl group of Glu413 placed at the
2. Results and discussion
2.1. Design
The
structure
of
the
phosphinic
dipeptide
hPheP[CH2]Phe (1, Fig. 1) bound to the active site of
LAP served as the lead compound to develop novel
inhibitors of leucine aminopeptidase. The overall struc-
tural motif, the pseudopeptide backbone as well as the
P1 hydrophobic phenylethyl side chain were conserva-
tively kept in all newly designed structures, whereas
the P10 position was suitably modified to obtain better
fit to the S10 pocket of LAP. These changes accom-
plished incorporation of the hydroxymethyl or amino-
methyl groups at the P10 proximity, namely they were
introduced at para position of the phenyl ring. The S10
pocket is partially open to solvent and based on the
docking mode of 5 (see below for details) the ortho
and meta substitutions would not result in additional
interactions with the protein. The other side of the phe-
nyl ring at P10 is approaching the residues of the S10
pocket: Ile421 and Ala333 and there is not enough space
to incorporate a functional group at the ortho and meta
positions. Therefore, para analogues were selected for
synthesis as only these were expected to form additional
contacts with LAP. Besides modification of the phenyl
ring at P10, the b carbon atom of the pseudophenylanine
YH
Modifications of
the P1' substituent
X
O
P
n
O HN
P
H2N
H2N
COOH
COOH
OH
OH
n = 0, 1
Y = O, NH
1, X = H (Ki = 66 nM, LAP)
(Ki = 276 nM, APM)
2, X = OH (Ki = 67 nM, LAP)
(Ki = 36 nM, APM)
Figure 1. The designed structural modifications of the lead compounds
(1 and 2)27 to optimize interaction of phosphinic dipeptides with the
S10 binding pocket of leucine aminopeptidase (LAP).

S. Vassiliou et al. / Bioorg. Med. Chem. 15 (2007) 3187–3200
3189
residue was formally replaced with nitrogen atom. Such
a replacement introduced the amino group of aromatic
character (aniline derivatives). On the other hand, inser-
tion of nitrogen gave aliphatic ones (extended benzyl-
amine derivatives), which might be important in terms
of their protonation (these amino groups differ in their
pKa values up to five orders of magnitude). Novel P10
residues are represented by p-hydroxymethyl- (4) or
p-aminomethylaniline (5), and p-hydroxymethyl- (7) or
p-aminomethylbenzylamine (8) structural fragments
(Fig. 2). They were synthesized together with the unsub-
stituted derivatives (3 and 6) as respective reference
compounds.
indicate that the general pattern of zinc complexation,
the P1 fragment interactions, as well as binding the
backbone through the network of hydrogen bonds with
Asp273, Lys262, Gly362 are fully retained, comparing
to 1 and 2 as well as to other known LAP inhibi-
tors.9,10,12,27 The presence of the modified side chain at
the P10 position opens possibilities for additional inter-
actions with the S10 pocket of LAP. The S10 pocket is
formed by the Ile421, Ala333 side chains from one side
and Asn330 together with the hydrophobic portion of
Asp332 side chain from the other side. At the same time,
this pocket is partially open to solvent, which causes
that part of the P10 substituent to be water exposed.
Therefore, the introduction of substituents at ortho
and meta positions of the phenyl ring would not provide
additional interactions with LAP. Based on the docking
mode, the amino/hydroxy group at the para position
should be able to form the hydrogen bond with the oxy-
gen atom of the side chain of Asn330 (Fig. 3, hydrogen
The designed structures were docked32 to the binding
site of leucine aminopeptidase obtained from the
X-ray structure of bovine lens LAP complexed with
phosphonic analogue of leucine,10 similarly as described
previously in the case of compounds 1 and 2.27 The
results of the analysis have been found promising.
Presumed interactions of 5 and other compounds
described here with LAP active site shown in Figure 3
˚
bond distance is about 2.9 A, hydrogen bond angle of
about 170ꢁ). Moreover, the secondary amino group in
b position of the pseudophenylalanine residue is in
˚
hydrogen bond distance (2.7 A) to the carbonyl oxygen
of Asp332. In this case, however, the geometry of hydro-
gen bond would be less favorable (hydrogen bond angle
of about 150ꢁ). To verify empirically such theoretical
arrangement, compounds 3–8 were chosen for synthesis
and for evaluation of their binding affinities toward
LAP and APM.
R
R
O HN
P
O HN
P
H N
2
H N
2
COOH
COOH
OH
OH
2.2. Chemistry
6, R = H
3, R = H
2.2.1. Retrosynthesis. The synthetic targets of this work
were molecules 3–8 of structures shown in Figure 2. The
designed compounds possess up to five functional
groups of various characters and thus their synthesis
cannot be straightforward. To face this challenge, a
short retrosynthetic analysis has been performed for
the simplest derivative 3 (Scheme 1). Michael addition
of phosphorus nucleophiles (suitably protected phosphi-
nic amino acid analogues) to appropriate carbon elec-
trophiles (usually acrylates) is the most commonly
applied method for the phosphinate pseudopeptidic
bond formation.33 In our case, the first considered idea
was the use of N-phenyl-a-aminoacrylate (N-phenylde-
hydroalanine ester, 9) as the key substrate (Scheme 1,
pathway A). To the best of our knowledge, this com-
pound of relatively simple structure has not been
reported in the literature yet. Various methods, includ-
ing elimination of appropriately substituted serine deriv-
atives, could be envisaged for its synthesis. Three other
possibilities are listed in Scheme 1, namely: arylation
of aniline with a-bromoacrylate 10, Knoevenagel reac-
tion of paraformaldehyde and monoalkyl N-phenylami-
nomalonate 11, and condensation of aniline with
pyruvate 12. All of them were preliminarily tested.
Not going into experimental details, these attempts
failed giving as the main products, respectively,
N-phenylaziridinecarboxylate in the reaction with the
use of a-bromoacrylate 10, N-phenylglycinate in Knoe-
venagel reaction, and a complex mixture in the case of
pyruvate 12 condensation. Because of the probable high
instability of the target molecule 9, it became obvious
7, R = CH OH
4, R = CH OH
2
2
8, R = CH NH
5, R = CH NH
2
2
2
2
Figure 2. Putative phosphinic dipeptide inhibitors of LAP containing
modified P10 substituents.
Figure 3. Modeled binding mode of 2-(4-aminomethylphenylamino)-
3-[(1-amino-3-phenylpropyl)hydroxyphosphinoyl]propionic acid 5 by
leucine aminopeptidase obtained using the LAP crystal structure10
(1lcp in PDB), and the Ludi computer program.27,32 The selected S10
pocket and the active site residues of LAP, interacting with the
dipeptide analogue, are presented. Hydrogen bonds and interactions of
the inhibitor with zinc ions are shown as black dashed lines.

3190
S. Vassiliou et al. / Bioorg. Med. Chem. 15 (2007) 3187–3200
B
A
O HN
P
O
P
NH2
O
H
N
H2N
O
COOH
Boc
R
OH
O
R-X
+
B
(Ar-X)
3
13, R = H
14, R = Me
A
Br COOMe
H
N
COOMe
α-aminophosphinate
10
+
component
9
H
N
COOMe
+
HOOC COOMe
15
Cbz
HN
COOMe
11
α-aminophosphinate
O
12
component
Scheme 1. The retrosynthesis of 3-[(1-amino-3-phenylpropyl)hydroxyphosphinoyl]-2-phenylaminopropionic acid (3). For the details, see Section 2.
Cbz
O
that much more general approach should be developed.
Elimination of serine derivatives provided an idea of
construction of versatile building blocks 13 and 14
(Scheme 1, pathway B) by Michael addition of the
appropriate phosphorus component to the dehydroala-
nine derivative 15. The availability of acid 13 and the
corresponding C-terminal ester 140 would allow multidi-
rectional diversification of the P1 position at the final
step of the synthesis using cross-coupling or alkylation
reactions, respectively.
O
P
H
N
O HN
P
H
N
H
Boc
OH
Boc
OH
O
a, b
c
16
17
Cbz
O HN
H
O
NH
2
H
N
N
P
O
P
OH
Boc
Boc
O
O
O
O
d, e
2.2.2. Building block synthesis. The synthesis of the de-
signed synthons 13 and 14 is outlined in Scheme 2. Careful
choice of the protecting groups for four functions present
in the molecules of phosphinic dipeptide precursors is cru-
cial, since their selective removal at the desired synthetic
step is an absolute requirement. Thus, by addition of
Boc-protected phosphinic homophenylalanine analogue
1634 to a-aminoacrylate 1535 (Cbz-protected dehydro-
alanine methyl ester) under standard conditions,36 phos-
phinic pseudodipeptide 17 was obtained in good yield
after chromatographic purification. It should be stressed
that acrylate 15 was used in almost equimolar amount
in relation to the phosphinic acid and was added very
slowly to the activated phosphinic silyl ester, otherwise
extensive polymerization occurred. Subsequent reaction
of the hydroxyphosphinyl function with 1-adamantyl
bromide36 gave fully protected derivative 18. Its saponifi-
cation followed by catalytic hydrogenation yielded the
desired compound 13, whereas synthon 14 was obtained
after hydrogenolysis over Pd/BaSO4. The latter one could
be stored for a short period of time in the form of free
amine, and for prolonged time as hydrochloric acid salt.
The usefulness of resulting building blocks 13 and 14
(in free amine forms) for parallel protocols was verified
in two reaction types: cross-coupling (to obtain com-
pounds 3–5) and nucleophilic substitution (leading to
compounds 6–8).
13
18
e
O
P
NH
2
H
N
O
Boc
O
O
14
Scheme 2. Preparation of the key building blocks 13 and 14, required
for their further diversification by cross-coupling or alkylation
reactions. Reagents and conditions: (a) hexamethyldisalazane,
110 ꢁC; (b) 15, 90 ꢁC, 77%; (c) 1-AdBr, Ag2O, CHCl3, reflux, 80%;
(d) NaOH/MeOH, rt, 16 h then HCl, pH 1, 92%; (e) H2, 10% Pd/
BaSO4, MeOH, 3.5 h, 100%.
the literature. A number of cleverly designed and useful
methods based on the Ullmann37 reaction using Cu
reagent and the Buchwald–Hartwig38 amination utiliz-
ing palladium catalyst have been developed. During
the last decade, several reports described mild methods
for catalytic amination of aryl halides using ligands like
a-amino acids39 or ethylene glycol40 in the case of the
Ullmann reaction and racemic BINAP41 in the case of
the Buchwald–Hartwig one. Attracted by their facility
and readiness we have tested the applicability of these
procedures for reaction of our synthon (13) with phenyl
2.2.3. Cross-coupling. Formation of the carbon–nitrogen
bond via cross-coupling reaction is well recognized in

S. Vassiliou et al. / Bioorg. Med. Chem. 15 (2007) 3187–3200
3191
iodide (19). Surprisingly, no cross-coupling product was
I
I
observed in the Pd-catalyzed reaction while in the case
of CuI applied as catalyst the desired product 20 was
formed although it was isolated only in low yield.
Attempts to optimize the latter reaction (Scheme 3) by
modification of the ligand and the solvent system
according to the literature39,40 are summarized in Table
1. When elevated temperatures were used, possibly
premature loss of the protective group(s) resulted in
mixture of unidentifiable reaction products, therefore
65 ꢁC was not exceeded.
a
b
COOH
OH
I
21
22
I
c, d
Boc
Br
N
H
23
24
Steric hindrance caused by the presence of the neighbor-
ing bulky adamantyl group and the relatively low reac-
tion temperature could explain the low yields of the
cross-coupling reaction. Lack of any accelerating effect,
observed by Ma39 for the a-amino acid assisted Ullmann
condensation, is in accordance with our data since the
structure of the P10 part of the phosphinic dipeptide
13 already represents this case.
Scheme 4. Preparation of p-substituted iodophenyl derivatives 22 and
24, substrates for cross-coupling with 13. Reagents and conditions: (a)
i-BuOCOCl, N-methylmorpholine, ꢀ10 ꢁC then NaBH4 and MeOH,
0 ꢁC, 70%; (b) PBr3, Et2O, 0 ꢁC ! rt, 3 h, 58%; (c) NH4OH, 5d; (d)
Boc2O, CH2Cl2, 0 ꢁC, 15 h, 72%.
amines.42 In order to find out the optimal conditions for
this reaction in terms of protective groups compatibility
and avoidance of formation of overalkylation byprod-
ucts, we performed a more detailed study using benzyl
bromide (27) as the alkylating agent of synthon 14
(Scheme 6). The obtained results are presented in Table 2.
Despite these not fully encouraging results, two addi-
tional iodobenzene derivatives 22 and 24 were synthe-
sized from a commercially available starting material
(Scheme 4). Reduction of p-iodobenzoic acid 21 with
NaBH4 provided p-iodobenzylalcohol 22 in good yield.
Bromination of the alcohol 22 followed by treatment
of the product 23 with ammonium hydroxide and subse-
quent protection of the amine with the use of di-tert-
butyl dicarbonate resulted in derivative 24.
Since high yield of the desired compound 28 was
achieved without detecting the overalkylation product
29, Cs2CO3 and gentle warming in dry DMF (entry A)
were concluded as recommended conditions in the next
experiments. The superiority of cesium carbonate used
as a base in N-alkylation of primary amines is well
known and has been described in the literature as
‘cesium effect’.43 Lack of overalkylation could be attri-
buted to the bulkiness of the neighboring adamantyl
group.
Reaction of the obtained iodo derivatives 22 and 24 with
13 in standard conditions (entry A, Table 1) furnished
the desired cross-coupling products 25 and 26 in 10%
and 8% yield, respectively (Scheme 5). Acidic cleavage
of the protective groups in all three blocked products
(20, 25, and 26) afforded the target aniline derivatives
3–5, which were finally purified by reverse-phase pre-
parative HPLC.
The appropriately p-substituted benzyl bromides (33
and 37) are commercially unavailable, thus they had to
be prepared prior to alkylation of the phosphinic dipep-
tide 14. Their synthesis is outlined in Scheme 7. Esterifi-
cation of p-bromomethylbenzoic acid 30 provided its
methyl ester 31, which was reduced to the corresponding
alcohol 32 with the use of DIBALH. Protection of the
hydroxyl group into the form of the tert-butyl ether pro-
vided the desired benzyl bromide 33. The p-amino deriv-
ative was prepared basing on the literature data.44
Initially, p-cyanobenzyl bromide 34 was hydrolyzed to
the corresponding alcohol 35. Reduction of the cyano
group furnished the amino alcohol, which was treated
subsequently with hydrobromic acid to provide bromide
36. Protection of the amino group with di-tert-butyl
dicarbonate yielded the desired bromide 37. Preliminary
experiments to reduce p-cyanobenzyl bromide in order
2.2.4. Nucleophilic substitution. N-Alkylation of primary
amines in the presence of inorganic bases is one of the
most general methods for the synthesis of secondary
O HN
P
H
N
PhI (19), CuI, base
OH
13
Boc
O
ligand
O
Ad
20
Scheme 3.
Table 1. Comparison of the effectiveness of the Ullmann cross-coupling reaction of the building block 13 with phenyl iodide performed in
recommended conditions39,40
Method
Base
Solvent
Ligand
Time (h)
Temperature (ꢁC)
Yield of 20 (%)
A
B
C
K2CO3
K2CO3
K2CO3
DMSO
DMSO
2-Propanol
—
L-Proline
OHCH2CH2OH
48
48
48
65
65
65
22
22
21

3192
S. Vassiliou et al. / Bioorg. Med. Chem. 15 (2007) 3187–3200
R
I
O HN
P
H
N
a
b
Boc
COOH
13
+
O
3-5
Ad
R
20, R = H
3, R = H
19, R = H
25, R = CH OH
4, R = CH OH
2
22, R = CH OH
2
2
26, R = CH NHBoc
5, R = CH NH
2
24, R = CH NHBoc
2
2
2
Scheme 5. Cross-coupling reaction with the use of 13 and final deprotection to yield the target phosphinic dipeptides 3–5. Reagents and conditions:
(a) CuI, K2CO3, 65 ꢁC, 48 h, 22% (20), 10% (25), 8% (26); (b) 50% TFA/CH2Cl2, rt, 3 h, then HPLC.
Br
Br
Br
Br
a
c
b
O HN
P
H
N
O
P
N
H
N
BnBr (27)
O
O
14
Boc
+
Boc
COOH
O
COOMe
O
OH
32
O
O
O
Ad
Ad
30
31
33
36
28
29
Br
OH
Br
OH
e
Scheme 6.
d
f
to obtain the corresponding amine in one-step proce-
dure resulted in p-methylbenzylamine 38.
CN
CN
35
NH
NH Br
3
2
34
e
g
The para substituted benzyl bromides 33 and 37 under-
went the reaction with precursor 14, under the condi-
tions described above, to provide the fully protected
phosphinic dipeptides 39 and 40 in good yield, similarly
to the non-substituted derivative 28 (Scheme 8). To
remove the protecting groups, pseudopeptides 28, 39,
and 40 were saponified and then treated with 50%
TFA/CH2Cl2 for 28 and 40, and 90% TFA for 39.
Preparative HPLC purification provided the target
phosphinic pseudodipeptides 6–8 as two pairs of
diastereoisomers.
Br
38
37
Boc
NH
2
N
H
Scheme 7. Preparation of p-substituted bromobenzyl derivatives 33
and 37, substrates for alkylation of 14. Reagents and conditions: (a)
MeOH, H2SO4, reflux, 5 h, 95%; (b) DIBALH, 0ꢁC, 4 h, 58%;
(c) i-butene, H2SO4, 0 ꢁC, 64 h, 63%; (d) Ba2CO3, H2O, 4 h, 89%; (e)
LiAlH4, THF, 0 ꢁC ! reflux, 4 h; (f) 46% HBr, 3.5 h, reflux, 15%; (g)
Boc2O, NaHCO3, H2O, 0ꢁC ! rt, 18 h, 99%.
2.3. Activity
Compounds 3–8 appeared to be moderate or weak
inhibitors of both leucine aminopeptidase and APM
(Table 3). They exhibited competitive type of inhibition
for both enzymes, with slow-binding mechanism of A
type determined for LAP and simple inhibition for
APM (with the exception of compounds 5 and 4-F2).
The mode of inhibition is, therefore, analogous to the
one observed earlier for aminophosphonates and
phosphinopeptides.27,28
For almost all phosphinic dipeptides satisfactory
separation of diastereoisomers was achieved upon their
reverse-phase HPLC purification. Thus, they were
collected separately (fractions F1 and F2, in the order
of elution) and subsequently tested individually giving
two Ki values. Obviously, one fraction corresponds to
the (RS,SR) pair of enantiomers, while the another
one to (RR,SS).
Table 2. Optimization of the conditions of alkylation of the building block 14 with benzyl bromide
Entry
Solvent
Base
Time (h)
Temperature (ꢁC)
Yield of 28 (%)
Yield of 29 (%)
Ratio 28/29
A
B
C
D
E
F
G
H
I
DMF
DMF
DMF
DMF
DMF
DMF
DMSO
Et2O
—
Cs2CO3
K2CO3
Li2CO3
DBU
6
6
45
45
45
45
rt
80
66
63
36
46
80
65
63
8
—
—
15
—
—
11
—
—
—
—
—
6
ꢁ4:1
—
6
Et3N
6
—
Cs2CO3
Cs2CO3
Et3N
6
rt
ꢁ8:1
—
6
rt
20
6
rt
45
—
—
Pyridine

S. Vassiliou et al. / Bioorg. Med. Chem. 15 (2007) 3187–3200
3193
R
Br
O HN
P
H
N
a
b, c
Boc
COOMe
14
+
O
6-8
Ad
R
28, R = H
6, R = H
27, R = H
39, R = CH O-t-Bu
7, R = CH OH
2
33, R = CH O-t-Bu
2
2
40, R = CH NHBoc
8, R = CH NH
2
37, R = CH NHBoc
2
2
2
Scheme 8. Alkylation of synthon 14 with benzyl bromides 27, 33, and 37, and final deprotection to yield the target phosphinic dipeptides 6–8.
Reagents and conditions: (a) Cs2CO3, DMF, 45 ꢁC, 6 h, 80% (28), 65% (39), 69% (40); (b) NaOH/MeOH, rt, 16–18 h then HCl, pH 1; (c) 50% TFA/
CH2Cl2, rt, 3 h for 28 and 40, 90% TFA/CH2Cl2, rt, 16 h for 39, then HPLC.
Table 3. Inhibition of cytosolic (LAP) and microsomal leucine
aminopeptidase (APM) by the dipeptide analogues 3–8 containing
There could be few reasons of such phenomenon. First,
the studied dipeptide analogues 3–8, possessing several
polar groups of varying properties, must be highly
modified amino substituents in their P10 positions (fractions F1 and F2
represent the enantiomeric pair (RS,SR) and (RR,SS) separated by
hydrated in aqueous media. Considering the inhibitor
means of reverse-phase HPLC)
affinity in terms of free binding energy, an unwanted loss
R
in entropy might occur because of necessity of breaking a
net of hydrogen bond with water molecules that can be
not counterbalanced with formation of novel interac-
tions of this type with the enzyme. Second, decreasing
affinity with increasing size of the P10 substituents (in
comparison to the lead compound 1) means that the
designed residues could be bound somewhat differently,
although theoretically predicted as well-fitting to the
S10 binding pocket. This is because this pocket is opened
to the solvent and the designed compounds might be
shifted a little toward the interface of the enzyme (thus
binding becomes not optimal) with their polar groups
(attached to aromatic ring) interfering with the solvent.
It has to be emphasized that the side chain of Asn330,
which was expected to form a hydrogen bond with
ligands, is solvent exposed and therefore its conforma-
tion might vary in solution and differ from the one
observed in the X-ray structure which was used for dock-
ing studies. Moreover, the detailed docking analysis indi-
cated relatively close placement of the zinc cation,
namely0Zn488, to the secondary amino group introduced
O HN
P
H N
2
COOH
OH
3-8
Compound
Ki (lM)
LAPa
F2
APM
F2
Entry
P10 substiuent (R)
F1
F1
3
0.57
0.45
NI
NI
OH
4
5
6
7
8
10.33
6.98
0.90
1.13a
NH
4.46
0.56a
2
10.06
4.21
2.78
12.37
1.42
2.62
NI
ꢁ1500
NI
˚
into P1 side chain (ꢁ3.7 A). This could result in strongly
OH
NH
unfavorable electrostatic repulsion, particularly in the
case of the probable protonation of the amino moiety,
and may also contribute to their shifting toward the
enzyme interface. This can be directly manifested by a
comparison of the binding affinities of 3 and 1. These
two differ only with the b carbon atom at P10 of 1 being
replaced by –NH– in 3, and this substitution causes
10-fold decrease in binding affinity (Table 3). The struc-
ture–activity relationship for the compounds studied
here is also ambiguous. For instance, the increase in
the length of the P10 fragment by a methyl group
decreases the affinity by about 20-fold for 6 versus 3,
while similar modification improves twice the affinity of
compounds bearing a polar group at the para position
of the phenyl ring (compare 7 vs 4 and 8 vs 5).
NI
2
19.90
2.66
a Competitive, slow-binding of A type mechanism of inhibition; NI, no
inhibition at 10–20 lM, Ki not calculated.
Generally, the synthesized compounds did not exhibit
improved activity toward LAP comparing to the potent
lead structure 1. The calculated Ki values vary in a quite
narrow range of 0.5–12 lM (Table 3) and are one–two
orders of magnitude higher than that found for com-
pound 1. Surprisingly, the simplest, non-substituted ani-
line derivative 3 appeared to be the most interesting
inhibitor. Compounds designed as superior to com-
pound 3, namely p-substituted benzylamine derivatives
7 and 8, were slightly better than corresponding aniline
derivatives 4 and 5.
Generally, there is no significant difference between the
affinity of the separated fractions F1 and F2. It seems
straightforward that the diastereoisomer corresponding
to the relative (L,L) configuration of the natural

3194
S. Vassiliou et al. / Bioorg. Med. Chem. 15 (2007) 3187–3200
substrates should be the most tightly bound. On the
other hand, it has been proven that even for more
extended phosphinic inhibitors of metalloproteases, the
stereochemistry is not a strictly discriminating factor.45
Thus, no additional efforts have been made to assume
the absolute configuration, nor to separate single diaste-
reoisomers of the studied analogues.
as it is with the S10 pocket of LAP, and therefore the
expectations driven from the ligand design studies might
not be reflected in experimentally determined properties.
3. Experimental
3.1. General
According to the theoretical predictions, the other than
para substitution in the P10 phenyl ring appeared
strongly discriminating for potency of the inhibitors.
Two ortho benzyl derivatives, additionally synthesized,
purified, and separated for fractions, exhibited very
poor activity toward LAP (no inhibition at 80 lM for
F1 and Ki = 247 lM for F2 of the o-hydroxymethyl
derivative, and no inhibition at 40 lM for F1 and
Ki = 21.6 lM for F2 of the o-aminomethyl one, data
not presented in Table 3). To some extent, this fact sup-
ports the assumption made by docking experiments that
only para derivatives are able to interact favorably with
residues forming the S10 binding pocket of LAP.
All of the compounds, for which analytical and spectro-
scopic data are given, were homogeneous by TLC. TLC
analyses were performed using silica gel plates (E.
Merck silica gel 60 F-254) and components were visual-
ized by the following methods: ultraviolet light absor-
bance, charring after spraying with a solution of
(NH4)HSO4 or after spraying with ninhydrin solution.
Column chromatography was carried out on silica gel
(E. Merck, 70–230 mesh). HPLC analyses were carried
out on: (i) a MZ-Analytical Column 250 · 4 mm, Kro-
masil, 100, C18, 5 lm, at a flow rate of 0.5 mL/min or
(ii) a semipreparative 250 · 8 mm, Kromasil, C18,
5 lm at a flow rate 3 mL/min. Solvent A: 10% CH3CN,
90% H2O, 0.1% TFA. Solvent B: 90% CH3CN, 10%
H2O, 0.09% TFA. The following gradients were used:
(1) t = 0 min (0% B), t = 29 min (20% B), t = 30 min
(50% B), t = 32 min (100% B), t = 34 min (100% B),
t = 38 (40 % B) for compounds 6 and 7. (2) t = 0 min
(0% B), t = 27 min (5% B), t = 30 min (50% B),
t = 32 min (100% B), t = 34 (100% B), t = 38 min (40%
B) for compound 8. (3) t = 0 min (0% B), t = 20 min
(35% B), t = 25 min (60% B), t = 32 min (100% B),
t = 34 (100% B), t = 38 min (40% B) for compounds 3
and 4. (4) t = 0 min (0% B), t = 20 min (20% B),
t = 35 min (30% B), t = 37 min (100% B), t = 39 (100%
B), t = 41 min (40% B) for compound 5. Eluted peaks
were detected at 254 nm. Given times (where more than
one) correspond to two pairs of diastereoisomers and
are counted in minutes.
There are some meaningful observations concerning the
selectivity of the studied phosphinates 3–8 toward LAP
and APM. Introduction of a simple hydrophobic P10
residue yielded very poor or no inhibitors (compound 3
and 6) of aminopeptidase N. In the case of LAP these
groups are favored, however, a size limit is observed
(compounds 3 being far more active than 6). APM
strictly favors polar groups at the P10 proximity, partic-
ularly the amino moieties. Moreover, in the case of APM
the size of the substituent seems to be much more limited
than in the case of LAP. Thus, only three compounds (4,
5, and 8) are potent inhibitors of the enzyme, whereas the
others are inactive (Table 3). These results support addi-
tionally the suggested importance of the Glu413 carboxyl
group placed at the bottom of the S10 enzyme pocket of
APM for inhibitor discrimination.27
1
2.4. Conclusions
All the compounds were characterized by H, 13C, and
31P NMR spectroscopy. H, 13C, and 31P NMR spectra
1
In summary, the presented results seem to have the most
significant synthetic impact. During last decade, phosphi-
nic pseudopeptides, considered as the analogues of the
high-energy transition states of the peptide bond hydroly-
sis,45–47 have been extensively applied for effective and
selective regulation of activity of various metalloproteas-
es.48–52 These achievements stimulate continuous interest
in development of novel strategies for their preparation,
including parallel and combinatorial synthesis.33 We
hope to give a contribution to this field by presenting
the synthesis of novel, amino derived building blocks 13
and 14 suitable for further diversification. Although the
obtained inhibition rates of cytosolic and microsomal
aminopeptidases by compounds 3–8 are moderate and
thus not fully satisfying, some valuable information con-
cerning structure–activity relationship for these two bio-
logically important enzymes have been also acquired.
The results have also shown that currently available
molecular modeling methods and tools are not able to
predict all possible effects associated with the binding of
inhibitor by the enzyme. In particular, this concerns
protein systems with solvent exposed ligand binding sites,
were recorded on a 200 MHz Mercury Varian spectrom-
eter. 13C and 31P NMR spectra are fully proton decou-
pled. 31P NMR chemical shifts are reported on d scale
(in ppm) downfield from 85% H3PO4.
ESI mass spectral analysis was performed on a mass
sprectrometer MSQ Surveyor, Finnigan, at the Labora-
tory of Organic Chemistry, University of Athens, using
direct sample injection. Negative or positive ion ESI
spectra were acquired by adjusting the needle and cone
voltages accordingly.
Commercially available reagents, solvents, and starting
materials were purchased from Aldrich, Merck, Sigma,
Labscan, and Fluka, and used without further purifica-
tion. THF was distilled over NaH.
3.2. Docking
Compounds described in this paper were docked by
applying the same procedure as we previously used to
calculate the binding mode of the phosphinic dipeptide

S. Vassiliou et al. / Bioorg. Med. Chem. 15 (2007) 3187–3200
3195
analogues.27 The X-ray structure of LAP-LeuP (1lcp in
PDB) served to design dipeptide analogues described in
Ref. 27, including compounds 1 and 2 (Fig. 1). Based on
the calculated binding mode of 127 new substituents
were introduced at the para position of the phenyl ring
at P10 portion using the fragment library from the
InsightII/Builder program. In addition the b carbon
atom at the P10 residue was replaced by –NH– moiety,
which resulted in compounds 4 and 5. These were
further optimized within the enzyme binding site
using the AMBER force field in Discover/InsightII
program.27,53 The backbone of the protein was frozen
during optimization due to lack of the structural
changes of the enzyme upon ligand binding observed
in different X-ray structures of LAP. The distance con-
strains between zinc ions and their ligands with the force
(2.81 g, 5.20 mmol) and 1-adamantylbromide (1.2 g,
6.20 mmol) were dissolved in CHCl3 (15 mL) and the
reaction mixture brought to reflux. Then, silver oxide
(1.2 g, 5.20 mmol) was added in five equal portions, over
50 min. This suspension was refluxed for additional 3 h.
Then, chloroform was removed; the residue was treated
with Et2O and filtered through a pad of Celite. The fil-
trate was concentrated and the residue was purified by
column chromatography using CHCl3/isopropanol
9.8:0.2 as the eluent to give compound 18 as solid. Yield:
80% (2.77 g), mp 65–67 ꢁC. 1H NMR (200 MHz,
CDCl3) d 1.46 (s, 9 H, C(CH3)3), 1.51–1.77 (m, 8H,
PCHCHH, PCHH, CHCH2CH of Ad), 1.97–2.45 (m,
12H, PCHCHH, PCHH, CHCO, CCH2 of Ad, CH of
Ad), 2.58–2.92 (m, 2H, CH2Ph), 3.74 (s, 3H, OCH3),
3
3.84–4.10 (m, 1H, PCH), 4.86–5.0 (d, JHH = 14 Hz,
2
˚
constants values of 500 kcal/mol · A were applied
1H, PCHNH), 5.13 (s, 2H, PhCH2O), 5.97–6.09 (d,
3JHH = 12 Hz, 1H, CH2CHNH), 7.10–7.45 (m, 10H,
aryl); 13C NMR (50 MHz, CDCl3) d 28.0 (C(CH3)3),
29.7 (PCHCH2), 30.3 (PCH2), 30.8 (CH of Ad), 31.8
during minimization procedure. The optimization was
performed using the conjugate gradient algorithm until
˚
the maximum derivative was <0.1 kcal/mol · A.
3
(d, JPC = 16 Hz,CH2Ph), 35.2 (CHCH2CH of Ad),
Compounds 6, 7, and 8 were obtained by introducing
the methyl group into the side chain of 3, 4, and 5,
respectively.
43.8 (CCH2 of Ad), 48.6 (PCH), 49.2 (CHCO), 52.2
(COOCH3), 66.5 (PhCH2O), 80.1 (C(CH3)3), 83.7
(POC), 125.6, 127.1, 127.7, 128.0, 129.0, 128.6, 129.0,
136.0, 140.8 (aryl), 155.8 (CONH), 156.4 (COOCH2Ph),
3
3.3. Preparation of synthons 13 and 14 for cross-coupling
and alkylation
171.5 (d, JPC = 12.9 Hz, COOCH3); 31P (81 MHz,
CDCl3) d 46.6, 47.9, 48.1, 48.5; ESMS m/z calcd for
C39H49N2O8P (M+H)+ 669.8, found 669.3.
3.3.1. 2-Benzyloxycarbonylamino-3-[(1-tert-butyloxycar-
bonylamino-3-phenylpropyl)hydroxyphosphinoyl]propionic
acid methyl ester (17). A suspension of compound 16
(2.46 g, 8.0 mmol) in hexamethyldisilazane (5 mL,
24.0 mmol) was heated at 110 ꢁC for 1 h under argon
atmosphere. After cooling to 90 ꢁC, acrylate 15
(2.08 g, 8.80 mmol) was added dropwise over 0.5 h and
the reaction mixture was allowed to cool to ca. 65 ꢁC,
and MeOH (5 mL) was added dropwise. After cooling
to room temperature, the volatile products were re-
moved in vacuo and the residue was dissolved in a mix-
ture AcOEt/Et2O 1:1 and treated with 1 M HCl and
brine. The organic layer was dried over Na2SO4 and
evaporated to dryness. The residue was purified by col-
umn chromatography using CH2Cl2/MeOH/CH3COOH
9:0.1:0.1 to 9:0.5:0.2 as the eluent to afford the com-
pound 17 as a white solid. Yield: 77% (3.3 g), mp 85–
90 ꢁC. 1H NMR (200 MHz, CDCl3) d 1.42 (s, 9H,
C(CH3)3), 1.62–1.90 (m, 2H, PCHCHH, PCHH),
1.98–2.29 (m, 3H, PCHCHH, PCHH, CHCO), 2.36–
2.79 (m, 2H, CH2Ph), 3.54 (s, 3H, OCH3), 3.67–3.81
(m, 1H, PCH), 4.62–4.87 (s, 1H, PCHNH), 5.09 (s,
2H, PhCH2O), 5.30–5.47 (s, 1H, CH2CHNH), 6.96–
7.45 (m, 10H, aryl); 13C NMR (50 MHz, CDCl3) d
28.3 (C(CH3)3), 28.4 (PCHCH2), 29.2 (PCH2), 32.1
(CH2Ph), 48.7 (d, 1JPC = 114.5 Hz, PCH), 49.6 (CHCO),
52.5 (COOCH3), 66.9 (PhCH2O), 80.1 (C(CH3)3), 125.8,
127.9, 128.3, 136.3, 141.0 (aryl), 156.2 (CONH), 156.4
3.3.3. 3-[(Adamantan-1-yloxy)-(1-tert-butyloxycarbon-
ylamino-3-phenylpropyl)phosphinoyl]-2-aminopropionic
acid methyl ester (14). Compound 18 (0.23 g, 0.34 mmol)
was dissolved in MeOH (15 mL) and H2O (2 mL), and
subjected to hydrogenolysis for 3.5 h in the presence of
10% Pd/BaSO4 as catalyst. The catalyst was removed
by filtration through a pad of Celite and the filtrate
was concentrated to dryness. MeOH was added to the
residue and concentrated to dryness. This procedure
was repeated twice to afford product 14 as a colorless
gum. The product was used immediately in the next
step. Yield: 100% (0.18 g). 1H NMR (200 MHz, CDCl3)
d 1.52 (s, 9H, C(CH3)3), 1.57–1.82 (m, 8H, PCHCHH,
PCHH, CHCH2CH of Ad), 1.91–2.38 (m, 12H,
PCHCHH, PCHH, CHCO, CCH2 of Ad, CH of Ad),
2.56–3.00 (m, 2H, CH2Ph), 3.77 (s, 3H, OCH3), 3.86–
3
4.29 (m, 1H, PCH), 5.15–5.25 (d, JHH = 10 Hz, 1H,
PCHNH), 6.18–6.35 (d, 3JHH = 10.6 Hz, 1H,
CH2CHNH), 7.15–7.40 (m, 5H, aryl); 13C NMR
(50 MHz, CDCl3) d 28.0 (C(CH3)3), 30.3 (PCHCH2),
30.5 (PCH2), 30.8 (CH of Ad), 32.1 (d,
3JPC = 5 Hz,CH2Ph), 35.3 (CHCH2CH of Ad), 43.9
(CCH2 of Ad), 49.0 (PCH), 51.8 (CHCO), 52.1
(COOCH3), 79.3 (C(CH3)3), 83.1 (POC), 125.6, 128.0,
128.1, 128.2, 129.0, 128.6, 129.0, 141.0 (aryl), 155.7
(CONH), 173.9 (COOCH3); 31P (81 MHz, CDCl3) d
46.2, 46.6, 48.1, 48.6, 49.3, 49.5.
3
(COOCH2Ph), 172.4 (d, JPC = 11.5 Hz, COOCH3);
31P (81 MHz, CDCl3) d 44.4 (br); ESMS m/z calcd for
3.3.4. 3-[(Adamantan-1-yloxy)-(1-tert-butyloxycarbon-
ylamino-3-phenylpropyl)phosphinoyl]-2-aminopropionic
acid methyl ester (13). Compound 18 (0.67 g, 1 mmol)
was dissolved in MeOH, and 4 M NaOH (2 mmol)
was added dropwise. The final concentration of the
NaOH was 0.4 M. The mixture was stirred at room tem-
C26H35N2O8P (M+H)+ 535.5, found 535.3.
3.3.2. 3-[(Adamantan-1-yloxy)-(1-tert-butyloxycarbon-
ylamino-3-phenylpropyl)phosphinoyl]-2-benzyloxycarbon-
ylaminopropionic acid methyl ester (18). Compound 17

3196
S. Vassiliou et al. / Bioorg. Med. Chem. 15 (2007) 3187–3200
perature for 16 h. Then, methanol was removed at re-
duced pressure and the residue was diluted with water
and acidified with 0.5 M HCl to pH 1 in an ice water
bath. This aqueous solution was extracted with AcOEt,
the organic layer was washed with water and brine,
dried over Na2SO4, and concentrated in vacuo to give
the crude product, which was purified by silica column
chromatography using CH2Cl2/methanol 9.5:0.5 as the
eluent to afford the acids as colorless gum. Yield: 92%.
1H NMR (200 MHz, CDCl3) d 1.40 (s, 9H, C(CH3)3),
1.45–1.75 (m, 8H, PCHCHH, PCHH, CHCH2CH of
Ad), 1.80–2.22 (m, 12H, PCHCHH, PCHH, CHCO,
CCH2 of Ad, CH of Ad), 2.38–2.93 (m, 2H, CH2Ph),
4.46–4.68 (m, 1H, PCH), 5.05 (s, 2H, PhCH2O), 6.21
0.30 mmol), L-proline (3.4 mg, 0.03 mmol), and iodo-
benzene (17 lL, 0.15 mmol) were added at room temper-
ature. The reaction mixture was heated at 65 ꢁC for
48 h. Compound 20 was obtained following the previ-
ously described work-up. Yield: 22% (20 mg).
(C) Phosphinic compound 13 (50 mg, 0.10 mmol) was
dissolved in 2-propanol (0.5 mL) and stirred under
argon. Then, CuI (1.9 mg,
0 01 mmol), K2CO3
(28 mg, 0.2 mmol), ethylene glycol (11 lL, 0.10 mmol),
and iodobenzene (11 lL, 0.10 mmol) were added at
room temperature. The reaction mixture was heated
at 65 ꢁC for 48 h. Compound 20 was obtained follow-
ing the previously described work-up. Yield: 21%
(12.5 mg).
3
(d, JHH = 11 Hz, 1H, CH2CHNH), 6.30 (PCHNH),
6.95–7.46 (m, 10H, aryl); 13C NMR (50 MHz, CDCl3)
d 28.2 (C(CH3)3), 30.5 (PCHCH2), 31.0 (PCH2), 31.1
(CH of Ad), 32.0 (CH2Ph), 35.4 (CHCH2CH of Ad),
43.9 (CCH2 of Ad), 49.1 (PCH), 49.7 (CHCO), 66.8
(PhCH2O), 79.6 (C(CH3)3), 84.8 (POC), 125.9, 127.3,
128.0, 128.3, 128.4, 136.0, 141.0 (aryl), 155.6 (CONH),
156.0 (COOCH2Ph), 172.4 (COOCH3); 31P (81 MHz,
CDCl3) d 48.3, 49.5, 49.7, 50.5. Subsequent hydrogena-
tion starting from the same quantity (0.22 g, 0.34 mmol)
and using the same procedure as described above for
compound 14 gave the analogue 13. Yield: 100%
1H NMR (200 MHz, CDCl3) d 1.27 (s, 9H, C(CH3)3),
1.37–1.9 (m, 8H, PCHCHH, PCHH, CHCH2CH of
Ad), 1.94–2.27 (m, 11H, PCHCHH, PCHH, CCH2 of
Ad, CH of Ad), 2.58–2.95 (m, 2H, CH2Ph), 3.41–3.61
(m, 1H, CHCO), 3.79–4.30 (m, 1H, PCH), 5.69
(PCHNH), 7.13–7.37, 7.45–7.62, 7.69–7.82 (m, 10H,
aryl); 13C NMR (50 MHz, CDCl3) d 29.3 (PCH2), 29.7
(C(CH3)3), 30.7 (PCHCH2), 31.2 (CH of Ad), 31.9
(CH2Ph), 35.5 (CHCH2CH of Ad), 44.1 (CCH2 of
Ad), 49.2 (PCH), 65.6 (CHCO), 79.4 (C(CH3)3), 83.4
(POC), 117.2, 126.0, 128.4, 128.8, 130.9, 140.5, 141.0
(aryl), 155.9 (CONH), 173.4 (COOH); 31P (81 MHz,
CDCl3) d 42.8, 43.6.
1
(0.18 g), mp 95–97 ꢁC. H NMR (200 MHz, CDCl3) d
1.52 (s, 9H, C(CH3)3), 1.56–1.88 (m, 8H, PCHCHH,
PCHH, CHCH2CH of Ad), 1.90–2.47 (m, 11H,
PCHCHH, PCHH, CCH2 of Ad, CH of Ad), 2.54–
3.04 (m, 2H, CH2Ph), 3.3–3.6 (m, 1H, CHCO), 3.79–
4.30 (m, 1H, PCH), 5.16 (PCHNH), 7.27 (s, 5H, aryl),
8.32 (COOH); 13C NMR (50 MHz, CDCl3) d 28.3
(C(CH3)3), 30.6 (PCHCH2), 30.7 (PCH2), 31.1 (CH of
Ad), 32.0 (CH2Ph), 35.5 (CHCH2CH of Ad), 45.1
(CCH2 of Ad), 49.2 (PCH), 68.1 (CHCO), 79.4
(C(CH3)3), 83.4 (POC), 126.1, 128.3, 128.5, 140.9 (aryl),
155.9 (CONH), 173.4 (COOH); 31P (81 MHz, CDCl3) d
49.7, 52.0, 53.5.
3.4.2. 4-Iodophenylmethanol (22). To a stirred solution of
4-iodobenzoic acid 21 (0.99 g, 4 mmol) in THF (20 mL)
at ꢀ10 ꢁC, N-methylmorpholine (0.40 g, 4 mmol) was
added, followed by isobutyl chloroformate (0.55 g,
4 mmol). After 10 min, NaBH4 (0.45 g, 12 mmol) was
added in one portion. MeOH (40 mL) is then added
dropwise to the mixture over a period of 10 min at
0 ꢁC. The solution is stirred for additional 10 min, and
then neutralized with 1 N HCl (8 mL). The organic sol-
vents are evaporated under reduced pressure and the
product is extracted with diethyl ether. The organic
phase is washed with 1 N HCl, 5% NaHCO3, H2O, dried
over Na2SO4, and the solvent is evaporated under re-
duced pressure. The residue was purified by column
chromatography using diethyl ether/light petroleum
ether 8:2 as the eluent to give 22 as white solid. Yield:
70% (0.65 g), mp 72–73 ꢁC. 1H NMR (200 MHz,
CDCl3) d 1.77 (br s, 1H, CH2OH), 4.64 (s, 2H, CH2OH),
7.09–7.70 (dd, 4H, J = 8.1 Hz, aryl). 13C NMR
(50 MHz, CDCl3) d 68.5 (CH2OH), 96.2 (CI), 128.9,
137.5, 139.8 (aryl).
3.4. Cross-coupling with the use of the building block 13
3.4.1. 3-[(Adamantan-1-yloxy)-(1-tert-butyloxycarbonyl-
amino-3-phenylpropyl)phosphinoyl]-2-phenylaminopropionic
acid (20). General procedures. (A) Phosphinic com-
pound 13 (55 mg, 0.11 mmol) was dissolved in DMSO
(0.5 mL) and stirred under argon. Then, CuI (2 mg,
0.011 mmol), K2CO3 (30 mg, 0.22 mmol), and iodoben-
zene (12 lL, 0.11 mmol) were added at room tempera-
ture. The reaction mixture was heated at 65 ꢁC for
48 h. The mixture was allowed to cool to room temper-
ature. Et2O and saturated solution of citric acid were
added to the mixture. The organic layer was separated,
washed with brine, and dried over Na2SO4. The solvent
was removed in vacuo to yield the crude product that
was purified by column chromatography using
CH2Cl2/MeOH 10:0 to 9.8:0.2 as the eluent. The prod-
uct 20 was obtained as deep yellow oil. Yield: 22%
(15 mg).
3.4.3. 1-Bromomethyl-4-iodobenzene (23). PBr3 (0.19 mL,
2 mmol) was added to an ice-salt cooled solution of 22
(0.47 g, 2 mmol) in diethyl ether (10 mL) and the reac-
tion mixture was stirred at room temperature for 3 h.
Then it was cooled with ice and quenched with water.
The organic phase was dried over Na2SO4 and the
solvent was evaporated. The residue was purified by
column chromatography using light petroleum ether as
the eluent to give 23 as white solid. Yield: 58%
(B) Phosphinic compound 13 (76 mg, 0.15 mmol) was
dissolved in DMSO (0.5 mL) and stirred under argon.
Then, CuI (3 mg, 0.015 mmol), K2CO3 (41.5 mg,
1
(0.34 g), mp 78–79 ꢁC. H NMR (200 MHz, CDCl3) d
4.41 (s, 2H, CH2Br), 7.10-7.70 (dd, 4H, J = 8.1 Hz, aryl).

S. Vassiliou et al. / Bioorg. Med. Chem. 15 (2007) 3187–3200
3197
13C NMR (50 MHz, CDCl3) d 32.7 (CH2Br), 94.4 (CI),
131.1, 137.6, 138.2 (aryl).
3.5. N-Alkylation of the building block 14
3.5.1. General procedure. To a stirred solution of com-
pound 14 (535 mg, 1 mmol, 1 equiv) in anhydrous
DMF (4 mL), cesium carbonate (164 mg, 0.5 mmol, 1
equiv) and the appropriate (non- or p-substituted) ben-
zyl bromide (1 mmol, 1 equiv) were added. The reaction
mixture was stirred at 45 ꢁC for 6 h and then treated
with H2O and Et2O. The organic layer was dried over
Na2SO4 and concentrated in vacuo. The residue was
purified by column chromatography using CH2Cl2/iso-
propanol 9.8:0.2 as the eluent to afford the desired
product.
3.4.4. 4-[(N-tert-Butyloxycarbonylamino)methyl]iodoben-
zene (24). A solution of p-iodobenzyl bromide 23 (0.39 g,
1.01 mmol) in 28% aqueous NH4OH (100 mL) was stir-
red at room temperature for 5 days. The reaction mix-
ture was extracted with CH2Cl2, and the extract was
washed with brine, dried (NaOH), and concentrated to
give the amino derivative, which was dissolved in
CH2Cl2 (5 mL). After cooling with ice, di-tert-butyl
dicarbonate (0.33 g 1.50 mmol) was added and the mix-
ture was stirred at room temperature for 15 h. The solu-
tion was diluted with dichloromethane, washed with
H2O, dried over Na2SO4, and evaporated under vac-
uum. The residue was purified by column chromatogra-
phy using diethyl ether/light petroleum ether 1:1 as the
eluent to give 24 as white solid. Yield: 72% (0.24 g),
3.5.2. 3-[(Adamantan-1-yloxy)-(1-tert-butyloxycarbon-
ylamino-3-phenylpropyl)phosphinoyl]-2-benzylamino-
1
propionic acid methyl ester (28). Yield: 80% H NMR
(200 MHz, CDCl3) d 1.29 (s, 9H, C(CH3)3), 1.44–1.72
(m, 8H, PCHCHH, PCHH, CHCH2CH of Ad), 1.87–
2.28 (m, 12H, PCHCHH, PCHH, CHCO, CCH2 of
Ad, CH of Ad), 2.51–2.90 (m, 2H, CH2Ph), 3.72 (s,
3H, OCH3), 3.74 (s, 2H, NHCH2), 3.86–4.10 (m, 1H,
1
mp 88–89 ꢁC. H NMR (200 MHz, CDCl3) d 1.43 (s,
9H, C(CH3)3), 4.21 (s, 2H, CH2), 4.85 (br s, 1H, NH),
6.99–7.65 (dd, 4H, J = 8.1 Hz, aryl). 13C NMR
(50 MHz, CDCl3) d 28.6 (CH3)3, 44.3 (CH2NH), 79.6
(C(CH3)3), 92.8 (CI), 129.5, 137.8, 138.9 (aryl), 156.0
(CO).
3
PCH), 4.95 (d, JHH = 10.6 Hz, 1H, PCHNH), 6.05 (d,
3JHH = 10.4 Hz, 1H, CH2CHNH), 7.06–7.42 (m, 10H,
aryl); 13C NMR (50 MHz, CDCl3) d 28.3 (C(CH3)3),
30.6 (PCHCH2), 30.7 (PCH2), 31.1 (CH of Ad), 31.8
(CH2Ph), 35.6 (CHCH2CH of Ad), 44.2 (CCH2 of
Ad), 52.1 (PCH), 52.2 (CHCO), 52.4 (COOCH3), 52.6
(NHCH2), 79.3 (C(CH3)3), 83.8 (POC), 125.9, 127.2,
128.3, 128.4, 129.0, 141.0 (aryl), 155.7 (CONH), 173.9
(COOCH3); 31P (81 MHz, CDCl3) d 45.8, 46.7, 47.5,
48.1, 48.4.
3.4.5. 3-[(Adamantan-1-yloxy)-(1-tert-butyloxycarbon-
ylamino-3-phenylpropyl)phosphinoyl]-2-(4-hydroxymeth-
ylphenylamino)propionic acid (25). Compound 25 was
obtained from 13 and 22 following the method A. Yield:
10%. 1H NMR (200 MHz, CDCl3) d 1.27 (s, 9H,
C(CH3)3), 1.37–1.9 (m, 9H, PCHCHH, PCHH,
CHCH2CH of Ad, OH), 1.94–2.27 (m, 11H, PCHCHH,
PCHH, CCH2 of Ad, CH of Ad), 2.58–2.95 (m, 2H,
CH2Ph), 3.41–3.61 (m, 1H, CHCO), 3.79–4.30 (m, 1H,
PCH), 4.69 (m, 2H, CH2OH), 5.69 (PCHNH), 7.13–
7.37, 7.45–7.62, 7.69–7.82 (m, 9H, aryl); 13C NMR
(50 MHz, CDCl3) d 29.3 (PCH2), 29.7 (C(CH3)3), 30.7
(PCHCH2), 31.2 (CH of Ad), 31.9 (CH2Ph), 35.5
(CHCH2CH of Ad), 44.1 (CCH2 of Ad), 49.2 (PCH),
65.6 (CHCO), 68.5 (CH2OH), 79.4 (C(CH3)3), 83.4
(POC), 117.2, 126.0, 128.4, 128.8, 130.9, 140.5, 141.0
(aryl), 155.9 (CONH), 173.4 (COOH); 31P (81 MHz,
CDCl3) d 43.6, 45.6, 46.7.
3.5.3. Methyl 4-bromomethylbenzoate (31). To a solution
of 4-bromomethylbenzoic acid 30 (0.5 g, 2.30 mmol) in
MeOH (5.6 mL), concentrated sulfuric acid (0.14 mL)
was added. The resulting mixture was refluxed for 5 h.
Then, the mixture was cooled to room temperature
and evaporated in vacuo. H2O (20 mL) was added to
the residue in an ice-water bath, and the resulting solid
was filtered and washed with cold water. The solid mate-
rial was partitioned between Et2O/AcOEt 1:1 and
Na2CO3. The organic layer was dried over Na2SO4,
and evaporated in vacuo to afford the product as pale
yellow oil. Yield: 93% (0.49 g). 1H NMR (200 MHz,
CDCl3) d 3.82 (s, 3H, CH3), 4.51 (s, 2H, CH2Br),
7.33–7.95 (dd, 4H, aryl); 13C NMR (50 MHz, CDCl3)
d 45.0 (CH2Br), 51.8 (CH3), 126.8, 128.2, 128.7, 129.3,
129.6 (aryl), 142.0 (CCH2Br), 166.1 (CO).
3.4.6. 3-[(Adamantan-1-yloxy)-(1-tert-butyloxycarbon-
ylamino-3-phenylpropyl)phosphinoyl]-2-[4-(tert-butyloxy-
carbonylaminomethyl)phenylamino]propionic acid (26).
Compound 26 was obtained from 13 and 24 following
the method A. Yield: 8%. 1H NMR (200 MHz,
CDCl3) d 1.27 (s, 18H, C(CH3)3), 1.37–1.9 (m, 8H,
PCHCHH, PCHH, CHCH2CH of Ad), 1.94–2.27
(m, 11H, PCHCHH, PCHH, CCH2 of Ad, CH of
Ad), 2.58–2.95 (m, 2H, CH2Ph), 3.41–3.61 (m, 1H,
CHCO), 3.79–4.30 (m, 1H, PCH), 4.27 (m, 2H,
CH2NHBoc), 5.69 (PCHNH), 7.13–7.37, 7.45–7.62,
7.69–7.82 (m, 9H, aryl); 13C NMR (50 MHz, CDCl3)
d 29.3 (PCH2), 29.7 (C(CH3)3), 30.7 (PCHCH2), 31.2
(CH of Ad), 31.9 (CH2Ph), 35.5 (CHCH2CH of
Ad), 44.1 (CCH2 of Ad), 44.38 (CH2NHBoc) 49.2
(PCH), 65.6 (CHCO), 79.4 (C(CH3)3), 83.4 (POC),
117.2, 126.0, 128.4, 128.8, 130.9, 140.5, 141.0 (aryl),
155.9 (CONH), 173.4 (COOH); 31P (81 MHz, CDCl3)
d 42.1, 43.9.
3.5.4. 4-Bromomethylphenylmethanol (32). To an ice-
cooled solution of 1 M DIBALH in hexane (5.4 ml,
5.40 mmol), toluene (1.2 mL) was introduced under ar-
gon atmosphere. A solution of 31 (0.49 g, 2.1 mmol) in
toluene (5 ml) was then added dropwise and the mixture
was stirred at 0 ꢁC for 4 h. The reaction mixture was
quenched using toluene/methanol 1:1 (5 mL), followed
by 2 M hydrochloric acid solution (5 mL). The solid alu-
minum salts were filtered, the organic layer was sepa-
rated, and the aqueous layer was extracted with
diethyl ether. The organic layer was dried over Na2SO4
and evaporated in vacuo. A mixture of diethyl ether/
light petroleum ether 1:1 was added to the white solid

3198
S. Vassiliou et al. / Bioorg. Med. Chem. 15 (2007) 3187–3200
residue. The precipitate was filtered and washed with
petroleum ether. Yield: 58% (0.24 g), mp 76–77 ꢁC. H
NMR (200 MHz, CDCl3) d 1.67 (br s, 1H, OH), 4.50
(s, 2H, CH2Br), 4.69 (s, 2H, CH2OH), 7.32–7.42 (m,
4H, aryl); 13C NMR (50 MHz, CDCl3) d 33.3 (CH2Br),
64.4 (CH2OH), 126.9, 127.1, 128.2, 129.1 (aryl), 136.8
(CCH2Br), 141.0 (CCH2OH).
was triturated with acetone to yield the product as a pale
1
yellow solid. Yield: 15% (0.15 g), mp 250-252 ꢁC. 1H
NMR (200 MHz, d6-DMSO)
d
3.98 (m, 2H,
CH2NH3Br), 4.68 (CH2Br), 7.31–7.44 (m, 4H, aryl),
7.97–8.21 (br s, 3H, NH3); 13C NMR (50 MHz, d6-
DMSO) d 34.0 (CH2Br), 41.9 (CH2NH3Br), 129.3,
129.5 (aryl), 134.0 (CCH2Br), 138.4 (CCH2N).
3.5.5. 1-Bromomethyl-4-tert-butyloxymethylbenzene (33).
To an ice-cooled solution of 32 (0.15 g, 0.75 mmol) in
CH2Cl2 (5 mL), a catalytic amount of concentrated sul-
furic acid was added. Then, a large excess of liquid
isobutene (20 mL) was added and the mixture was stir-
red at room temperature for 64 h. After careful removal
of isobutene, the reaction mixture was partitioned be-
tween CH2Cl2 and Na2CO3, the organic layer was
washed with H2O and brine, dried over Na2SO4, and
concentrated in vacuo. The resulting colorless oil was
purified by column chromatography using light petro-
leum ether/diethyl ether 5:1 as the eluent to afford the
3.5.8. (4-Bromomethylbenzyl)carbamic acid tert-butyl
ester (37). To a solution of 36 (85 mg, 0.30 mmol) in
dioxane (2.5 mL) and H2O (2.5 mL), Boc2O (1.32 g,
6.0 mmol) was added at 0 ꢁC. Then, NaHCO3
(29.4 mg, 0.35 mmol) was added at this temperature
and the resulting mixture was stirred for 18 h at room
temperature. The mixture was partitioned between
Et2O and H2O. The aqueous phase was extracted with
Et2O and the combined organic phases were dried over
Na2SO4 and concentrated in vacuo. The residue was
purified by column chromatography using light petro-
leum ether/diethyl ether 3:2 as the eluent to afford the
1
1
product as colorless oil. Yield: 63% (0.12 g). H NMR
product as colorless oil. Yield: 99% (90 mg). H NMR
(200 MHz, CDCl3) d 1.30 (s, 9H, C(CH3)3), 4.45 (s,
2H, CH2Br), 4.50 (s, 2H, CH2O), 7.30–7.35 (m, 4H, ar-
yl);13C NMR (50 MHz, CDCl3) d 27.6 (C(CH3)3), 33.5
(CH2Br), 63.6 (CH2O), 73.4 (C(CH3)3), 127.6, 128.9
(aryl), 136.5 (CCH2Br), 140.2 (CCH2O).
(200 MHz, CDCl3) d 1.49 (s, 9H, C(CH3)3), 4.27 (s,
2H, CH2NH), 4.46 (s, 2H, CH2Br), 4.93 (br s, 1H,
NH), 7.19–7.34 (dd, 4H, aryl); 13C NMR (50 MHz,
CDCl3) d 28.3 (C(CH3)3), 33.2 (CH2NH), 44.2 (CH2Br),
79.5 (C(CH3)3), 127.7, 129.2 (aryl), 136.7 (CCH2N),
139.3 (CCH2Br), 155.8 (CO).
3.5.6. 4-Hydroxymethylbenzonitrile (35). 4-Bromo-
methylbenzonitrile 34 (1.4 g, 7.14 mmol) was refluxed
in deionized water (43 mL) with barium carbonate
(2.8 g, 14.20 mmol) for 4 h. Then, the mixture was
cooled to room temperature, and the solid byproduct
was removed by filtration. The filtrate was extracted
with CH2Cl2 twice. The organic phases were combined,
dried over Na2SO4 and evaporated in vacuo. The residue
was purified by column chromatography using light
petroleum ether/diethyl ether 2:3 as the eluent to afford
35 as white solid. Yield: 89 % (0.85 g), mp 49–50 ꢁC. 1H
NMR (200 MHz, CDCl3) d 2.68 (s, 1H, OH), 4.67 (s,
2H, CH2), 7.36–7.60 (dd, 4H, aryl); 13C NMR
(50 MHz, CDCl3) d 63.5 (CH2), 110.3 (CCN), 118.7
(CN), 126.8, 132.0 (aryl), 146.5 (CCH2OH).
3.5.9. 3-[(Adamantan-1-yloxy)-(1-tert-butyloxycarbon-
ylamino-3-phenylpropyl)phosphinoyl]-2-(4-tert-butyloxy-
methylbenzylamino)propionic acid methyl ester (39).
Yield: 65%. 1H NMR (200 MHz, CDCl3) d 1.24 (s,
9H, CH2OC(CH3)3), 1.40 (s, 9H, OCOC(CH3)3), 1.44–
1.74 (m, 8H, PCHCHH, PCHH, CHCH2CH of Ad),
1.86–2.47 (m, 12H, PCHCHH, PCHH, CHCO, CCH2
of Ad, CH of Ad), 2.57–2.80 (m, 2H, CH2Ph), 3.69 (s,
3H, OCH3), 3.72 (s, 2H, NHCH2), 3.86–4.10 (m, 1H,
3
PCH), 4.40 (s, 2H, CH2O), 5.18 (d, JHH = 7.8 Hz, 1H,
3
PCHNH), 6.22 (d, JHH = 9.4 Hz, 1H, CH2CHNH),
7.28–7.38 (m, 9H, aryl); 13C NMR (50 MHz, CDCl3) d
27.5 (CH2OC(CH3)3), 28.2 (OCOC(CH3)3), 29.5
(PCHCH2), 30.5 (PCH2), 31.0 (CH of Ad), 32.3
(CH2Ph), 35.5 (CHCH2CH of Ad), 44.0 (CCH2 of
Ad), 49.2 (PCH), 50.0 (CHCO), 52.1 (COOCH3), 55.3
(NHCH2), 63.7 (CH2O), 73.2 (CH2OC(CH3)3), 79.4
(C(CH3)3), 83.5 (POC), 125.8, 126.7, 127.2, 127.4,
128.2, 128.3, 137.4, 138.9, 141.3 (aryl), 155.7 (CONH),
174.4 (COOCH3); 31P (81 MHz, CDCl3) d 46.1, 47.2,
47.6, 48.4, 48.6.
3.5.7. 4-Bromomethylbenzylamine hydrobromide (36).
p-Cyanobenzyl alcohol 35 (0.42 g, 3.15 mmol) was dis-
solved in dry THF (8 mL) and was added dropwise to
LiAlH4 (0.30 g, 7.88 mmol) in dry THF (8 mL) over
15 min at 0 ꢁC under argon. After the end of the addi-
tion, the reaction mixture was refluxed for 4 h. Then,
the solvent was removed in vacuo and 1 M NaOH
(10 mL) was added dropwise at 0 ꢁC. The resulting mix-
ture was treated with AcOEt (3 · 30 mL), the organic
layers were combined and extracted with 1 M HCl (2 ·
70 mL), and the solvent was removed using rotary evap-
oration. The residue was dissolved in H2O (3 mL) and
the solution was neutralized with 1 M NaOH (5 mL).
The solution was extracted with AcOEt (5 · 30 mL),
the combined organic phases were dried over Na2SO4
and concentrated in vacuo to yield the a-amino-a0-hy-
droxy-p-xylene as a pale yellow solid. This solid was dis-
solved in H2O (3 mL) and HBr 46% (4.6 mL) was added.
The resulting solution was refluxed for 3.5 h. Solvent
was removed under reduced pressure and the residue
3.5.10. 3-[(Adamantan-1-yloxy)-(1-tert-butyloxycarbon-
ylamino-3-phenylpropyl)phosphinoyl]-2-[4-(tert-butyloxy-
carbonylaminomethyl)benzylamino]propionic acid methyl
1
ester (40). Yield: 69%. H NMR (200 MHz, CDCl3) d
1.46 (s, 18H, C(CH3)3), 1.50–1.81 (m, 8H, PCHCHH,
PCHH, CHCH2CH of Ad), 1.86–2.30 (m, 12H,
PCHCHH, PCHH, CHCO, CCH2 of Ad, CH of Ad),
2.36–2.95 (m, 2H, CH2Ph), 3.72 (s, 3H, CH3), 3.74 (s,
2H, NHCH2), 3.86–4.10 (m, 1H, PCH), 4.41 (s, 2H,
3
CH2NHBoc), 5.00 (BocNH), 5.98 (d, JHH = 10.6 Hz,
1H, CH2CHNH), 7.08–7.47 (m, 9H, aryl); 13C NMR
(50 MHz, CDCl3) d 27.3 (C(CH3)3), 29.6 (PCHCH2),
30.6 (PCH2), 31.1 (CH of Ad), 32.2 (CH2Ph), 35.5

S. Vassiliou et al. / Bioorg. Med. Chem. 15 (2007) 3187–3200
3199
(CHCH2CH of Ad), 36.0 (CH2NHBoc), 44.1 (CCH2 of
Ad), 51.3 (PCH), 51.8 (CHCO), 52.1 (COOCH3), 55.3
(NHCH2), 79.6 (C(CH3)3), 83.6 (POC), 125.9, 127.5,
128.3, 137.7, 141.2 (aryl), 155.8 (CONH), 173.9
(COOCH3); 31P (81 MHz, CDCl3) d 46.1, 48.3, 48.7.
ESMS m/z calcd for C20H27N2O4P (M+H)+ 391.40,
found 391.18.
3.6.5. 3-[(1-Amino-3-phenylpropyl)hydroxyphosphinoyl]-
2-benzylaminopropionic acid (6). HPLC: tR = 26.9 and
1
28.7 min using gradient 1. H NMR (200 MHz, D2O)
3.6. Deprotection
d 1.80–2.48 (m, 4H, PCHCHH, PCHH, PCHCHH,
PCHH), 2.55–2.93 (m, 2H, CH2Ph), 3.06–3.29 (m, 1H,
CHCO), 3.53–3.96 (m, 1H, PCH), 4.08–4.51 (m, 2H,
NHCH2), 7.12–7.53 (m, 10H, aryl); 13C NMR
3.6.1. General procedure for removal of Boc, t-Bu, and Ad
protecting group. Initially, in dipeptides 28, 39, and 40
the C-terminal methyl ester was removed by saponifica-
tion in same conditions as for compound 18 to give 13.
Then, 20, 25, 26, 28, and 40 were treated with a mixture
50% TFA/CH2Cl2 while 39 with 90% TFA/CH2Cl2 and
one drop of a scavenger (triisopropylsilane). The result-
ing solution was stirred for 3 h (50% TFA) or for 16 h
(90% TFA) at room temperature. Then, the mixture
was concentrated in vacuo. CH2Cl2 was added to the
residue, and the solvent was removed under reduced
pressure (3 times). A mixture of Et2O/petroleum ether
1:1 was added to the residue. The white precipitate
was filtered and washed with Et2O. The product was
purified with semi-preparative RP-HPLC using suitable
gradient.
1
(50 MHz, D2O) d 27.5 (d, JPC = 89 Hz, PCH2), 31.7
(d, JPC = 8 Hz, CH2Ph), 38.8 (CHCH2), 43.7 (PCH),
2
48.4 (CHCO), 50.3 (NHCH2), 126.8, 128.6, 129.0,
129.5, 130.0, 130.1, 130.5, 140.6 (aryl), 195.7 (COOH);
31P (81 MHz, D2O) d 31.6,39.2; ESMS m/z calcd for
C19H24N2O4P (MꢀH)+ 375.38, found 375.26.
3.6.6. 3-[(1-Amino-3-phenylpropyl)hydroxyphosphinoyl]-2-
(4-hydroxymethylbenzylamino)propionic acid (7). HPLC:
tR = 17.4 and 18.9 min using gradient 1. 1H NMR
(200 MHz, D2O) d 1.86–2.22 (m, 2H, PCHCHH,
PCHH), 2.26–2.48 (m, 2H, PCHCHH, PCHH), 2.60–
2.90 (m, 2H, CH2Ph), 3.06–3.27 (m, 1H, CHCO),
3.83–4.04 (m, 1H, PCH), 4.14–4.44 (m, 2H, NHCH2),
4.52 (s, 2H, CH2OH), 7.16–7.50 (m, 9H, aryl); 13C
1
3.6.2. 3-[(1-Amino-3-phenylpropyl)hydroxyphosphinoyl]-
2-phenylaminopropionic acid (3). HPLC: tR = 21.1 and
NMR (50 MHz, D2O) d 27.3 (d, JPC = 128 Hz,
PCH2), 30.3 (CH2Ph), 31.6 (d, 2JPC = 8.35 Hz, CHCH2),
49.1 (PCH), 50.0 (NHCH2), 51.0 (CHCO), 63.4
(CH2OH) 126.7, 128.2, 128.6, 129.0, 129.5, 129.8,
130.3, 140.3, 142.0 (aryl), 195.7 (COOH); 31P
(81 MHz, D2O) d 30.8, 31.3; ESMS m/z calcd for
C20H26N2O5P (MꢀH)+ 405.4, found 405.3.
1
21.4 min using gradient 3. H NMR (200 MHz, D2O)
d 1.75–2.25 (m, 4H, PCHCHH, PCHH, PCHH,
PCHCHH), 2.57–2.80 (m, 2H, CH2Ph), 2.85–2.95 (m,
1H, CHCO), 2.97–3.15 (m, 1H, PCH), 7.20–7.43 (m,
10H, aryl); 13C NMR (50 MHz, D2O) d 27.5 (PCH2),
31.0 (CH2Ph), 38.8 (CHCH2), 43.8 (PCH), 49.0
(CHCO), 110.8, 126.6, 128.7, 129.0, 129.5, 130.5, 140.6
(aryl), 195.7 (COOH); 31P (81 MHz, D2O) d 32.2, 34.4;
ESMS m/z calcd for C20H27N2O4P (MꢀH)+ 361.35,
found 361.01.
3.6.7. 2-(4-Aminomethylbenzylamino)-3-[(1-amino-3-phe-
nylpropyl)hydroxyphosphinoyl]propionic acid (8). HPLC:
tR = 14.9 and 16.1 min using gradient 2. 1H NMR
(200 MHz, D2O) d 1.81–2.22 (m, 4H, PCHCHH,
PCHH, PCHCHH, PCHH), 2.24–2.48 (m, 2H, CH2Ph),
2.56–2.89 (m, 1H, CHCO), 3.10–3.30 (m, 1H, PCH), 4.0
(s, 2H, CH2NH2), 4.21–4.48 (m, 2H, NHCH2), 6.94–
7.47 (m, 9H, aryl); 13C NMR (50 MHz, D2O) d 27.3
(d, 1JPC = 100 Hz, PCH2), 29.3 (CH2Ph), 31.8 (CHCH2),
36.9 (PCH), 42.6 (CHCO), 49.7 (NHCH2), 50.3
(CH2NH2) 126.7, 128.5, 128.9, 129.7, 130.7, 131.3,
134.3, 140.5 (aryl), 195.7 (COOH); 31P (81 MHz, D2O)
d 31.3, 31.6; ESMS m/z calcd for C20H27N2O4P
(MꢀH)+ 404.42, found 404.3.
3.6.3. 3-[(1-Amino-3-phenylpropyl)hydroxyphosphinoyl]-2-
(4-hydroxymethylphenylamino)propionic acid (4). HPLC:
tR = 20.7 and 21.2 min using gradient 3. 1H NMR
(200 MHz, D2O) d 1.75–2.25 (m, 4H, PCHCHH,
PCHH, PCHH, PCHCHH), 2.57–2.80 (m, 2H, CH2Ph),
2.85–2.95 (m, 1H, CHCO), 2.97–3.15 (m, 1H, PCH),
3.45–3.60 (m, 2H, CH2OH), 7.20–7.43 (m, 9H, aryl);
13C NMR (50 MHz, D2O) d 27.5 (PCH2), 31.0 (CH2Ph),
38.8 (CHCH2), 43.8 (PCH), 49.0 (CHCO), 68.0
(CH2OH), 110.8, 126.6, 128.7, 129.0, 129.5, 130.5,
140.6 (aryl), 195.7 (COOH); 31P (81 MHz, D2O) d
34.2, 35.9; ESMS m/z calcd for C21H29N2O5P (MꢀH)+
391.15, found 390.01.
3.7. Enzymatic assays
For the details of the kinetic studies: the enzymes, prep-
aration, activation, assays of LAP and APM activity,
and the Ki value determination, as well as computational
data for inhibitor docking, see the Ref. 27.
3.6.4. 2-(4-Aminomethylphenylamino)-3-[(1-amino-3-
phenylpropyl)hydroxyphosphinoyl]propionic acid (5).
HPLC: tR = 30.2 and 31.0 min using gradient 4. 1H
NMR (200 MHz, D2O) d 1.75–2.25 (m, 4H, PCHCHH,
PCHH, PCHH, PCHCHH), 2.57–2.80 (m, 2H, CH2Ph),
2.85–2.95 (m, 1H, CHCO), 2.97–3.15 (m, 1H, PCH),
3.44–3.60(m, 2H, CH2NH2), 7.20–7.43 (m, 9H, aryl);
13C NMR (50 MHz, D2O) d 27.5 (PCH2), 31.0 (CH2Ph),
38.8 (CHCH2), 43.8 (PCH), 46.3 (CH2NH2), 49.0
(CHCO), 110.8, 126.6, 128.7, 129.0, 129.5, 130.5, 140.6
(aryl), 195.7 (COOH); 31P (81 MHz, D2O) d 32.2, 34.4;
Acknowledgments
The financial support within Polish-Greek Scientific and
Technological International Cooperation Joint Project
for the Years 2004/2005 is gratefully acknowledged. This
work was also partially financed by Polish Ministry of
Education and Science.

3200
S. Vassiliou et al. / Bioorg. Med. Chem. 15 (2007) 3187–3200
29. Turner, A. J. In Handbook of Proteolytic Enzymes;
References and notes
Barrett, A. J., Rawlings, N. D., Woessner, J. F., Eds.;
Academic Press: Amsterdam, 2002; pp 996–1000.
1. Taylor, A. Trends Biochem. Sci. 1993, 18, 167–171.
2. Taylor, A. FASEB J. 1993, 7, 290–298.
3. Current Topics: Aminopeptidases in Health and Disease.
Biol. Pharm. Bull. 2004, 27, 760–780.
4. Stra¨ter, N.; Lipscomb, W. N. In Handbook of Proteolytic
Enzymes; Barrett, A. J., Rawlings, N. D., Woessner, J. F.,
Eds.; Academic Press: Amsterdam, 2002; pp 1384–1389.
5. Lowther, W. T.; Matthews, B. W. Chem. Rev. 2002, 102,
4581–4607.
6. Burley, S. K.; David, P. R.; Taylor, A.; Lipscomb, W. N.
Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 6878–6882.
7. Burley, S. K.; David, P. R.; Sweet, R. M.; Taylor, A.;
Lipscomb, W. N. J. Mol. Biol. 1992, 224, 113–140.
8. Kim, H.; Burley, S. K.; Lipscomb, W. N. J. Mol. Biol.
1993, 230, 722–724.
9. Kim, H.; Lipscomb, W. N. Biochemistry 1993, 32, 8465–
8478.
10. Stra¨ter, N.; Lipscomb, W. N. Biochemistry 1995, 34, 9200–
9210.
11. Kim, H.; Lipscomb, W. N. Adv. Enzymol. Relat. Areas
Mol. Biol. 1994, 68, 153–213.
12. Stra¨ter, N.; Lipscomb, W. N. Biochemistry 1995, 34,
14792–14800.
´
30. Chen, H.; Roques, B. P.; Fournie-Zaluski, M.-C. Bioorg.
Med. Chem. Lett. 1999, 9, 1511–1516.
31. Bauvois, B.; Dauzonne, D. Med. Res. Rev. 2006, 26, 88–
130.
32. Insight 2000 Molecular Modelling Program Package;
Accelrys, 2000.
33. Yiotakis, A.; Georgiadis, D.; Matziari, M.; Makaritis, A.;
Dive, V. Curr. Org. Chem. 2004, 8, 1135–1158.
34. Georgiadis, D.; Matziari, M.; Vassiliou, S.; Dive, V.;
Yiotakis, A. Tetrahedron 1999, 55, 14635–14648.
35. Photaki, I. J. Am. Chem. Soc. 1963, 85, 1123–1126.
´ˇ
36. Yiotakis, A.; Vassiliou, S.; Jiracek, J.; Dive, V. J. Org.
Chem. 1996, 61, 6601–6605.
37. Ullmann, F. Ber. Dtsch. Chem. Ges. 1903, 36, 2382–2384;
for review, see Lindley, J. Tetrahedron 1984, 40, 1433–1456.
38. Hartwig, J. F. Angew. Chem., Int. Ed. 1998, 37, 2046–
2067.
39. Ma, D.; Zhang, Y.; Yao, J.; Wu, S.; Tao, F. J. Am. Chem.
Soc. 1998, 120, 12459–12467.
40. Kwong, F. Y.; Klapars, A.; Buchwald, S. L. Org. Lett.
2002, 4, 581–584.
41. Wolfe, J. P.; Buchwald, S. L. J. Org. Chem. 2000, 65,
1144–1157.
42. Salvatore, R. N.; Yoon, C. H.; Jung, K. W. Tetrahedron
2001, 57, 7785–7811.
13. Stra¨ter, N.; Sun, L.; Kantrowitz, E. R.; Lipscomb, W. N.
Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 11151–11155.
14. Schurer, G.; Horn, A. H. C.; Gedeck, P.; Clark, T.
¨
43. Salvatore, R. N.; Nagle, A. S.; Jung, K. W. J. Org. Chem.
2002, 67, 674–683.
44. Gavin, J. A.; Garcia, M. E.; Benesi, A. J.; Mallouk, T. E.
J. Org. Chem. 1998, 63, 7663–7669.
45. Gall, A. L.; Ruff, M.; Kannan, R.; Cuniasse, P.; Yiotakis,
A.; Dive, V.; Rio, M.-C.; Basset, P.; Moras, D. J. Mol.
Biol. 2001, 307, 577–686.
46. Grams, F.; Dive, V.; Yiotakis, A.; Yiallouros, I.; Vassi-
liou, S.; Zwilling, R.; Bode, W.; Sto¨cker, W. Nat. Struct.
Biol. 1996, 3, 671–675.
J. Phys. Chem. B 2002, 106, 8815–8830.
15. Erhardt, S.; Jaime, E.; Weston, J. J. Am. Chem. Soc. 2005,
127, 3654–3655.
16. York, I. A.; Goldberg, A. L.; Mo, X. Y.; Rock, K. L.
Immunol. Rev. 1999, 172, 49–66.
17. Jo¨sch, C.; Klotz, L.-O.; Sies, H. Biol. Chem. 2003, 384,
213–218.
18. Cappiello, M.; Lazzarotti, A.; Buono, F.; Scaloni, A.;
´
D’Ambrosio, C.; Amodeo, P.; Mendez, B. L.; Pelosi, P.;
Del Corso, A.; Mura, U. Biochem. J. 2004, 378, 35–44.
19. Pulido-Cejudo, G.; Conway, B.; Proulx, P.; Brown, R.;
Izaguirre, C. A. Antiviral Res. 1997, 36, 167–177.
20. Beninga, J.; Rock, K. L.; Goldberg, A. L. J. Biol. Chem.
1998, 273, 18734–18742.
21. Taylor, A.; Daims, M.; Lee, J.; Surgenor, T. Curr. Eye
Res. 1982, 2, 47–56.
22. Umezawa, H. Recent Results Cancer Res. 1980, 75, 115–
125.
23. Gupta, S. K.; Aziz, M.; Khan, A. A. Indian J. Pathol.
Microbiol. 1989, 32, 301–305.
24. Grembecka, J.; Kafarski, P. Mini-Rev. Med. Chem. 2001,
1, 133–144.
´
47. Selkti, M.; Tomas, A.; Gaucher, J.-F.; Prange, T.;
Fournie-Zaluski, M.-C.; Chen, H.; Roques, B.-P. Acta
´
Cryst. D 2003, 59, 1200–1205.
48. Kafarski, L.; Lejczak, B. In Aminophosphonic and Amino-
phosphinic Acids. Chemistry and Biological Activity;
Kukhar, V. P., Hudson, H. R., Eds.; John Wiley:
Chichester, 2000; pp 173–203, 407–442.
´ ´ˇ
49. Collinsova, M.; Jiracek, J. Curr. Med. Chem. 2000, 7, 629–
647.
50. Dive, V.; Lucet-Levannier, K.; Georgiadis, D.; Cotton, J.;
Vassiliou, S.; Cuniasse, P.; Yiotakis, A. Biochem. Soc.
Trans. 2000, 28, 455–460.
51. Dive, V.; Georgiadis, D.; Matziari, M.; Makaritis, A.;
Beau, F.; Cuniasse, P.; Yiotakis, A. Cell. Mol. Life Sci.
2004, 61, 2010–2019.
52. Cuniasse, P.; Devel, L.; Makaritis, A.; Beau, F.; Georgi-
adis, D.; Matziari, M.; Yiotakis, A.; Dive, V. Bichemie
2005, 87, 393–402.
25. Giannousis, P. P.; Bartlett, P. A. J. Med. Chem. 1987, 30,
1603–1609.
26. Lejczak, B.; Kafarski, P.; Zygmunt, J. Biochemistry 1989,
28, 3549–3555.
27. Grembecka, J.; Mucha, A.; Cierpicki, T.; Kafarski, P.
J. Med. Chem. 2003, 46, 2641–2655.
53. Discover, Molecular Modelling Program Package; Molec-
ular Simulation 1997.
28. Dra˛g, M.; Grembecka, J.; Pawelczak, M.; Kafarski, P.
Eur. J. Med. Chem. 2005, 40, 764–771.