Preparation of benzylphosphonates via a palladium(0)-catalyzed cross-coupling of H-phosphonate diesters with benzyl halides. Synthetic and mechanistic studies

Article abstract of DOI:10.1039/b9nj00585d

We have developed a new, efficient method for the synthesis of benzylphosphonate and benzylphosphonothioate diesters via a palladium(0)- catalyzed cross-coupling reaction between benzyl halides and H-phosphonate or H-phosphonothioate diesters, using Pd₂(dba)₃(CHCl ₃) as a palladium source and Xantphos as a supporting ligand. Some mechanistic aspects of these reactions were investigated using ³¹P NMR spectroscopy.

Full text of DOI:10.1039/b9nj00585d

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/njc | New Journal of Chemistry
Preparation of benzylphosphonates via a palladium(0)-catalyzed
cross-coupling of H-phosphonate diesters with benzyl halides.
Synthetic and mechanistic studieswzy
a
a
a
ab
´
Gaston Laven, Marcin Kalek, Martina Jezowska and Jacek Stawinski*
Received (in Montpellier, France) 21st October 2009, Accepted 11th December 2009
First published as an Advance Article on the web 12th February 2010
DOI: 10.1039/b9nj00585d
We have developed a new, efficient method for the synthesis of benzylphosphonate and
benzylphosphonothioate diesters via a palladium(0)-catalyzed cross-coupling reaction between
benzyl halides and H-phosphonate or H-phosphonothioate diesters, using Pd2(dba)3(CHCl3)
as a palladium source and Xantphos as a supporting ligand. Some mechanistic aspects of
these reactions were investigated using 31P NMR spectroscopy.
H-phosphonate diesters catalyzed by Pd(0). Despite the vast
Introduction
interest in the area of palladium-catalyzed cross-coupling
using heteroatom nucleophiles,13 only a few examples of a
P–C(sp3) bond formation have been reported so far.14
Analogues of naturally occurring phosphorus-containing
compounds possessing P–C bonds (C-phosphonates and
C-phosphinates) exhibit enhanced stability to hydrolysis and
enzymatic digestion, and therefore, have attracted attention as
potential pesticides and therapeutics.1 A prominent group
among these compounds are benzylphosphonates, that show
a vast array of biological activity,2,3 and are frequently used
as prodrugs.3 In the antisense/antigene research area, the
recently synthesized benzylphosphonate-modified oligo-
nucleotides turned out to be superior inhibitors against the
hepatitis-C virus to the corresponding phosphorothioates or
methylphosphonates.4
For the formation of the P–C(sp3) bond, the most common
approaches are probably those involving the Michaelis–
Arbuzov5 and Michaelis–Becker6 reactions. Although quite
general in scope, these methods usually require harsh reaction
conditions7 that are unsuitable for the synthesis of biologically
interesting, complex molecules.8 Due to this, for the prepara-
tion of oligonucleoside benzylphosphonates, the only method
of practical value makes use of benzylphosphonous dichloride
as a precursor, but it is lengthy and, because of a facile ligand
exchange at the phosphorus centre, complex reaction mixtures
are usually formed.9
Results and discussion
Synthesis of benzylphosphonate and benzylphosphonothioate
diesters
In a preliminary account of this work,15 we reported on the
preparation of benzylphosphonates using a palladium(0)-
catalyzed cross-coupling of H-phosphonate diesters and
benzyl halides. By screening various ligands, we found that
only large bite angle phosphines efficiently promoted the
formation of benzylphosphonates, and the best catalytic
system consisted of Pd(OAc)2/Xantphos and N,N-diisopropyl-
ethylamine (DIPEA) as a base.
Although the Pd(OAc)2/Xantphos system worked well,
further studies showed that the results were not completely
reproducible (variations in yields), and we traced this back to
the reduction step of the Pd(II) precursor.16 Since the reduction
has to be carried out in the presence of a controlled amount of
water,15,16 partial hydrolysis of the substrates may occur due
to the high lability of H-phosphonate diesters under basic
conditions.
Recent studies in our10,11 and other12 laboratories have
shown that the application of transition metal chemistry
enables a mild synthesis of aryl- and vinylphosphonates, even
in the case of sensitive nucleotide analogues.
We envisaged that the synthesis of benzylphosphonates
would be simpler to conduct if the reduction step could be
omitted by using Pd(0) as a palladium source. Since Pd(PPh3)4
was not suitable for this purpose,15 we tried tris(dibenzylide-
Herein, we describe our synthetic and mechanistic investi-
gations aimed at broadening the scope of methods for P–C
bond formation by cross-coupling of benzyl halides with
neacetone)dipalladium(0)(chloroform)
[Pd2(dba)3(CHCl3)]
and Xantphos as a supporting ligand. We found that cross-
coupling of diethyl H-phosphonate with benzyl bromide
catalyzed by Pd2(dba)3(CHCl3) and Xantphos in THF (reflux)
showed the same efficiency and kinetics as that with Pd(OAc)2/
Xantphos (see the ESIy), and thus, this catalyst system (Fig. 1)
has been adopted for the present studies.17
a Department of Organic Chemistry, Arrhenius Laboratory,
Stockholm University, S-106 91 Stockholm, Sweden
b Institute of Bioorganic Chemistry, Polish Academy of Sciences,
Noskowskiego 12/14, 61-704 Poznan, Poland.
E-mail: js@organ.su.se; Fax: +46 8 15 49 08
w This article is part of a themed issue on Biophosphates.
z This paper is dedicated to Professor Wojciech J. Stec on the occasion
of his 70th birthday.
y Electronic supplementary information (ESI) available: Further
experimental and characterisation details. See DOI: 10.1039/b9nj00585d
In Table 1, the results of cross-coupling of various benzyl
halide derivatives with diethyl H-phosphonate are summarised.
The optimal reaction conditions consisted of using 1.2 equiv.
of a benzyl halide relative to diethyl H-phosphonate, 1.2 equiv.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
New J. Chem., 2010, 34, 967–975 | 967

View Article Online
Table 1 Synthesis of diethyl benzylphosphonate derivatives
Fig. 1 A catalyst system for the synthesis of benzylphosphonates.
R–X
Conv. (%)a Yield (%)b
Entry
of N,N-diisopropylethylamine (DIPEA) in the presence
of 5% of the Pd catalyst, and 5% of the supporting ligand
in THF, under reflux. These, in most cases, secured a complete
conversion (31P NMR spectroscopy) into the corresponding
benzylphosphonates within 4 h, and only for the 5-chloro-
methylfurane derivative (entry 10), the reaction time had to be
extended to 6 h with a simultaneous increasing of the catalyst
load to 8% Pd. Consistent with our previous findings,
no visible differences in reactivity have been observed between
Cl vs. Br derivatives, and electronic features of the substituents
in the aromatic ring did not appreciably affect the rates of the
cross-coupling reactions investigated.
1
2
3
4
5
1
2
3
4
5
99
99
99
99
99
89
88
87
96
90
Compared with our previously developed reaction condi-
tions [Pd(OAc)2/Xantphos],15 using Pd2(dba)3(CHCl3)
permitted us, in most cases, to lower the catalyst load from
10 to 5 mol% of Pd, and to reduce the amount of the
supporting ligand from 20 to 5 mol%. Also, the overall
reproducibility of the cross-coupling reactions in terms of
yields was significantly improved with this new catalyst
system.
6
6
7
8
99
96
99
86
92
92
7
To demonstrate the generality of the developed synthetic
protocol, various H-phosphonate diesters with diverse
structural features were subjected to a cross-coupling reaction
with benzyl chloride. The results are summarised in Table 2.
As was observed for diethyl H-phosphonate, cross-coupling
of asymmetrically substituted dialkyl, alkyl glyceryl and alkyl
cholesteryl H-phosphonate diesters with benzyl chloride
(Table 2, entries 1–3) occurred smoothly, affording the corres-
ponding benzylphosphonates 11–13 in good isolated yields.
Previously, we showed that benzylation of dithymidyl
H-phosphonate diesters occurred stereospecifically with, most
likely, retention of configuration at the phosphorus center.15
To address a possible issue of coordination of palladium by
functional groups present in complex organic molecules, we
next investigated benzylation of dinucleoside H-phosphonate
diesters bearing four standard heterocyclic bases present in
DNA (adenine, cytosine, guanine and thymine derivatives;
Table 2, entries 4–7). No deterioration of the catalyst system
could be observed, as judged from similar kinetics of the cross-
coupling reaction as those for simple diethyl H-phosphonate.
In all instances, the benzylation of dinucleoside H-phosphonates
was uneventful, and the corresponding benzylphosphonates
14–17 were obtained as a ca. 1 : 1 mixture of the diastereomers
in high yields.
8c
9c
9
99
99
90
10d
10
99d
i = 2.5 mol% Pd2(dba)3(CHCl3), 5 mol% Xantphos, 1.2 equiv.
N,N-diisopropylethylamine, THF, 4 h under reflux.a Determined by
31P NMR spectroscopy. Isolated yield. 2.5 equiv. of the base was
b
c
d
used. 4 mol% Pd2(dba)3(CHCl3), 8 mol% Xantphos, 6 h.
cross-coupling with benzyl chloride. In the absence of a
Pd(0) catalyst, no diethyl benzylphosphonothioate formation
could be observed upon reflux overnight of equimolar
amounts of diethyl H-phosphonothioate and benzyl chloride
in THF, in the presence of N,N-diisopropylethylamine
(1.3 equiv.). However, using the reaction conditions deve-
loped for the cross-coupling of H-phosphonate diesters
[Pd2(dba)3(CHCl3) + Xantphos], clean formation of the
corresponding benzylphosphonothioate diesters 18 and 19
(Table 3, entries 1 and 3, respectively), was observed.
Efficient cross-coupling of H-phosphonate diesters with
benzyl halides prompted us to investigate P-benzylation of
H-phosphonothioate diesters as a possible entry to new
phosphate analogues, namely benzylphosphonothioates
(reaction scheme in Table 3).
A striking feature of these cross-coupling reactions was,
however, a significantly longer reaction time than that for
H-phosphonate derivatives (18–19 h vs. 4 h), even after
increasing the catalyst load to 10% Pd (Table 3, entries 1
and 3). In contrast to these, when benzyl bromide was used as
As representative compounds, diethyl and nucleoside ethyl
H-phosphonothioate diesters (Table 3) were subjected to
a
reactant, the cross-coupling to the corresponding
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
968 | New J. Chem., 2010, 34, 967–975

View Article Online
Table 2 Synthesis of diverse benzylphosphonate diesters
Entry
1
Conv. (%)a
Yield (%)b
11
12
99
88
2
3
99
99
74
88
13
14
4
5
99
99
88
85
15
6
7
16
17
99
99
80
81
R = t-butyldiphenylsilyl; R1 = 4,40-dimethyoxytrityl; Thy = thymin-1-yl; CytBz = N4-benzoylcytosin-1-yl; Guaibu = N2-isobutyrylguanin-9-yl;
AdeBz = N6-benzoyladenin-9-yl. i = 2.5 mol% Pd2(dba)3(CHCl3), 5 mol% Xantphos, 1.2 equiv. N,N-diisopropylethylamine, THF, 4 h under
b
reflux.a Determined by 31P NMR spectroscopy. Isolated yield.
benzylphosphonothioates 18 and 19 became very fast and the
reactions went to completion within 2 h (Table 3, entries 2
and 4).
P–H bond in H-phosphonothioate diesters is more acidic than
that of the corresponding H-phosphonate derivatives,18 one
should, in principle, expect a faster reaction for H-phosphono-
thioate diesters, due to more efficient generation of a
phosphorus nucleophile under the reaction conditions. It
seems that the reactivity pattern of H-phosphonothioate
Previously, we observed similar differences in reactivity
between H-phosphonate and H-phosphonothioate diesters in
Pd-catalyzed cross-coupling arylation reactions.11 Since the
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
New J. Chem., 2010, 34, 967–975 | 969

View Article Online
Table 3 Cross-coupling of H-phosphonothioate diesters with benzyl halides
Entry
1
R–X
Time/h
19
Conv. (%)a
Yield (%)b
18
18
99
88
2
3
2
99
99
79
78
19
19
18
4
2
99
83
R = t-butyldiphenylsilyl; Thy = thymin-1-yl. i = Entries 1 and 3: 2.5 mol% Pd2(dba)3(CHCl3), 5 mol% Xantphos; entries 2 and 4: 5 mol%
b
Pd2(dba)3(CHCl3), 10 mol% Xantphos; 1.2 equiv. N,N-diisopropylethylamine, THF, under reflux.a Determined by 31P NMR spectroscopy.
Isolated yield.
diesters with benzyl halides follows that of aryl halides,11 i.e.
the catalytic cycle is most efficient for electrophilic substrates
that undergo rapid oxidative addition to Pd(0) complexes
(also, vide infra).
To gain some insight into mechanistic aspects of these
cross-coupling reactions, 31P NMR experiments were carried
out (Fig.
2 and 3). To this end, the catalyst pre-
cursor, Pd2(dba)3(CHCl3), was dissolved in THF at 40 1C,
and an equimolar amount of Xantphos was added. The 31P
NMR spectrum revealed the rapid formation of a Pd(0)–
Xantphos complex, as was apparent from the appearance of
two resonances at ca. 8.5 and 10.5 ppm (Fig. 2a). Upon
addition of benzyl bromide (1 equiv.), these signals disap-
peared within 5 min, and a new resonance at ca. 11.5 ppm,
assigned to an oxidative addition product of benzyl bromide
to the Pd(0)–Xantphos complex [a benzylpalladium(II)
complex], emerged (Fig. 2b). When this reaction mixture was
then treated with diethyl H-phosphonate and N,N-diiso-
propylethylamine (5 equiv. of both reactants), the intensity
of the signal at ca. 11.5 ppm was gradually decreasing in
time with the simultaneous formation of the elimi-
nation product, diethyl benzylphosphonate (signal at ca.
26 ppm; Fig. 2c). After 1 h at 40 1C, ca. 24% of the
benzylphosphonate product was formed, but heating for
additional 10 min at 60 1C brought the reaction to completion
(Fig. 2d and e).
Mechanistic aspects of the Pd-catalysed benzylphosphonate
synthesis
Benzyl bromides are usually much more reactive in oxidative
addition to Pd(0) complexes than the corresponding
chlorides,19 and thus, the similar reaction times observed for
the Pd-catalyzed benzylation of H-phosphonate diesters with
various benzyl halides suggested that oxidative addition was
probably not the rate determining step of the catalytic cycle.
The slower reaction of electron-deficient benzyl halides vs.
electron-rich ones15 also lends support for this hypothesis.20
Since Stockland et al.21 reported that reductive elimination
from metal phosphonate complexes is significantly accelerated
in the presence of large bite angle phosphine ligands, the high
efficiency of Xantphos (and to a lesser extent, other bidentate
ligands15) in the investigated reactions pointed to the kinetic
importance of later steps in the catalytic cycle.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
970 | New J. Chem., 2010, 34, 967–975

View Article Online
An analogous 31P NMR experiment was also carried out
with benzyl chloride, instead of benzyl bromide, and
fragments of the corresponding spectra are shown in Fig. 3.
It is clearly visible that benzyl chloride underwent significantly
slower oxidative addition than benzyl bromide, and after
30 min at 40 1C, only traces of a benzylpalladium(II) complex
(signal at ca. 11.5 ppm) could be detected (Fig. 3b).
Upon heating of the reaction mixture, the intensity of the
signal at ca. 11.5 ppm was gradually increasing, with a
simultaneous broadening of the signals due to the Pd(0)–
Xantphos complex. When the benzylpalladium(II) complex
became the major product (Fig. 3d), diethyl H-phosphonate
and N,N-diisopropylethylamine (5 equiv. of both reactants)
were added to the reaction mixture. This, after 1 h at 40 1C,
resulted in 52% conversion of the benzylpalladium(II) complex
into diethyl benzylphosphonate (signal at ca. 26 ppm; Fig. 3e
and f).
By comparing the spectra in Fig. 2b and 3b, it became
apparent that oxidative addition of benzyl bromide to the
Pd(0)–Xantphos complex was significantly faster than that of
benzyl chloride, but the added diethyl H-phosphonate reacted
faster with the benzylpalladium(II) complex produced from
benzyl chloride than with that from benzyl bromide (Fig. 2d
vs. Fig. 3e). Thus, the observed similar reaction times for
cross-coupling of benzyl chlorides and benzyl bromides with
H-phosphonate diesters were apparently due to the fact that
a faster ligand exchange/reductive elimination step in the
instance of benzyl chloride derivatives made up for the slow
oxidative addition step. In light of this, it seemed likely that
the turnover limiting step for the Pd(0)-catalyzed cross-
coupling of benzyl chlorides with H-phosphonates was
apparently the oxidative addition step, while for the corres-
ponding bromides, it was the reaction of H-phosphonate with
a benzylpalladium(II) complex.
Fig. 2 31P NMR spectra showing the oxidative addition of equimolar
amounts of benzyl bromide to Pd2(dba)3(CHCl3)/Xantphos (1 : 2 mol
ratio), followed by the reaction of added diethyl H-phosphonate
(5 equiv.) and N,N-diisopropylethylamine (5 equiv.).
In this context, the structural features of benzylpalladium(II)
complexes, produced in the oxidative addition step, were also
pertinent. Although for benzylpalladium(II) complexes a
structure analogous to that of arylpalladium(II) complexes
was initially proposed22 (Z1-type complexes), recent crystallo-
graphic data provided evidence for an Z3-coordination of the
benzyl group.23,24 In solution, however, a rapid Z3–Z1–Z3
isomerization was proposed for such complexes, with the
dominant form being Z3-benzylpalladium(II).24,25
coordination of the PQO function of an H-phosphonate
diester to Z1-benzylpalladium(II) (A) on the way to benzyl-
(phosphonate)palladium(II) complex B, rather than a nucleo-
philic attack of the phosphite anion on Z3-benzylpalladium(II)
complex to afford the final product (see Scheme 1). Although
due to rapid Z3–Z1-equilibria in benzylpalladium(II)
complexes28 it was difficult to find out which form of these
complexes reacted with H-phosphonate diesters (only one
signal observed in the 31P NMR spectra, see Fig. 2 and 3),
the higher reactivity of benzylpalladium(II) complexes
generated from benzyl chlorides vs. benzyl bromides pointed
to the kinetic importance of Z1-benzylpalladium(II) forms.
This would be in line with the order of reactivity of aryl-
(halide)palladium(II) complexes, for which higher rates of
ligand substitution were observed for chloride vs. bromide
derivatives.31
Since Z3-benzylpalladium(II) complexes are structurally
related to Z3-allylpalladium complexes, it was assumed that
nucleophilic substitution with nitrogen, oxygen or carbon
nucleophiles occurs analogously to the Tsuji–Trost reaction,26
i.e. via an Z3-benzylpalladium intermediate.24,25,27,28 Although
this probably is the case for heteroatom and carbon nucleo-
philes with a negative charge or a lone electron pair on
the nucleophilic atom, the situation can be different for
H-phosphonate nucleophiles. In this instance, the nucleophilic
species, a phosphite anion (generated in situ by the added
base), is usually present only in tiny amounts29 and, as we
demonstrated for arylpalladium(II) complexes, the rate
determining step for ligand substitution is coordination of
the PQO group to the Pd(II) complex.30
In light of the above, we tentatively formulated a catalytic
cycle for the Pd-catalyzed cross-coupling of benzyl halides
with H-phosphonate diesters as depicted in Scheme 1.
The cycle is analogous to that generally assumed for the
synthesis of arylphosphonates10,30–32 and involves the
We suggest that a similar mechanism also operates for
benzylpalladium(II) complexes, and the reaction involves
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
New J. Chem., 2010, 34, 967–975 | 971

View Article Online
Scheme
1
A
catalytic cycle for palladium(0)-mediated benzyl-
phosphonate formation.
Fig. 2, but using PPh3 instead of Xantphos as the supporting
ligand (Fig. 4). When benzyl bromide was added to the in situ
generated catalyst system [from Pd2(dba)3(CHCl3) and PPh3,
Fig. 4a], the slow formation of a benzylpalladium(II) complex
(signal at ca. 22 ppm), accompanied by benzylation of the
phosphine ligand (benzyltriphenylphosphonium salt, signal at
ca. 23 ppm), was observed (Fig. 4b). The reaction was almost
complete after ca. 2 h (Fig. 4c), and the addition of diethyl
H-phosphonate and N,N-diisopropylethylamine initiated
formation of the product, diethyl benzylphosphonate (signal
at ca. 25.5 ppm, Fig. 4d). It took 2 h at 40 1C, and an
additional 10 min at 60 1C, to bring this reaction to comple-
tion. During this time, the signal due to benzylated ligand
(ca. 23 ppm) remained unchanged (Fig. 4e and f).
Fig. 3 31P NMR spectra showing the oxidative addition of equimolar
amounts of benzyl chloride to Pd2(dba)3(CHCl3)/Xantphos (1 : 2 mol
ratio), followed by the reaction of added diethyl H-phosphonate
(5 equiv.) and N,N-diisopropylethylamine (5 equiv.).
generation of benzylpalladium(II) complex A, followed by a
ligand exchange step to produce benzyl(phosphonate)-
palladium(II) complex B, from which a reductive elimination
of the product benzylphosphonate diesters occurs. Although
in the case of H-phosphonate diesters, it is most likely the
Z1-benzylpalladium(II) species A that undergoes a ligand
exchange to produce intermediate B, for more reactive
phosphorus nucleophiles (e.g. alkali metal phosphite salts),
the nucleophilic attack may occur directly on the
benzylic carbon of the Z3-benzylpalladium(II) complex (see
Scheme 1).
As is apparent from these experiments, oxidative addition of
benzyl bromide to the Pd(0) complex in the presence of PPh3
was significantly slower than that in the presence of Xantphos.
Also, formation of the product from the benzylpalladium(II)
complex was much slower in the presence of PPh3 than in the
reaction with Xantphos. These pointed to a beneficial role of
Xantphos as the supporting ligand in the benzylphosphonate
synthesis, that not only accelerated the reductive elimination,
but also facilitated the oxidative addition step.
A similar catalytic cycle probably also operates for the
cross-coupling of H-phosphonothioate diesters, as judged
from the 31P NMR experiments on the formation of diethyl
benzylphosphonothioate diester from diethyl H-phosphono-
thioate and benzyl bromide (see the ESIy).
As a final part of these investigations, we wanted to find out
why triphenylphosphine (PPh3) was a highly inefficient ligand
in the Pd-catalyzed synthesis of benzylphosphonates using
benzyl halides and H-phosphonate diesters.15 To this end,
we carried out a 31P NMR experiment, analogous to that in
Concerning a competing benzylation of PPh3 by benzyl
bromide, since this reaction occurred to a significant extent
with a stoichiometric ratio of reactants, one can anticipate
that under catalytic reaction conditions, this may completely
inactivate the supporting ligand. Although using benzyl
chloride would lessen this problem, the oxidative addition of
benzyl chloride to Pd(0) in the presence of PPh3 was found
to be very sluggish, and no benzylpalladium(II) formation
could be detected in THF at 40 1C after 2.5 h (results not
shown).
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
972 | New J. Chem., 2010, 34, 967–975

View Article Online
wide range of benzyl derivatives and H-phosphonate diesters,
and is also applicable to the synthesis of benzyl-
phosphonothioate analogues of phosphate esters. The 31P NMR
studies revealed that for a cross-coupling involving benzyl
bromides, the turnover-limiting step of the catalytic cycle is
probably the ligand exchange/reductive elimination step,
while for less reactive benzyl chlorides, it is the oxidative
addition. Xantphos, as a supporting ligand, seems to facilitate
both the oxidative addition and the reductive elimination
step.
Since the underlying chemistry of the method is different
from that of the Michaelis–Arbuzov and the Michaelis–Becker
reactions, this new protocol expands the range of synthetic
methods available for the preparation of biologically impor-
tant C-phosphate analogues.
Experimental
General
All reagents were of analytical grade, obtained from
commercial suppliers and were used without further purifica-
tion. Dinucleoside H-phosphonates33 and H-phosphono-
thioate diesters34 were synthesized from the appropriate
H-phosphonate35 and H-phosphonothioate36 monoesters
and alcohols according to standard coupling procedures.
THF was dried using a VAC solvent purifier system. All
reactions were carried out using standard Schlenk techniques.
Column chromatography was performed on silica gel
(DAVISIL, 35–70 micron). 1H, 13C and 31P NMR spectra
were recorded at 400, 100 and 162 MHz, respectively, in
CDCl3 on a Bruker Avance II 400 MHz spectrometer. 1H
NMR spectra were referenced to internal tetramethylsilane
(TMS, 0.00 ppm) or CHCl3 (7.26 ppm), 13C NMR spectra to
the center-line of CDCl3 (77.16 ppm), and 31P NMR spectra to
external 2% H3PO4 in D2O. Coupling constants, J, are
reported in Hz. For integration of the 31P-NMR signals, a
long pulse delay (3 s) and a broadband proton decoupling
during the acquisition time, were used. For characterisation
data and NMR spectra of products in Tables 1–3, see the ESI.y
Fig. 4 31P NMR spectra showing the oxidative addition of equimolar
amounts of benzyl bromide to Pd2(dba)3(CHCl3)/PPh3 (1 : 4 mol
ratio), followed by the reaction of added diethyl H-phosphonate
(5 equiv.) and N,N-diisopropylethylamine (5 equiv.).
In separate experiments, we also checked a possible
benzylation of the Xantphos ligand with benzyl halides
(for details, see the ESIy). With 10 equiv. of benzyl chloride
in THF at 60 1C, no benzylation of Xantphos could be
detected after reaction overnight. In contradiction to this,
under analogous reaction conditions, benzyl bromide effected
monobenzylation of Xantphos with similar kinetics to those of
benzylation of PPh3 (t1/2 = 95 and 100 min, respectively).
Apparently, due to the fast oxidative addition of benzyl
bromide to Pd(0)–Xantphos complexes, no benzylation of
Xantphos could be detected in the experiment depicted in
Fig. 2, in contrast to an analogous experiment with PPh3 in
Fig. 4 that showed significant benzylation of the supporting
ligand.
Syntheses
General procedure for the synthesis of benzylphosphonates
1–9 and 11–13. Pd2(dba)3(CHCl3) (0.025 mmol, 26 mg),
Xantphos (0.050 mmol, 29 mg), solid benzyl halide
(1.2 mmol), solid H-phosphonate diester (1.0 mmol) and
THF (5 mL) were placed in a two-neck flask equipped with
a reflux condenser. The apparatus was filled with nitrogen by
applying two cycles of vacuum/N2. The mixture was heated to
reflux and N,N-diisopropylethylamine (for compounds 1–7
and 10–13: 1.2 mmol, 155 mg, 210 mL; for compounds 8 and
9: 2.2 mmol, 310 mg, 420 mL), liquid benzyl halide (1.2 mmol)
and liquid H-phosphonate diester (1.0 mmol) were added. The
mixture was refluxed for 4 h, concentrated under vacuum, and
the product purified by silica gel chromatography or by
extraction (see the ESIy for details).
Conclusions
We have developed a new, general method for efficient pre-
paration of benzylphosphonate diesters via the Pd(0)-
catalyzed cross-coupling reaction of benzyl chlorides or bromides
with H-phosphonate diesters using Pd2(dba)3(CHCl3)/Xantphos
as a catalyst system. The reaction is rather general, accepts a
Ethyl 5-diethylphosphonomethylfuran-2-carboxylate (10).
The Reaction was carried out similarly to that described for
compounds 1–9 and 11–13, except for the use of 0.04 mmol
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
New J. Chem., 2010, 34, 967–975 | 973

View Article Online
(41.4 mg) Pd2(dba)3(CHCl3), 0.08 mmol Xantphos (46 mg),
and 6 h heating under reflux.
7 Some simple benzylphosphonate derivatives can be prepared under
milder reaction conditions using modified versions of this reaction,
for example, photo-Arbuzov rearrangement (S. Ganapathy,
B. B. V. S. Sekhar, S. M. Cairns, K. Akutagawa and
W. G. Bentrude, J. Am. Chem. Soc., 1999, 121, 2085–2096), Lewis
acid catalyzed reactions (P. Y. Renard, P. Vayron, E. Leclerc,
A. Valleix and C. Mioskowski, Angew. Chem., Int. Ed., 2003, 42,
2389–2392), or silicon-mediated Michaelis–Arbuzov rearrange-
ment (P. Y. Renard, P. Vayron and C. Mioskowski, Org. Lett.,
2003, 5, 1661–1664).
8 D. Witt and J. Rachon, Phosphorus, Sulfur Silicon Relat. Elem.,
1994, 91, 153–164; A. Kers, J. Stawinski, L. Dembkowski and
A. Kraszewski, Tetrahedron, 1997, 53, 12691–12698; R. J. Cohen,
D. L. Fox, J. F. Eubank and R. N. Salvatore, Tetrahedron Lett.,
2003, 44, 8617–8621; J. J. Heynekamp, W. M. Weber,
L. A. Hunsaker, A. M. Gonzales, R. A. Orlando, L. M. Deck
and D. L. Vander Jagt, J. Med. Chem., 2006, 49, 7182–7189;
L. M. Klingensmith, E. R. Strieter, T. E. Barder and
S. L. Buchwald, Organometallics, 2006, 25, 82–91.
General procedure for the synthesis of dinucleoside benzylphos-
phonates 14–17. Pd2(dba)3(CHCl3) (0.0125 mmol, 13 mg),
Xantphos (0.025 mmol, 15 mg), a suitable protected dinucleo-
side H-phosphonate (0.5 mmol) and THF (5 mL) were placed
in a two-neck flask equipped with a reflux condenser. The
apparatus was filled with nitrogen by applying two cycles of
vacuum/N2, and N,N-diisopropylethylamine (0.6 mmol,
78 mg, 105 mL) and benzyl chloride (0.6 mmol, 76 mg,
69 mL) were added. The mixture was refluxed for 4 h,
concentrated under vacuum and purified by silica gel chromato-
graphy using pentane–EtOAc (1 : 1, v/v) with a gradient of
MeOH (2.5–10%).
9 S. Amberg and J. W. Engels, Helv. Chim. Acta, 2002, 85, 2503–2517.
10 T. Johansson and J. Stawinski, Chem. Commun., 2001, 2564–2565.
11 G. Laven and J. Stawinski, Coll. Symposium Series, 2005, 7, 195–199.
´
12 S. Abbas and C. J. Hayes, Synlett, 1999, 1124–1126.
13 D. Prim, J.-M. Campagne, D. Joseph and B. Andrioletti,
Tetrahedron, 2002, 58, 2041–2075.
14 K. Bravo-Altamirano, Z. H. Huang and J. L. Montchamp, Tetra-
hedron, 2005, 61, 6315–6329; L. Coudray and J. L. Montchamp,
Eur. J. Org. Chem., 2008, 4101–4103.
Diethyl benzylphosphonothioate (18). The reaction was
carried out starting from benzyl bromide and diethyl
H-phosphonothioate similarly to the procedure described for
compounds 1–9 and 11–13, except for 2 h heating under reflux.
The product was purified by silica gel chromatography using
pentane–EtOAc (20 : 1, v/v).
50-O-(tert-Butyldiphenylsilyl)thymidin-30-yl ethyl benzyl-
phosphonothioate—mixture of diastereoisomers (19). The
reaction was carried out starting from benzyl bromide and
50-O-(tert-butyldiphenylsilyl)thymidin-30-yl ethyl H-phosphono-
thioate, similarly to the method described for compounds
14–17, except for 3 h heating under reflux. The product was
purified by silica gel chromatography, using a linear gradient
of MeOH (0–10%) in CH2Cl2.
15 G. Laven and J. Stawinski, Synlett, 2009, 225–228.
´
16 C. Amatore, A. Jutand and A. Thuilliez, Organometallics, 2001, 20,
3241–3249.
17 In separate experiments we attempted to react benzyl chloride
with diethyl H-phosphonate in the presence 1.3 equiv. of
N,N-diisopropylethylamine, and benzyl chloride with triethyl
phosphite in the presence of Pd/Xantphos (3%) (THF, under
reflux, overnight). No detectable amounts of diethyl benzyl-
phosphonate formation could be observed under these reaction
conditions.
18 E. S. Petrov, M. I. Terekhova, I. G. Malakhova, E. N. Tsvetkov,
A. I. Shatenshtein and M. I. Kabachnik, Zh. Obsh. Khim., 1979,
49, 2410–2414; A. G. Matveeva, M. I. Terekhova, I. M. Aladzheva,
I. L. Odinets, E. S. Petrov, T. A. Mastryukova and
M. I. Kabachnik, Zh. Obsh. Khim., 1993, 63, 607–610; R. Wallin,
M. Kalek, A. Bartoszewicz, M. Thelin and J. Stawinski,
Phosphorus, Sulfur Silicon Relat. Elem., 2009, 184, 908–916.
19 D. Milstein and J. K. Stille, J. Am. Chem. Soc., 1979, 101,
4992–4998; A. Gillie and J. K. Stille, J. Am. Chem. Soc., 1980,
102, 4933–4941.
Acknowledgements
Financial support from the Swedish Natural Science Research
Council is gratefully acknowledged.
References
20 J. K. Stille and K. S. Y. Lau, Acc. Chem. Res., 1977, 10, 434–442;
C. Amatore and A. Jutand, J. Organomet. Chem., 1999, 576,
254–278.
21 R. A. Stockland, A. M. Levine, M. T. Giovine, I. A. Guzei and
J. C. Cannistra, Organometallics, 2004, 23, 647–656.
22 P. Fitton, J. E. McKeon and B. C. Ream, J. Chem. Soc. D, Chem.
Commun., 1969, 370–371.
1 R. Engel, Chem. Rev., 1977, 77, 349–367; R. Engel, The use of
carbon–phosphorus analogue compounds in the regulation
of biological processes, in Handbook of Organophosphorus
Chemistry, ed. R. Engel, Marcel DekkerNew York, 1992,
pp. 559–600.
2 K. Yoshino, T. Kohno, T. Uno, T. Morita and G. Tsukamoto,
J. Med. Chem., 1986, 29, 820–825; K. Goto, S. Nakamura,
Y. Morioka, M. Kondo, S. Naito and K. Tsutsumi, Chem. Pharm.
Bull., 1996, 44, 547–551.
23 B. Liegault, J.-L. Renaud and C. Bruneau, Chem. Soc. Rev., 2008,
´
37, 290–299.
24 R. Kuwano, Synthesis, 2009, 1049–1061.
25 A. M. Johns, J. W. Tye and J. F. Hartwig, J. Am. Chem. Soc.,
2006, 128, 16010–16011.
26 G. Consiglio and R. M. Waymouth, Chem. Rev., 1989, 89,
257–276; C. G. Frost, J. Howarth and J. M. J. Williams,
Tetrahedron: Asymmetry, 1992, 3, 1089–1122; Y. Tamaru, Eur. J.
Org. Chem., 2005, 2647–2656.
27 J.-Y. Legros, M. Toffano and J.-C. Fiaud, Tetrahedron, 1995, 51,
3235–3246; R. Kuwano and Y. Kondo, Org. Lett., 2004, 6,
3545–3547.
28 A. M. Johns, M. Utsunomiya, C. D. Incarvito and J. F. Hartwig,
J. Am. Chem. Soc., 2006, 128, 1828–1839.
29 J. Stawinski, Some Aspects of H-Phosphonate Chemistry, in
Handbook of Organophosphorus Chemistry, ed. R. Engel, Marcel
Dekker, New York, 1992, pp. 377–434.
30 M. Kalek and J. Stawinski, Organometallics, 2008, 27, 5876–5888.
31 M. Kalek and J. Stawinski, Organometallics, 2007, 26, 5840–5847.
32 C. Amatore and A. Jutand, Acc. Chem. Res., 2000, 33, 314–321.
3 J. P. Krise and V. J. Stella, Adv. Drug Delivery Rev., 1996, 19,
287–310; Q. Dang, Y. Liu, R. M. Rydzewski, B. S. Brown,
E. Robinson, P. D. van Poelje, T. J. Colby and M. D. Erion,
Bioorg. Med. Chem. Lett., 2007, 17, 3412–3416.
4 M. Alt, S. Eisenhardt, M. Serwe, R. Renz, J. W. Engels and
W. H. Caselmann, Eur. J. Clin. Invest., 1999, 29, 868–76;
T. J. Lehmann and J. W. Engels, Bioorg. Med. Chem., 2001, 9,
1827–1835.
5 A. Michaelis and R. Kaehne, Chem. Ber., 1898, 31, 1048–1055;
A. K. Bhattacharya and G. Thyagarajan, Chem. Rev., 1981, 81,
415–430; Methoden der organischen Chemie (Houben-Weyl),
ed. E. Muller, George Thieme Verlag, Stuttgart, 1964, vol. XII/1,
¨
p. 433.
6 A. Michaelis and T. Becker, Chem. Ber., 1897, 30, 1003–1009;
Methoden der organischen Chemie (Houben-Weyl), ed. E. Muller,
¨
George Thieme Verlag, Stuttgart, 1964, vol. XII/1, p. 446;
R. Waschbusch, J. Carran, A. Marinetti and P. Savignac,
¨
Synthesis, 1997, 727–743.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
974 | New J. Chem., 2010, 34, 967–975

View Article Online
33 J. Stawinski and R. Stromberg, Di- and Oligonucleotide Synthesis
¨
34 R. Zain and J. Stawinski, J. Org. Chem., 1996, 61, 6617–6622.
35 J. Jankowska, M. Sobkowski, J. Stawinski and A. Kraszewski,
Tetrahedron Lett., 1994, 35, 3355–3358.
36 J. Stawinski, M. Thelin, E. Westman and R. Zain, J. Org. Chem.,
1990, 55, 3503–3506.
Using H-Phosphonate Chemistry, in Oligonucleotide Synthesis:
Methods and Applications, ed. P. Herdewijn, Humana Press,
Totowa, NJ, 2004, vol. 288, pp. 81–100; T. Johansson and
J. Stawinski, Bioorg. Med. Chem., 2001, 9, 2315–2322.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
New J. Chem., 2010, 34, 967–975 | 975