Influence of the nucleobase and anchimeric assistance of the carboxyl acid groups in the hydrolysis of amino acid nucleoside phosphoramidates

Article abstract of DOI:10.1002/chem.201102279

Nucleoside phosphoramidates (NPs) are a class of nucleotide analogues that has been developed as potential antiviral/antitumor prodrugs. Recently, we have shown that some amino acid nucleoside phosphoramidates (aaNPs) can act as substrates for viral polymerases like HIV-1 RT. Herein, we report the synthesis and hydrolysis of a series of new aaNPs, containing either natural or modified nucleobases to define the basis for their differential reactivity. Aqueous stability, kinetics, and hydrolysis pathways were studied by NMR spectroscopy at different solution pD values (5-7) and temperatures. It was observed that the kinetics and mechanism (P-N and/or P-O bond cleavage) of the hydrolysis reaction largely depend on the nature of the nucleobase and amino acid moieties. Aspartyl NPs were found to be more reactive than Gly or β-Ala NPs. For aspartyl NPs, the order of reactivity of the nucleobase was 1-deazaadenine>7- deazaadenine>adenine>thymine≥3-deazaadenine. Notably, neutral aqueous solutions of Asp-1-deaza-dAMP degraded spontaneously even at 4°C through exclusive P-O bond hydrolysis (a 50-fold reactivity difference for Asp-1-deaza-dAMP vs. Asp-3-deaza-dAMP at pD 5 and 70°C). Conformational studies by NMR spectroscopy and molecular modeling suggest the involvement of the protonated N3 atom in adenine and 1- and 7-deazaadenine in the intramolecular catalysis of the hydrolysis reaction through the rare syn conformation. Touching (nucleo)base: A dual intramolecular catalytic influence is demonstrated by the nucleobase and carboxyl groups in the chemical hydrolysis of amino acid nucleoside phosphoramidate prodrugs (see scheme). The replacement of the adenine N1 or N7 atoms instead of the N3 atom is shown to have a conformational role in which the protonated N3 is crucial in regulating the kinetics and mechanism of nucleotide (P-N pathway) versus nucleoside (P-O pathway) formation. Copyright

Full text of DOI:10.1002/chem.201102279

FULL PAPER
DOI: 10.1002/chem.201102279
Influence of the Nucleobase and Anchimeric Assistance of the Carboxyl Acid
Groups in the Hydrolysis of Amino Acid Nucleoside Phosphoramidates
Munmun Maiti,[a] Servaas Michielssens,[b] Natalia Dyubankova,[a] Mohitosh Maiti,[a]
Eveline Lescrinier,[a] Arnout Ceulemans,[b] and Piet Herdewijn*[a]
Abstract: Nucleoside phosphorami-
dates (NPs) are a class of nucleotide
analogues that has been developed as
potential antiviral/antitumor prodrugs.
Recently, we have shown that some
amino acid nucleoside phosphorami-
dates (aaNPs) can act as substrates for
viral polymerases like HIV-1 RT.
Herein, we report the synthesis and hy-
drolysis of a series of new aaNPs, con-
taining either natural or modified nu-
cleobases to define the basis for their
differential reactivity. Aqueous stabili-
ty, kinetics, and hydrolysis pathways
were studied by NMR spectroscopy at
different solution pD values (5–7) and
temperatures. It was observed that the
nine>7-deazaadenine>adenine>thy-
mineꢀ3-deazaadenine. Notably, neu-
tral aqueous solutions of Asp-1-deaza-
dAMP degraded spontaneously even at
À
kinetics and mechanism (P N and/or
À
P O bond cleavage) of the hydrolysis
À
reaction largely depend on the nature
of the nucleobase and amino acid moi-
eties. Aspartyl NPs were found to be
more reactive than Gly or b-Ala NPs.
For aspartyl NPs, the order of reactivi-
ty of the nucleobase was 1-deazaade-
48C through exclusive P O bond hy-
drolysis (a 50-fold reactivity difference
for Asp-1-deaza-dAMP vs. Asp-3-
deaza-dAMP at pD 5 and 708C). Con-
formational studies by NMR spectros-
copy and molecular modeling suggest
the involvement of the protonated N3
atom in adenine and 1- and 7-deaza-
Keywords: NMR spectroscopy
·
AHCTUNGTRENNaGUN denine in the intramolecular catalysis
of the hydrolysis reaction through the
rare syn conformation.
nucleosides · nucleotides · physico-
chemical properties · prodrugs
Introduction
prodrug also has to be stable in the extracellular medium so
that it is not degraded outside the cell.
Over the past several decades, significant progress has been
made to improve the therapeutic potency of nucleoside ana-
logues for the treatment of viral infections, which has led to
the development of nucleotide prodrugs.[1] The important
feature of most nucleotide prodrugs is that the negatively
charged phosphate group remains masked outside the cell,
allowing permeation through the lipophilic cell membrane,
and the active nucleotide is processed in the cell by chemi-
cal and/or enzymatic degradation.[2,3] A suitable nucleotide
Nucleoside 5’-phosphoramidates (NPs) constitute one
class of these prodrug candidates, pioneered by Jones et al.
in the early 1980s,[4] and later developed by McGuigan
et al.[5] A few years later, Wagner et al. showed that aromat-
ic amino acid nucleoside phosphoramidates (aaNPs), con-
taining a negatively charged phosphate group, could exhibit
greater potency than the parent nucleoside.[6,7] Studies on
the activation pathways of aaNP triester prodrugs reveal
that inside the cell, esterase first catalyzes hydrolysis of the
carboxylic ester linkage. This results in the cleavage of the
P-O-aryl group by intramolecular attack of the carboxylate
group on the phosphorus atom, giving rise to a cyclic struc-
ture that is subsequently hydrolyzed to form the nucleoside
5’-phosphoramidate.[8] A phosphoramidase enzyme is be-
lieved to be involved in the release of the nucleoside 5’-
[a] M. Maiti, Dr. N. Dyubankova, M. Maiti, Prof. Dr. E. Lescrinier,
Prof. Dr. P. Herdewijn
Laboratory of Medicinal Chemistry
Rega Institute for Medical Research
Katholieke Universiteit Leuven
Minderbroedersstraat 10, 3000 Leuven (Belgium)
Fax : (+32)16-337340
monophosphate (5’-NMP)[7] by P N bond cleavage, bypass-
À
ing the first crucial rate-limiting 5’-monophosphorylation
step in the kinase pathway. The 5’-NMP then undergoes
consecutive 5’-di- and 5’-triphosphorylation by cellular kin-
ases, which subsequently produce their therapeutic effects as
a result of DNA chain termination/polymerase inhibition.[3,9]
A considerable challenge, however, is to develop nucleo-
tide analogues that are direct substrates/inhibitors of viral
polymerases, thereby bypassing the entire cellular kinase ac-
tivation cascade. Recently, we have shown that some aaNPs,
in particular, l-Asp-dAMP (1, Figure 1) and l-His-dAMP,
can act as direct substrates for HIV-1 reverse transcriptase
E-mail: Piet.Herdewijn@rega.kuleuven.be
[b] S. Michielssens, Prof. Dr. A. Ceulemans
Laboratory of Quantum Chemistry
Department of Chemistry and INPAC Institute for
Nanoscale Physics and Chemistry
Katholieke Universiteit Leuven
Celestijnenlaan 200F, 3001 Heverlee (Belgium)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201102279. It contains the de-
tailed synthetic schemes, procedures, and molecular characterization
data.
Chem. Eur. J. 2012, 18, 857 – 868
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
857

Figure 1. Chemical structures and atom numbering of the amino acid nucleoside phosphoramidates (aaNPs).
(HIV-1 RT).[10] Notably, it was observed that the presence of
the free carboxylic acid groups of the l-aspartic acid leaving
group is essential for recognition by HIV-1 RT.[11,12] These
carboxylic groups, however, are also a source of chemical in-
stability in the NPs. Therefore, it is important to study the
mechanism of chemical degradation of NPs to allow the
design of new, modified NPs with improved properties. Re-
cently, the kinetics of the hydrolysis of l-alanine methyl
esters of NP prodrugs was reported and a mechanism for
the hydrolysis reaction in both acidic and basic solutions
was proposed.[13–15]
However, the physicochemical properties and structural
determinants that influence the chemical stability and hy-
drolysis mechanism of this class of compounds are not
known in detail. We have synthesized a series of aaNPs, con-
taining either natural or modified nucleobases (Figure 1).
Aqueous stability, kinetics, and reaction pathways of hydrol-
ysis have been studied in pD 5–7 at different temperatures
(25, 40, and 708C) by using NMR spectroscopy. The pKa
values, protonation sites, and conformation in solution were
investigated by NMR spectroscopy for compounds 1, 21, 32,
and 33. Experimental and theoretical studies have shown
that the nucleobase, as well as the amino acid moiety, has a
large influence on the chemical stability, kinetics, and hy-
drolysis mechanism. These data contribute to the under-
standing of the determinants that govern the stability and
hydrolysis kinetics of aaNPs and may assist in the future
design of new therapeutic analogues.
part was carried out with sodium hydroxide (0.4m) in a
methanol/water solution at room temperature. An example
of this type of synthesis is shown in Scheme 1a for Asp-
dAMP (1).
The synthesis of the 1-deazaadenine analogues is shown
in Scheme 1b. The synthesis of 2-deoxy-3,5-di-O-para-toluo-
yl-a-d-ribofuranosyl chloride (10) was carried out by use of
a three-step reported method[17] starting from 2-deoxy-a/b-
d-ribose (Scheme 1b). Compound 15 was synthesized ac-
cording to a previously described method starting from com-
mercially available 2,3-diaminopyridine (11), with optimiza-
tion of some of the reaction steps to improve the yield.[18]
Sodium salt glycosylation of 15 with 10 is a stereospecific re-
action, yielding only the N9 b-isomer, which was then con-
verted into 2’-deoxy-1-deazaadenosine (18, purine number-
ing is followed, see Figure S1 in the Supporting Information)
by deprotection and reduction of the nitro group. Com-
pound 18 was phosphorylated[19] at the 5’-hydroxyl group
with POCl3 to obtain 19. Compound 19 was then converted
into phosphoramidates 21–23 in two steps, as shown in
Scheme 1a.
The 7-deazaadenine compound (27) was synthezised ac-
cording to the method described by Minakawa et al.[20]
Scheme 1c depicts the synthetic strategy followed to obtain
l-Asp-7-deaza-dAMP (32). Solid–liquid phase-transfer gly-
cosylation[21] of 27 with 10 was performed without amino
group protection by using tris(3,6-dioxaheptyl)amine (TDA-
1)/KOH, which afforded 28 with preferred b selectivity
(57%). Compound 28 was separated from its isomeric mix-
ture with 28a and 28b by flash column chromatography.
Compound 30 was converted into phosphoramidate 32, as
described in Scheme 1a.
Results and Discussion
Synthesis: The synthesis of methyl ester protected aaNPs
(Figure 1) was performed according to the method described
by Wagner et al.,[7,16] starting from 5’-NMP (Scheme 1a). Di-
methyl l-aspartic acid is used for the synthesis of aspartyl
NPs. Deprotection of the methyl esters of the amino acid
The synthesis of 2’-deoxy-3-deazaadenosine was per-
formed according to the protocol described by Seela et al.[22]
5’-Phosphorylation, DCC-assisted coupling, and methyl ester
deprotection were performed to obtain Asp-3-deaza-dAMP
(33) by a method analogous to that described for the Asp-1-
858
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 857 – 868

Hydrolysis of Amino Acid Nucleoside Phosphoramidates
FULL PAPER
Scheme 1. a) The dicyclohexylcarbodiimide (DCC)-assisted coupling and methyl ester deprotection. b) The synthesis of 1-deazaadenine-containing phos-
phoramidate analogues (21–23). For detailed reaction schemes, including the reagents and conditions, see Schemes S1 and S2 in the Supporting Informa-
tion. c) The synthesis of Asp-7-deaza-dAMP (32). For a detailed scheme, including the reagents and conditions, see Scheme S3 in the Supporting Infor-
mation. d) The synthesis of Asp-3-deaza-dAMP (33).
or Asp-7-deaza analogues. The synthetic strategy is briefly
shown in Scheme 1d.
tial use of aaNPs as substrates for thermostable polymerases
and/or application in the polymerase chain reaction (PCR).
1
Hydrolysis was monitored by 31P and H 1D NMR spec-
Hydrolysis kinetics: Most of the aaNPs were found to be
stable in neutral to basic solutions (pD 7–10; for 21 pD 8–
10) in D2O at room temperature for more than 24h, due to
the presence of monoanionic phosphoramidate groups.
Since the hydrolysis kinetics of aaNPs (1, 21, and 32) were
slow at room temperature in acidic solution (pD 5), the re-
action was studied at 708C. Knowledge of the chemical sta-
bility at high temperature is important in view of the poten-
troscopy over the range pD 5–7 (Figure 2a). The reactions
are expected to follow pseudo-first-order kinetics with re-
spect to phosphoramidate concentration, since D2O was
present in large excess and there was no significant change
in the pD of the solution after hydrolysis. The first-order ki-
netics (as reported previously for other phosphor-
AHCTUNGTRENNUNG
amidates)[13,23,24] is also apparent from the exponential decay
in the concentration of phosphoramidate (Figure 3, top left)
and from the corresponding linear transformed data
Chem. Eur. J. 2012, 18, 857 – 868
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
859

P. Herdewijn et al.
Influence of the amino acid and nucleobase on reactivity: In
general, rate constants from Table 1 indicate that aspartyl
NPs containing both a- and b-COOH groups (1, 21, and 4)
hydrolyzed faster than those containing Gly (with an a-
COOH) and b-Ala (with a b-COOH) NPs (Figure S4 in the
Supporting Information). For example, hydrolysis of Asp-1-
deaza-dAMP (21) was 28 and 47 times faster than that of
Gly-1-deaza-dAMP (22) and b-Ala-1-deaza-dAMP (23), re-
spectively, at pD 6. Glycinyl NPs showed moderate reactivi-
ty, whereas the b-Ala NPs were the least reactive in the
series. The order of reactivity of the amino acids is Asp>
Gly>b-Ala.
À
À
The aaNPs were hydrolyzed either by P N or P O bond
cleavage or by both pathways simultaneously (Figure S5 in
the Supporting Information). The hydrolysis products, as
well as the reaction pathway, were identified in the H–31P
1
TOCSY spectrum (Figure 2b). The b-Ala phosphoramidates
À
(3, 6, and 23) hydrolyzed solely by P N bond cleavage,
which was independent of the nature of the nucleobase (Fig-
Figure 2. a) A kinetic profile of the hydrolysis reaction of Asp-1-deaza-
dAMP (21) at pD 7.0 (T=708C), depicted by a stacked plot of the 31P
spectra at different time intervals, showing the disappearance of the
parent 31P peak (ꢁ7 ppm) and gradual increase in another 31P peak
ure S6 in the Supporting Information). Hydrolysis of the
Gly NPs, however, underwent P N and/or P O cleavage,
depending on the nucleobase present. In Gly-1-deaza-
dAMP (22), P O bond hydrolysis dominated, whereas the
opposite trend was observed for Gly-dAMP (2). Conversely,
in the presence of thymine, that is, Gly-TMP (5), hydrolysis
À
À
(ꢁ2 ppm). The asterisk ( ) indicates the free phosphate peak. b) The
*
1H–31P TOCSY spectrum during hydrolysis of 21 at pD 6.0 (T=708C),
À
À
depicting the formation of inorganic phosphate by P O bond cleavage.
Inorganic phosphate (indicated with an asterisk ( )) shows no correlation
*
with any protons, whereas the phosphoramidate 31P atom shows correla-
tions with Ha and Hb of the aspartyl group, along with H4’, H5’, and
H5’’ of the sugar moiety.
(Figure 3, top right). Tempera-
ture- and pH-dependent kinet-
ics were observed for all of the
aaNPs, that is, faster kinetics
under acidic conditions (Fig-
ure S2 in the Supporting Infor-
mation) and at higher tempera-
tures (Figure S3 in the Support-
ing Information). The acid-cata-
lyzed hydrolysis demonstrates a
first-order dependence on the
hydronium
ion
concentra-
tion,[13,23,25] as shown by the ex-
ponential decay curve in the
pD–rate
profile
(Figure 3,
bottom left) and from the cor-
responding linear relationship
(Figure 3, bottom right). The
overall first-order rate constants
for the hydrolysis of aaNPs are
presented in Table 1. Hydrolysis
of 21 was also performed at
408C and the rate constants for
this reaction are provided in
Table S1 in the Supporting In-
formation.
Figure 3. Kinetic plots for the acid-catalyzed hydrolysis of amino acid nucleoside phosphoramidates at pD 5
and 708C. The exponential decay of phosphoramidates with time (top left) and the corresponding linear rela-
tionships (top right, time vs. ln C plot) in the hydrolysis reaction, indicating first-order kinetics with respect to
phosphoramidate concentration (C). The exponential (bottom left, pD vs. first-order rate constant (k)) and
linear (bottom right, pD vs. ln k) pD–rate profiles for the hydrolysis of the nucleoside phosphosphoramidates
at 708C, showing first-order dependence on the hydronium ion concentration.
860
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 857 – 868

Hydrolysis of Amino Acid Nucleoside Phosphoramidates
FULL PAPER
Table 1. The first-order rate constants (k, ꢁ10À4 minÀ1) for the hydrolysis of the amino acid nucleoside phos-
phoramidates at 708C (pD 5.0–7.0). “Slow” indicates that the kinetics of hydrolysis are very slow at 708C.
phoryl aspartic acid in compari-
son with other N-phosphoryl
amino acids containing only a-
or b-COOH groups was report-
ed. It was shown that the a-
COOH group is involved in the
nucleophilic attack on phospho-
rus[26,27] in a neighboring-group
participation (NGP) reaction
mode during the hydrolysis, re-
À
À
À
À
P O>P N means that the P O bond cleavage is observed more than P N bond cleavage. Half life (t1/2) and
standard error of the mean (SEM) are provided within parentheses and as ÆSEM, respectively.
Compound
Mechanism
Rate constant, k [ꢁ10À4 minÀ1
]
pD 5.0
pD 6.0
pD 7.0
6
À
À
Asp-dAMP (1)
P O>P N
À
P O
54Æ7 (t1/2 =128 min)
17
226Æ6
36Æ2
slow
3
Asp-1-deaza-dAMP (21)
Asp-7-deaza-dAMP (32)
Asp-3-deaza-dAMP (33)
Asp-TMP (4)
Gly-dAMP (2)
Gly-1-deaza-dAMP (22)
Gly-TMP (5)
b-Ala-dAMP (3)
b-Ala-1-deaza-dAMP (23)
b-Ala-TMP (6)
388Æ56 (t1/2 =18 min)
48Æ0.5
À
À
P O>P N
80Æ4 (t1/2 =87 min)
8
slow
2
À
À
P O<P N
8Æ1 (t1/2 =918 min)
À
À
P O<P N
13Æ0.2 (t1/2 =516 min)
À
À
P O<P N
14
62
5
13
46
8
slow
8
slow
slow
slow
slow
slow
slow
À
sulting in the P O bond cleav-
À
À
P O>P N
age.[13] The presence of an addi-
tional b-COOH group was sug-
gested to have an extra catalytic
role through intramolecular hy-
drogen bonding.[26,28] These
roles suggest that in aspartyl
À
À
P N
1
5
5
À
P N
À
P N
À
P N
slow
À
by only P N bond cleavage was observed, despite the pres-
ence of an a-COOH group. In the aspartyl NP series, for
Asp-TMP (4) and Asp-3-deaza-dAMP (33), cleavage of the
and glycinyl NPs the a-COOH group is involved in the P O
bond hydrolysis. All of the aaNPs containing only the b-
COOH group (b-Ala) undergo acid-catalyzed P N bond hy-
À
À
P N bond dominated, whereas for Asp-dAMP (1) and Asp-
drolysis without the involvement of any intramolecular cat-
alysis due to the absence of an a-carboxyl group. All of
these observations are consistent with our kinetic data.
À
7-deaza-dAMP (32), predominant cleavage of the P O
bond was detected. The most important characteristic ob-
À
served for Asp-1-deaza-dAMP (21) was exclusive P O bond
hydrolysis (Figure 2).
pKa and protonation sites: The pKa values and protonation
sites of aspartyl adenine and 1-, 3-, and 7-deazaadenine NPs
(1, 21, 32, and 33) were determined with the aim of charac-
terizing the reactive protonated tautomeric form(s) under
the studied conditions in order to explain any conformation-
al preferences based on electrostatic interactions and for ac-
curate molecular modeling, which may find a rationale for
the observed reactivity differences. Nucleobase pKa values
were determined from the pH-dependent chemical shift per-
turbation (CSP) of the nucleobase protons. To obtain the
pKa values of the phosphoramidate nitrogen for compounds
1 and 21, the corresponding 15N-labeled aspartyl NPs were
synthesized. The CSP of the 15N atom (Figure 4a) showed a
typical sigmoidal response (Figure 4b). The pKa value was
determined from analysis of the Hill plot (Figure 4c).
Table 2 presents the pKa values and protonation sites of as-
partyl NPs.
During hydrolysis of the aspartyl and glycinyl NPs (except
5), the corresponding nucleoside was obtained as one of the
products and its formation is not due to the hydrolysis of
the nucleotide intermediate. The nucleoside was formed di-
À
rectly by P O bond cleavage from the parent phosphorami-
date. This was verified by studying the stability of the re-
spective nucleotides under the same hydrolytic conditions,
for which no nucleoside product was formed.
The nature of the nucleobase was observed to have a pro-
nounced effect on the (stability) kinetics (Figures 3 and S4
in the Supporting Information) and mechanism of the hy-
drolysis reaction. For aspartyl NPs, the reactivity of adenine
or the 1- or 7-deazaadenine analogues was higher than for
the thymine or 3-deazaadenine analogues. Asp-1-deaza-
dAMP (21: t1/2 =18 min) hydrolyzed 50 times faster than
Asp-3-deaza-dAMP (33: t1/2 =918 min) at pD 5 and 708C.
The effect of 1-deazaadenine on reactivity was the
Table 2. Protonation sites and pKa values of aspartyl nucleoside phosphoramidates (1,
21, 32, and 33) obtained by NMR spectroscopic pH-metric titration. The protonation
site in the nucleobase of 21, 32, and 33 was determined by using their corresponding
nucleoside (18, 29, and 2’-deoxy-3-deazaadenosine, respectively). The pKa of the phos-
phoramidate nitrogen in 1 and 21 was determined by using 15N-labeled aspartyl nu-
cleoside phosphoramidates. Marker atoms used for pKa determination are indicated
within parentheses; n.d.=not determined.
highest and kinetic acceleration was observed even
with the Gly and b-Ala analogues. The general
order of reactivity of the nucleobases is 1-deazaade-
nine>7-deazaadenine>adenine>thymineꢀ3-de-
ACHTUNGTRENNUNGazaACHTUNGTRENNUNGadenine. From the rate constants for the hydrol-
ysis of the Asp adenine and deazaadenine conju-
gates, it is evident that the N3 atom of the adenine Compound entry
base has a catalytic role, and hence the 1-deaza an-
Protonation
site
pKa
À
Nucleobase
P N
nitrogen
alogue is the most reactive and the 3-deaza ana-
logue the least.
2’-deoxy-1-deazaadenosine (18)
N3
N1
N1
N1
N3
N1
N1
4.6 (C2)
–
–
–
2’-deoxy-7-deazaadenosine (29)
2’-deoxy-3-deazaadenosine
Asp-dAMP (1)
4.7Æ0.1 (H2/C2)
7.5 (H2)
À
The accelerated kinetics with preferential P O
bond cleavage in some of the aaNPs containing the
aspartyl group and adenine/deazaadenine seem to
have a dual role in the intramolecular catalysis. Pre-
viously, the higher reactivity of N-diisopropylphos-
3.6 (H2)
4.3 (15N)
4.1 (15N)
n.d.
Asp-1-deaza-dAMP (21)
Asp-7-deaza-dAMP (32)
Asp-3-deaza-dAMP (33)
5.5Æ0.3 (H1/H2)
4.2 (H2)
7.4Æ0.1 (H2)
n.d.
Chem. Eur. J. 2012, 18, 857 – 868
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
861

P. Herdewijn et al.
The pKa values of the 3-de-
ACHUTNGRENNUaG zaACHTUNTGNERaNUGN denine moiety in 33 (ca.
pKa =7.4) and the 1-deazaade-
nine moiety in 21 (ca. pKa =
5.5) are much higher than that
of the 7-deazaadenine moiety
in 32 (pKa =4.2) and even
higher than that of the adenine
moiety in 1 (pKa =3.6). There-
fore, it is clear that the popula-
tion of the protonated species
of 33 and 21 is higher than that
of 32 and 1 in acidic solution
(pD 5–7). Although the popula-
tion of protonated forms of 33
is higher than that of 21, based
on their pKa values, compound
33 was much more stable than
21, which decomposed rapidly.
Therefore, it is evident that the
pKa data alone cannot explain
their reactivity.
However, the base protona-
tion site for Asp-1-deaza-
dAMP (21) is the N3 atom,
whereas it is N1 for the others,
and this may have a role in ex-
plaining the exceptional reactiv-
ity of 21. During the pH-metric
titration of aspartyl adenine
and deazaadenine NPs, as well
as their corresponding nucleo-
1
Figure 4. a) An overlay of the H–15N HSQC (long-range) spectra of (15N) Asp-dAMP (1), depicting the chem-
ical shift perturbation (CSP) of the phosphoramidate nitrogen in an NMR spectroscopic pH-metric titration
from pD 9.0 to 2.7 at 258C. b) The corresponding sigmoidal response of the CSP. c) A Hill-plot analysis of the
sigmoidal data (b) is shown as an example. A similar analysis was performed for the determination of all of
the pKa values. d) The observed deshielding of the H8 proton in 1 during the pH-metric titration, shown by a
sigmoidal plot. e) The unusual shielding (a wrong way chemical shift) of the H8 proton in Asp-1-deaza-dAMP
(21) in the pH-metric titration.
sides,
a significant difference
was observed for the CSP of
Determination of the protonation site of the nucleobase
in the deazaadenine analogues (21, 32, and 33) was carried
out by using the corresponding nucleosides (18 and 29). In
the H8 proton of 21. A shielding effect on the H8 proton (a
wrong way chemical shift, Figures 4e and S10 in the Sup-
porting Information) was observed during the base protona-
tion of 21, whereas the usual deshielding effect on the H8
proton (Figures 4d and S11 in the Supporting Information)
occurred in all other aspartyl adenine and deazaadenine
NPs and nucleosides (even for 2’-deoxy-1-deazaadenosine).
This suggests that, owing to the protonation, the environ-
ment of the purine ring in 21 differs from that of the aspar-
tyl adenine and 3- and 7-deazaadenine NPs, possibly indicat-
ing a conformational change in the nucleobase that positions
H8 in 21 away from the deshielding zone of the phosphate
group.
1
the titration of 18, the C2 carbon in the H–13C HSQC spec-
trum showed upfield CSP, whereas the C8 carbon remains
unchanged (Figure S7 in the Supporting Information), clear-
ly indicating that the protonation site is N3, with a pKa of
4.6. Similarly, the protonation site for 2’-deoxy-3-deazaade-
nosine was found to be N1, with a pKa of 7.5. However, in
the case of 29, the 1H–13C HSQC spectrum alone cannot
conclusively provide the protonation site in the pyrimidine
ring because protonation either at N1 or N3 leads to shield-
ing of the C2 carbon (Figure S8 in the Supporting Informa-
1
tion). For this reason, H–13C HMBC spectra were recorded
À
À
to monitor the CSP of the C4 and C6 carbon atoms. Al-
though upfield shifts of both the C4 and C6 carbon atoms
were observed (Figure S9 in the Supporting Information),
the total change in CSP was 5.6 ppm for C6, whereas that
for C4 was much smaller (Dd=1.8 ppm). This suggests that
the protonation likely occurs at N1 with a pKa of 4.7. As re-
ported previously, the N1 atom is the protonation site for 2’-
deoxyadenosine (pKa =3.8).[29]
Possible mechanism of the acid-catalyzed P N and P O
bond hydrolysis: In the acid-catalyzed hydrolysis of aaNPs,
À
À
only P N bond hydrolysis for b-Ala NPs, simultaneous P N
and P O cleavage for Asp and Gly NPs (except for 5), and
only P O bond hydrolysis for Asp-1-deaza-dAMP were ob-
served. This suggests the involvement of at least two inde-
À
À
À
À
pendent reaction mechanisms for P N and P O bond hy-
drolysis that are operating together or independently de-
pending on the molecular nature of the aaNPs. To explain
862
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 857 – 868

Hydrolysis of Amino Acid Nucleoside Phosphoramidates
FULL PAPER
À
the hydrolysis pathways, we propose a reaction mechanism
phosphoramidates.[13,31,33,34] P N bond hydrolysis most prob-
ably proceeds through the N-protonated zwitterionic form
(III) because attack by water will be more favorable for the
zwitterionic phosphate than the monoanionic form and the
leaving group can depart as a neutral amine. The mechanism
À
À
for P N and P O bond hydrolysis based on the reported lit-
erature.
The possible protonated forms of aaNPs and their tauto-
meric equilibrium in solutions at different pH values are
shown in Scheme 2a. The phosphoramidate monoanion (IV)
can in principle be protonated either at the amide nitrogen
or at the phosphoryl oxygen. Experimental evidence has
been reported that favors N protonation (vs. O protonation)
in solution,[25,30] although some authors have provided theo-
retical arguments[23,31] for partial O protonation and for the
existence of an equilibrium between the N- and O-protonat-
ed forms.
À
of P N bond hydrolysis (Scheme 2b) possibly involves a
“loose” SN2-like transition state (TS) in which the P N
bond cleavage is far more advanced than the O P bond for-
mation, that is, both the entering and leaving groups are
present in the transition state,[25,32,35] representing a border-
line case between an associative and dissociative mecha-
nism.
À
À
The Asp and Gly NPs are also (partly or fully) hydrolyzed
À
À
The acid-catalyzed P N bond hydrolysis of various non-
nucleosidic phosphoramidates has been documented previ-
by P O bond cleavage that is triggered by the intramolecu-
lar attack of the a-carboxyl group on the phosphate center.
ously[15,23,32] and it has recently been discussed for nucleoside
A plausible P O bond-hydrolysis mechanism is depicted in
À
Scheme 2c. The mechanism
may involve the O-protonated
tautomer (I) since the attack of
a neutral phosphate by the a-
carboxyl oxyanion will be more
favorable. However, it has re-
cently been shown, by NMR
spectroscopy of the 18O-labeled
compound, that the neighboring
carboxylate anion can also
attack
a monoanionic phos-
phate group.[36] Initial attack of
the a-carboxyl oxyanion on
phosphorus gives rise to
a
penta-coordinated, five-mem-
bered cyclic trigonal bipyrami-
dal (tbp) phosphorane inter-
mediate in which the carboxy
oxygen, as the entering nucleo-
phile, takes an apical posi-
tion,[13] and hence, the nitrogen
atom, as a member of the five-
membered ring, is forced to
remain in an equatorial posi-
tion.[37] The 5’-O-linked nucleo-
side moiety occupies the re-
maining apical position and
hence departs with a proton
transfer to the 5’-oxygen. In
contrast to the protonated
À
amine leaving group (P N hy-
drolysis), a nucleoside is expect-
ed to be a rather poor leaving
À
group (P O hydrolysis) because
it is difficult to protonate the
5’-oxygen under the studied
conditions (pD 5–7). For the re-
moval of a neutral nucleoside
leaving group a proton transfer
is necessary. The departure of
the nucleoside from the phos-
Scheme 2. a) The possible protonated forms of the nucleoside phosphoramidates and their tautomers based on
the experimental pKa and pH in solution. The possible mechanisms for the acid-catalyzed hydrolysis of the
À
amino acid nucleoside phosphoramidates: b) a mechanism for P N bond hydrolysis and c) a mechanism for
À
P O bond hydrolysis.
Chem. Eur. J. 2012, 18, 857 – 868
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
863

P. Herdewijn et al.
phorane intermediate yields a cyclic phosphoramidate that
is anticipated to be hydrolytically unstable at 708C, and thus
is subsequently hydrolyzed into the amino acid and inorgan-
ic phosphate.
transfer occurs without a water molecule the barrier of this
step is 35.9 kcalmolÀ1. The most stable reactant is the zwit-
terionic form with the nitrogen protonated and the phos-
phate deprotonated, but since the cyclic intermediate
(having a penta-valent phosphorus atom with a protonated
À
Influence of the amino acid and nucleobase from a theoreti-
cal perspective:
nitrogen) for P O bond hydrolysis is not stable with this
tautomer the reaction must occur via tautomer
I
(Scheme 2c) with the proton on the phosphate group. There
is an energy penalty of 5.7 kcalmolÀ1 for this tautomeriza-
tion step. The b-carboxyl group might have an enhancing
entropic effect through a structural role, as can be seen in
PO_ts2_base, in which a hydro-
Amino acid: The influence of the amino acid was studied by
using the model compound, asp-monomethylphosphate
shown in Figure 5 in which the nucleoside part of the mole-
gen bond between the 3’-hy-
droxyl group of the sugar
À
moiety and/or the N H group
of Asp and the b-carboxyl
group places the a-carboxyl in
the correct position for attack
on phosphorus.
À
The cleavage of the P N
bond occurs via the most stable
tautomer III (Scheme 2b) and
is not influenced by the amino
acid because it is geometrically
impossible to form a cyclic five-
membered intermediate with
nitrogen and the carboxyl
group of the amino acid at the
À
apical positions. The P N bond
hydrolysis occurs through
a
concerted mechanism in which
À
À
P N bond cleavage and O P
bond formation occur simulta-
neously. First, the N protona-
À
tion lengthens the P N bond
from 1.6 to 1.81 ꢂ, before the
À
formation of the O P bond
starts. In the transition state
À
(PN_ts_wat) the P N bond is
further extended to 2.0 ꢂ with
simultaneous partial formation
À
À
Figure 5. An overview of the mechanisms with the lowest energy barriers for P O and P N bond hydrolysis.
The red line indicates the role of the 1-deazaadenine base. All energy values are in kcalmolÀ1
.
À
of the O P bond (1.99 ꢂ). Fi-
nally, moving from the transi-
À
cule has been replaced by a methyl group and only the
effect of the carboxyl groups is investigated. Previous
work[34] on monomethylphosphate has determined that the
tion state to the product state the O P bond is fully formed
À
(1.75 ꢂ) and the P N bond is cleaved. This is different from
the mechanism predicted for N-methyl phosphoramidate,[34]
in which a more dissociative mechanism is preferred. This
can be explained by the presence of an extra methyl group
on phosphate that disfavors the formation of a metaphos-
phate intermediate.
À
barrier for hydrolysis of the P O bond is about 40 kcal
molÀ1. In asp-monomethylphosphate the a- and b-carboxyl
groups offer the possibility for intramolecular catalysis. As
shown in Figure 5, the barrier is reduced to 25.7 kcalmolÀ1
when a five-membered ring is formed by the attack on phos-
phorus by the a-carboxyl group (Scheme 2c). Formation of
a six-membered ring by the attack of the b-carboxyl group
is 5 kcalmolÀ1 higher in energy. Catalytic water molecules
are of major importance in this reaction (Figure 5), with the
highest energy barrier in the presence of water being
25.7 kcalmolÀ1 for PO_ts2_wat, whereas when the proton
Nucleobase: The effect of the nucleobase is most pro-
nounced for the 1-deaza analogues. To further catalyze the
hydrolysis reaction the base must reduce the highest barrier,
which is 25.7 kcalmolÀ1 for PO_ts2_wat. Calculations for
Asp-1-deaza-dAMP (21) show that when the base is proton-
ated at N3 (expected for 1-deazaadenine), the syn and anti
864
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 857 – 868

Hydrolysis of Amino Acid Nucleoside Phosphoramidates
FULL PAPER
conformations of the base are equal in free energy. This pro-
tonation can favor the possibility of forming a hydrogen
bond between the N3-H and the phosphate group. Hydro-
gen bonding between the base and the catalytic water mole-
between the syn and anti conformations of the nucleobase.
The JH1’-C4 coupling constant of approximately 2 Hz for
3
Asp-dAMP (1) was observed at both pD values, indicating a
predominantly anti orientation (Figure S14b in the Support-
ing Information).[38] The anti conformation of the nucleo-
base was further supported by the observed 1H–1H NOE
cross peaks (Figure S14c in the Supporting Information).
However, in case of Asp-1-deaza-dAMP (21), a significant
conformational difference in the nucleobase orientation was
observed in acidic and basic solutions. In basic solutions
cule, as shown in PO_ts2_base (Figure 5), reduces the barri-
À1
À
er for P O cleavage by 6.6 kcalmol . Although, there is
also an alternative possibility with a slightly higher energy
in which the protonated N3 of the nucleobase can directly
be involved in hydrogen bonding with the 5’-O of the nu-
cleoside part, thereby effectively delivering a proton to the
leaving nucleoside. Together these may explain the in-
3
(pD 8), the JH1’-C4 coupling constant was approximately
À
creased rate of the P O bond hydrolysis for Asp-1-deaza-
dAMP compared to Asp-dAMP, Asp-7-deaza-dAMP, and
Asp-3-deaza-dAMP.
2.0 Hz, indicating a predominantly anti conformation of the
nucleobase (Figure 6a), which is further supported by the
observed NOE cross peaks. The H8 proton showed cross
peaks with the sugar and aspartyl protons, whereas the H2
proton did not show any correlation (Figure 6b, left). More-
over, in the HOESY spectrum, the H8 proton showed a cor-
relation with phosphorus, whereas H2 did not (Figure 6c,
left). These data clearly indicate that the nucleobase orien-
tation in 21 in basic solution is in the anti domain. On the
Conformational study by NMR spectroscopy: To investigate
the role of conformation in determining the kinetics and
mechanism of P O bond hydrolysis, as suggested by molec-
ular modeling, a comparative conformational study in solu-
tion for Asp NPs containing adenine and deazaadenines was
performed in acidic and basic solutions (pD 5 and 8). The
deoxyribose sugar ring in all of the conjugates adopts a typi-
cal S-type (South, 2’-endo) puckering in D2O, irrespective of
the pD of the solution, as indicated by the appearance of a
triplet H1’ peak (Figure S12A in the Supporting Informa-
À
3
other hand, in acidic condition (pD 5), the JH1’-C4 coupling
constant was not possible to determine because the C4
carbon peak became too broad (Figure 6a), possibly indicat-
ing conformational dynamics across the glycosyl bond. How-
ever, from the ROESY and HOESY spectra, the existence
of some molecules in the syn conformation was identified
due to the presence of H2 proton cross peaks in addition to
those of H8 (Figure 6b and c, right). The nucleobase orien-
tation in 32 and 33 was akin to 21, that is, some population
of the syn conformation was present in acidic solution (Fig-
ures S15 and S16 in the Supporting Information). The
change in nucleobase conformation from anti to syn in
acidic solution can be rationalized on the basis of possible
electrostatic interactions between the positively charged pyr-
imidine ring and the negatively charged phosphate and/or
carboxylic acid groups.
In acidic solution, for deazaadenine-containing analogues
the syn conformers are detected, whereas this was not so for
the adenine analogue. This could be explained on the basis
of the higher nucleobase pKa values of the deazaadenines
compared with adenine, for which the existence of more
protonated forms is expected in acidic solutions. As dis-
cussed earlier, both the kinetic data and the molecular mod-
eling suggest the involvement of a protonated N3 atom in
3
tion, JH1’-H2’/H2’’ =7 Hz). In all of the Asp NPs, the nucleotide
part adopts the orientation (at both pD values) observed in
the nucleotide residue of regular DNA duplexes. The JC4’-P
(9.3 Hz) coupling constant was determined from the H-de-
3
1
coupled carbon spectrum, indicating that the b torsion angle
(C4’-C5’-O5’-P) is within the trans region.[38] Moreover, a
4
small JH4’-P coupling constant (ꢁ1–2 Hz) was detected by
comparing the 31P-coupled and -decoupled H4’ proton peak
(Figure S12b in the Supporting Information). This was fur-
ther confirmed by the appearance of a small cross peak in
the 1H–31P HETCOR spectra (Figure S12c in the Supporting
Information). The observation of this long-range coupling
(4JH4’-P) signified a planar W-shaped conformation of the mo-
lecular fragment H4’-C4’-C5’-O5’-P with a b torsion angle in
the anti region and g torsion angle (C3’-C4’-C5’-O5’) in the
gauche+ region.[38] The aspartyl moiety in all of these ana-
logues adopts a similar conformation at both pD values as
no significant change in the 3J homo- and heteronuclear cou-
pling constants was observed (Table S2 in the Supporting In-
formation). The aspartyl moiety was found to be close to
the sugar residue, irrespective of the solution pD, as shown
by the NOE correlation between the H3’ and Ha/Hb of Asp
that was observed in the ROESY spectra (Figure S13 in the
Supporting Information). This is somewhat in agreement
with the proposed molecular model, in which hydrogen
bonding between the b-COOH and the 3’-OH was shown.
To determine the existence of a syn conformation of the
nucleobase, as suggested by the molecular modeling study,
À
catalyzing the P O bond hydrolysis. The protonated N3
atom in adenine and 1- and 7-deazaadenine might be in-
volved in hydrogen bonding through a water molecule (or
directly) with the 5’-O of aaNPs, for which it can act to sta-
À
bilize the TS of the P O bond hydrolysis or by effectively
delivering a proton to the nucleoside leaving group through
the rare syn conformation. Based on the protonation site of
the nucleobase, the N3 protonated form of Asp-1-deaza-
dAMP is expected to be more abundant than for Asp-7-
deaza-dAMP and Asp-dAMP, for which the N3 protonated
form is rare because N1 is the predominant protonation site.
For Asp-3-deaza-dAMP, on the other hand, there is no pos-
sibility for N3 protonation. This may in general explain the
1H-coupled 13C, H–1H ROESY and H–31P HOESY experi-
ments were performed. The glycosyl torsion angle c was de-
1
1
3
3
fined by the JH1’-C4 (and JH1’-C8) coupling constant and based
on the H–1H and H–31P NOE interactions to discriminate
1
1
Chem. Eur. J. 2012, 18, 857 – 868
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
865

P. Herdewijn et al.
Figure 6. a) The 1H-coupled 13C spectra of Asp-1-deaza-dAMP (21), showing the splitting pattern of the C4 carbon peak at pD 5 and 8, indicating the
3
change in the JH1’-C4 coupling constant due to conformational changes in the nucleobase in basic and acidic solution. b) An expansion of the ROESY
spectrum of 21 (left: at pD 8), showing NOE cross peaks between H8 and the sugar protons and Ha of the aspartyl group, whereas H2 shows no NOE
cross peaks with any protons, indicating the predominantly ꢃanti-likeꢄ conformation of the nucleobase moiety in basic solution. In contrast at pD 5
(right), the H2 proton shows NOE cross peaks with the sugar and aspartyl protons in addition to those of the H8 proton, indicating some population of
the ꢃsynꢄ conformation in acidic conditions. c) The expansion of the 1H–31P HOESY spectrum of 21 (left: at pD 8), depicting the NOE cross peak be-
tween the H8 and 31P in which no cross peak appeared between H2 and 31P, whereas at pD 5 (right), a small cross peak appeared between H2 and 31P,
along with that with the H8 proton, indicating proximity between the H2 proton and phosphorus in acidic conditions. All spectra were recorded at 108C
in D2O.
À
order of reactivity of the nucleobase in catalyzing the P O
bond hydrolysis of Asp and Gly NPs.
tial to identify the determinants that govern selectivity for
À
À
P N/P O bond hydrolysis.
In this study, we have attempted to identify the structural
factors that modulate the moleculeꢄs physicochemical prop-
erties and influence the kinetics and mechanism of the hy-
drolysis reaction. The experimental and theoretical studies
have revealed a significant reactivity difference based on the
nature of the amino acid and nucleobase. The presence of
Conclusion
Amino acid nucleoside phosphoramidates can be used in
the antiviral prodrug strategy and as direct substrates for
viral polymerases. The ꢃprodrugꢄ approach relies on the suc-
cessful delivery of the nucleotide moiety into the cell either
by chemical or enzymatic hydrolysis of the nucleoside phos-
À
only a b-carboxyl group triggered selective P N bond cleav-
age, whereas the presence of an a-carboxyl group made the
À
P O bond more labile and this lability was further enhanced
À
phoramidate by selective P N bond cleavage. When func-
by the presence of both an a- and a b-carboxyl group. The
rate of the hydrolysis reaction was further accelerated by
the presence of the adenine, 1-, or 7-deazaadenine base for
tioning as a direct substrate for a polymerase, the phosphor-
À
amidate should be stable enough to resist the unwanted P
N and P O bond cleavage before it reaches the enzymeꢄs
pocket. In both of these approaches, unnecessary release of
À
À
P O bond hydrolysis. The 1-deazaadenine nucleobase, in
combination with a- and b-carboxylic acid groups, made the
À
À
the nucleoside by means of P O bond cleavage should be
avoided. To establish the aforementioned criteria, it is essen-
molecule exceptionally prone to P O bond hydrolysis. Ki-
netic data, molecular modeling, and conformational studies
866
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 857 – 868

Hydrolysis of Amino Acid Nucleoside Phosphoramidates
FULL PAPER
enhancement and gradient coherence selection for 8 scans, 128/2048 com-
plex data points, 100/10 ppm spectral width in t1 and t2, respectively. The
pH titration studies were performed over a pD range of 1.5<pD<12.
The pH values were obtained by using the equation: pH=pDÀ0.4.[40]
The pKa determination is based on a linear regression analysis of the Hill
suggest the possibility for the involvement of the protonated
N3 atom in adenine and 1- and 7-deazaadenine in the intra-
À
molecular catalysis of the P O bond hydrolysis through the
syn conformation.
Based on the hydrolysis reaction preferences (P N vs. P
O) of these compounds, it could, for example, be expected
plot[41] by using the equation: pH=log
A
À
À
sents the fraction of protonated species. The value of a was calculated
from the change in the chemical shift relative to the neutral state at a
given pH divided by the total change in chemical shift between the neu-
tral and fully protonated states.
À
À
that 3-deaza (P N>P O) phosphoramidates would more
favorably deliver nucleoside monophosphates compared
with 1- or 7-deaza compounds, which would preferentially
deliver nucleosides through chemical degradation. Recently,
it was shown that 2’-deoxyadenosine-5’-iminodiacetyl phos-
phoramidate (IDA-dAMP, Figure S17A in the Supporting
Information) is a more efficient substrate for the HIV-1 RT
than Asp-dAMP (1).[12] Notably, its hydrolysis study (Fig-
ure S17B in the Supporting Information) revealed that de-
spite the presence of two a-carboxyl groups it hydrolyzed
Determination of protonation sites: To determine the protonation site of
the nucleobase the pH-dependent 13C chemical shifts were monitored by
using 1H–13C HSQC and 1H–13C HMBC NMR spectroscopy. The
1H–13C HSQC spectra were recorded with sensitivity enhancement and
1
gradient coherence selection. To measure the H–13C correlations of only
the nucleobase part, 1JCH was set to 180 Hz and 4 scans, 128/2048 complex
data points, and 80/10 ppm spectral widths in t1 and t2 were used, respec-
tively. The 1H–13C HMBC spectra were measured with 4 scans and 128/
2048 complex data points, and 100/10 ppm spectral widths in t1 and t2, re-
spectively. Delays were optimized for the transfer in nucleobases by
using 1JCH =145 Hz in the low-pass filter and 3JCH =8 Hz for long-range
correlations.
À
exclusively by P N bond cleavage (Figure S17C and D in
the Supporting Information). This could be due to the in-
creased basicity of the phosphoramidate nitrogen, leading to
favorable protonation and release of iminodiacetate by
attack of a water molecule. This example and other data
from this study show that the properties of the nucleobase,
as well as those of the amino acid leaving group, have to be
Theoretical calculations: All calculations were performed by using the
Gaussian 09 package.[42] Visualization was done with VMD.[43] The struc-
tures were optimized with the B3LYP method[44] with the DGDZVP[45]
basis set. Solvation corrections by using continuum solvent were per-
formed during the optimization process by using PCM.[46] Every station-
ary point was checked by frequency analysis. Zero-point energy correc-
tions by using B3LYP were scaled[47] with a factor of 0.9877.
À
À
considered to explain the P N/P O bond hydrolysis of nu-
cleoside phosphoramidates.
Conformational studies by NMR spectroscopy: 1D 31P-coupled and -de-
coupled 1H spectra were recorded with 32 scans and 66k FIDs at 258C.
Proton-coupled 1D 13C spectra with gated 1H decoupling were acquired
by using 1024 scans with a spectral width of 70 ppm and 100k FIDs. The
1H–31P HETCOR[48] spectrum was acquired at 258C with 128 scans, 2048
Experimental Section
1
data points in the H dimension (t2) and 512 increments in the 31P dimen-
sion (t1) over spectral widths of 3500 and 4050 Hz, respectively. The
ROESY[49] experiments in D2O (ROESY-spin-lock pulse=300 ms) were
performed at 108C with a sweep width of 10 ppm in both dimensions, 64
scans, 2048 data points in t2 and 512 FIDs in t1. All of the above experi-
ments were repeated in both acidic (pD=5.0) and basic (pD=8.0) solu-
tion. 1H–31P 2D heteronuclear NOE (HOESY) experiments[50] were per-
formed by using the Bruker pulse program hoesyph. The WALTZ-65
composite pulse sequence for phosphorus decoupling (1H as the observed
nuclei) was used during acquisition. Spectra were recorded in the phase-
sensitive mode by using the States-TPPI method with 512 (t1) increments
and 2 K (t2) data points, 64 scans, 16 dummy scans, and D1=1.5 s. The
Kinetic study of the hydrolytic reaction: An approximately 20 mm sample
was prepared in D2O and the pD of the sample was adjusted by the addi-
tion of a small volume (a few mL) of HCl or NaOH solutions in D2O
(0.1m/0.5m). The progress of the reaction was monitored by using a
Bruker Avance II 500 MHz NMR spectrometer and the kinetic cycle was
recorded by using the Bruker Icon NMR automation program. The cycle
began by recording 1H (NS=16) and 31P (NS=128) NMR spectra at
258C. Each round of the kinetic cycle consisted of a heating step at 40 or
708C for 25 min inside the NMR instrument, the sample was then cooled
to 258C and 1H and 31P spectra were recorded. The peaks that appeared
in the 1H and 31P NMR spectra were integrated and the extent of reac-
tion (%) was calculated from the peak integral. Rate constants were de-
termined by plotting concentration (%) versus time (min) and fitting the
data points to an exponential decay. The reproducibility of the kinetic ex-
periments was verified by repeating the hydrolysis study for both com-
pounds 1 and 21 twice. No significant discrepancy in the rate constants
was observed. The average rate constant and SEM were calculated by
1
mixing time was set to 0.8 s. Sweep widths were 10 and 40 ppm in the H
and 31P dimensions, respectively. The data sets were Fourier transformed
by using a shifted sine-bell window function in both dimensions and pro-
cessed into a 2 Kꢁ0.5 K matrix without prior zero filling.
1
considering both the disappearance of the H and 31P peaks of the parent
aaNP and the appearance of the product peaks in the 1H and 31P spectra
wherever feasible. The products formed after hydrolysis were identified
by 1H–31P TOCSY,[39] acquired with 64 scans, 2048 data points in the
proton dimension (t2) and 128 increments in the phosphorus dimension
(t1) over a sweep width of 10 and 40 ppm, respectively. The residual
HOD peak was suppressed by low-power on resonance presaturation.
Acknowledgements
This work was financed by KU Leuven grants GOA, IDO. We are in-
debted to Prof. Jef Rozenski for providing MS and HRMS data and Prof.
Roger Busson for helpful discussions. We thank Luc Baudemprez for
NMR spectroscopy technical assistance, Dr. Elisabetta Groaz and Jolien
Claessens for technical support, and Chantal Biernaux for editorial help.
pKa determination by NMR titration: Samples were dissolved in D2O
(0.5 mL) and the sampleꢄs pD was adjusted directly inside the NMR
tube. The pD was measured at room temperature by using the Hamilton
1
SPINTRODE electrode. H and 31P NMR spectra were recorded at 258C
[1] a) C. Meier, Synlett 1998, 233–242; b) C. R. Wagner, V. V. Iyer, E. J.
McIntee, Med. Res. Rev. 2000, 20, 417–451; c) E. De Clercq, Nat.
Rev. Drug Discov. 2002, 1, 13–25; d) E. De Clercq, J. Clin. Virol.
2004, 30, 115–133; e) Y. Mehellou, J. Balzarini, C. McGuigan,
ChemMedChem 2009, 4, 1779–1791.
for each pH point by using 32 and 64 scans, respectively. The pKa of the
phosphoramidate nitrogen (15N-labeled 1 and 21) was determined by
monitoring the 15N chemical shift by using 1H–15N HSQC (long-range)
experiments on a Bruker Avance II 600 MHz spectrometer. All 1H–15N
HSQC spectra in the pH-metric titration were recorded with sensitivity
[2] C. L. Freel Meyers, R. F. Borch, J. Med. Chem. 2000, 43, 4319–4327.
Chem. Eur. J. 2012, 18, 857 – 868
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
867

P. Herdewijn et al.
[3] C. R. Wagner, S. L. Chang, G. W. Griesgraber, H. Song, E. J. Mc-
Intee, C. L. Zimmerman, Nucleosides Nucleotides 1999, 18, 913–919.
[29] Z. Kazimierczuk, J. A. Vilpo, F. Seela, Helv. Chim. Acta 1992, 75,
2289–2297.
ACHTUNGTRENNUNG
[4] a) R. N. Hunston, M. Jehangir, A. S. Jones, R. T. Walker, Tetrahe-
dron 1980, 36, 2337–2340; b) A. S. Jones, C. McGuigan, R. T.
Walker, J. Balzarini, E. de Clercq, J. Chem. Soc. Perkin Trans. 1
1984, 1471–1474; c) A. S. Jones, C. McGuigan, R. T. Walker, J.
Chem. Soc. Perkin Trans. 1 1985, 199–202.
[5] a) K. G. Devine, C. McGuigan, T. J. OꢄConnor, S. R. Nicholls, D.
Kinchington, AIDS 1990, 4, 371–373; b) D. Cahard, C. McGuigan,
J. Balzarini, Mini Rev. Med. Chem. 2004, 4, 371–381.
[6] a) T. W. Abraham, C. R. Wagner, Nucleosides Nucleotides 1994, 13,
1891–1903; b) E. J. McIntee, R. P. Remmel, R. F. Schinazi, T. W.
Abraham, C. R. Wagner, J. Med. Chem. 1997, 40, 3323–3331.
[7] T. W. Abraham, T. I. Kalman, E. J. McIntee, C. R. Wagner, J. Med.
Chem. 1996, 39, 4569–4575.
[30] a) P. Haake, T. Koizumi, Tetrahedron Lett. 1970, 11, 4849–4850;
b) K. E. DeBruin, A. G. Padilla, D. M. Johnson, Tetrahedron Lett.
1971, 12, 4279–4280; c) T. Koizumi, P. Haake, J. Am. Chem. Soc.
1973, 95, 8073–8079; d) M. P. Gamcsik, S. M. Ludeman, E. M. Shul-
manroskes, I. J. Mclennan, M. E. Colvin, O. M. Colvin, J. Med.
Chem. 1993, 36, 3636–3645.
[31] M. Ora, K. Mattila, T. Lonnberg, M. Oivanen, H. Lonnberg, J. Am.
Chem. Soc. 2002, 124, 14364–14372.
[32] S. J. Benkovic, E. J. Sampson, J. Am. Chem. Soc. 1971, 93, 4009–
4016.
[33] M. Ora, M. Murtola, S. Aho, M. Oivanen, Org. Biomol. Chem. 2004,
2, 593–600.
[34] S. Michielssens, N. Tien Trung, M. Froeyen, P. Herdewijn, M. Tho
Nguyen, A. Ceulemans, Phys. Chem. Chem. Phys. 2009, 11, 7274–
7285.
[35] W. P. Jencks, M. Gilchrist, J. Am. Chem. Soc. 1964, 86, 1410–1417.
[36] Y. Kajihara, S. Nishigaki, D. Hanzawa, G. Nakanishi, R. Okamoto,
N. Yamamoto, Chem. Eur. J. 2011, 17, 7645–7655.
[8] a) D. Saboulard, L. Naesens, D. Cahard, A. Salgado, R. Pathirana, S.
Velazquez, C. McGuigan, E. De Clercq, J. Balzarini, Mol. Pharma-
col. 1999, 56, 693–704; b) T. F. Chou, J. Baraniak, R. Kaczmarek, X.
Zhou, J. Cheng, B. Ghosh, C. R. Wagner, Mol. Pharm. 2007, 4, 208–
217.
[9] a) S. Chang, G. W. Griesgraber, P. J. Southern, C. R. Wagner, J. Med.
Chem. 2001, 44, 223–231; b) J. Balzarini, Pharm. World Sci. 1994,
16, 113–126.
[37] F. H. Westheimer, Acc. Chem. Res. 1968, 1, 70–78.
[38] S. S. Wijmenga, B. N. M. van Buuren, Prog. Nucl. Magn. Reson.
Spectrosc. 1998, 32, 287–387.
[10] O. Adelfinskaya, P. Herdewijn, Angew. Chem. 2007, 119, 4434–
4436; Angew. Chem. Int. Ed. 2007, 46, 4356–4358.
[11] a) O. Adelfinskaya, M. Terrazas, M. Froeyen, P. Marliere, K. Nauwe-
laerts, P. Herdewijn, Nucleic Acids Res. 2007, 35, 5060–5072; b) A.
Giraut, N. Dyubankova, X. P. Song, P. Herdewijn, Chembiochem
2009, 10, 2246–2252.
[12] A. Giraut, X. P. Song, M. Froeyen, P. Marliere, P. Herdewijn, Nucle-
ic Acids Res. 2010, 38, 2541–2550.
[13] M. Ora, J. Ojanpera, H. Lohnberg, Chem. Eur. J. 2007, 13, 8591–
8599.
[14] A. Leisvuori, Y. Aiba, T. Lonnberg, P. Poijarvi-Virta, L. Blatt, L.
Beigelman, H. Lonnberg, Org. Biomol. Chem. 2010, 8, 2131–2141.
[15] T. Lonnberg, M. Ora, H. Lonnberg, Mini-Rev. Org. Chem. 2010, 7,
33–43.
[16] J. G. Moffatt, H. G. Khorana, J. Am. Chem. Soc. 1961, 83, 649–658..
[17] I. Dhimitruka, J. SantaLucia, Jr., Synlett 2004, 0335–0337.
[18] D. D. Boehr, A. R. Farley, F. J. LaRonde, T. R. Murdock, G. D.
Wright, J. R. Cox, Biochemistry 2005, 44, 12445–12453.
[19] M. Yoshikawa, T. Kato, T. Takenish, Bull. Chem. Soc. Jpn. 1969, 42,
3505–3508.
[39] G. W. Kellogg, J. Magn. Reson. 1992, 98, 176–182.
[40] P. K. Glasoe, F. A. Long, J. Phys. Chem. 1960, 64, 188–190.
[41] S. Chatterjee, W. Pathmasiri, O. Plashkevych, D. Honcharenko, O. P.
Varghese, M. Maiti, J. Chattopadhyaya, Org. Biomol. Chem. 2006, 4,
1675–1686.
[42] Gaussian 09, Revision A.1, M. J. Frisch, G. W. Trucks, H. B. Schlegel,
G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V.
Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X.
Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Son-
nenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,
M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven,
J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J.
Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J.
Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar,
J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A.
Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, ꢅ.
Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian,
Inc., Wallingford CT, 2009.
[20] N. Minakawa, Y. Kawano, S. Murata, N. Inoue, A. Matsuda, Chem-
biochem 2008, 9, 464–470.
[21] F. Seela, B. Westermann, U. Bindig, J. Chem. Soc. Perkin Trans. 1
1988, 697–702.
[22] F. Seela, H. Rosemeyer, S. Fischer, Helv. Chim. Acta 1990, 73, 1602–
1611.
[43] W. Humphrey, A. Dalke, K. Schulten, J. Mol. Graph. 1996, 14, 33–
38.
[44] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[45] N. Godbout, D. R. Salahub, J. Andzelm, E. Wimmer, Can. J. Chem.
1992, 70, 560–571.
[23] A. W. Garrison, C. E. Boozer, J. Am. Chem. Soc. 1968, 90, 3486–
3494.
[24] L. M. Qiang, S. X. Cao, X. Y. Zhao, R. Y. Liu, J. H. Liu, J. S. Lu,
Y. F. Zhao, Chin. J. Chem. 2007, 25, 1559–1562.
[25] J. Rahil, P. Haake, J. Am. Chem. Soc. 1981, 103, 1723–1734.
[26] Z. Z. Chen, B. Tan, Y. M. Li, Y. F. Zhao, Y. F. Tong, J. F. Wang, J.
Org. Chem. 2003, 68, 4052–4058.
[46] a) B. Mennucci, J. Tomasi, J. Chem. Phys. 1997, 106, 5151–5158;
b) S. Miertus, E. Scrocco, J. Tomasi, Chem. Phys. 1981, 55, 117–129.
[47] M. P. Andersson, P. Uvdal, J. Phys. Chem. A 2005, 109, 2937–2941.
[48] V. Sklenꢆr, H. Miyashiro, G. Zon, H. T. Miles, A. Bax, Febs Lett.
1986, 208, 94–98.
[49] A. Bax, D. G. Davis, J. Magn. Reson. 1985, 63, 207–213.
[50] C. Yu, G. C. Levy, J. Am. Chem. Soc. 1984, 106, 6533–6537.
[27] B. Tan, M. C. Lee, M. Cui, T. Liu, Z. Z. Chen, Y. M. Li, Y. Ju, Y. F.
Zhao, K. X. Chen, H. L. Jiang, J. Mol. Struct. THEOCHEM 2004,
672, 51–60.
[28] Y.-M. Li, D.-q. Zhang, H.-w. Zhang, G.-J. Ji, Y.-f. Zhao, Bioorg.
Chem. 1992, 20, 285–295.
Received: July 25, 2011
Revised: November 4, 2011
Published online: December 15, 2011
868
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 857 – 868