α-Aminoalkylphosphonates as a tool in experimental optimisation of P1 side chain shape of potential inhibitors in S1 pocket of leucine- and neutral aminopeptidases

Article abstract of DOI:10.1016/j.ejmech.2005.02.011

The synthesis and biological activity studies of the series of structurally different α-aminoalkylphosphonates were performed in order to optimise the shape of the side chain of the potential inhibitors in S1 pocket of leucine aminopeptidase [E.C.3.4.11.1]. Analysis of a series of compounds with aromatic, aliphatic and alicyclic P1 side chains enabled to find out the structural features, optimal for that fragment of inhibitors of LAP. The most active among all investigated compounds were the phosphonic analogues of homo-tyrosine (K_i = 120:nM) and homo-phenylalanine (K_i = 140:nM), which even as racemic mixtures were better inhibitors in comparison with the best till now-phosphonic analogue of l-leucine (230 nM). Additional comparison of the inhibitory activity obtained for aminopeptidase N (APN, E.C.3.4.11.2) give insight into structural preferences of both enzymes.

Full text of DOI:10.1016/j.ejmech.2005.02.011

European Journal of Medicinal Chemistry 40 (2005) 764–771
www.elsevier.com/locate/ejmech
Original article
a-Aminoalkylphosphonates as a tool in experimental optimisation
of P1 side chain shape of potential inhibitors in S1 pocket of leucine-
and neutral aminopeptidases
Marcin DrTg a,*, Jolanta Grembecka a, Małgorzata Pawełczak b, Paweł Kafarski a,*
a Institute of Organic Chemistry, Biochemistry and Biotechnology, University of Technology, Wybrzeze Wyspianskiego 27, 50-370 Wroclaw, Poland
b Faculty of Chemistry, University of Opole, Opole, Oleska 48, Poland
Received 26 October 2004; received in revised form 14 February 2005; accepted 24 February 2005
Available online 20 April 2005
Abstract
The synthesis and biological activity studies of the series of structurally different a-aminoalkylphosphonates were performed in order to
optimise the shape of the side chain of the potential inhibitors in S1 pocket of leucine aminopeptidase [E.C.3.4.11.1]. Analysis of a series of
compounds with aromatic, aliphatic and alicyclic P1 side chains enabled to find out the structural features, optimal for that fragment of
inhibitors of LAP. The most active among all investigated compounds were the phosphonic analogues of homo-tyrosine (Ki = 120 nM) and
homo-phenylalanine (Ki = 140 nM), which even as racemic mixtures were better inhibitors in comparison with the best till now-phosphonic
analogue of L-leucine (230 nM). Additional comparison of the inhibitory activity obtained for aminopeptidase N (APN, E.C.3.4.11.2) give
insight into structural preferences of both enzymes.
© 2005 Elsevier SAS. All rights reserved.
Keywords: Aminophosphonates; Inhibitor; Leucine aminopeptidase; Aminopeptidase N
1. Introduction
ine and neutral aminopeptidases being the most promising
targets) [5].
Aminopeptidases are enzymes that catalyse the removal
of amino acids from the N-terminus of a peptide or protein.
These enzymes have been found in bacteria, yeast, plants,
animal and human tissues, what is considerably ascribed to
their broad substrate specificity [1]. They play such a biologi-
cal functions as protein maturation, activation and degrada-
tion of bioactive peptides, hormone level regulation and con-
trol of cell proliferation [2,3]. It is well known, that elevated
concentration of these enzymes was observed in such patho-
logical disorders as cancer and cataracts [1,3]. In the case of
the carcinogenesis, the methionine aminopeptidases has been
shown to be a promising target for development of anti-
angiogenesis agents [4].Additionally, the importance of these
enzymes arises from the observation, that some of them might
play the role in the apoptosis of cancer cells, what makes
them interesting targets in oncological research (with methion-
Aminopeptidases can be subdivided into two major groups,
with the first one being responsible for the hydrolysis of amino
acids with hydrophobic side chains from N-terminus of the
peptide or protein, and the second, which specifically remove
other amino acid residues [3].
Within the first group, leucine aminopeptidase (LAP,
E.C.3.4.11.1) is one of the most detailed characterised enzyme
with respect to its sequence, structure and mechanism of
action. Human LAP plays an important role in the early stages
of HIV infection, where the elevated concentration of this
enzyme was observed [6]. The unusual activity of this enzyme
has also been observed in such pathological disorders as can-
cer, inflammation of liver or eye cataract [7–9]. Because of
this leucine aminopeptidase is often called the stress enzyme.
Substantial insight into the structure of the LAP derives
from X-ray structures of the enzyme from bovine lens deter-
mined in its native form and complexed with transition state
inhibitors such as: bestatin, amastatin, phosphonic analogue
of L-leucine and L-leucinal coordinated to the active site.
* Corresponding author. Tel.: +48 71 320 4027; fax: +48 71 328 4064.
E-mail address: marcin.drag@pwr.wroc.pl (M. DrTg).
0223-5234/$ - see front matter © 2005 Elsevier SAS. All rights reserved.
doi:10.1016/j.ejmech.2005.02.011

M. DrTg et al. / European Journal of Medicinal Chemistry 40 (2005) 764–771
765
Bovine lens leucine aminopeptidase is a hexamer with a mass
of 324 kDa, where every from six subunits is identical and
contains three zinc ions. Two of them are located in active
site of the enzyme and play a catalytic role, and third one is
situated apart from the active site and most likely plays a struc-
tural function. LAP hydrolyses all of the substrates with
N-terminal L-amino acid and glycine, but kinetic preference
of the enzyme is towards L-leucine and L-phenylalanine, and
to a less extend, to other hydrophobic substrates [10–15].
The inhibitory activity of the inhibitors of LAP is rou-
tinely measured using commercially available enzyme from
porcine kidney. Unfortunately, X-ray structure of this enzyme
has not been determined so far. In the search for the potential
inhibitors of LAP this fact seems to have no significant mean-
ing since both enzymes (from bovine lens and from porcine
kidney) reveal high first-order sequence homology and simi-
lar kinetic properties, what suggests close structural features
of their active centres [1].
Several sets of various inhibitors of leucine aminopepti-
dase have been reported so far. On the basis of their structure,
they can be divided into two major groups. First one consti-
tute analogues of short peptides, among which analogue of
dipeptide-bestatine, and tetrapeptide-amastatine, are the best
known ones [16,17]. The second group represent analogues
of amino acids, with L-leucinal (Ki = 60 nM), the boronic acid
analogue of L-leucine (Ki = 130 nM), phosphonic analogue
of L-leucine (Ki = 230 nM) and hydroxamic analogue of
D-leucine (Ki = 1.3 µM) being the most effective (Fig. 1)
[18–22].
All of these inhibitors are transition state analogues of LAP,
able to bind to two zinc ions present in the active site of the
enzyme. According to the substrate specificity, the
L-enantiomers at the N-terminus of polypeptide are stronger
cooridinated to the active site of the enzyme than their
D-enantiomers. Additionaly, the common feature of all pre-
sented inhibitors is the hydrophobic character of the side
chain, where leucine isobutyl moiety is complexed by LAP.
The present results are the continuation of our previous
studies on the inhibition of leucine aminopeptidase, where
the crystal structure of L-leucine phosphonic acid coordi-
nated in active centre of bovine lens leucine aminopeptidase
was chosen as a lead compound in the search for the optimal
structure of the inhibitors [23]. In the current study we have
mainly focused on the experimental optimisation of the
P1 side chain shape of the inhibitor located at S1 site of the
enzyme, paying only marginal attention to the stereochemis-
try, because this feature has been the subject of earlier detailed
investigations. Since similar comopunds have been reported
as inhibitors of structurally closely related aminopeptidase N
(APN, EC 3.4.11.2), we have performed additional studies to
compare the structural requirements of S1 pocket of LAP ver-
sus the same pocket of APN [21].
2. Results
1-Aminoalkanephosphonic acids have been obtained as
racemic mixtures according to the well established proce-
dures. In the case of racemic compounds, the Oleksyszyn or
modified Arbuzov methodologies have been applied. Enan-
tiopure 1-aminoalkylphosphonates have been synthesised
using Hamilton–Walker reaction (Table 1) [24–26]. In the case
of 1-amino-3-phenylbutanephosphonic acid (15), the enzy-
matic assay was performed using the mixture of two pair of
enantiomers obtained in 1:0.7 ratio (stereochemistry not deter-
mined) (Fig. 2).
Seven aromatic, three aliphatic and eight alicyclic ami-
noalkylphosphonates were used for optimisation of the
P1 fragment of potential LAP inhibitors. These structures were
selected according to the expected change of their inhibitory
activity against LAP due to the carbon chain length, steric
bulk, presence or absence of methylene bridge between ami-
nophosphonate and hydrophobic portions of the molecule, as
well as taking into consideration overall shape of the phos-
phonate. The introduction of the polar, hydroxy groups at the
termini of long, hydrophobic chains was chosen to check the
hypothesis that additional hydrogen bonds with Asp365 and
Ala451 residues located at the bottom of S1 cavity might
improve the binding properties of the potential inhibitor [23].
All of the 1-aminoalkylphosphonates tested in vitro using
porcine kidney leucine aminopeptidase appeared to be slow-
binding, competitive inhibitors of the enzyme with Ki values
in the range of 0.12 and 798 µM. The mechanism of their
binding is a slow one-step process, as indicated from the lin-
ear dependence of the apparent first-order rate constant for
slow-binding versus inhibitor concentration The lowest activ-
ity was found for 1(RS)-amino-1-methyl-1-iso-propyl-
methanephosphonic acid (2) (Ki = 798 µM), an analogue od
a-methylvaline. Both enantiomers of 1-amino-2-ethyl-
buthanephosphonic acid (3/4) (Ki = 243 and 119 µM for R
and
S enantiomer respectively) and 1(RS)-amino-3-
propylpenthanephosphonic acid (5) (Ki = 33 µM) also rank
among weak LAP inhibitors. Better results were obtained for
alicyclic analogues since 1(RS)-amino-1-cyclopropyl-
methanephosphonic acid (6) (Ki = 84.5 µM), 1(RS)-amino-
1-cyclobuthylmethanephosphonic acid (7) (Ki = 22.7 µM),
1(RS)-amino-4-cyclohexyl-buthanephosphonic acid (20)
(Ki = 9.57 µM), both enantiomers of 1-amino-1-cyclohexyl-
methanephosphonic acid (8/9) (Ki = 7.89 and 6.96 µM for R
Fig. 1. The structures of the inhibitors of leucine aminopeptidase: bestatine,
amastatine, L-leucinal, L-leucine boronic acid, L-L euP and D-leucine
hydroxamic acid.

766
M. DrTg et al. / European Journal of Medicinal Chemistry 40 (2005) 764–771
Table 1
Entry Structure
Stereomer Ki
Ki
The structures of 1-aminophosphonic acids and their experimental inhibi-
tion constants (Ki) towards porcine kidney leucine aminopeptidase and micro-
somal aminopeptidase N (NI—no inhibition up to final concentration of inhi-
bitor 2 mM)
(µM)-LAP (µM)-APN
16
1RS
6.35
18.5
Entry Structure
Stereomer Ki
Ki
(µM)-LAP (µM)-APN
1
1R
0.2321
798
5321
17
18
19
1RS
1RS
1RS
0.21
0.14
0.12
37.1
15.9
23.7
2
1RS
NI
3/4
1R/1S
1RS
243/119
33.0
138/NI
NI
5
20
21
1RS
1RS
9.57
0.33
41.6
6
1RS
84.5
22.7
178
3.69
7
1RS
77.4
and
S enantiomer respectively) and 1(RS)-amino-2-
8/9
1R/1S
7.89/6.96 161/NI
cyclopentylethanephosphonic acid (10) (Ki = 6.92 µM) are
moderate inhibitors of LAP, whereas 1(RS)-amino-2-
cyclohexylethanephosphonic acid (11) (Ki = 0.75 µM) and
1(RS)-amino-3-cyclohexylpropanephosphonic acid (17)
(Ki = 0.21 µM) rank amongst the most active compounds.Aro-
matic analogues appeared to be much more effective inhibi-
tors of the enzyme. Thus, 1(RS)-amino-2-(4-hydroxy-
phenyl)ethanephosphonic (12) acid and 1(RS)-amino-5-
phenylpentanephosphonic acid (21) have the same value of
Ki equal 0.33 µM, close to that obtained for the mixture of
diastereomers of 1(RS)-amino-3(RS)-phenylbutane-
phosphonic acid (15) (Ki = 0.26 µM). The most active amongst
compounds studied in this work were 1(RS)-amino-3-
phenylpropanephosphonic acid (18) (Ki = 0.14 µM) and
1(RS)-amino-3-(4-hydroxyphenyl)propanephosphonic acid
(19) (Ki = 0.12 µM). It is worth to note, that compounds 17,
18 and 19, even as racemic mixtures are better inhibitors of
pkLAP than the stereochemically favoured phosphonic acid
analogues of L-leucine (1) (Ki = 0.23 µM) and L-phenylalanine
(Ki = 0.42 µM).
10
11
12
1RS
1RS
1RS
6.92
0.75
0.33
47.7
54.7
170
13
14
15
1RS
NI
168
NI
1RS
1.37
0.26
All of the investigated aminoalkylphosphonates appeared
to be only moderate inhibitors of aminopeptidase N with Ki
values in micromolar range and competitive type of inhibi-
tion. The slow-binding was observed only in the case of com-
pounds 7, 13 and 20 and was also found as the slow one-step
process. The most active of them was 1(RS)-amino-5-
phenylpentanephosphonic acid (21) with Ki value equal
1RS/3RS
36.9

M. DrTg et al. / European Journal of Medicinal Chemistry 40 (2005) 764–771
767
Fig. 2. Procedures and conditions of synthesis of a-aminoalkylphosphonates.
3.69 µM.Almost half of the tested by us aminoalkylphospho-
nates (7, 10, 11, 15, 16, 17, 18, 19 and 20) had Ki value in the
narrow range between 15.9 µM and 77.4 µM. The rest of the
compounds had the inhibition constants higher than 100 µM
(3, 6, 8, 9, 12 and 13) or did not reveal any inhibitory activity
toward APN (2, 4, 5 and 14).
lay in the observation that this pocket of blLAP has a large
volume and thus is able to bind far more bulky side chains
than those found in natural substrates (Leu, Phe, Ala). Addi-
tionally, computer-aided modelling of this pocket resulted in
discovery of potent low-molecular inhibitors of this enzyme
differing in placement of their side chains in S1 pocket in
both enzymes [30,31].
Among all the molecules tested in this work, the most
active inhibitor of pkLAP appeared to be 1(RS)-amino-3-(4-
hydroxyphenyl)-propanephosphonic acid (19) (Ki = 0.12 µM),
homologue of phosphonic acid analogue of tyrosine. We have
obtained this compound in order to test the hypothesis
advanced after computer-aided design, which suggested
improvement of the binding abilities upon formation of addi-
tional hydrogen bond between hydroxy group on aromatic
ring of 19 and Asp365 located deeply in the S1 pocket of
LAP [23]. However, 1(RS)-amino-3-phenylpropane-
phosphonic acid (18) (Ki = 0.14 µM), a homologue of phos-
phonic acid of phenylalanine appeared to be an equipotent
inhibitor of pkLAP. This clearly showed that formation of the
expected hydrogen bond between Asp365 and phenolic moi-
ety of 19 does not form. More advanced computer modelling
using GRID programme have shown that Asp365 is involved
in formation of complex hydrogen bond network with water
molecules present at the enzyme-solvent interface and the phe-
nolic group of 19 disturbs this network [32].
The potent inhibitory activity of compounds 18 and 19
towards pkLAP, the prolongation of the side chain with one
additional methylene group resulted in substantial improve-
ment of inhibitory activity (racemic mixture) if comparing to
L-PheP (Ki = 0.42 µM). Similar homologue, namely 1(RS)-
amino-3(RS)-phenylbutanephosphonic acid (15) reveals a
little bit lower inhibitory activity towards pkLAP, however it
is worth to note that it was applied as the mixture of four
diastereomers (Ki = 0.26 µM). The prolongation of the ali-
phatic side chain by additional methylene group caused slight
drop in the inhibitory activity, which is well illustrated by
affinity of 1(RS)-amino-5-phenylpentanephosphonic acid (21)
(Ki = 0.33 µM) towards the enzyme. The same pattern of struc-
ture–activity relationship was observed in the case of APN
3. Discussion and conclusions
The structure of leucine aminopeptidase is specific accord-
ing to the defined species, but not to tissues. LAP isolated
from pig kidney have been routinely used for biological inves-
tigations. Unfortunately till now there are no reports describ-
ing X-ray structure of this enzyme. Therefore, crystal struc-
ture of LAP isolated from bovine lens in its free form as well
as complexed with some inhibitors (bestatine, L-LeuP, amasta-
tine and L-leucinal) is routinely used for designing of new
inhibitors [23]. The fact, that the biological studies are per-
formed using pig LAP, and potent inhibitors are theoretically
evaluated using bovine LAP seems to have not influence on
overall effect, because of the results from immunological test,
which had proven that first-order structure of bovine and pig
LAP reveal 91.5% of homogenity [1]. Both aminopeptidases
(blLAP and pkLAP), like the aminopeptidases from Aeromo-
nas Proteolytica (AAP) and Streptomyces griseus (SAP), pre-
fer hydrophobic residues at NH2-terminal position [27–29].
The S1 pocket of blLAP, adjacent to the active site of the
enzyme, has hydrophobic character and plays crucial role in
binding side chains of substrates and potent inhibitors. It is
made up of Met270, Ala451, Thr359, Gly362 and Met454
[23]. Far less is known about three-dimensional shape of ami-
nopeptidase N binding sites since its crystal structure is not
available. Our previous studies indicated, however, that there
is substantial difference in requirements of S1 pocket of the
studied enzymes [21].
The main goal of the investigations presented in this paper
was to obtain more detailed experimental insight into the
structural preferences of S1 pocket of both enzymes pointing
out similarities and differences. The basis for our research

768
M. DrTg et al. / European Journal of Medicinal Chemistry 40 (2005) 764–771
although the studied compounds exhibited substantially lower
affinities towards this enzyme.
analogue of valine with reported inhibition constants towards
LAP of 0.15 and 12 µM for R and S enantiomers respectively
[21]. Compound 5 containing more bulky substuituents at
a-carbon atom appeared to be more potent towards LAP than
3/4 and inactive towards APN showing that structure–
activity relationship is far more complicated and does not
allow simple reasoning. Most probably, the presence of addi-
tional bulky hydrophobic group located at a-carbon atom dis-
tores coordination of amino and phosphonic moieties to the
Zn ions in the active site of LAP by non-optimal binding of
hydrophobic part of the molecule in S1 pocket, and thus show-
ing that the propoper design of this part of molecule is vital
for inhibitory potency. The lack of intuitive reasoning in
inhibitor design, which is based on simple comparison of
chemical structures is also well illustrated by the differences
in activity of structurally related compounds 13 (being active
towards APN and inactive towards LAP) and 14 (active
towards LAP and inactive towards APN) towards both
enzymes.
The presented results clearly show, that the maximal activ-
ity of the potential LAP andAPN inhibitors increase with the
length of the side chain located in S1 pocket, reaching opti-
mum for three atoms separating aromatic fragment of the mol-
ecule and its aminoalkylphosphonate portion. Compounds
having b and c non-substitued methylene groups are prefer-
ably bound (Fig. 3).
The observable decrease on the Ki drastically dimnishes
when this key feature is abused by the incorporation of the
additional substituents at b and c positions as well as elonga-
tion or shortening of the side chain with slight, although
important differences observed in the case of both enzymes
(Fig. 4).
The replacement of aromatic group by various alicyclic
ones enables to observe the influence of their shape and size
on the Ki of pkLAP. In opposition to the rigid aromatic ring,
the alicyclic one can undergo a series of conformational
transformations, which might influence the activity of the
tested inhibitors. The most active towards pkLAP appeared
1(RS)-amino-3-cyclohexylpropanephosphonic acid (17)
(Ki = 0.21 µM), which is structural analogue of 18, in which
phenyl group was substituted by cyclohexyl moiety. Small
difference in the inhibition constant between them (18
favoured of about 70 nM) suggests, that cyclohexyl group
has similar affinity to S1 pocket of the enzyme as planar aro-
matic ring. This assumption seems to be well supported by
the Ki found for 1(RS)-amino-2-cyclohexylethanephosphonic
acid (11) (Ki = 0.75 µM), which is nearly equipotent with
phosphonic analogue of L-PheP (Ki = 0.42 µM and 15 µM
for R and S enantiomers, respectively). The replacement of
cyclohexyl group by cyclopentyl one resulted in significant
decrease of the affinity, as illustrated by low activity of com-
pound 16 (Ki = 6.35 µM). Surprisingly, this minute change in
the chemical structure is drastic enough, that it does not allow
to dock this fragment of the molecule effectively into the
S1 pocket of LAP. Interestingly, compound 16 ranks amongst
the best inhibitors of APN being nearly equipotent (Ki
= 18.5 µM) with the second most active and structurally close
compound 18. These findings also clearly show that there are
minute, although important perferences of S1 pockets of both
enzymes and that the pocket of APN seem to be more spa-
cious than this of LAP.
1-Amino-1-cyclohexylmethanephosphonic acid (8/9)
(Ki = 7.89 and 6.96 µM for R and S enantiomers respec-
tively) is quite weak inhibitor of both aminopeptidases stud-
ied. However, there are substantial differences in its potency
if considering stereochemistry. Equipotency of both enanti-
omers towards LAP indicates the lack of stereochemical pref-
erentions of the enzyme towards analogues lacking methyl-
ene arm separating aminoalkylphosphonate portion of the
molecule and its hydrophobic fragment. Substantially differ-
ent activity was seen in the case of APN, where strong stere-
ochemical preference towards R isomer (corresponding to
L-configuration) was observed.
The presented analysis, when using 1-aminoalkyl-
phosphonic acids as molecular probes gave substantial insight
into the preferences of LAP S1 binding site and provides some
useful information regarding further inhibitor design. Addi-
tionally, acceptable agreement of some of the theoretical data
obtained by the analysis of bovine lens LAP with the experi-
mental data presented here for porcine kidney LAP well con-
firms the similarity of both enzymes [33]. Thus, the proper
design of the part of inhibitor bound in the S1 pocket is vital
for the activity of inhibitors towards both enzymes, however,
it is more important in the case of LAP than APN.
Aminoalkylphosphonates containing alkyl side chains,
designed to mimic opened forms of cycloalkyl rings and thus
being more flexible, ranked amongst the weakest inhibitors
of pkLAP. 1-Amino-2-ethylbuthanephosphonic acid (3/4)
(Ki = 243 and 119 µM for R and S enantiomer respectively)
is far less active towards pkLAP than compounds 8/9 and 7
and similarly as in the case of R and S enantiomers of 8/9
they revealed only minute stereochemical preference. This
compound was, however, nearly equipotent with 8/9 towards
APN and also showed significant stereochemical preference,
thus following the same structure–activity pattern. These find-
ings are a little bit astounding if taking into condsideration
strong structural relation of compound 3/4 to phosphonic acid
4. Experimental protocols
4.1. General methods
Unless otherwise noted, all reagents were obtained from
commercial suppliers and used without further purification.
NMR spectra were recorded on a Bruker Avance DRX
300 MHz instrument, operating at 300.13 MHz (1H),
121.499 MHz (31P) and 75.46 MHz (13C). Measurements
were made in D2O (99.8 at. % D) solutions at temperature
300 K, all solvents were supplied by Dr. Glaser AG (Basel,

M. DrTg et al. / European Journal of Medicinal Chemistry 40 (2005) 764–771
769
trophotometrically at 280 nm, assuming A2800.1% = 0.83 cm–1
and molecular weight of the enzyme equal 255.000 Da [21].
Porcine kidney microsomal aminopeptidase (APN) was
also obtained from Sigma Chemical Co. and treated identical
as described earlier [21]. The inhibitory activity of the enzyme
was measured at 25 °C in 50 mM potassium phosphate buffer
(pH = 7.2) using L-leucine-4-nitroanilide as substrate
(Km = 0.52 mM). The final volume (2 ml) of the assay mix-
ture contained 0.1 ml of the substrate solution in DMSO, 5 µl
of theAPN solution in final concentration of 4 µg ml–1, 0.5 ml
of the inhibitor solution and 1395 ul of phosphate buffer. The
concentration of theAPN was determined spectrophotometri-
cally at 280 nm, assuming A2800.1% = 1.63 cm–1 and molecu-
lar weight of the enzyme equal 280.000 Da [21]. The pos-
sible error in the Ki value in all inhibitors do not exceed 10%.
General method of synthesis of 1-aminoalkylphosphonates
5, 6, 7, 10, 11, 12, 16, 17, 19, 20 and 21.
0.1 mol of the appropriate carboxylic acid and 50 ml of
toluene was placed in a flask (250 ml) equipped with mag-
netic stirrer. After dissolving of the acid, 0.3 mol of the thio-
nyl chloride was added to the reaction mixture. The mixture
was refluxed the for 6–7 h with stirring. The volatile com-
pound were removed by rotary evaporation. To the remained
reaction mixture was added 50 ml of hexane and heated again
until boiling. The repetition of the rotary evaporation gave
the desired, crude carbonyl chloride (quantitative and quali-
tative yield), which was immediately used to the next step
without future purification. 0.1 mol of the carbonyl chloride
was dissolved in 50 ml of toluene and placed in the flask
(250 ml). The mixture was cooled to 0 °C and 0.2 mol of
triethyl phosphonate was added very slowly with stirring (dur-
ing this step the temperature can not exceed 7–8 °C). The
mixture was left for a night, and after the volatile products
were removed by rotary evaporation. To the remained keto-
phosphonate, 0.1 mol of hydroxylamine hydrochloride dis-
solved in 100 ml methanol containing 0.1 mol of piridine was
added. The mixture was left for a night, and after the volatile
products were removed by rotary evaporation. To the remain-
ing mixture, 100 ml of ice cold water solution of HCl (20 ml
of 8 M HCl in 80 ml of distilled water) was added. The extrac-
tion with chloroform (3 × 80 ml), washing of the combined
organic layers with ice cold water (3 × 100 ml) and brine
(100 ml), drying with MgSO4, filtration and evaporation of
the volatile products gave the desired a-oxymephosphonate.
The mixture was dissolved in formic acid (100 ml dried ear-
lier for 3 days in CuSO4 and distilled). To the intensively
stirred mixture, 0.4 mol of Zn was added in small portions.
The mixture was left for the night, the zinc formate filtered
off, and from the filtrate the volatile products removed by
rotary evaporation. To the remained oil, 100 ml of the 10 M
HCl was added and refluxed for 8 h. After the removing of
the volatile products, the oil was dissolved in 100 ml of etha-
nol and pyridine or propylene oxide was added till the pH
reached 6. The precipitated 1-aminophosphonic acid was fil-
tered off and recrystallised twice from water and ethanol. The
deprotection of the hydroxy group in the aromatic ring (in
Fig. 3
. Structure–activity relationships of the experimental inhibitory acti-
vity of 1-aminoalkylphosphonates toward pkLAP.
Fig. 4. The key elements of the structure of the optimised inhibitor of pkLAP.
Switzerland). Chemical shifts are reported in parts per mil-
lion relative to TMS, and coupling constants are reported in
Hertz. Spectrophotometric measurements were performed
using thermostated Specord M40 (Carl Zeiss, Jena, Ger-
many).
Porcine kidney leucine aminopeptidase was purchased
from Sigma Chemical Co. and activated according to the pro-
cedure described by Andersson et al. [18]. Enzymatic reac-
tions were routinely carried out in 2.85 ml of 7.5 mM trietha-
nol amine-HCl buffer, pH 8.4, containing 15 µg ml–1 of
porcine kidney leucine lminopeptidase, 0.1 ml substrate-L-
leucine-4-nitroanilide and 0.1 ml of water solution of inhibi-
tor (various concentrations). Reactions were monitored at
400 nm with substrate concentration varying from 0.2 to
0.8 mM. The concentration of the LAP was determined spec-

770
M. DrTg et al. / European Journal of Medicinal Chemistry 40 (2005) 764–771
the substrate carboxylic acid, the hydroxy group should be
protected in the form of methoxy) was performed by reflux-
ing in 40% HBr for 10 h and repeating whole procedure as in
the case of HCl hydrolysis.
1(RS)-amino-3-propylpenthanephosphonic acid (5): (yield:
46%), M.p.: 262–264 °C; mmax (KBr)/cm–1 2933, 1622,
1168 and 1023; dP (D2O + DCl) 17.29; dH (D2O + DCl) 0.67
(6 H, t, 3JHH = 6.6 Hz, 2 × CH3CH2CH2), 0.99–1.43 (8 H, m,
2 × CH3CH2CH2), 1.75 (1 H, m, CHCHP), 3.30 (1 H, dd,
3JHH = 7.20 Hz, 2JHP = 16.0 Hz, CHP); dC (D2O + DCl) 13.1
(s, CH3CH2CH2), 19.4 (d, 4JCP = 7.4 Hz, CH3CH2CH2), 31.7
(d, 3JCP = 9.4 Hz CH3CH2CH2), 36.27 (s, CHCHP), 50.5 (d,
1JCP = 148.5 Hz, CHP).
1(RS)-amino-1-cyclopropylmethanephosphonic acid (6):
(yield: 72%), M.p.: 247–248 °C; mmax (KBr)/cm–1 2916, 1650,
1176 and 1029; dP (D2O + DCl) 15.16; dH (D2O + DCl) 0.11
(2 H, m, CH2, cyclopropyl.), 0.37 (2H, m, CH2, cyclopro-
pyl.), 0.7 (1 H, m, CH, cyclopropyl.), 2.29 (1 H, dd,
3JHH = 11.1 Hz, 2JHP = 14.1 Hz, CHP); dC (D2O + DCl) 3.6
(d, 3JCP = 12.9 Hz, 2 × CH2, cyclopropyl.), 8.68 (s, cyclopro-
pyl.), 53.4 (d, 1JCP = 151.4 Hz, CHP).
2934, 1631, 1516, 1492, 1457, 1170 and 1025; dP
(D2O + DCl) 17.09; dH (D2O + DCl) 1.85–2.08 (2H, m,
CH2CH2CHP), 2.61 (2H, m, CH2CHP), 3.33 (1H, dt,
2
3JHH = 7.26 Hz, JHP = 16.9 Hz, CHP), 6.70 (2H, d,
3
3JHH = 8.3 Hz, arom.AB,), 7.04 (2H, d, JHH = 8.3 Hz, aro-
m.AB,); dC (D2O + DCl) 28.99 (s, CH2CH2CHP), 29.41 (s,
2JCP = 8.0 Hz, CH2CHP), 46.60 (d, 1JCP = 150.2 Hz, CHP),
114.65 (s, arom.), 128.87 (s, arom.), 131.33 (s, arom.), 152.8
(s, arom.).
1(RS)-amino-4-cyclohexylbuthanephosphonic acid 20:
(yield: 36%), M.p.: 271–273 °C; mmax (KBr)/cm–1 2923, 1024,
1651 and 1174; dP (D2O + DCl) 16.88; dH (D2O + DCl) 0.67–
1.72 (17H, m, cyclohexyl. + CH2CH2CH2CHP), 3.34 (1H,
dt, 3JHH = 7.44 Hz, 2JHP = 13.9 Hz, CHP); dC (D2O + DCl)
22.98 (s, CH2CH2CHP), 23.59 (s, cyclohexyl.), 23.97 (s,
cyclohexyl.), 30.07 (s, CH2CH2CH2CHP), 30.25 (s,
CH2CH2CHP), 30.37 (s, cyclohexyl.), 34.35 (s, cyclo-
hexyl.), 46.39 (d, 1JCP = 152.63 Hz, CHP).
1(RS)-amino-5-phenylpentanephosphonic acid (21):
(yield: 26%), M.p.: 274–276 °C; mmax (KBr)/cm–1 2931, 1643,
1532, 1496, 1453, 1159 and 1028; dP (D2O + DCl) 16.72; dH
(D2O + DCl) 1.02–1.51 (6H, m, CH2CH2CH2CHP), 2.18
(2H, t, 3JHH = 7.40 Hz, CH2CH2CH2CH2CHP), 2.97 (1H, dt,
1(RS)-amino-1-cyclobuthylmethanephosphonic acid (7):
(yield: 66%), M.p.: 251–253 °C; mmax (KBr)/cm–1 2972, 1651,
1171 and 1024; dP (D2O + DCl) 15.31; dH (D2O + DCl) 0.9–
1.36 (6H, m, 3 × CH2, cyclobuthyl.), 1.95 (1H, m, CH,
cyclobuthyl.), 2.71 (1H, dd, 3JHH = 10.8 Hz, 2JHP = 12.9 Hz,
CHP); dC (D2O + DCl) 17.02 (s, cyclobuthyl.), 25.10 (s,
2
3JHH = 7.20 Hz, JHP = 13.6 Hz, CHP), 6.74–6.90 (5H, m,
3
arom.); dC (D2O + DCl) 24.50 (d, JCP = 9.1 Hz,
CH2CH2CHP), 27.33 (s, CH2CH2CH2CHP), 29.8 (s,
CH2CHP), 34.16 (s, CH2CH2CH2CH2CHP), 47.7 (d,
1JCP = 150.2 Hz, CHP), 125.6 (arom.), 128.3 (arom.), 128.5
(arom.), 142.5 (arom.).
Method of synthesis of 1-aminoalkylphosphonates 1, 3, 4,
8 and 9. These compounds have been synthesised according
to the method of synthesis described earlier by Hamilton et
al. [26].
3
cyclobuthyl), 25.27 (d, JCP = 12.8 Hz, CH, cyclobuthyl.),
32.64 (d, 2JCP = 2.9 Hz, CH2CHP), 53.4 (d, 1JCP = 147.3 Hz,
CHP).
1(RS)-amino-2-cyclopentylethanephosphonic acid (10):
(yield: 29%), spectral data identical as published earlier [34].
1(RS)-amino-2-cyclohexylethanephosphonic acid (11):
(yield: 42%), spectral data identical as published earlier [34].
1(RS)-amino-2-(4-hydroxyphenyl)ethanephosphonic acid
(12): (yield: 41%), spectral data identical as published earlier
[35].
1(R)-amino-3-methylbutanephosphonic acid (1): (yield:
29%), spectral and optical activity data identical as published
earlier [26].
1-Amino-2-ethylbuthanephosphonic acid (3/4): (yield:
36% and 33% for compounds 3 and 4 respectively); M.p.:
247–248 °C; mmax (KBr)/cm–1 2928, 1608, 1164 and 1024; dP
(D2O + DCl) 17.61; dH (D2O + DCl) 0.48 (3H, d,
1(RS)-amino-3-cyclopentylpropanephosphonic acid (16):
(yield: 48%), M.p.: 266–268 °C; mmax (KBr)/cm–1 2950, 1648,
1175 and 1038; dP (D2O + DCl) 17.26; dH (D2O + DCl) 0.90–
1.79 (13H, m, cyclopentyl. + CH2CH2CHP), 3.34 (1H, dt,
3JHH = 7.7 Hz, 2JHP = 14.9 Hz, CHP); dC (D2O + DCl) 23.78
(s, cyclopentyl.), 25.97 (s, cyclopentyl.), 30.51 (d,
3
3JHH = 7.6 Hz, CH2CH3), 0.51 (3H, d, JHH = 7.6 Hz,
CH2CH3), 0.85–1.02 (2H, m, CH2CH3), 1.17 (2H, m,
CH2CH3), 1.29 (1H, m, CHCH2CH3), 1.53 (1H, dd,
3JHH = 4.4 Hz, 2JHP = 16.0 Hz, CHP); dC (D2O + DCl) 10.28
(s, 2 × CH3CH2CH), 21.31 (d, 3JCP = 32 Hz, CH3CH2CH),
40.10 (s, CH3CH2CH), 49.67 (d, 1JCP = 147.4 Hz, CHP).
1-Amino-1-cyclohexylmethanephosphonic acid (8/9):
(yield: 42% and 54% for compounds 8 and 9 respectively);
spectral data and optical activity identical as published ear-
lier [26].
Method of synthesis of 1-aminoalkylphosphonates 2, 13,
14, 15 and 18. These compounds have been synthesised
according to the method of synthesis described earlier by Ole-
ksyszyn et al. [24].
3
2JCP = 8.7 Hz, CH2CHP), 30.92 (d, JCP = 12.9 Hz,
CH2CH2CHP), 37.96 (s, cyclopentyl.), 47.26 (d,
1JCP = 151.09 Hz, CHP).
1(RS)-amino-3-cyclohexylpropanephosphonic acid (17):
(yield: 52%), M.p.: 271–273 °C; mmax (KBr)/cm–1 2922, 1651,
1173 and 1024; dP (D2O + DCl) 17.01; dH (D2O + DCl) 0.63–
1.71 (15H, m, cyclohexyl. + CH2CH2CHP), 3.32 (1H, dt,
3JHH = 7.2 Hz, 2JHP = 13.6 Hz, CHP); dC (D2O + DCl) 22.30
(s, cyclohexyl.), 22.93 (s, cyclohexyl.), 29.42 (s, CH2CHP),
3
29.61 (d, JCP = 14.5 Hz, CH2CH2CHP), 33.72 (s, cyclo-
1
hexyl.), 35.06 (s, cyclohexyl.), 45.78 (d, JCP = 152.8 Hz,
CHP).
1(RS)-amino-1-methyl-1-iso-propylmethanephosphonic
acid 2: (yield: 52%), spectral data identical as published ear-
lier [36].
1(RS)-amino-3-(4-hydroxyphenyl)propanephosphonic
acid (19): (yield: 36%), M.p.: 261–266 °C; mmax (KBr)/cm–1

M. DrTg et al. / European Journal of Medicinal Chemistry 40 (2005) 764–771
771
[7] H. Umezawa, Recent Results Cancer Res. 75 (1980) 115–125.
[8] A. Taylor, M.A. Daims, J. Lee, T. Surgenor, Curr. Eye Res. 2 (1982)
47–56.
[9] J. Beninga, K.L. Rock, A.L. Goldberg, J. Biol. Chem. 273 (1998)
18734–18742.
[10] S.K. Burley, P.R. David, W.N. Lipscomb, Proc. Natl. Acad. Sci. USA
88 (1991) 6916–6920.
[11] N. Sträter, W.N. Lipscomb, Biochemistry 34 (1995) 14792–14800.
[12] N. Sträter, W.N. Lipscomb, Biochemistry 34 (1995) 9200–9210.
[13] S.K. Burley, P.R. David, R.M. Sweet, A. Taylor, W.N. Lipscomb, J.
Mol. Biol. 224 (1992) 113–140.
[14] H. Kim, S.K. Burley, W.N. Lipscomb, J. Mol. Biol. 230 (1993)
722–724.
[15] H. Kim, W.N. Lipscomb, Biochemistry 32 (1993) 8465–8478.
[16] H. Umezawa, T. Aoyagi, H. Suda, M. Hamada, T. Takeuchi, J. Anti-
biot. 29 (1976) 97–99.
[17] T.Aoyagi, H. Tobe, F. Kojima, M. Hamada, T. Takeuchi, H. Umezawa,
J. Antibiot. 31 (1976) 636–638.
[18] L. Andersson, T.C. Isley, R. Wolfenden, Biochemistry 21 (1982)
4177–4180.
[19] A.B. Shenvi, Biochemistry 25 (1986) 1286–1291.
[20] P.P. Giannoussis, P.A. Bartlett, J. Med. Chem. 30 (1987) 1603–1609.
[21] B. Lejczak, P. Kafarski, J. Zygmunt, Biochemistry 28 (1989) 3549–
3555.
[22] S.H. Wilkes, J.M. Prescott, J. Biol. Chem. 258 (1983) 13517–13521.
[23] J. Grembecka, W.A. Sokalski, P. Kafarski, J. Comp. Aid. Design. 14
(2000) 531–544.
[24] J. Oleksyszyn, R. Tyka, P. Mastelerz, Synthesis (Mass.) (1978) 479–
480.
[25] J. Kowalik, P. Kupczyk-Sobotkowska, P. Mastalerz, Synthesis (Mass.)
(1981) 57–58.
[26] R. Hamilton, B. Walker, B.J. Walker, Tetrahedron Lett. 36 (1995)
4451–4454.
[27] R. Holz, Coord. Chem. Rev. 232 (2002) 5–26.
[28] A. Spungin, S. Blumberg, Eur. J. Biochem. 183 (1989) 471–477.
[29] D. Ben-Meir, A. Spungin, R. Ashkenazi, S. Blumberg, Eur. J. Bio-
chem. 212 (1993) 107–112.
1(RS)-amino-1-(4-(iso-propyl)phenyl)methanephos-
phonic acid (13): (yield: 42%), spectral data identical as pub-
lished earlier [37].
1(RS)-amino-1-(4-(N,N-dimethylamino)phenyl)methane-
phosphonic acid (14): (yield: 16%), M.p.: 237–238 °C; mmax
(KBr)/cm–1 3372, 1172, 2961, 1651, 1552, 1511, 1457 and
1055; dP (D2O + DCl) 18.66; dH (D2O + DCl) 4.17 (6H, s, 2
2
× CH3), 5.93 (1H, d, JHP = 22.8 Hz, CHP), 7.26 (s, 4H,
arom.); dC (D2O + DCl) 46.24 (s, 2 × CH3), 59.2 (d,
1JCP = 150.18 Hz, CHP), 116.6 (s, arom.), 127.92 (s, arom.),
134.53 (d, 3JCP = 14.2 Hz, arom.), 144.8 (s, arom.).
1(RS)-amino-3-(RS)-phenylbutanephosphonic acid (15):
(yield: 52%), M.p.: 250–256 °C; mmax (KBr)/cm–1 2941, 1039,
1633, 1586, 1534, 1495, 1457 and 1174; dP (D2O + DCl)
17.28 and 17.43 (ratio 1:0.7); dH (D2O + DCl) 1.03 (3H, d,
3JHH = 7.24 Hz, CH3, diast. 1), 1.14 (3H, d, 3JHH = 7.24 Hz,
CH3, diast. 0.7), 1.65–2.08 (1H, m, CHCH2CHP, 2 diast.),
2.52 (2H, m, CH2CH2CHP, 2 diast.), 3.34 (1H, m, CHP,
2 diast.), 7.11–7.29 (5H, m, aromat., 2 diast.); dC (D2O + DCl)
22.31 (s, CH3, diast. 1), 22.35 (s, CH3, diast. 0.7), 32.11 (d,
3JCP = 4.2 Hz, CH2CH2CHP, diast. 1), 32.89 (d,
3JCP = 4.75 Hz, CH2CH2CHP, diast. 0.7), 40.05 (s, CH3, diast.
1), 41.17 (s, CH3, diast. 0.7), 51.38 (d, 1JCP = 148.5 Hz, CHP,
diast. 1), 51.38 (d, 1JCP = 148.5 Hz, CHP, diast. 0.7), 126.12
(arom., diast. 0.7), 127.10 (arom., diast. 1), 127.84 (arom.,
diast. 1), 128.25 (arom., diast. 0.7), 129.18 (arom., diast. 0.7),
130.42 (arom., diast. 1), 146.32 (arom., diast. 1), 147.02
(arom., diast. 0.7).
1(RS)-amino-3-phenylpropanephosphonic acid (18):
(yield: 65%), spectral data identical as published earlier [38].
[30] M. Drag, J. Grembecka, P. Kafarski, Phosphorus, Sulfur and Silicon
177 (2002) 1591–1595.
References
[31] M. Drag, M. Pawelczak, P. Kafarski, Chirality 15 (2003) S104–S107.
[32] P.J. Goodford, J. Med. Chem. 28 (1985) 849–857.
[33] J. Grembecka, A. Mucha, T. Cierpicki, P. Kafarski, J. Med. Chem. 46
(2003) 2641–2655.
[34] L. Maier, D. Diel, Phosphorus, Sulfur and Silicon 62 (1991) 15–27.
[35] L. Maier, Phosphorus, Sulfur and Silicon 53 (1990) 43–67.
[36] M. Soroka, Liebigs Ann. Chem. 4 (1990) 331–334.
[37] A. Kabatschnik, I. Medved, IASKA6, Izv. Akad. Nauk. SSSR Ser.
Khi. 868 (1953) 769–774.
[1] A. Taylor, Trends Biochem. Sci. 18 (1993) 167–172.
[2] A. Taylor, FASEB J. 7 (1993) 290–298.
[3] A. Taylor, in: Molecular Biology Intelligence Unit, Co, R.G. Landes
Co., Austin, TX, 1996, pp. 1–219.
[4] W.T. Lowther, D.A. McMillen, A.M. Orville, B.W. Matthews, Proc.
Natl. Acad. Sci. USA 95 (1998) 12153–12157.
[5] K. Sekine, H. Fuji, F. Abe, Leukemia 13 (1999) 720–734.
[6] G. Pulido-Cejudo, B. Conway, P. Proulx, R. Brown, C.A. Izaguirre,
Antiviral Res. 36 (1997) 167–177.
[38] A. Ryglowski, P. Kafarski, Tetrahedron 52 (1996) 10685–10692.