Hydrolysis of the Phosphonamidate Bond in Phosphono Dipeptide Analogues - The Influence of the Nature of the N-Terminal Functional Group

Article abstract of DOI:10.1002/ejoc.200300469

Phosphonamidate pseudodipeptides, designed as transition state analogue inhibitors of leucine aminopeptidase, revealed unexpected instability in aqueous solutions of pH values varying from acidic up to highly basic. This reaction has been studied in some

Full text of DOI:10.1002/ejoc.200300469

FULL PAPER
Hydrolysis of the Phosphonamidate Bond in Phosphono Dipeptide
Analogues — the Influence of the Nature of the N-Terminal Functional Group
Artur Mucha,*[a] Jolanta Grembecka,[a] Tomasz Cierpicki,[b] and Paweł Kafarski[a]
Keywords: Phosphonamidates / NMR spectroscopy / Peptides / Bioorganic chemistry
Phosphonamidate pseudodipeptides, designed as transition
of the free N-terminal amino moiety is required for the occur-
state analogue inhibitors of leucine aminopeptidase, re- rence of rapid P−N bond decomposition; N-protected derivat-
vealed unexpected instability in aqueous solutions of pH
values varying from acidic up to highly basic. This reaction
ives are significantly more stable. The mechanisms of the
cleavage of the phosphonamidate bond have been proposed
has been studied in some detail by means of NMR spectro- and discussed. The observed instability of unblocked P−N
scopy and it was found that the phosphonamidate stability
depended strongly on the solution pH, relying on the pro-
tonation state of two crucial functional groups of the molec-
ule — the phosphonamidate and amino moieties. Protonation
peptides may be the cause of the substantial limitations of
their practical application as aminopeptidase inhibitors.
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2003)
Introduction
Phosphonamidates (1, X ϭ NH), phosphinates (1, X ϭ
CH2) and phosphonates (1, X ϭ O) form three important
families of phosphorus analogues of peptides (Figure 1)
that resemble the tetrahedral transition state of the amide
bond hydrolysis.[1Ϫ5] These moieties are able to coordinate
the zinc atom present in the active sites of metalloproteases
and to block its function in the process of hydrolysis. Thus,
their chemistry continues to attract considerable
interest[6Ϫ10] and provides a wide variety of potent inhibi-
tors of proteases.[11Ϫ15]
Phosphonamidates seem to be the closest analogues
among all three types of isosters in terms of their resem-
blance to the transition state resulting in the inhibitory ef-
fect. This has been adequately illustrated in the literature
for an endopeptidase thermolysin, a prototypical member
of the zinc-containing metalloprotease family, and its Cbz-
Glyψ[P(O)(OH)X]Leu-Leu inhibitor (Figure 1, compound
2).[16Ϫ19] Although the mode of interaction of the phos-
phonate (X ϭ O) with the enzyme, revealed by the crystal
structure, is virtually identical to the corresponding phos-
phonamidate (X ϭ NH),[2] the latter binds approximately
1000-fold more tightly to thermolysin (Ki ϭ 9.1 nm).[16]
This observation was explained by the difference in binding
free energy due to the specific hydrogen bonding. Surpris-
ingly, the phosphinic analogue (X ϭ CH2), for which the
crystal structure is not available, which presumably interacts
in a similar manner, exhibits only a moderately lower ac-
tivity than the PϪN one.[17,18] It has become obvious that
complex effects both in the active site and in solution, in-
cluding differences in the basicity of the oxygen ligands,
electrostatic repulsions and desolvation processes, can
dramatically influence the enzyme-inhibitor adduct
formation.[17Ϫ19] Taking all these factors into account,
phosphonamidates seem to be the best candidates as they
possess only a single disadvantage associated with their
limited stability.
Figure 1. Peptides analogues with an amide bond replaced by a
phosphorus-containing moiety (1): Cbz-Glyψ[P(O)(OH)X]Leu-Leu
inhibitors of thermolysin — phosphorus analogues of N-benzyl-
oxycarbonyl-glycyl-leucyl-leucine {2, ψ[P(O)(OH)X] depicts
pseudopeptide bond}
[a]
Institute of Organic Chemistry, Biochemistry and
Biotechnology, Wrocław University of Technology,
´
Wybrzez˙e Wyspianskiego 27, 50Ϫ370 Wrocław, Poland
Fax: (internat.) ϩ 48-71/328-4064
E-mail: artur.mucha@pwr.wroc.pl
Institute of Biochemistry and Molecular Biology, University
of Wrocław
[b]
Recently we have focused our interests on phosphon-
amidate pseudodipeptides, developing new, potent inhibi-
Tamka 2, 50-137 Wrocław, Poland
Eur. J. Org. Chem. 2003, 4797Ϫ4803
DOI: 10.1002/ejoc.200300469
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4797

A. Mucha, J. Grembecka, T. Cierpicki, P. Kafarski
FULL PAPER
tors of leucine aminopeptidase.[20] Leucine aminopeptidase Results and Discussion
(LAP, E.C.3.4.11.1) is one of the first discovered and the
most widely studied aminopeptidase with respect to se-
quence, structure and mechanism of action.[21Ϫ25] Similar
to other enzymes of this family it is of significant biological
and medical importance due to its key role in protein modi-
fication. What’s more, the altered activity of human LAP
Chemistry
The synthesis of phosphonamidate pseudopeptides is
complex due to the presence of diverse functionalities in
one molecule, and thus the necessity for their selective pro-
tection and deprotection, side reactions accompanying
PϪN bond formation via phosphonochloridates[6Ϫ10] and
the instability of the forming phosphonamidate
bond.[11,17,33,35,36,39Ϫ42] The strategy applied in this work is
briefly illustrated in Scheme 1. The starting α-amino- and
α-hydroxyphosphonate monoesters (3 and 4) were con-
verted into the PϪN peptides 5Ϫ7 by a standard approach
based on aminolysis of the appropriate phosphonochlorid-
ates with the amino ester. After purification of the blocked
pseudopeptides, obtained as diastereoisomeric mixtures, the
protecting groups were removed under mild conditions. The
procedure ended with hydrolysis of the carboxylate ester
groups in basic solution in order to conserve the phosphon-
amidate moiety intact. The final compounds 8Ϫ10 were ob-
tained as mixtures of enantiomers or diastereoisomers and
stored as their dilithium salts. It was shown that α-amino
analogues 8 and 9 are stable in solution of pH above 12 (no
decomposition observed within three days at room tem-
perature). The α-hydroxy analogue 10 was synthesised in
order to study the influence of the basicity of the N-ter-
minal group on the stability of the phosphonamidate moi-
ety. The details of the synthetic procedure, together with
inhibitor design and results of kinetic assays towards LAP,
have been published in a separate paper.[20]
has
been
associated
with
several
pathological
disorders.[26Ϫ32] In our recent work, the application of a
molecular modelling methodology followed by synthesis of
a set of rationally designed phosphinic pseudodipeptides re-
sulted in the development of the most potent organo-
phosphorus inhibitors of LAP reported so far.[20] However,
the most promising among them, namely the phosphon-
amidate analogues containing free N-terminal amino group,
were found to be extremely unstable in aqueous solutions,
due to PϪN bond cleavage at pH values below 12. This
unexpected property excludes their use in enzymatic studies,
although theoretically they were predicted to be even more
potent than the corresponding phosphinates.
The results concerning the stability of the phosphon-
amidate moiety obtained by various authors are quite con-
tradictory. Although such peptidomimetics have been quite
frequently used as peptide isosters, exhibiting a strong in-
hibitory activity against metalloproteases,[11Ϫ13,16Ϫ19,33Ϫ38]
the PϪN bond is considered to be unstable in numerous
reports.[11,17,33,35,36,39Ϫ42] Bartlett, for example, in his orig-
inal paper, reported the half-life of N-terminally protected
phosphonamidate dipeptide Cbz-Glyψ[P(O)(OH)NH]Phe
in aqueous solution at pH 6.2 to be 4 h, whereas it leng-
thens dramatically even under slightly basic conditions.[33]
An excellent illustration of pH-dependent hydrolysis was
presented by Christianson and Lipscomb in a crystallo-
graphic study of carboxypeptidase A complexed with the
same compound.[43,44] All these data concern N-protected
analogues for which neutral pH seems to be the limit for
staying intact. However, PϪN lability is correlated to both
pH and the chemical structure of the molecule. Thus, the
presence of a free terminal amino group in phosphon-
amidates was found to enhance dramatically the suscepti-
bility of such analogues to hydrolysis.[45] This fact, along
with our observations,[20] makes the infrequently reported Scheme 1. Reagents and conditions: (a) SOCl2; (b) ()-
HCl·H2NCH(R2)COOMe, NEt3; (c) H2, Pd/C; (d) LiOH/MeOH
examples of their application as enzyme inhibitors clearly
unreliable.[46,47] Such incompatible and scarce literature
data concerning PϪN pseudodipeptides with an unprotec-
ted α-amino group, encouraged us to investigate this prob-
Stability Studies of Phosphonamidate Dipeptide Analogues
lem. To the best of our knowledge, no comprehensive stud-
ies on the pH-dependent behaviour of such derivatives has
In order to evaluate the influence of all functional groups
so far been performed and presented. The second reason present in the molecule on the PϪN bond stability, and to
was to establish key functional and structural features of suggest the mechanism of this process, several phosphona-
their lability in terms of designing new LAP inhibitors un- midate derivatives were studied. Thus, the fully blocked di-
affected by physiological pH, but still bearing groups cru- peptides, those with certain functional groups selectively de-
cial for efficient binding to the aminopeptidase. Here we protected as well as fully deprotected phosphonamidates
present the results of this study performed on structurally were investigated. Furthermore, the effects of the replace-
diverse phosphonamidates, with special emphasis on those ment of the α-amino group by hydroxyl, as well as the sta-
containing a free N-terminal amino group, and propose bility of compounds lacking the C-terminal carboxyl moi-
mechanisms for the cleavage.
ety were additionally examined. Studies on the dependence
4798
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2003, 4797Ϫ4803

Hydrolysis of the Phosphonamidate Bond in Phosphono Dipeptide Analogues
FULL PAPER
of phosphonamidate hydrolysis rate on pH were performed LeuP concentrations were monitored by performing the
using 31P NMR spectroscopy.
range of 31P NMR measurements at certain time intervals
(Figure 3).
Variously Substituted Derivatives
The totally substituted phosphonamidate peptides 5Ϫ7
appear to be stable over almost the whole range of pH.
Only under strongly acidic conditions (pH Ͻ 1, elevated
temperature or elongated reaction time) is the PϪN bond
partially susceptible to acidolysis, but approximately to the
same extent as the carbamate bond of the benzyloxycar-
bonyl group. This observation is in good accordance with
literature data.[48] Selective removal of the Cbz group from
compound 5 did not significantly affect its stability. Only
side reactions, including carboxylate ester hydrolysis and
cyclisation of the molecule in slightly basic conditions, pro-
vided compounds 11 and 12, while the PϪN bond remained
intact (Figure 2). The cyclisation reaction seems to be lim-
ited to a narrow pH range (pH ϭ 7Ϫ9) and is not observed
in the strongly basic conditions necessary for LiOH-as-
sisted deprotection.
Figure 3. 31P NMR assisted monitoring of phosphonamidate bond
hydrolysis of Leuψ[P(O)(OH)NH]Gly (8) at starting pH 11.0 (the
process is accompanied by pH lowering, which also causes a change
of the chemical shift values)
The integration of the signals allowed calculation of the
ratio of the phosphonamidate to LeuP reported as a percent
of the unhydrolysed phosphonamidate. These values were
used for evaluation of the kinetic parameters of these reac-
tions, which proceeded according to typical ‘‘pseudo’’ first-
order kinetics. The rates of Leuψ[P(O)(OH)NH]Gly hy-
drolysis as a function of pH (between 10 to 12) are shown
in Figure 4.
Figure 2. Side products observed upon N-deprotection of phos-
phonamidate 5 followed by treatment with water/methanol basic
solution (pH 7Ϫ9)
The calculated rate constants for hydrolysis of phosphon-
Selective hydrolysis of the methyl esters in compounds 5 amidates 8 and 9 are compared in Figure 5. Their values
and 6 led to N-benzyloxycarbonyl derivatives — representa- apparently decrease with a rise of pH. There is also a visible
tives of the most extensively studied phosphononamidate steric effect in this reaction since leucyl-leucine analogue 9
peptide analogues[11Ϫ13,16Ϫ19,33Ϫ36] — that are generally decomposes about two times slower than the leucyl-glycine
considered as stable in conditions of enzymatic assays. This analogue 8. Compound 9 was synthesised as mixture of two
stability was fully confirmed in our studies when using Cbz- diastereoisomers (, and ,). Well separated 31P NMR
Leuψ[P(O)(OH)NH]Gly as a representative example. This signals allowed us to follow the hydrolysis of each isomer
compound showed hydrolytic resistance within three days individually, however, no effect of influence of their con-
at pH 8.5 at room temperature. A decrease in pH to 7.0 figuration on the rate of cleavage was observed. Thus, the
resulted in slow PϪN bond cleavage and release of Cbz- summary effect of decomposition of 9 is presented in Fig-
LeuP (the phosphonic analogue of N-benzyloxycarbonyl ure 5 and 6.
leucine) and glycine. This shows that the protonation of the
The obtained results indicate explicitly that such a fast
molecule, which occurs in neutral aqueous solutions, gives decomposition of deprotected phosphonamidates excludes
the opportunity to form a far better leaving group and is the possibility of determining their binding affinities
responsible for the observed cleavage.
towards leucine aminopeptidase. For LAP activity the opti-
mal pH is 8.5, at which more than 90% of 8 and 9 hydrolyze
within 10 minutes. The presence of a free α-amino group,
however, is crucial for efficient inhibition of this protease.
LAP exhibits aminoexopeptidase activity and there is no
space in the active site to bind a bulky N-protecting group.
The presence of the unsubstituted phosphonic group coordi-
nating zinc atoms is also necessary for strong interaction
with the LAP active site. This restricts the application of
such phosphonamidate dipeptide analogues, although they
were designed as the most promising inhibitors of LAP.
Fully Deprotected Phosphonamidates
Cbz-Leuψ[P(O)(OH)NH]Gly appears to be significantly
more stable than the corresponding fully deprotected ana-
logue Leuψ[P(O)(OH)NH]Gly (8). Although the lability of
phosphonamidate compounds containing a free terminal
amino group has been mentioned in the literature,[45] no
details concerning the conditions and the reasons for their
instability were given. Our NMR studies showed that both
Leuψ[P(O)(OH)NH]Gly (8) and Leuψ[P(O)(OH)NH]Leu
(9) are stable only in strongly basic media, and a decrease
in the pH to below 12 results in PϪN bond hydrolysis,
yielding LeuP and the appropriate amino acid, namely Gly
Influence of the C-Terminal Carboxylate Group
The simultaneous presence of two deprotected moiet-
or Leu. The decrease in phosphonamidate and increase in ies — the phosphonamidate and the N-terminal amino
Eur. J. Org. Chem. 2003, 4797Ϫ4803
www.eurjoc.org  2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 4799

A. Mucha, J. Grembecka, T. Cierpicki, P. Kafarski
FULL PAPER
Figure 6. Influence of the C-terminal carboxylate moiety on
‘‘pseudo’’ first-order rate constants for the hydrolysis of phosphon-
amidate analogues as a function of pH: Leuψ[P(O)(OH)NH]Gly
(8; ∆), Leuψ[P(O)(OH)NH]Gly (9; ᭺), Phgψ[P(O)(OH)-
NH]Bzl (ᮀ)
ined, however it appeared not to be significant. Thus, a
phosphonamidate
lacking
this
moiety,
namely
Phgψ[P(O)(OH)NH]Bzl (obtained by standard deprotec-
tion of the N-protected compound available from previous
study[48]), revealed no significant stability differences in
comparison to Leuψ[P(O)(OH)NH]Gly (8) and Leuψ-
[P(O)(OH)NH]Leu (9) in terms of both reaction kinetics
and calculated values of ‘‘pseudo’’ first-order rate con-
stants (Figure 6).
Replacement of the N-Terminal Amino Moiety by a
Hydroxy One
Figure 4. Time-dependent hydrolysis of the phosphonamidate bond
of Leuψ[P(O)(OH)NH]Gly (8) as exponential decrease of the sub-
strate concentration (top), and linear in log scale (bottom) at vari-
ous starting pH values (᭿: pH 11.7; ᭜: pH 11.1; •: pH 10.9; ᭡: pH
10.7; —: pH 10.4)
In order to investigate the presumed importance of the
basicity of the free α-amino group on the PϪN bond hy-
drolysis, derivative 10, containing an α-hydroxyl group in-
stead of the α-amino one (depicted as LeuOH
-
ψ[P(O)(OH)NH]Gly), was synthesised (Scheme 1). Stability
studies and the calculations of the percent of non-hydro-
lysed phosphonamidate were performed in a similar manner
to that described above and indicated that this
compound is significantly more stable than Leuψ-
[P(O)(OH)NH]Gly (8), with less than 10% of this com-
pound decomposed after seven days at pH 8.5. The hydroly-
sis also proceeded according to ‘‘pseudo’’ first-order kine-
tics. However, the rate constant for hydrolysis of compound
10 at pH 8.5 (k ϭ 6.4 ϫ 10Ϫ3 1/min) is about 1000 times
lower than the corresponding value for 8 at pH 11.0. Thus,
the hydroxy analogue 10, similarly to the widely studied N-
benzyloxycarbonyl derivatives,[11Ϫ13,16Ϫ19,33Ϫ36] appears to
be stable enough to be able to evaluate its inhibitory activity
at physiological pH. These observations indicate that the
presence of a basic terminal amino group is essential for
such an unexpectedly rapid hydrolysis.
Figure 5. Comparison of rate constants calculated for hydrolysis of
Leuψ[P(O)(OH)NH]Gly (8; ∆) and Leuψ[P(O)(OH)NH]Leu (9; ᭺)
upon varying the pH
The Mechanism of Phosphonamidate Bond Hydrolysis
one — was proved to be responsible for the dramatically
The results obtained from the stability studies on phos-
rapid process of PϪN cleavage. The influence of the C-ter- phonamidate dipeptide analogues suggest that the simul-
minal carboxylate group on the hydrolysis was also exam- taneous presence of the deprotected α-amino and phos-
4800
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2003, 4797Ϫ4803

Hydrolysis of the Phosphonamidate Bond in Phosphono Dipeptide Analogues
FULL PAPER
phonamidate functions make these compounds stable only
An alternative mechanism might be proposed analogous
under strongly basic conditions. Furthermore, the removal to that described in the literature reaction for cleavage com-
of the C-terminal carboxylic moiety does not affect the pounds containing a trivalent phosphorus atom,[49] namely
PϪN bond hydrolysis rate. It is worthwhile to note that the the reaction proceeds via a four-membered transition state.
studied compounds are stable at pH values greater 12.0, It involves binding of the proton of the attacking nucleoph-
when all the functional groups of the molecule exist in the ile (here water) by the leaving group, which is associated
deprotonated state. The hydrolysis of the PϪN bond pro- with coordination of the electron pair of the water oxygen
ceeds very rapidly upon lowering the pH: the half-life of by the electrophilic phosphorus atom (Figure 7, c). Such
Leuψ[P(O)(OH)NH]Gly (120 minutes) at pH 11.1 decreases a cyclic transition state mechanism has been proposed for
to below 10 minutes at pH 10.4. This pH range corresponds substitution of PIII phosphoramidates. In the case of the
to the protonation of the N-terminal amino group. The compounds studied here, the free α-amino group can play
compounds lacking a free α-amino group are significantly an additional, important role in stabilizing the presumable
more stable, as shown for the hydroxy- and Cbz-blocked transition state. This mechanism seems to better explain the
derivatives LeuOHψ[P(O)(OH)NH]Gly and Cbz-Leu- observed instability of phosphonamidate dipeptides.
ψ[P(O)(OH)NH]Gly respectively, with a lack of decompo-
The third alternative — a unimolecular pathway of PϪN
sition products (Ͻ 5%) observed within three days at pH bond hydrolysis — would involve formation of the meta-
8.5. Based on these results some alternative pathways for phosphate (Figure 7, d). Conversion of the phosphonamid-
the PϪN bond hydrolysis process might be proposed.
ate moiety into an uncharged and planar metaphosphonate
The first one is the classical nucleophilic substitution enables the nucleophilic attack of the water molecule and
(SN2 type) at phosphorus in which the protonation of the phosphonic amino acid analogue formation. Although the
cleaved amino leaving group of the C-terminal amino acid dissociative mechanism is not often considered in the litera-
is indispensable. We suggest that the protonated N-terminal ture, the metaphosphate intermediate has been postulated
amino group compensates the negative charge of the phos- as being involved in the phosphorylation of alcohols by
phonamidate moiety by the electrostatic interactions, and phosphoramidic acid monoesters.[50] This mechanism seems
facilitates nucleophilic attack of an oxygen atom of the nu- to explain more favourably the observed phenomenon for
cleophile (Figure 7, a). A water molecule rather than a hy- the phosphonamidates with a blocked N-terminal amino
droxyl group is a better candidate for the nucleophile in this group. However, it looks less probable for free phosphona-
process because it is able to deliver a proton to the leaving midates because the formation of the metaphosphonate
amino group. The resulting five-coordinate transition state would require proton transfer to the leaving amino group
undergoes cleavage with the simultaneous proton transfer at pH 11.
to the leaving group either from the neighbouring N-ter-
minal α-amino group or from a water nucleophile.
Experimental Section
General: Unless otherwise stated, materials were obtained from
commercial suppliers (SigmaϪAldrich, Fluka, Merck) and used
without purification. Isovaleryl aldehyde and thionyl chloride were
freshly distilled. Triethylamine was distilled and stored over potass-
ium hydroxide. Ethanol- and water-free chloroform was obtained
by passing the solvent through basic alumina followed by its distil-
lation from P2O5. Anhydrous benzene was obtained by its distil-
lation from P2O5 and stored over sodium. Column chromatography
was performed on silica gel 60 (70Ϫ230 mesh).
Melting points were determined on a Boetius apparatus and were
not corrected. IR spectra were recorded in KBr pellets or in film
on a PerkinϪElmer System 2000 FT IR spectrometer. Proton and
Figure 7. Proposed pathways of PϪN bond hydrolysis (see dis-
phosphorus NMR spectra were recorded on a Bruker DRX spec-
cussion in the text)
1
trometer operating at 300.13 MHz for H and 121.50 MHz for 31P.
Measurements were made in CDCl3, [D6]DMSO or D2O solutions.
Proton chemical shifts are reported in relation to tetramethylsilane
used as internal standard. 31P NMR spectra were obtained with
For derivatives containing the Cbz dipeptides or depsi-
peptides containing the hydroxyl group (10), the hydrolysis
is observed at pH values below 7.0. Under these conditions
the protonation of the phosphonamidate NH group occurs,
which enhances the nucleophilic attack of the water mol-
ecule (Figure 7, b). Thus, PϪN bond hydrolysis of phos-
phonamidate dipeptide analogues takes place when the to-
tal charge of the functional groups attached to Cα, N-ter-
minal amino and the phosphonamidate, is equal to zero.
1
the use of broad-band H decoupling; chemical shifts are reported
in relation to 85% H3PO4 (sealed capillary). Microanalyses were
performed by Instrumental Analysis Unit of the Institute of Or-
ganic Chemistry, Biochemistry and Biotechnology, Wrocław Uni-
versity of Technology.
Methyl Hydrogen [1-(N-Benzyloxycarbonylamino)-3-methylbu-
tyl]phosphonate (3): White crystals; m.p. 118Ϫ120 °C. 1H NMR
(CDCl3): δ ϭ 0.85 (d, J ϭ 6.7 Hz, 6 H, 2 ϫ CH3), 1.47 (m, 2 H,
Eur. J. Org. Chem. 2003, 4797Ϫ4803
www.eurjoc.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4801

A. Mucha, J. Grembecka, T. Cierpicki, P. Kafarski
FULL PAPER
CH2), 1.64 (m, 1 H, CH), 3.63 (d, J ϭ 10.8 Hz, 3 H, OCH3), 4.11 4.87 (d each, J ϭ 11.1 Hz, 2 H together, CH2O), 7.33 (m, 5 H,
(m, 1 H, CHP), 5.05 (s, 2 H, CH2O), 7.26 (m, 5 H, C6H5), 10.33 C6H5) ppm. 31P NMR: δ ϭ 28.78 and 30.75 (4:3 ratio) ppm. IR
(s, 1 H, OH) ppm. 31P NMR: δ ϭ 29.56 ppm.
(film): ν˜ ϭ 3210 cmϪ1 (NH), 2955 (CH), 1755 (CϭO), 1260 (Pϭ
O), 1210, 1155 and 1045 (CϪO, PϪO). C17H28NO5P (357.4): calcd.
C 57.13, H 7.90, N 3.92, P 8.67; found C 57.31, H 8.09, N 3.75,
P 8.57.
Ethyl Hydrogen (1-Benzyloxy-3-methylbutyl)phosphonate (4): Yel-
lowish glass. 1H NMR ([D6]DMSO): δ ϭ 0.71 and 0.84 (d each,
J ϭ 6.4 Hz, 3 H and 3 H, 2 ϫ CH3), 1.13 (t, J ϭ 6.9 Hz, 3 H,
OCH2CH3), 1.44 and 1.58 (m each, 1 H and 1 H, CH2), 1.73 (m, Removal of the Protecting Groups
1 H, CH), 3.47 (m, 1 H, CHP), 3.96 (m, 2 H, POCH2), 4.47 and
Dilithium N-[(1-Amino-3-methylbutyl)oxyphosphinyl]glycinate (8).
4.86 (d each, J ϭ 11.3 Hz, 1 H and 1 H, CH2O), 7.30 (m, 5 H,
General Procedure: To remove the benzyloxycarbonyl group the
peptide 5 (0.19 g, 0.5 mmol) was dissolved in MeOH (5 mL) and
hydrogenated in the presence of 10% Pd/C catalyst (0.1 g). Hydro-
gen was passed through the solution by gentle bubbling. The reac-
tion, followed by TLC as well as by 31P NMR spectroscopy, pro-
ceeded quantitatively and was complete within 2 h. The catalyst
was filtered off and the solution was evaporated to dryness giving
a colourless oil. Removal of the methyl groups was achieved by
alkaline hydrolysis of methyl N-[(1-amino-3-methylbutyl)methoxy-
phosphinyl]glycinate in a 1.5  LiOH/MeOH solution (2 mL, 3
equiv. of LiOH were used). After evaporation to dryness at room
temperature the racemic sample was used directly for tests. Due
to its instability at pH lower than 12 the deprotected peptide was
C6H5) ppm. 31P NMR: δ ϭ 19.41 ppm.
Synthesis of the Phosphonamidates
Methyl N-{[1-(NЈ-Benzyloxycarbonylamino)-3-methylbutyl]-
methoxyphosphinyl}glycinate (5). General Procedure: The monome-
thyl ester 3 (0.63 g, 2 mmol) was dissolved in CHCl3 (10 mL) and
thionyl chloride was added (0.22 mL, 3 mmol). The solution was
stirred for 2 h at room temperature and refluxed for an additional
2 h. The volatile components were evaporated under reduced press-
ure, the residue was treated with benzene (10 mL) and the solvents
evaporated again. The resulting phosphonochloridate was dis-
solved in CHCl3 (10 mL) and added dropwise to a mixture of gly-
cine methyl ester hydrochloride (0.25 g, 2 mmol) and triethylamine
(0.70 mL, 5 mmol) in CHCl3 (15 mL) cooled in an ice bath. The
resulting solution was left at room temperature overnight and then
washed successively with 1  NaOH, water, 5% HCl, water, 1 
Na2CO3 and brine (20 mL of each). The organic layer was dried
over Na2SO4. After removing the solvent the crude peptide was
purified by column chromatography using an EtOAc/hexane (3:1)
solution as the eluent, giving a white crystalline product as a dia-
1
characterised only by NMR spectroscopy as its dilithium salt. H
NMR (D2O): δ ϭ 0.83 and 0.89 (d each, J ϭ 6.6 Hz, 3H and 3 H,
2 ϫ CH3), 1.37 (m, 2 H, CH2), 1.75 (m, 1 H, CH), 2.74 (m, 1 H,
CHP), 3.42 (d, J ϭ 8.0 Hz, 2 H, NCH2) ppm. 31P NMR: δ ϭ
29.39 ppm.
Dilithium N-[(1-Amino-3-methylbutyl)oxyphosphinyl]-(L)-leucinate
(9): Mixture of diastereoisomers in a 3:2 ratio. 1H NMR (D2O):
δ ϭ 0.86 (m, 12 H, 4 ϫ CH3), 1.29Ϫ1.75 (m, 6 H, 2 ϫ CH2CH),
stereomeric mixture in 3:2 ratio. Yield: 0.43 g (56%); m.p. 62Ϫ79
1
°C. H NMR (CDCl3): δ ϭ 0.85 (m, 6 H, 2 ϫ CH3), 1.51 (m, 2 H, 2.72 (m, 1 H, CHP), 3.72 (m, 1 H, NCH) ppm. 31P NMR: δ ϭ
CH2), 1.66 (m, 1 H, CH), 3.05 and 3.15 (m each, 1 H together,
PNH), 3.47Ϫ3.76 (m, 8 H, 2 ϫ OCH3 and NCH2), 4.01 (m, 1 H,
CHP) 4.95 and 5.24 (d each, J1 ϭ 10.0, J2 ϭ 10.1 Hz, 1 H together,
NH), 5.03 (AB system, J ϭ 12.2 Hz, 2 H, CH2O), 7.27 (m, 5 H,
C6H5) ppm. 31P NMR: δ ϭ 31.39 and 32.72 (2:3 ratio) ppm. IR
(KBr): ν˜ ϭ 3295 and 3225 cmϪ1 (NH), 3065 and 2950 (CH), 1760
and 1700 (CϭO), 1545 (δNH), 1270 (PϭO), 1210, 1150, 1060 and
1035 (CϪO, PϪO). C17H27N2O6P (386.4): calcd. C 52.84, H 7.04,
N 7.25, P 8.02; found C 52.75, H 7.39, N 7.17, P 7.74.
26.58 and 26.99 (2:3 ratio) ppm.
Dilithium
N-[(1-Hydroxy-3-methylbutyl)oxyphosphinyl]glycinate
(10): Racemic mixture. 1H NMR (D2O): δ ϭ 0.81 and 0.86 (d each,
J ϭ 6.6 Hz, 3 H and 3 H, 2 ϫ CH3), 1.36 and 1.49 (m each, 1 H
and 1 H, CH2), 1.70 (m, 1 H, CH), 3.41 (d, J ϭ 10.1 Hz, 2 H,
NCH2), 3.63 (m, 1 H, CHP) ppm. 31P NMR: δ ϭ 25.26 ppm.
Stability Studies: The fully deprotected phosphonamidates were
stored in dry form as their lithium salts in the presence of excess
LiOH. The stability studies were carried out in D2O or H2O/D2O
(10% D2O) solutions. Concentrations of the phosphonamidates
ranged between 7 and 40 mm. Titration with 0.5  HCl was used
to obtain the appropriate starting pH (range 6Ϫ12). All NMR
measurements were carried out at 300 K. Decomposition of the
phosphonamidates was followed by 31P NMR spectroscopy with
proton decoupling. Furthermore, two dimensional TOCSY, 1H-31P
HMQC and 1H-31P HMQC-TOCSY spectra were acquired to
identify the products of decomposition. The decrease in phosphon-
amidate concentration was monitored by recording a set of 31P
NMR spectra at set time intervals. The integration of the phos-
phorus spectra allowed us to obtain the percentage of non-hydro-
lysed phosphonamidate. ‘‘Pseudo’’ first-order rate constants were
calculated as the best fitting to experimental values.
Methyl N-{[1-(NЈ-Benzyloxycarbonylamino)-3-methylbutyl]-
methoxyphosphinyl}-(L)-leucinate (6): Colourless oil obtained as a
diastereomeric mixture in 4:4:3:2 ratio (purification by column
1
chromatography using EtOAc/hexane, 3:2); yield: 0.46 g (52%). H
NMR (CDCl3): δ ϭ 0.86 (m, 12 H, 4 ϫ CH3), 1.47 (m, 4 H, 2 ϫ
CH2), 1.64 (m, 2 H, 2 ϫ CH), 3.00 (m, 1 H, PNH), 3.61 (m, 6 H,
2 ϫ OCH3), 3.99 (m, 2 H, NCH and CHP), 5.04 (m, 3 H, NH and
CH2O), 7.26 (m, 5 H, C6H5) ppm. 31P NMR: δ ϭ 30.45, 30.96,
31.90 and 32.23 (4:4:3:2 ratio) ppm. IR (film): ν˜ ϭ 3295 and 3220
cmϪ1 (NH), 2955 (CH), 1745 and 1720 (CϭO), 1540 (δNH), 1265
(PϭO), 1215, 1150, 1110 and 1050 (CϪO, PϪO). C21H35N2O6P
(442.5): calcd. C 57.00, H 7.97, N 6.33, P 7.00; found C 56.98, H
7.90, N 6.19, P 6.72.
Methyl N-[(1-Benzyloxy-3-methylbutyl)ethoxyphosphinyl]glycinate
(7): Colourless oil obtained as a diastereomeric mixture in a 4:3
ratio (purification by column chromatography using EtOAc/
CH2Cl2, 2:1); yield: 0.32 g (45%). 1H NMR (CDCl3): δ ϭ 0.79,
0.81 and 0.93 (d each, J ϭ 6.6 Hz, 6 H together, 2 ϫ CH3), 1.33
and 1.37 (t each, J ϭ 7.0 Hz, 3 H together, OCH2CH3), 1.51 and
1.69 (m each, 2 H together, CH2), 1.84 (m, 1 H, CH), 3.08 and 3.20
(m each, 1 H together, PNH), 3.70 (s, 3 H, OCH3), 3.72Ϫ3.93 (m,
3 H, NCH2 and CHP), 4.13 (m, 2 H, POCH2), 4.56, 4.64, 4.71 and
Acknowledgments
This work has been supported by KBN grant 6 PO4B 02 18.
[1]
H. M. Holden, D. E. Tronrud, A. F. Monzingo, L. H. Weaver,
B. W. Matthews, Biochemistry 1987, 26, 8542Ϫ8553.
[2]
D. H. Tronrud, H. M. Holden, B. W. Matthews, Science 1987,
235, 571Ϫ574.
4802
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2003, 4797Ϫ4803

Hydrolysis of the Phosphonamidate Bond in Phosphono Dipeptide Analogues
FULL PAPER
[3]
F. Grams, V. Dive, A. Yiotakis, I. Yiallouros, S. Vassiliou, R.
Zwilling, W. Bode, W. Stöcker, Nat. Struct. Biol. 1996, 3,
671Ϫ675.
A.-L. Gall, M. Ruff, R. Kannan, P. Cuniasse, A. Yiotakis, V.
Dive, M.-C. Rio, P. Basset, D. Moras, J. Mol. Biol. 2001, 307,
577Ϫ586.
J. L. Radkiewicz, M. A. McAllister, E. Goldstein, K. N. Houk,
J. Org. Chem. 1998, 63, 1419Ϫ1428.
P. Kafarski, B. Lejczak, in Aminophosphonic and Aminophos-
phinic Acids. Chemistry and Biological Activity (Ed.: V. P. Ku-
khar, H. R. Hudson), John Wiley & Sons, Chichester, 2000,
pp. 173Ϫ203.
R. Hirschmann, K. M. Yager, C. M. Taylor, J. Witherington,
P. A. Sprengeler, B. W. Phillips, W. Moore, A. B. Smith III, J.
Am. Chem. Soc. 1997, 119, 8177Ϫ8190.
W. P. Malachowski, J. K. Coward, J. Org. Chem. 1994, 59,
7616Ϫ7624.
W. P. Malachowski, J. K. Coward, J. Org. Chem. 1994, 59,
7625Ϫ7634.
A. Mucha, P. Kafarski, F. Plenat, H.-J. Cristau, Tetrahedron
1994, 50, 12743Ϫ12754.
P. A. Bartlett, C. K. Marlowe, P. E. Giannousis, J. E. Hanson,
Cold Spring Harbor Symp. Quant. Biol. 1987, 52, 83Ϫ90.
P. A. Bartlett, C. K. Marlowe, Science 1987, 235, 569Ϫ571.
P. Kafarski, B. Lejczak, in Aminophosphonic and Aminophos-
phinic Acids. Chemistry and Biological Activity (Ed.: V. P. Ku-
khar, H. R. Hudson), John Wiley & Sons, Chichester, 2000,
pp. 407Ϫ442.
zymes (Eds.: A. J. Barrett, N. D. Rawlings, J. F. Woessner),
Academic Press, Amsterdam, 2002, pp. 1384Ϫ1389.
H. Umezawa, Rec. Res. Cancer Res. 1980, 75, 115Ϫ125.
S. K. Gupta, M. Aziz, A. A. Khan, Indian J. Pathol. Microbiol.
1989, 32, 301Ϫ305.
A. Taylor, M. Daims, J. Lee, T. Surgenor, Curr. Eye Res. 1982,
2, 47Ϫ56.
A. Taylor, M. J. Brown, M. A. Daims, J. Cohen, Invest.
Ophthalmol. Vis. Sci. 1983, 24, 1172Ϫ1180.
A. Taylor, T. Surgenor, D. K. Thomson, R. J. Graham, H.
Oettgen, Exp. Eye Res. 1984, 38, 217Ϫ229.
C. S. Scott, M. Davey, A. Hamilton, D. R. Norfolk, Blut 1986,
52, 297Ϫ303.
G. Pulido-Cejudo, B. Conway, P. Proulx, R. Brown, C. A. Izag-
uirre, Antiviral Res. 1997, 36, 167Ϫ177.
N. E. Jacobsen, P. A. Bartlett, J. Am. Chem. Soc. 1981, 103,
654Ϫ657.
K. Yamauchi, S. Ohtsuki, M. Kinoshita, Biochim. Biophys.
Acta 1985, 827, 275Ϫ282.
J. E. Hanson, A. P. Kaplan, P. A. Bartlett, Biochemistry 1989,
28, 6294Ϫ6305.
R. L. Elliott, N. Marks, M. J. Berg, P. S. Portoghese, J. Med.
Chem. 1985, 28, 1208Ϫ1216.
E. D. Thorsett, E. E. Harris, E. R. Peterson, W. J. Greenlee, A.
A. Patchett, E. H. Ulm, T. C. Vassil, Proc. Natl. Acad. Sci. U.
S. A. 1982, 79, 2176Ϫ2180.
H. Barelli, V. Dive, A. Yiotakis, J.-P. Vincent, F. Checler, Bio-
chem. J. 1992, 287, 621Ϫ625.
M. Hariharan, R. J. Motekaites, A. E. Martell, J. Org. Chem.
1975, 40, 470Ϫ473.
N. E. Jacobsen, P. A. Bartlett, J. Am. Chem. Soc. 1983, 105,
1613Ϫ1619.
[26]
[27]
[4]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[14]
[15]
´ˇ
M. Collinsova, J. Jiracek, Curr. Med. Chem. 2000, 7, 629Ϫ647.
V. Dive, K. Lucet-Levannier, D. Georgiadis, J. Cotton, S. Vassi-
liou, P. Cuniasse, A. Yiotakis, Biochem. Soc., Trans. 2000, 28,
455Ϫ460.
P. A. Bartlett, C. K. Marlowe, Biochemistry 1983, 22,
4618Ϫ4624.
K. A. Mookthiar, C. K. Marlowe, P. A. Bartlett, H. E. Van
Wart, Biochemistry 1987, 26, 1962Ϫ1965.
D. Grobelny, U. B. Goli, R. E. Galardy, Biochemistry 1989,
28, 4948Ϫ4951.
B. P. Morgan, J. M. Scholtz, M. D. Balinger, I. D. Zipkin, P.
A. Bartlett, J. Am. Chem. Soc. 1991, 113, 297Ϫ307.
J. Grembecka, A. Mucha, T. Cierpicki, P. Kafarski, J. Med.
Chem. 2003, 46, 2641Ϫ2655.
A. Taylor, Trends Biochem. Sci. 1993, 18, 167Ϫ171.
A. Taylor, FASEB J. 1993, 7, 290Ϫ298.
N. E. Jacobsen, P. A. Bartlett, J. Am. Chem. Soc. 1983, 105,
1619Ϫ1626.
A. Yiotakis, A. Lecoq, A. Nicolaou, J. Labadie, V. Dive, Bio-
chem. J. 1994, 303, 323Ϫ327.
D. W. Christianson, W. N. Lipscomb, J. Am. Chem. Soc. 1986,
108, 545Ϫ546.
D. W. Christianson, W. N. Lipscomb, J. Am. Chem. Soc. 1988,
110, 5560Ϫ5565.
[16]
[17]
[18]
[19]
[20]
K. Yamauchi, S. Ohtsuki, M. Kinoshita, J. Org. Chem. 1984,
49, 1158Ϫ1163.
W. Yuan, B. Munoz, C.-H. Wong, J. Z. Haeggstoem, A. Wetter-
holm, B. Samuelsson, J. Med. Chem. 1993, 36, 211Ϫ220.
K.-W. Yang, J. J. Brandt, L. L. Chatwood, M. W. Crowder,
Bioorg. Med. Chem. Lett. 2000, 10, 1085Ϫ1088.
A. Mucha, P. Kafarski, Tetrahedron 2002, 58, 5855Ϫ5863.
F. Mathis, Phosphorus, Sulfur and Silicon 1976, 1, 109Ϫ128.
M. J. P. Harger, C. Pater, Chem. Commun. 1999, 931Ϫ932.
Received July 28, 2003
[21]
[22]
[23]
H. Kim, W. N. Lipscomb, Adv. Enzymol. Relat. Areas Mol.
Biol. 1994, 68, 153Ϫ213.
N. Sträter, W. N. Lipscomb, Biochemistry 1995, 34,
14792Ϫ14800.
[48]
[49]
[50]
[24]
[25]
N. Sträter, W. N. Lipscomb, in Handbook of Proteolytic En-
Eur. J. Org. Chem. 2003, 4797Ϫ4803
www.eurjoc.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4803