Reactive intermediates in the H-phosphonate synthesis of oligonucleotides

Article abstract of DOI:10.1039/c2ob07130d

The formation of H-phosphonate diesters is an important step in the synthesis of oligonucleotides. Using diphenylchlorophosphate as the activator for the coupling step is often accompanied by side reactions as a result of self 'capping' and other reactions of the reactive intermediate. In the absence of base, the activation of ethyl H-phosphonate with diphenylchlorophosphate probably occurs through the intermediate formation of bis diethyl pyro-di-H-phosphonate rather than the expected diphenyl ethyl pyro-H-phosphonate. Pyridine acts as a nucleophilic catalyst converting diphenylchlorophosphate to its pyridinium adduct. Several side and unwanted reactions are quantified so that conditions to minimise these can be identified.

Full text of DOI:10.1039/c2ob07130d

View Online / Journal Homepage
Organic &
Biomolecular
Chemistry
This article is part of the
OBC 10th anniversary
themed issue
All articles in this issue will be gathered together
online at
www.rsc.org/OBC10

Dynamic Article Links
Organic &
Biomolecular
Chemistry
Cite this: DOI: 10.1039/c2ob07130d
www.rsc.org/obc
PAPER
Reactive intermediates in the H-phosphonate synthesis of oligonucleotides†‡
Nicholas Powles, John Atherton and Michael I. Page*
Received 19th December 2011, Accepted 24th April 2012
DOI: 10.1039/c2ob07130d
The formation of H-phosphonate diesters is an important step in the synthesis of oligonucleotides. Using
diphenylchlorophosphate as the activator for the coupling step is often accompanied by side reactions as a
result of self ‘capping’ and other reactions of the reactive intermediate. In the absence of base, the
activation of ethyl H-phosphonate with diphenylchlorophosphate probably occurs through the
intermediate formation of bis diethyl pyro-di-H-phosphonate rather than the expected diphenyl ethyl
pyro-H-phosphonate. Pyridine acts as a nucleophilic catalyst converting diphenylchlorophosphate to its
pyridinium adduct. Several side and unwanted reactions are quantified so that conditions to minimise
these can be identified.
In general, coupling reactions involving phosphorus(III)
are much more reactive than those with phosphorus(V), both
Introduction
Synthetic oligonucleotides are used as substrates for polymerase
chain reactions, DNA sequencing and mutagenesis experiments
and modified ones are of interest as potential antisense drugs, the
first of which, vitravene, was for the treatment of cytomegalo-
virus-induced retinitis in AIDS patients.1 Since the discovery of
RNA interference (RNAi) as a means to silence the expression
of specific genes, small interfering RNA (siRNA) oligonucleo-
tides have been recognized as powerful tools for targeting and
inhibiting therapeutically important mRNAs.2 This discovery
has created a high demand for synthetic oligonucleotides as
potential therapeutics and has spurred a renaissance in the devel-
opment of rapid, efficient methods for solid-phase RNA syn-
thesis.3 There are many challenges to the large scale
manufacture of synthetic oligonucleotides and the current cost of
a large scale manufactured drug is about £1000 per gram and is
dependent on the structure, length, sequence and nature of side
chain modifications.
towards water and 5′ nucleosidic hydroxyl groups. Currently,
most oligonucleotides are synthesised using phosphate tri- or
di-esters, phosphoramidites or H-phosphonates. Ribonucleoside
H-phosphonates7–9 are a simple and efficient way of producing
diester linkages and dinucleoside H-phosphonates with very
high yields (Scheme 1).10 The coupling step requires a dehydrat-
ing agent or activator,11 such as an acyl chloride or diphenyl-
chlorophosphate, which is then often accompanied by side
reactions as a result of self ‘capping’ and other reactions of the
reactive intermediate.12 H-phosphonate nucleosides are relatively
stable compounds even in the presence of water, however as
only the activator reacts with any water present during coupling,
the production of oligonucleotides using H-phosphonate
chemistry is becoming an attractive option.
H-Phosphonates exist in two tautomeric forms – the three
coordinate phosphite and the usually more stable four coordinate
H-phosphonate (Scheme 2), the fraction of phosphite species in
With the exception of the phosphoramidite methodology,4
the synthesis of oligonucleotides is usually from monomeric
units using phosphorylating agents such as phosphites and phos-
phonates with a dehydrating agent.5 Coupling reactions can be
carried out on a solid supported resin with the 3′-hydroxyl of
the first nucleoside base in the sequence bound directly to an
insoluble support such as glass beads, silica, polystyrene resin
derivative or swellable polymers.6
Department of Chemical and Biological Sciences, University of
Huddersfield, Queensgate, Huddersfield HD1 3DH, UK.
E-mail: m.i.page@hud.ac.uk; Tel: +01484 472534
†This article is part of the Organic & Biomolecular Chemistry 10th
Anniversary issue.
‡Electronic supplementary information (ESI) available. See DOI:
10.1039/c2ob07130d
Scheme 1
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem.

bis diethyl pyro-di-H-phosphonate (7) and diphenylphosphate
(DPP) and a small amount of tetraphenylpyrophosphate (TPPP)
which is more slowly formed. The 1H decoupled 31P NMR
analysis showed no peaks attributable to diphenyl ethyl pyro-H-
phosphonate (3), although it is presumably formed as an inter-
mediate which then reacts with ethyl H-phosphonate to produce
bis diethyl pyro-di-H-phosphonate (7) and DPP. TPPP is formed
through the reaction of DPP with DPCP (Scheme 3).
The question then arises as to whether any of these products
are the reactive phosphonylating agents in the coupling
reaction, particularly the relative activities of (7) and (3) towards
alcohols. Addition of excess tetrahydrofurfuryl alcohol to the
product mixture described above produced tetrahydrofurfuryl
ethyl H-phosphonate diester (4) and an equal molar increase in
the amount of ethyl H-phosphonate present (Scheme 4). This is
indicative of the novel H-phosphonate (7) being the active phos-
phonylating agent.
Scheme 2
ethyl H-phosphonate in aqueous solution is estimated to be
below 10−9 at most pHs, although above pH 13 it is the domi-
nant form of diethyl H-phosphonate, the pKa of which is 12.7.13
Proton abstraction of H-phosphonates to produce the phosphite
anion has been studied by the oxidation of H-phosphonates and
by H–D exchange.14 However, despite much qualitative work on
the H-phosphonate coupling step there is little quantitative infor-
mation, especially about the competing processes:15 the over-
activation of the H-phosphonates leading to unreactive phosphite
triesters16,17 and the phosphorylation of the 5′-hydroxyl of the
nucleoside to be coupled.17 Herein, we provide quantitative
mechanistic information on the reaction rates of the competing
processes using diphenylchlorophosphate as the activator, with
and without base, to establish the active phosphonylating agent
and investigate the variables which determine whether the
desired coupling or side reactions dominate.
During these reactions, none of the unwanted phosphorylation
of the hydroxyl group occurred and no phosphite triesters (6)
were formed. The rates of formation of the various species were
determined from 1H decoupled 31P NMR time scans of a
1 : 1 mixture of DPCP and triethylammonium ethyl H-phospho-
nate in CDCl3 at 25 °C with triphenylphosphate as an internal
standard. After 100 min, the reaction mixture gives approxi-
mately 85% bis diethyl pyro-di-H-phosphonate (7) and DPP and
the second order rate constant for formation of (7) is 6.8 ×
Results and discussion
1
The major loss in yield in the coupling of 3′ protected nucleo-
sides and H-phosphonate nucleosides using diphenylchlorophos-
phate (DPCP) as the activator to form the H-phosphonate diester
is due to: (i) ‘over-reaction’ of the presumed intermediate or
product diester with the activator, leading to a phosphite triester
(1)18 and (ii) 5′-phosphorylation of the 3′ protected nucleoside
by DPCP to give (2) (Scheme 1). The proportions of these
unwanted and desired products are determined by the relative
rates of the reactions and so we investigated the kinetics and
mechanisms of the reactions involved.
10−3 M−1 s−1. During the reaction, the H decoupled 31P NMR
chemical shift for ethyl H-phosphonate anion increases from
4.5 ppm to 6.7 ppm, indicating a change from the anionic salt to
its protonated form. Presumably, the DPP formed is a stronger
acid than ethyl H-phosphonic acid in CDCl3, and so DPP exists
as its anion and ethyl H-phosphonate is present as its un-
dissociated acid, and so its rate of reaction with DPCP decreases.
The suspected reactive intermediate, bis diethyl pyro-di-H-
phosphonate (7) was prepared from ethyl H-phosphonic acid
in benzene using dicyclohexylcarbodiimide as the coupling
agent.22
The formation of the unwanted phosphite triesters (1)
(Scheme 1) occurs particularly under basic conditions.12,19 The
initial reaction in the coupling of H-phosphonates promoted by
acid chlorides is thought to be activation of the H-phosphonate
monoester by acylation to form a ‘H-phosphonate anhydride’
intermediate,20,21 which subsequently reacts with the 5′-hydroxyl
of the 3′ protected nucleoside to form the desired H-phosphonate
diester (Scheme 1). However, either this diester product or the
intermediate could react further with the activator to form an
‘over-activated’ species which subsequently reacts with the 3′
protected nucleoside to produce phosphite triester by-products
(1).16,17 Our initial assumption was that similar reactions could
occur using DPCP as the activator by initial phosphorylation
rather than acylation. Thus the coupling of tetrahydrofurfuryl
alcohol with ethyl H-phosphonate using DPCP as the activator
would give initially diphenyl ethyl pyro-H-phosphonate (3) as
the reactive intermediate, which could then react either with the
alcohol to give the desired diester (4) or further with DPCP to
give an ‘over-phosphorylated’ intermediate (5) which could react
with the alcohol to give phosphite triesters (6) (Scheme 3).
However, in the absence of alcohol, ethyl H-phosphonate
triethylammonium salt in CDCl3 reacts with 1 equivalent of
diphenylchlorophosphate (DPCP) to give the novel symmetrical
The 1H decoupled 31P NMR spectrum of the diastereo-
isomeric (7) shows two singlets due to the racemic mixture of
the two enantiomers and the meso form. The equilibrium
between (7) and DPP to diphenyl ethyl pyro-H-phosphonate (3)
and ethyl H-phosphonic acid (Scheme 5) depends markedly on
whether the free acids or their salts are involved. With the
former, aliquots of DPP added to (7) in CDCl3 produced increas-
ing amounts of (3) and ethyl H-phosphonate, from which an
equilibrium constant K = 0.37 was obtained. The reaction is
reversible as confirmed by adding ethyl H-phosphonic acid
to the mixture which regenerated bis diethyl pyro-di-H-phospho-
nate (7). The equilibrium constants for the free acids in d3-aceto-
nitrile and d6-benzene both give K = 0.15, showing little
dependence on solvent and all show bis diethyl pyro-di-H-phos-
phonate (7) as the more stable derivative. With salts, the overall
equilibrium is shifted even more towards bis diethyl pyro-di-H-
phosphonate (7) and DPP triethylammonium salt, with no (3)
detectable, as expected from our earlier results.
These observations suggest that the activation of ethyl
H-phosphonate with DPCP probably occurs through the inter-
mediate formation of bis diethyl pyro-di-H-phosphonate (7)
rather than diphenyl ethyl pyro-H-phosphonate (3). The rate
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2012

Scheme 3
Scheme 7
Scheme 8
Scheme 4
Reactions in the presence of base
Scheme 5
The large scale production of H-phosphonate diesters using
DPCP as an activator is usually carried out in the presence of
base but does produce unwanted reactions including the phos-
phorylation of the 5′-hydroxyl of the nucleoside to be coupled
(Scheme 1). The rate of phosphorylation of hydroxyl groups by
DPCP was initially investigated using benzyl alcohol as a model
1
by monitoring H decoupled 31P NMR spectra. Although there
is no reaction between benzyl alcohol and DPCP (1 : 1 ratio) in
d3-acetonitrile, the addition of 1 equivalent of pyridine to the
solution resulted in total conversion of DPCP to benzyldiphenyl-
phosphate. If traces of water were present small amounts of
diphenylphosphate (DPP), from hydrolysis of DPCP, and tetra-
phenylpyrophosphate (TPPP), from the reaction of DPCP with
DPP, were formed.
Scheme 6
The phosphorylation of tetrahydrofurfuryl alcohol by DPCP
in DMF (Scheme 7) with pyridine produces pyridinium hydro-
chloride and so the rate of reaction was measured by conduc-
tance. However, there is a non-linear dependence of the latter on
the concentration of pyridinium hydrochloride. The data could
not be fitted to the Ostwald or Kohlrausch equations but a quad-
ratic conductance model based on two equilibria (Scheme 8)
and, using [Py]tot = [Py] + [PyH+] + [PyHCl]ip, where ip rep-
resents the non-conducting ion-pair, eqn (1) can be derived.
The concentration of pyridinium hydrochloride [PyH+] is
related to the conductance Λ of the solution through ε, the
specific conductivity of PyH+ ions.
of ethanolysis of (7) in CDCl3 at 25 °C is too fast to follow, but
that of iso-propanol, although complete within two minutes,
1
could be followed using H- decoupled 31P NMR time scans,
with triphenylphosphate as an internal standard and using only
1 scan per spectrum acquisition rather than the normal 16,
to give an approximate second order rate constant of 7.8 ×
10−2 M−1 s−1. In summary, the activation of ethyl H-phospho-
nate by DPCP in the absence of base proceeds through the inter-
mediate formation of diphenyl ethyl pyro-H-phosphonate (3)
which is in equilibrium with the thermodynamically more stable
symmetrical bis diethyl pyro-di-H-phosphonate (7) (Scheme 6).
Furthermore, no over-activation arises in the coupling of DPCP
with ethyl H-phosphonate in the absence of base to give either
tri-ester formation (6) or phosphorylation of the alcohol by
DPCP.
p
p
p
2
ðꢁð1 þ KaÞ þ ðð1 þ KaÞ þ 4½Pyꢀt=KdissÞÞKdiss
½PyHþꢀ ¼
2
ð1Þ
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2012

Scheme 9
increasing molar equivalents of pyridine which also gives a
decreasing concentration of TPPP. The optimum coupling ratios
found for diethyl H-phosphonate synthesis were used for the
coupling of DMT-dA-Bz-H-phosphonate triethylammonium salt
and HO-dC-Bz-OLev in d7-DMF at 25 °C using DPCP and pyri-
dine (Scheme 9) gave a very fast reaction in which the only
Fig. 1 The change in conductance with time for the alcoholysis of
DPCP (0.05 M) with tetrahydrofurfuryl alcohol (1.4 × 10−3 M) and pyri-
dine (0.1 M) in DMF at 25 °C.
1
phosphorus component observed in the H decoupled 31P NMR
spectrum is the product DMT-dA-Bz-dC-Bz-OLev formed in
99% yield.
A calibration graph was determined by adding hydrogen
chloride in DMF to 0.1 M pyridine in DMF. With excess pyri-
dine Ka can be neglected and the pyridine concentration taken as
constant. The concentration of conducting ions is then approxi-
mated by eqn (2) using λo = 7.54 × 10−7 S, Kdiss = 6.96 ×
10−4 M and ε = 0.077 S M−1. Using a pKa of pyridinium hydro-
chloride in DMF of 3.0,23 eqn (2)
Catalysis by pyridine could be due either to nucleophilic cata-
lysis through the formation of an active phosphoryl pyridinium
adduct from the reaction of pyridine with DPCP or to general
base catalysis of hydroxyl group attack on the active phosphony-
lating agent. Triethylamine is unlikely to act as a nucleophilic
catalyst because of steric hindrance but the addition of the base
to the reaction of ethyl H-phosphonate triethylammonium salt
and DPCP in CDCl3 increases the yield of bis diethyl pyro-di-H-
phosphonate (7). However, with excess base (7) is unstable and
is converted to a phosphite species. The reaction of ethyl
H-phosphonate triethylammonium salt with ten equivalents of
DPCP in CDCl3 containing 22% v/v of triethylamine gives,
after two minutes of reaction, a phosphite species thought to be
2,4,6-triethoxy-1,3,5-trioxa-2,4,6-triphosphinane (tris-ethylmeta-
phosphite) (8).25 This probably explains the poor yield obtained
when using pre-activation26 which encourages metaphosphite
formation and any phosphite by-products formed from the alco-
holysis of the metaphosphite. The hydrolysis of tris-ethyl meta-
phosphite (8) gives only ethyl H-phosphonate, whilst alcoholysis
gives ethyl H-phosphonate, the H-phosphonate diester (4) and a
phosphite triester (6), indicating that the alcohol hydroxyl group
preferentially reacts at phosphorus(III) rather than phosphorus(V).
Reaction at phosphorus(V) would produce only H-phosphonate
diester, with no phosphite triesters. Metaphosphite formation
could occur during oligonucleotide synthesis on solid supports
using high loadings on porous resins.26
p
ðꢁKdiss
þ
ðKd2iss ꢁ 4½HClꢀt KdissÞÞ
½PyHþꢀ ¼
ð2Þ
2
gives an excellent fit to the calibration curve for pyridinium
hydrochloride. A similar calibration for hydrogen chloride in
DMF was undertaken, so that the rate of phosphonylation could
be measured in the absence of base, using ε = 4.0 × 10−4 S M−1
and a pKa = 3.40 for hydrogen chloride in DMF.24 The rate of
alcoholysis of DPCP (0.05 M) by tetrahydrofurfuryl alcohol
with pyridine in DMF at 25 °C was followed by conductance
(Fig. 1) and the second order rate constant found to be
2.4 × 10−3 M−1 s−1 compared with that for water = 1.66 ×
10−2 M−1 s−1
.
Alcoholysis of DPCP is slower than hydrolysis and so, for
example, the reaction of equimolar quantities of water, benzyl
alcohol and DPCP gave a ratio of 7 : 1 diphenyl phosphate to
phosphate triester. Although DMF is often used for H-phospho-
nate coupling reactions, its unsuitability was also confirmed by
another side reaction of DPCP with DMF to give a reactive Vils-
meier intermediate, chloroiminium ion, leading to formylation of
nucleoside residues.
Although the coupling of ethyl H-phosphonate triethylammo-
nium salt with ethanol to form diethyl H-phosphonate using
DPCP activation occurs in d7-DMF and CDCl3 without base,
there is a faster reaction with added pyridine. Using 0.14 M pyri-
dine the yield of diethyl H-phosphonate from ethyl H-phospho-
nate (0.1 M) and ethanol (0.15 M) in d7-DMF at 25 °C also
depends on the amount of DPCP and increases linearly from 10
to 70% as the DPCP increases from 0.02 to 0.14 M. However,
with increasing molar equivalents of DPCP above 0.9 DPCP, the
initial product diphenyl phosphate is increasingly converted to
tetraphenylpyrophosphate (TPPP) presumably through its slower
reaction with DPCP, compared with ethyl H-phosphonate
(Scheme 3). The yield of diethyl H-phosphonate increases with
Using one equivalent each of DPCP, ethyl H-phosphonate
triethylammonium salt and triethylamine in CDCl3 there is no
ethyl H-phosphonate anion remaining after 2 min but large quan-
tities of DPP which subsequently reacts with DPCP to form
TPPP, which in turn reacts further and its concentration decreases
with time (Fig. 2).
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2012

Scheme 10
Table 1 The third order rate constants for the reaction of DPCP with
DPP to give TPPP in CDCl3 at 25 °C, the pKa of the bases are in water
Base
pKa
1.12
5.25
6.65
6.9
k3/M−2 s−1
4-Cyanopyridine
Pyridine
2,6-Lutidine
4-Methoxypyridine
Triethylamine
2.00 × 10−4
2.04 × 10−2
2.00 × 10−3
8.91 × 10−2
3.80 × 10−3
Fig. 2 Concentration–time plots for the reaction of DPCP (0.24 M)
and ethyl H-phosphonate triethylammonium salt (0.23 M) with triethyl-
amine (0.23 M) and triphenylphosphate (0.14 M) as internal standard in
CDCl3 measured by 1H decoupled 31P NMR at 298 K.
10.75
DPCP reacted completely within 100 min but there is no
metaphosphite produced. Before the first spectrum was obtained,
bis-diethyl pyro-di-H-phosphonate is formed and an unknown
compound with a 1H decoupled 31P NMR chemical shift of
−2 ppm is produced the concentration of which then slowly
reduces. In addition to base facilitating formation of active phos-
phonylating agents, it could also have a catalytic role in the
coupling step and so the reactions of the individual components
with base were investigated. Both DPCP and TPPP show no
detectable reaction with pyridine or triethylamine in CDCl3, but
bis-diethyl pyro-di-H-phosphonate (7) and diphenyl ethyl pyro-
H-phosphonate (3) with pyridine both form the unknown com-
1
pound with a H decoupled 31P chemical shift of −2 ppm and a
small amount of tris-ethyl metaphosphite (8). The compound
with a 1H decoupled 31P chemical shift of −2 ppm (Fig. 2)
could be the pyridinium adduct of ethyl H-phosphonate (9)
formed directly from the reaction of bis-diethyl pyro-di-H-
phosphonate with pyridine,27 but mass spectral analysis to trap
the intermediate proved inconclusive. Although the reaction of
pyridine with bis-diethyl pyro-di-H-phosphonate gives the com-
Scheme 11
DPCP gives a single signal of chemical shift of −12.8 ppm
assumed to be the DMAP–DPCP adduct (10).28 The hydrogens
ortho to the pyridine nitrogen shift from 6.47 to 7.36 upon
phosphorylation, compared with protonation which causes a
shift to 6.91.
1
pound with a H decoupled 31P chemical shift of −2 ppm but
no ethyl H-phosphonate anion, with diphenyl ethyl pyro-H-
phosphonate the former is formed with an equivalent amount of
DPP. Although H-phosphonate–pyridinium adducts are reported
in the literature the evidence is far from conclusive.27
Pyridine does catalyse the reaction of diphenyl phosphate and
DPCP to give TPPP and the corresponding second order rate
constant shows a first order dependence on the concentration of
pyridine and the overall rate law is given by eqn (3).
The third order rate constants for the formation of TPPP using
a variety of substituted pyridines and triethylamine are given in
Table 1. The third order rate constants increase with the basicity
of the substituted pyridine and, excluding the sterically hindered
lutidine, generate a linear Bronsted plot of log k3 against the pKa
of the base (in water) giving an apparent Bronsted βnuc of 0.46.
With increasing basicity of the substituted pyridine, the equili-
brium constant for formation of the adduct is expected to
increase but the rate of the second step would decrease so the
observed value is compatible with rate limiting displacement of
pyridine by attack of DPP on the adduct.
Rate of TPPP formation ¼ k3½DPCPꢀ½DPPꢁꢀ½Pyridineꢀ
þ k2uncat½DPCPꢀ½DPPꢁꢀ
ð3Þ
The third order rate constant, k3, is 2 × 10−2 M−2 s−1 and
the second order uncatalysed rate constant, k2uncat, is 8.9 ×
10−3 M−1 s−1. The overall third order term is compatible with
nucleophilic catalysis (Scheme 10) with a pre-equilibrium
formation of the pyridinium adduct which then reacts with DPP
in the rate limiting step. In support of this proposal, the 1H
decoupled 31P NMR of an equimolar mixture of DMAP and
Conclusions
The coupling reaction of ethyl H-phosphonate and tetrahydrofur-
furyl alcohol promoted by DPCP and pyridine in CDCl3 is given
in Scheme 11. DPCP reacts initially with pyridine to form a reac-
tive intermediate (9) that is attacked by ethyl H-phosphonate to
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem.

Triethylammonium ethyl H-phosphonate
generate diphenyl ethyl pyro-H-phosphonate (3) which in turn
reacts either with pyridine to form another reactive intermediate
or with ethyl H-phosphonate to give bis diethyl pyro-di-H-phos-
phonate (7). It is these two activated H-phosphonate derivatives
that then react with the alcohol to form the desired diester
product (4). However, although added base does increase the
overall rate of reaction it also increases the amount of unwanted
side products.
Ammonium ethyl H-phosphonate (12.96 g, 0.1 mol) was dis-
solved in acetonitrile (300 ml) and triethylamine (75 ml,
0.54 mol) added. The reaction was stirred for 24 h under nitro-
gen. The mixture was then evaporated under reduced pressure,
acetonitrile (100 ml) was added and evaporated. Triethylamine
was removed under vacuum 1 mmHg for 24 h. 1H decoupled 31
P
1
NMR δ (CDCl3) 4.56 (s, 1P); H NMR δ (CDCl3) 7.6 (s, 0.5H),
6.05 (s, 0.5H), 3.92 (q, 2H), 3.1 (t, 6H), 1.34 (t, 9H), 1.28
(t, 3H).
Experimental
Diphenylalkylphosphates
Tetrabutylammonium ethyl H-phosphonate
The alcohol (46.4 mmol), pyridine (5.55 g, 70.2 mmol) and
DCM (50 ml) were cooled to ∼0 °C over a nitrogen atmosphere.
Diphenylchlorophosphate (15.04 g, 55.5 mmol) was added over
5 min using a syringe and septa. The reaction was stirred for
5 min before quenching and washing with water (2 × 50 ml).
The DCM phase was dried over magnesium sulphate and evap-
orated under reduced pressure. To remove pyrophosphates the
residue was dissolved in DCM (50 ml), washed with acid (50 ml
at pH 2), then base (50 ml sat. sodium bicarbonate solution)
before a final water wash (50 ml). The organic phase was dried
with magnesium sulphate and evaporated to dryness.
Ammonium ethyl H-phosphonate (2.02 g, 15.9 mmol) was dis-
solved in acetonitrile (50 ml). Tetrabutylammonium chloride
(4.49 g, 16 mmol) was added and the reaction left stirring over-
night. DCM (20 ml) and water (20 ml) was added and stirred for
24 h. The DCM phase was then separated and evaporated under
reduced pressure. 1H decoupled 31P NMR δ (CD3CN) 1.61
(s, 1P); H NMR δ (CD3CN) 7.6 (s, 0.5H), 5.6 (s, 0.5H), 3.7
(q, 2H), 3.1 (t, 8H), 1.6 (m, 8H), 1.3 (m, 8H), 1.2 (t, 3H), 0.9
(t, 12H).
1
Ethyl H-phosphonic acid
Diphenylbenzylphosphate. (10.41 g) White solid. GC-MS
analysis: 5.50 min (M + 1, 341). 1H decoupled 31P NMR δ
(CD3CN) −10.9 (s, 1P); 1H NMR δ (CD3CN) 7.0–7.5 (m, 15H),
5.26 (d, J3 = 8.3 Hz, 2H).
Hydrogen chloride in ether (2 M, 100 ml, 0.2 mol) was added to
ammonium ethyl H-phosphonate (15 g, 0.12 mol) in DCM
(100 ml) and left stirring overnight under a nitrogen atmosphere.
After filtering and evaporation under reduced pressure, trace
amounts of ethyl H-phosphonate ammonium salt were removed
with piperidine on a resin (12.5 g, 94.7%). MS (+ve ion,
CD3CN): 111.2; 1H decoupled 31P NMR: δ (CDCl3) 8.02
Diphenylisopropylphosphate. (12.1 g) White crystals. GC-MS
analysis: (DCM) (M + 1, 293); 1H decoupled 31P NMR: δ
(CD3CN) −11.8 (s, 1P); 1H NMR δ (CD3CN) 7.0–7.5 (m, 10H),
4.90 (m, 1H, J3 = 6.3 Hz), 1.36 (dd, 6H, J3 = 6.14 Hz, J4 =
0.88 Hz).
1
(s, 1P); H NMR: δ (CDCl3) 12.40 (s, 1H), 7.66 (s, 1H), 4.12
(q, 2H), 1.36 (t, 3H).
Bis diethyl pyro-di-H-phosphonate29
Diphenylethylphosphate. (13.16 g) White crystals. GC-MS
analysis: 4.41 min (M + 1, 279); 1H decoupled 31P NMR δ
(CD3CN) −11.0 (s, 1P); 31P coupled NMR δ (CD3CN) −11.0 (t,
Ethyl H-phosphonate free acid (2.01 g, 18.2 mmol) in benzene
(5 ml) was placed in oven dried glassware and a nitrogen
flushed filtration apparatus. Dicyclohexylcarbodiimide (2.23 g,
10.7 mmol) in benzene (10 ml) was added over 30 min then
stirred for 1 h before filtering under nitrogen. The filter cake was
washed with benzene (3 ml) and the product stored in the fridge
as the benzene solution for later use. For NMR analysis bis
diethyl pyro-di-H-phosphonate solution (0.5 ml) was evaporated
under reduced pressure in a Schlenk flask. 1H decoupled 31P
NMR: δ (CDCl3) −3.1 (d, 2P); MS (+ve ion, CD3CN): (M + 1,
203). All pyrophosphonates were manipulated in a sealed nitro-
gen glove box.
1
J3 = 8.4 Hz); H NMR: δ (CD3CN) 7.0–7.5 (m, 10H), 4.35 (m,
2H, J3 = 7.02 Hz, J4 = 1.32 Hz), 1.36 (dt, 3H, J3 = 7.02Hz, J4 =
1.32 Hz).
Ammonium ethyl H-phosphonate
To a concentrated ammonia solution (675 ml, 13.87 mol) was
slowly added diethyl H-phosphonate (129 ml, 1 mol) with stir-
ring. The flask was sealed and stirred at room temperature for
24 hours, before evaporation to dryness. The resulting thick
clear solution was suspended in acetonitrile (500 ml) upon
which a crystalline white solid formed. The solid was filtered
and dried by desiccation overnight to yield solid ammonium
DMT-dA-Bz-H-phosphonate triethylammonium salt
1
ethyl H-phosphonate (120.6 g, yield 96.8%). H decoupled 31P
1,2,4-Triazole (12.3 g, 0.175 mol), triethylamine (24.6 g,
0.242 mol) and tetrahydrofuran (317 ml) were cooled to below
−10 °C under a nitrogen atmosphere. Phosphorus trichloride
(5.3 ml, 30.8 mmol) in THF (10 ml) was added over 30 min
with a THF wash (10 ml). The reaction was stirred for 30 min,
1
NMR δ (D2O) 6.77 (s, 1P,) H decoupled 31P NMR δ (D2O +
HCl) 7.27 (s, 1P). 31P coupled NMR δ (D2O) 8.69 and 4.85 (dt,
P–H 621.04 Hz, J3 = 8.47 Hz); 1H NMR δ (D2O) 7.64 (s, 0.5H),
5.53(s, 0.5H), 3.78 (q, 2H), 1.13 (t, 3H).
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2012

DMT-dA-Bz-OH (10 g, 15.2 mmol) in THF (263 ml) was then
added over 3 h, again maintaining the temperature at −10 °C.
The reaction was then left stirring until HPLC analysis showed
less than 1% starting material. The reaction mixture was warmed
to room temperature and water (80 ml) and triethylamine
(80 ml) were added. A quench of DCM (500 ml), water
(750 ml) and acetic acid (1 ml) adjusted to pH 7 using triethyl-
amine was then prepared. The reaction mixture was then
quenched into the DCM–water maintaining the pH at 6–8 using
acetic acid with a final pH of 7. The aqueous/organic layers were
separated and the water phase washed with DCM (200 ml).
After combination the DCM layer was dried over magnesium
sulphate before concentrating under vacuum. The off-white solid
(9.3 g) was purified by dissolving in a minimum of DCM and
loaded onto a packed A60 silica column (150 g) and eluting
with DCM–methanol–triethylamine mixtures and analysing by
HPLC. Evaporation of solvent gave a white solid with water
was added to the cell and the datalog started. After one hour pyr-
idine (0.79 g, 10 mmol) was added, after a further hour tetra-
hydrofurfuryl alcohol (10.3 mg, 0.1 mmol); after two hours
a further amount of tetrahydrofurfuryl alcohol was added
(20.6 mg, 0.2 mmol). After a total of five hours the datalog was
stopped and the data retrieved. For the hydrolysis of DPCP:
DMF (100 ml) containing water (1.3%w/w, 0.78 M) was added
to the flask and the initial conductance value noted. DPCP
(2.7 g, 10 mmol) was added to the cell and the datalog started.
After one hour an additional aliquot of DPCP (2.7 g, 10 mmol)
was added to the cell. After a further hour of data collection
the datalog was stopped and results retrieved. For competitive
alcoholysis and hydrolysis of DPCP: DPCP (92.7 mg,
0.34 mmol) was added to a mixture of benzyl alcohol (40.5 mg,
0.37 mmol) and water (6.7 mg, 0.37 mmol) in d3-acetonitrile
1
(0.75 ml) and analysed by H decoupled 31P NMR: δ (d3-aceto-
nitrile) −4.7 ppm (s, 0.609P, DPCP), −9.9 ppm (s, 0.322P,
DPP), −10.9 ppm (s, 0.046P, benzyldiphenylphosphate),
−24.4 ppm (s, 0.024P, TPPP).
1
content 0.24%. H decoupled 31P NMR: δ 3.70 ppm (s, 1P) MS
(acetonitrile, +ve): M + 1, 722, (acetonitrile, −ve): M − 1, 720.
Alcoholysis of DPCP
DPCP activation
DPCP (199.2 mg, 0.7 mmol) was added to a solution of benzyl
alcohol (95.9 mg, 0.9 mmol) in CD3CN (1.5 ml) and analysed
by NMR. Pyridine (70.3 mg, 0.9 mmol) was added and the
sample re-analysed by NMR.
DPCP (65.1 mg, 0.24 mmol, 0.32 M) was added to a solution
of ethyl H-phosphonate triethylammonium salt (52.6 mg,
0.25 mmol, 0.33 M) in CDCl3 (0.75 ml) and monitored by
1H decoupled 31P NMR: δ (CDCl3) 4.4 ppm (s, 0.94P, ethyl
H-phosphonate), −3.05 ppm (d, 0.48P, bis diethyl pyro-di-H-
phosphonate), −4.7 ppm (s, 1P, DPCP), −11.6 ppm (s, 0.29P,
DPP), −25.0 ppm (s, 0.06P, TPPP). 1H NMR: δ (CDCl3)
8.0 ppm (s, 0.5H, P–H bis diethyl pyro-di-H-phosphonate),
7.6 ppm (s, 0.5H, P–H ethyl H-phosphonate), 6.9–7.4 ppm (m,
aromatic protons DPCP, DPP and TPPP), 6.1 ppm (s, 0.5H, P–H
bis diethyl pyro-di-H-phosphonate), 5.9 ppm (s, 0.5H, P–H ethyl
H-phosphonate), 4.3 ppm (m, 4H, OCH2CH3 bis diethyl pyro-
di-H-phosphonate), 4.0 ppm (m, 2H, OCH2CH3 ethyl H-phos-
phonate), 3.0 ppm (q, 6H, HN+(CH2CH3)3), 1.4 ppm (m, 6H,
OCH2CH3 from bis diethyl pyro-di-H-phosphonate), 1.3 ppm
(t, 9H, HN+(CH2CH3)3).
DPCP and pyridine
DPCP (202.0 mg, 0.7 mmol) was added to a solution of pyridine
1
(74.9 mg, 0.9 mmol) in CD3CN (1.5 ml) and analysed by H
decoupled 31P NMR.
TPPP and benzyl alcohol
Benzyl alcohol (88.2 mg, 0.8 mmol) was added to a solution of
TPPP (0.34 g, 0.7 mmol) in CD3CN (1.5 ml) and analysed by
NMR, then repeated with pyridine (63.5 mg, 0.8 mmol).
Ethyl H-phosphonate triethylammonium salt (31.7 mg,
0.15 mmol, 0.2 M) was then added to a solution of DPCP
(51.6 mg, 0.19 mmol, 0.25 M) in CDCl3 (0.75 ml) and moni-
tored by 1H decoupled 31P NMR.
Pyrophosphoric acid and benzyl alcohol
Benzyl alcohol (87.2 mg, 1.1 mmol) was added to pyrophos-
phoric acid (0.19 g, 1.1 mmol) in CD3CN (1.5 ml) and analysed
by NMR; then repeated with pyridine (87.2 mg, 1.1 mmol).
DPCP (51.6 mg, 0.19 mmol, 0.25 M) was added to a solution
of ethyl H-phosphonate triethylammonium salt (38.0 mg,
0.18 mmol, 0.24 M) and triphenylphosphate (31.3 mg,
0.1 mmol, 0.13 M) in CDCl3 (0.75 ml), left for two hours and
Alcoholysis of DPCP
1
analysed by H decoupled 31P NMR. After analysis tetrahydro-
DPCP (81.4 mg, 0.3 mmol) was added to benzyl alcohol
(43.7 mg, 0.4 mmol) in d7-DMF (0.75 ml) and analysed by
NMR; then repeated with pyridine (31.7 mg, 0.4 mmol).
furfuryl alcohol (19.6 mg, 0.19 mmol, 0.25 M) was added and
the sample re-analysed by 1H decoupled 31P NMR. Pyridine
(15.1 mg, 0.19 mmol, 0.25 M) was added and the mixture again
analysed by 1H decoupled 31P NMR: δ (CDCl3) 6.8 ppm
(s, 0.84P, ethyl H-phosphonate), −3.0 ppm (d, 1.59P, bis diethyl
pyro-di-H-phosphonate), −4.7 ppm (s, 1.06P, DPCP), −11.6 ppm
(s, 0.90P, DPP), −17.1 ppm (s, 1.0P, triphenylphosphate),
Conductance studies
Glassware was oven dried and the conductance cell purged with
nitrogen prior to use. The conductance was calibrated using KCl
(0.745 g) in water (1 l) giving a conductance of 1413 mS. For
the alcoholysis of DPCP : DMF (100 ml) was added to the flask
and the initial conductance value noted, DPCP (2.7 g, 10 mmol)
1
−25.0 ppm (s, 0.37P, TPPP). H decoupled 31P NMR: δ (CDCl3
+ tetrahydrofurfuryl alcohol) 9.2 ppm (s, 0.40P, tetrahydrofur-
furyl ethyl H-phosphonate diester), 8.6 ppm (s, 0.47P, tetra-
hydrofurfuryl ethyl H-phosphonate diester), 7.0 ppm (s, 1.50P,
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem.

1
ethyl H-phosphonate), −4.7 ppm (s, 1.01P, DPCP), −11.6 ppm
(s, 0.97P, DPP), −17.1 ppm (s, 1.0P, triphenylphosphate),
−25.0 ppm (s, 0.38P, TPPP). 31P coupled NMR: δ (CDCl3 +
tetrahydrofurfuryl alcohol) 11.3 and 6.9 ppm (dq, P–H coupling
706.6 Hz, J3 = 9.36 Hz, tetrahydrofurfuryl ethyl H-phosphonate
diester), 10.8 and 6.5 ppm (dq, P–H coupling 704.3 Hz, J3 =
9.14 Hz, tetrahydrofurfuryl ethyl H-phosphonate diester), 9.1
and 5.0 ppm (dt, P–H coupling 686.5 Hz, ethyl H-phosphonate).
1H NMR: δ (CDCl3 + tetrahydrofurfuryl alcohol) 7.73 and
5.95 ppm (d, P–H coupling 706.61 Hz, tetrahydrofurfuryl ethyl
H-phosphonate diester), 7.69 and 5.93 ppm (d, P–H coupling
704.3 Hz, tetrahydrofurfuryl ethyl H-phosphonate diester), 7.60
and 5.89 ppm (d, P–H coupling 686.8 Hz, ethyl H-phosphonate).
1H decoupled 31P NMR: δ (CDCl3 + tetrahydrofurfuryl alcohol
+ pyridine) 9.2 ppm (s, 0.53P, tetrahydrofurfuryl ethyl H-phos-
phonate diester), 8.6 ppm (s, 0.67P, tetrahydrofurfuryl ethyl
H-phosphonate diester), 7.0 ppm (s, 1.0P, ethyl H-phosphonate),
−4.7 ppm (s, 0.81P, DPCP), −11.6 ppm (s, 1.18P, DPP),
−17.1 ppm (s, 1.0P, triphenylphosphate), −25.0 ppm (s, 0.35P,
TPPP). MS (+ve, acetonitrile): (M + 1, 195, tetrahydrofurfuryl
ethyl H-phosphonate diester).
6.89 ppm (m), 3.09 ppm (s, 6H, N+(CH3)2 DMAP-DPCP); H
decoupled 31P NMR: δ (DPCP and DMAP in CDCl3) −5.0 ppm
(s, 3.09, DPCP), −13.1 ppm (s, 11.53, DMAP-DPCP),
−25.1 ppm (s, 1.0, TPPP). H NMR: δ (DMAP-HCl in CDCl3)
14.63 ppm (broad s, 1H, DMAP-HCCl), 8.15 ppm (s, 2H, HA
DMAP-HCl), 6.91 ppm (s, 2H, HB DMAP-HCl), 3.31 ppm
(s, 6H, N+(CH3)2 DMAP-HCl).
1
Abbreviations
DPCP diphenylchlorophosphate
DPP
diphenylphosphate
TPPP tetraphenylpyrophosphate
Acknowledgements
We are grateful to Avecia Biotechnology for support.
References
1 The Vitravene Study Group, Am. J. Opthalmol., 2002, 133, 467.
2 S. L. Beaucage, Curr. Opin. Drug Discovery Dev., 2008, 11, 203.
3 J. Tian, K. Ma and I. Saaem, Mol. BioSyst., 2009, 5, 714.
4 M. H. Caruthers, Science, 1985, 230, 281.
1H decoupled 31P NMR timescan kinetics
DPCP (51.6 mg, 0.19 mmol, 0.25 M) was added to a solution of
triethylammonium ethyl H-phosphonate (38.0 mg, 0.18 mmol,
0.24 M) and triphenylphosphate (31.3 mg, 0.1 mmol, 0.13 M)
in CDCl3 or d3-acetonitrile (0.75 ml). The sample was then ana-
5 G. Zon, New J. Chem., 2010, 34, 795.
6 R. B. Merrifield, Angew. Chem., Int. Ed. Engl., 1985, 24, 799.
7 R. H. Hall, A. Todd and R. F. Webb, J. Chem. Soc., 1957, 3291.
8 B. C. Froehler and M. D. Matteucci, Tetrahedron Lett., 1986, 27, 469.
9 B. C. Froehler, P. G. Ng and M. D. Matteucci, Nucleic Acids Res., 1986,
14, 5399.
1
lysed continuously by H decoupled 31P NMR. All individual
10 E. de Vroom, C. E. Dreef, H. van den Elst, G. A. van der Marel and
J. H. van Boom, Recl. Trav. Chim. Pays-Bas, 1988, 107, 592.
11 H. G. Khorana, Science, 1979, 203, 614.
NMR spectra were 16 scans per spectra.
12 P.J. Garegg, T. Regberg, J. Stawinski and R. Stromberg, Nucleosides,
Nucleotides Nucleic Acids, 1987, 6, 655.
Equilibrium studies
13 P. J. Guthrie, Can. J. Chem., 1979, 57, 236.
DPP (156.7 mg, 0.62 mmol) was dissolved in the chosen solvent
(0.4 ml) and aliquots (50 μl) were added to a solution of bis
diethyl pyro-di-H-phosphonate containing triphenylphosphate
(17.6 mg, 0.05 mmol) in the same solvent (0.6 ml) under a nitro-
gen atmosphere. The sample was then analysed by 1H decoupled
31P NMR and 31P coupled–31P coupled NMR COSY. The
position of DMAP–DPCP adduct formation was determined
from DPCP (78.0 mg, 0.29 mmol) in CDCl3 (0.75 ml) analysed
by both 1H and 1H decoupled 31P NMR. DMAP (55.mg,
14 Z. Luz and B. Silver, J. Am. Chem. Soc., 1961, 83, 4518.
15 J. Stawinski and A. Kraszewski, Acc. Chem. Res., 2002, 35, 952.
16 R. Stromberg, Nucleosides, Nucleotides Nucleic Acids, 2003, 22, 1.
17 Y. V. Dankov, N. V. Batchikova, N. V. Scaptsova, E. S. Besidsky and
A. V. Azhayev, Bioorg. Khim., 1988, 14, 615.
18 P. J. Garegg, T. Regberg, J. Stawinski and R. Stromberg, Nucleosides,
Nucleotides Nucleic Acids, 1987, 6, 283.
19 P. J. Garegg, I. Lindh, T. Regberg, J. Stawinski, R. Stromberg and
C. Henrichson, Tetrahedron Lett., 1986, 27, 4051.
20 B. C. Froehler and M. D. Mattuecci, Nucleosides Nucleotides, 1987, 6,
287.
21 P. J. Garegg, T. Regberg, J. Stawinski and R. Stromberg, Chemica
Scripta, 1986, 26, 59.
22 A. Zwierzak and A. Koziara, Tetrahedron, 1967, 23, 2243.
23 M. Breant and G. Demanga-Guerin, Bull. Soc. Chim. Fr., 1969, 2935.
24 R. L. Beniot and C. Buisson, Electrochim. Acta, 1973, 18, 105.
25 P. J. Garegg, J. Stawinski and R. Stromberg, J. Org. Chem., 1987, 52,
284.
26 J. H. Atherton, Avecia Biotechnology, personal communication.
27 I. Y. Dubey, V. A. Efimov, T. V. Lyapina and D. M. Fedoryak, Bioorg.
Khim., 1992, 18, 911; S. Sigurdsson and R. Stromberg, J. Chem. Soc.,
Perkin Trans. 2, 2002, 1682.
1
0.45 mmol) in CDCl3 (0.75 ml) analysed by H NMR. DPCP
(0.16 g, 0.57 mmol) was added to a solution of DMAP
(62.5 mg, 0.51 mmol) in CDCl3 (0.75 ml) and analysed by both
1H and 1H decoupled 31P NMR. DMAP hydrochloride
(80.9 mg, 0.51 mmol) in CDCl3 (0.75 ml) analysed by 1H
NMR: δ (DPCP in CDCl3) 7.32 ppm (m, phenyl protons DPCP)
1H decoupled 31P NMR: δ (DPCP in CDCl3) −4.6 ppm (s, 1P,
1
DPCP), −24.9 ppm (s, 0.02P, TPPP); H NMR: δ (DMAP in
CDCl3) 8.22 ppm (d, 2H, HA DMAP), 6.47 ppm (d, 2H, HB
DMAP), 2.97 ppm (s, 6H, N(CH3)2 DMAP). 1H NMR: δ
(DPCP and DMAP in CDCl3) 8.23 ppm (broad s, 2H, HA
DMAP-DPCP), 7.36 ppm (broad s, 2H, HB DMAP-DPCP),
28 V. F. Zarytova, E. M. Ivanova, D. G. Knorre, A. V. Lebedev,
A. I. Rezvukhin and E. V. Yarmolinskaya, Dokl. Akad. Nauk. SSSR.,
1979, 248, 1124.
29 A. Zwierzak and A. Koziara, Tetrahedron, 1967, 23, 2243.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2012