Ester prodrugs of acyclic nucleoside thiophosphonates compared to phosphonates: Synthesis, antiviral activity and decomposition study

Article abstract of DOI:10.1016/j.ejmech.2013.02.039

9-[2-(Thiophosphonomethoxy)ethyl]adenine [S-PMEA, 8] and (R)-9-[2-(Thiophosphonomethoxy)propyl]adenine [S-PMPA, 9] are acyclic nucleoside thiophosphonates we described recently that display the same antiviral spectrum (DNA viruses) as approved and potent phosphonates PMEA and (R)-PMPA. Here, we describe the synthesis, antiviral activities in infected cell cultures and decomposition study of bis(pivaloyloxymethoxy)-S-PMEA [Bis-POM-S-PMEA, 13] and bis(isopropyloxymethylcarbonyl)-S-PMPA [Bis-POC-S-PMPA, 14] as orally bioavailable prodrugs of the S-PMEA 8 and S-PMPA 9, in comparison to the equivalent "non-thio" derivatives [Bis-POM-PMEA, 11] and [Bis-POC-PMPA, 12]. Compounds 11, 12, 13 and 14 were evaluated for their in vitro antiviral activity against HIV-1-, HIV-2-, HBV- and a broad panel of DNA viruses, and found to exhibit moderate to potent antiviral activity. In order to determine the decomposition pathway of the prodrugs 11, 12, 13 and 14 into parent compounds PMEA, PMPA, 8 and 9, kinetic data and decomposition pathways in several media are presented. As expected, bis-POM-S-PMEA 13 and bis-POC-S-PMPA 14 behaved as prodrugs of S-PMEA 8 and S-PMPA 9. However, thiophosphonates 8 and 9 were released very smoothly in cell extracts, in contrast to the release of PMEA and PMPA from "non-thio" prodrugs 11 and 12.

Full text of DOI:10.1016/j.ejmech.2013.02.039

Accepted Manuscript
Ester prodrugs of acyclic nucleoside thiophosphonates compared to phosphonates :
Synthesis, Antiviral activity and Decomposition Study
Loïc Roux, Stéphane Priet, Nadine Payrot, Clément Weck, Maëlenn Fournier, Fabien
Zoulim, Jan Balzarini, Bruno Canard, Karine Alvarez
PII:
S0223-5234(13)00146-3
10.1016/j.ejmech.2013.02.039
EJMECH 6034
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 6 December 2012
Revised Date: 30 January 2013
Accepted Date: 26 February 2013
Please cite this article as: L. Roux, S. Priet, N. Payrot, C. Weck, M. Fournier, F. Zoulim, J. Balzarini, B.
Canard, K. Alvarez, Ester prodrugs of acyclic nucleoside thiophosphonates compared to phosphonates :
Synthesis, Antiviral activity and Decomposition Study, European Journal of Medicinal Chemistry (2013),
doi: 10.1016/j.ejmech.2013.02.039.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Highlight 1
The RT of HIV is an important target for antiretroviral drugs such as nucleosi(ti)de RT
inhibitors.
Highlight 2
We have designed new ANPs: α-thiophosphonates S-PMEA and S-PMPA active against
HIV-1, HIV-2 and HBV.
Highlight 3
To improve pharmacokinetic, we have designed, prodrug forms: bis-POM-S-PMEA and bis-
POC-S-PMPA.
Highlight 4
Prodrug forms increase antiviral efficacy against HIV-1, HIV-2, HBV and HSV.
Highlight 5
We established the decomposition pathway of prodrug forms in cell extracts.

ACCEPTED MANUSCRIPT
Graphical Abstract.

ACCEPTED MANUSCRIPT
Ester prodrugs of acyclic nucleoside thiophosphonates
compared to phosphonates : Synthesis, Antiviral activity
and Decomposition Study
a
a
a
a
b
Loïc Roux , Stéphane Priet , Nadine Payrot , Clément Weck , Maëlenn Fournier , Fabien
b
c
a
a*
Zoulim , Jan Balzarini , Bruno Canard , and Karine Alvarez .
Abstract.
9-[2-(Thiophosphonomethoxy)ethyl]adenine
[S-PMEA,
8]
and
(R)-9-[2-
(Thiophosphonomethoxy)propyl]adenine [S-PMPA, 9] are acyclic nucleoside thiophosphonates we
described recently that display the same antiviral spectrum (DNA viruses) as approved and potent
phosphonates PMEA and (R)-PMPA.. Here, we describe the synthesis, antiviral activities in infected cell
cultures and decomposition study of bis(pivaloyloxymethoxy)-S-PMEA [Bis-POM-S-PMEA, 13] and
bis(isopropyloxymethylcarbonyl)-S-PMPA [Bis-POC-S-PMPA, 14] as orally bioavailable prodrugs of the
S-PMEA 8 and S-PMPA 9, in comparison to the equivalent „non-thio“ derivatives [Bis-POM-PMEA, 11]
and [Bis-POC-PMPA, 12] . Compounds 11, 12, 13 and 14 were evaluated for their in vitro antiviral
activity against HIV-1-, HIV-2-, HBV- and a broad panel of DNA viruses, and found to exhibit moderate
to potent antiviral activity.. In order to determine the decomposition pathway of the prodrugs 11, 12, 13
and 14 into parent compounds PMEA, PMPA, 8 and 9, kinetic data and decomposition pathways in
several media are presented. As expected, bis-POM-S-PMEA 13 and bis-POC-S-PMPA 14 behaved as
prodrugs of S-PMEA 8 and S-PMPA 9. However, thiophosphonates 8 and 9 were released very smoothly
in cell extracts, in contrast to the release of PMEA and PMPA from „non-thio“ prodrugs 11 and 12.
1

ACCEPTED MANUSCRIPT
(Footnotes) * Corresponding author. Tel.: +33 491 825 571; fax: +33 491 266 720.
E-mail address: Karine.Alvarez@afmb.univ-mrs.fr (K. Alvarez).
a: Laboratoire d’Architecture et Fonction des Macromolécules Biologiques, UMR CNRS 7257, Equipe
"
Medicinal Chemistry and Structural Virology", Université Aix-Marseille, Parc scientifique de Luminy,
63 av. de Luminy, 13288 Marseille Cedex 9, France.
1
b: INSERM, U871, Université de Lyon, Hospices Civils de Lyon, 69003 Lyon, France.
c: Laboratory of Virology and Chemotherapy, Rega Institute for Medical Research, K.U. Leuven, B-3000
Leuven, Belgium.
Keywords: Nucleotide analog; Prodrug, Thiophosphonate; Antiviral activity; Metabolism.
Abbreviations: HIV, human immunodeficiency virus; HBV, hepatitis B virus; RT, reverse transcriptase;
dNTP,
phosphonomethoxyethyl]adenine; S-PMEA, 9-[2-thiophosphonomethoxyethyl]adenine; (R)-PMPA, (R)-
-[2-phosphonomethoxypropyl]adenine; (R)-S-PMPA, (R)-9-[2-thiophosphonomethoxypropyl]adenine;
bis-POC-PMPA, (R)-9-{2-[O,O’-bis(isopropoxycarbonyloxymethyl)phosphonomethoxy]propyl}adenine;
deoxynucleotide;
ANP,
acyclic
nucleoside
phosphonate;
PMEA,
9-[2-
9
bis-POC-S-PMPA,
(R)-9-{2-[O,O’-
bis(isopropoxycarbonyloxymethyl)thiophosphonomethoxy]propyl}adenine;
bis-POM-PMEA,
9-{2-
[
O,O’-bis(pivaloyloxymethyl)phosphonomethoxy]ethyl}adenine;
bis-POM-S-PMEA,
9-{2-[O,O’-
bis(pivaloyloxymethyl)thiophosphonomethoxy]ethyl}adenine.
1
. Introduction.
Acyclic nucleoside phosphonates (ANPs) [1-3] are outstanding molecules in the fight against DNA virus
and retrovirus infections. Three of those compounds have been formally licensed for the treatment of
HCMV infections [4] in AIDS patients (HPMPC, cidofovir, Vistide®), chronic HBV infections [1, 5, 6]
(PMEA, Adefovir dipivoxil, Hepsera®, (R)-PMPA, Tenofovir disoproxil fumarate, Viread®) and HIV
infections (Tenofovir disoproxil fumarate, Viread®). Tenofovir is also available in a combination
formulation with emtricitabine (Truvada®), or emtricitabine and efavirenz (Atripla®) or emtricibatine
2

ACCEPTED MANUSCRIPT
and rilpivirine (Complera®/Eviplera®) for the treatment of AIDS. In addition to these indications, there
are various other clinical conditions in which ANPs have proven to be active [1, 7].
ANPs are nucleotide analogs, and thus a phosphonate group is attached to the nucleoside moiety. Hence
ANPs: 1) should resist any attack by esterases, phosphatases or any catabolic enzymes, thus enhancing
physiological stability and half-life in biological fluids and cells, and 2) require only two phosphorylation
steps to be converted into their diphosphate active form [8], thus circumventing the often rate-limiting
first phosphorylation step of nucleoside analogs.
With respect to their mode of action, these nucleoside phosphonate analogs are phosphorylated by cellular
kinases into nucleoside phosphonate diphosphates after endocytosis-mediated penetration into the
infected cells. While they are generally poor substrates for cellular polymerases, these nucleotide analogs
are, on the other hand, fairly well incorporated into the growing viral DNA chain by viral reverse
transcriptase (RT) or polymerase. As these analogs do not possess a free 3’-OH group, their incorporation
results in a termination of DNA synthesis that is eventually responsible for the antiviral effect [9-11].
Given the antiviral potency, activity spectrum and long-term safety of ANPs, could their performance still
be improved? To address this question, various structural modifications have been attempted[1],
including structural diversity at both the side chain [12-14], the heterocyclic moiety [15-20] and the
phosphonate. Among phosphonate moiety modifications, two strategies have been explored. One consists
of designing prodrug adducts [21-26], such as alkoxyalkyl ester or phosphoramidate or cycloSaligenyl
derivatives to enhance bioavailability, and the other consists of modifying the environment of the
phosphorus atom of the phosphonate moiety itself [27-30]. The latter strategy has been explored to target
the resistance problem specifically, and to circumvent the loss of nucleotide analogue efficacy. Indeed,
the phosphonate moiety represents the site of the polymerization reaction during nucleotide incorporation
into the growing DNA chain and its modification may cause disturbances (electronic, steric effects) which
could be advantageous in binding affinity and/or catalytic efficiency.
We previously reported [28, 29] the synthesis of the α-thiophosphonates S-PMEA 8 and S-PMPA 9,
bearing a sulfur atom at the α-position of acyclic nucleoside phosphonate (Figure 1). Target compounds 8
3

ACCEPTED MANUSCRIPT
and 9 exhibit robust antiretroviral activity in cells infected by HIV-1 (LAI strain, and viral isolates from
subtypes A to G), HIV-2, HBV and a moderate activity against herpes simplex viruses and vaccinia virus.
We showed that their diphosphate forms S-PMEApp and S-PMPApp are readily incorporated by wild-
type (WT) HIV-1 RT into DNA and act as DNA chain terminators [28, 29].
Because these novel acyclic nucleoside thiophosphonates display promising antiviral activities, we
decided to improve their pharmacokinetic properties and related efficacy by the synthesis of prodrug
forms to make them potential candidates for HIV and HBV therapy.
A prodrug is a pharmacologically inactive compound that is converted to the active form of the drug by
endogenous enzymes or intracellular metabolism, to overcome problems associated with stability, toxicity,
lack of specificity and/or limited bioavailability. To be clinically useful a prodrug should fulfill
pharmacokinetic and pharmacodynamic profiles which are mainly dependent on the kinetics of its
conversion to the parent drug (sensitivity of the linkage between the prodrug moiety and the parent drug
to hydrolysis) and on its membrane permeation properties (lipophilicity).
Prodrugs of ANPs have been largely described [31, 32] and in particular, alkoxyalkyl ester prodrugs,
which are the forms currently employed in the licensed Adefovir dipivoxil (Hepsera®) and Tenofovir
disoproxil fumarate (Viread®). Historically, Srinivas et al. [33] described the first anti-HIV evaluation of
the 9-{2-[O,O’-bis(pivaloyloxymethyl)phosphonomethoxy]ethyl}adenine “bis-POM-PMEA” or Adefovir
dipivoxil. Antiviral activity was increased 30-fold, but pivaloyl-containing compounds generated pivaloic
acid during release of the parent drug, which constituted a potential toxicity concern [34]. To overcome
this issue, “bis-POC” prodrugs were developed [11]. The anti-HIV activity of (R)-9-{2-[O,O’-
bis(isopropoxycarbonyloxymethyl)phosphonomethoxy]propyl}adenine “bis-POC-PMPA” or Tenofovir
disoproxil was 100-fold higher than that of PMPA.
We describe here an efficient strategy to synthesize the alkoxyalkyl ester prodrugs: bis-POM-S-PMEA 13
and bis-POC-S-PMPA 14 via a thionation reaction of the corresponding “non-thio” derivatives bis-POM-
PMEA 11 and bis-POC-PMPA 12 (Figure 1). This thionation approach was also applied for the synthesis
of the parent thiophosphonates S-PMEA 8 and S-PMPA 9. In this study, we also evaluate the antiviral
4

ACCEPTED MANUSCRIPT
activity against a broad panel of viruses and the decomposition pathway in several media of the “thio”-
prodrugs 13 and 14 in comparison with the “non-thio”-prodrugs 11 and 12.
2
. Results and Discussion.
2
.1 Chemistry
We previously published [27, 28] a synthetic pathway to obtain the thiophosphonates S-PMEA 8 and S-
PMPA 9. We developed an original method to synthesize a H-phosphinate intermediate [27] that can
easily serve as a precursor to generate phosphorus-modified nucleotides. This key step was applied to
synthesize α-boranophosphonates [27], α-selenophosphonates and α-thiophosphonates [28] derived from
PMEA and PMPA. However, this exploratory method was not feasible for large-scale synthesis, limited
by the overall low yield. So, we decided to develop a new synthetic strategy allowing us to obtain target
compounds and prodrugs derived thereof using a common strategy, thereby limiting the number of steps
(Scheme 1).
The first step of the synthesis uses a methodology largely described [27, 28, 35-37] and consists of
substituting adenine in order to obtain 9-(2-hydroxyethyl)adenine 1 for the S-PMEA series and (R)-9-(2-
hydroxypropyl)adenine 2 for the S-PMPA series.
The following step, which consists of alkylating alcohols 1 and
2
with diethyl[[(p-
toluenesulfonyl)oxy]methyl]phosphonate 3 [38] in DMF in the presence of alkoxide base, is the limiting
step characterized by its low yield and side-product formation which makes purification difficult. Indeed,
two side-products were identified following the reaction by HPLC/MS, the N7-regioisomer and the
elimination product. The formation of the N7-regioisomer has been reported by Hakimelahi, G. H. [35]
and we also observed its formation using sodium tert-butoxide as alkoxide base. Indeed, The synthesis of
compounds 4 and 5 with 1.1 eq of sodium tert-butoxide leads to a N-7/N-9 ratio 15/85 and 40/60
respectively. Regioisomers were not efficiently separated by purification on silica gel chromatography.
We hypothesize that the elimination product results from a β-elimination reaction of the compound 5
which contains ether function. Under the action of strong bases such as organolithium or organosodium
5

ACCEPTED MANUSCRIPT
and to a lesser extent organomagnesium [39], hydrogen atom in beta position of the oxygen can be
removed (Reformatsky reaction analogy). Elimination reaction leads to an alkene and an alcoholate.
In order to improve regioselectivity and efficiency, and to reduce side product formation, we tried several
conditions of temperature (RT, 70 °C), solvent (DMF, NMP) and bases (tBuONa, tBuOK, NaH, LiH) [27,
3
5, 37, 40]. The alkylation can be effected also with magnesium tert-butoxide [41-43] and demonstrated
to be superior to the current manufacture production method. However, the strategy described is one-pot
synthesis and diethylphosphonates 4 and 5 were not isolated. According to the recent literature [42, 43],
alkylation of compounds 1 and 2 was accomplished in dimethylformamide at 35 °C during 20 h in the
presence of 1.5 equiv of magnesium tert-butoxide to lead to the corresponding diethylphosphonates 4 and
5
[17, 27], with 52% and 53% yields, respectively.
The new approach we designed consists then to transfer a sulfur atom to diethylphosphonates 4 and 5
forming diethylthiophosphonates 6 and 7 using an efficacious thionation reagent (e.g., Belleau reagent,
Lawesson reagent, Davy reagent) [44-47]. The Davy reagent in dry benzene at 90°C lead to the
corresponding diethylthiophosphonates 6 and 7 with 43% and 44% yields, respectively. Thionation of 4
3
1
and 5 proved to be fast and went to completion within 2h. Inspection of the decoupling P NMR spectra
showed that the signals from 4 (δ: 21.56 ppm) and 5 (δ: 22.11 ppm) were replaced by new resonances at
8
8.02 ppm for 6 and 88.41 ppm for 7.
Methyl and ethyl are current protecting groups of phosphonate diesters. However, while they can be
removed by treatment with trimethylsilyl iodide or bromide, this method fails or gives low yields with
thiophosphonate derivatives [48]. Deprotection of diethyl α,α-difluoromethylenephosphonothioic acid
via a thiono-thiolo rearrangement followed by Pd-catalysed deallylation has been reported [49]. Our goal
was to identify reaction conditions capable of generating thiophosphonic acids in a one step process.
Since nucleophilic displacement is rate-determining for TMSBr-mediated hydrolysis, the use of TMSCl
in conjugation with several exogenous nucleophiles (NaBr, NaI) was optimised for the hydrolysis of
diethylthiophosphonates 6 and 7 [42]. Compounds 6 and 7 were treated with 40 equivalents of sodium
bromide and trimethylsilyl chloride in N-methylpyrrolidone (NMP) at 90°C, during 12h, followed by
6

ACCEPTED MANUSCRIPT
hydrolysis with triethylammonium bicarbonate 0,05 M (TEAB). The deprotection reaction led to the
formation of thiophosphonates 8 and 9 with 40% and 48% yields, respectively, after reverse phase
chromatography purification.
In summary, a more efficient process has been developed to prepare thiophosphonates 8 and 9. Even if
the the overal yield remains equivalent (10%), notable features include key alkylation using magnesium
tert-butoxide, reduction of purification steps using reverse-phase chromatography, thionation of ANPs
never described and efficient transsilylation of S-ANPs never reported.
To obtain the prodrug forms [29, 30] bis-POM-S-PMEA 13 and bis-POC-S-PMPA 14 we used a
similar synthesis pathway involving a thionation reagent to introduce the sulfur atom on the alkoxyalkyl
ester phosphonate. The reaction involved the hydrolysis of the diethylphosphonates 4 and 5 followed by
the conversion into the alkoxyalkyl phosphonate diesters 11 and 12 by base-promoted alkylation. After
treatment of 4 and 5 with two equivalents of sodium bromide and four equivalents of trimethylsilyl
chloride in NMP at 60°C during 12h, phosphonate diacids were obtained but not isolated. After ensuring
the complete removal of excess of TMSCl by co-distillation with ethyl acetate, they were directly
alkylated with triethylamine and the appropriate pivaloyloxymethyl chloride (POMCl) or
chloromethylisopropyl carbonate 10 [50] (POCCl), in NMP at 60°C during 6h. Bis-POM-PMEA 11 and
bis-POC-PMPA 12 were isolated by silica chromatography in 27% and 33% yield, respectively. The last
step was the thionation reaction with the Davy Reagent in benzene at 80°C during 30 min. The bis-POM-
S-PMEA 13 and bis-POC-S-PMPA 14 were then obtained after silica gel chromatography with 29% and
3
2% yields, respectively.
2
.2 Anti-HIV activity of thiophosphonate prodrugs 13 and 14 compared to phosphonate prodrugs 11 and
1
B
2 against wild-type (WT) HIV- 1 (III ) and HIV-2 (ROD) and cytotoxicity in CEM cell cultures.
To evaluate the activity of the target compounds, a cell-based antiviral drug susceptibility assay was
performed using WT HIV-1 (III
B
) and HIV-2 (ROD) (Table 1). CEM cell cultures were treated with
7

ACCEPTED MANUSCRIPT
increasing concentrations of compounds 11, 12, 13 and 14, and infected with 100 TCID50 (50% tissue
culture infectious doses) of HIV-1 (III
RT) activity in cell culture medium was measured to determine the antiviral properties of the compounds
expressed as 50% reduction of viral replication (half maximal effective concentration or EC ) and virus-
B
) and HIV-2 (ROD). On day 3 post infection, reverse transcriptase
(
50
induced cytopathicity was microscopically recorded (Table 1). Concomitantly, the antimetabolic effect
50% cytotoxic concentration, CC50) of each compound was evaluated in the drug-exposed CEM cell
(
cultures.
Under the assay conditions, all compounds significantly inhibited the replication of HIV-1 and HIV-2 in
CEM cell cultures (Table 1). The alkoxyalkyl ester prodrugs bis-POM-PMEA 11, bis-POC-PMPA 12,
bis-POM-S-PMEA 13 and bis-POC-S-PMPA 14 exhibited markedly increased antiviral activity against
HIV-1 and HIV-2 compared to PMEA (60- and 110-fold, respectively [28]), PMPA (300- and 540-fold,
respectively [28]), S-PMEA 8 and S-PMPA 9 (170- to 60-fold, respectively [28]). The prodrug strategy
leads to an important decrease of EC50 compared to parent compounds [28]. Nevertheless the
thiophosphonate prodrug 14 tends to display an increased EC50 (5- to 10-fold lower antiviral activity) as
compared to the phosphonate prodrug bis-POC-PMPA 12.
As described in the literature [11, 33, 51], cytotoxicity of the POM prodrugs also markedly increased.
Indeed, as shown in Table 1, bis-POM-PMEA 11 and bis-POM-S-PMEA 13 inhibited the CEM cell
growth at an CC50 of 1.6 and 2 µM, respectively. Consequently, the bis-POM prodrugs showed a
substantially lower selectivity than the bis-POC prodrugs with selectivity ratios for the POM series of ≥
1
0-35 and ≥ 270-1900 for the POC series. Both prodrug series are hydrolyzed to formaldehyde, carbon
dioxide and the parent compound [11]. Bis-POM prodrugs also generate pivalic acid, whereas the bis-
POC prodrugs release isopropyl alcohol. It is possible that the accumulation of pivalic acid contributes to
the increased cytotoxicity.
The increased antiviral activity of alkoxyalkyl ester prodrugs of the thiophosphonates bis-POM-S-PMEA
1
3 and bis-POC-S-PMPA 14 and the phosphonate compounds bis-POM-PMEA 11 and bis-POC-PMPA
8

ACCEPTED MANUSCRIPT
1
2, can be attributed to increased cellular permeation of the prodrugs followed by effective intracellular
conversion to the parent compounds, which undergo subsequent phosphorylation to their antivirally
active diphosphate metabolite.
To conclude anti-HIV evaluation, thiophosphonates 13 and 14 are slightly less active (4 to 10 fold) than
phosphonate compounds Adefovir Dipivoxil 11 and Tenofovir Disoproxil 12.
2
.3 Anti-HBV activity of thiophosphonate prodrugs 13 and 14 compared to phosphonate prodrugs 11 and
2 and cytotoxicity in Huh7 cell cultures.
1
Since prodrug forms of PMEA (Adefovir Dipivoxil) and PMPA (Tenofovir Disoproxil) are clinically
approved to inhibit HBV replication in infected patients, the compounds were tested in a cell-based
antiviral drug susceptibility assay. The inhibition of WT HBV replication was assessed in Huh7 cell
cultures treated with increasing concentrations of compounds 11, 12, 13 and 14. The amount of viral
DNA produced in the cells was quantified to evaluate the antiviral properties of the compounds as the
5
0% reduction of viral DNA (EC50) (Table 2). Concomitantly, the cytotoxic effect of each compound was
evaluated as well.
Under the assay conditions, all compounds significantly inhibited the multiplication of HBV in huh-7 cell
cultures (Table 2). None of the tested compounds exhibited any cytotoxicity at a concentration as high as
2
5 µM. Surprisingly, the alkoxyalkyl ester prodrugs bis-POM-PMEA 11, bis-POC-PMPA 12, bis-POM-
S-PMEA 13 and bis-POC-S-PMPA 14 exhibited only a small increase of efficacy against HBV compared
to PMEA (4-fold [28]), PMPA (5-fold [28]), S-PMEA 8 (3-fold [28]) and S-PMPA 9 (3-fold [28]). The
prodrug strategy leads to a moderate decrease of EC50 compared to parent compounds [28].
To conclude anti-HBV evaluation, thiophosphonates 13 and 14 displayed the same activity (1.2 to 1.7
fold) than phosphonates 11 and 12.
9

ACCEPTED MANUSCRIPT
2
.4 Evaluation of thiophosphonate prodrugs 13 and 14 compared to phosphonate prodrugs 11 and 12
against other DNA viruses in HEL cell culture.
To further explore the antiviral selectivity of the compounds, the antiviral activity testing against a panel
of DNA and RNA viruses was performed with herpes simplex virus type 1 and 2 (HSV-1 KOS, HSV-1
TK- KOS and HSV-2 G), and vaccinia virus. The inhibition of viral replication was assessed in HEL cell
cultures treated with increasing concentrations of compounds 11, 12, 13 and 14. The virus-induced
cytopathic effect (CPE) was recorded microscopically at 3 days post infection and the antiviral activity
(
EC ) was expressed as the percentage of inhibition of the CPE compared with the untreated controls
50
(Table 3).
Under the assay conditions, except for bis-POC-S-PMPA 14 which is inactive up to 50 µM, all
compounds inhibited the multiplication of HSV-1 and HSV-2 in HEL cell cultures (Table 3). None of the
tested compounds exhibited any cytotoxicity at a concentration as high as 50 µM.
The alkoxyalkyl ester prodrugs bis-POM-S-PMEA 13, bis-POM-PMEA 11 and bis-POC-PMPA 12
exhibited increased potency compared to S-PMEA 8 (6- to 10-fold [28]), PMEA (100- to 400-fold [28])
and PMPA (> 20-fold [28]). However, the increase of antiviral efficacy was less pronounced for
thiophosphonate prodrugs than phosphonates prodrugs. The alkoxyalkyl ester prodrugs 11 and 13
exhibited an efficacy equivalent to that of cidofovir and showed a better activity than acyclovir against
thymidine kinase-deficient HSV-1 (TK- KOS).
As described previously [28], S-PMPA 9 displayed a moderate activity against HSV-1 and HSV-2
(EC50 169 ± 13 µM and 48 ± 0.0 µM, respectively) in HEL cell cultures, while PMPA was inactive at 200
µM. Surprisingly, no antiviral activity up to 50 µM was observed against any of the DNA viruses tested
with the alkoxyalkyl ester prodrug bis-POC-S-PMPA 14, while the bis-POC-PMPA 12 diplayed a good
activity against HSV-1 and HSV-2 (EC 5.8 µM) in HEL cell cultures.
50
1
0

ACCEPTED MANUSCRIPT
While S-PMEA 8 showed modest inhibitory properties against vaccinia virus in contrast to the inactivity
of PMEA at 200 µM, surprisingly the alkoxyalkyl ester prodrug of PMEA 11 was 7 times more active
than S-PMEA 13. It is also intriguing that the prodrug derivative of PMEA was much more active than
the corresponding prodrug of S-PMEA.
2
.5 Decomposition Studies and Metabolism.
It has been previously [52] demonstrated that the mechanism by which the bis-POM-PMEA 11 and bis-
POC-PMPA 12 are converted to the corresponding PMEA and PMPA parent drugs involves, in a first
step, a carboxyesterase-mediated decomposition process followed by a spontaneous C-O bond breakage
and thus giving rise to the corresponding phosphonate monoesters “mono-POM-PMEA” and “mono-
POC-PMPA”, respectively, with the release of formaldehyde and pivalic acid or isopropyl alcohol. In a
second step, the corresponding phosphonate monoesters are converted to the expected parent compounds
PMEA and PMPA either following a decomposition process similar to that involved in the first step
and/or by phosphodiesterase-mediated hydrolysis.
In order to determine whether bis-POM-S-PMEA 13 and bis-POC-S-PMPA 14 exert their biological
effects via intracellular delivery of the corresponding S-PMEA 8 and S-PMPA 9, we decided to establish
the decomposition pathways of 13 and 14 in cell extracts and to determine their stability in various media.
The decomposition products, pathways and kinetics of thiophosphonates S-PMEA 8, S-PMPA 9, mono-
POC-S-PMPA 16, bis-POM-S-PMEA 13, bis-POC-S-PMPA 14 and phosphonates PMEA, PMPA, mono-
POC-PMPA 15, bis-POM-PMEA 11 and bis-POC-PMPA 12 were determined upon incubation at 37°C in
culture medium (RPMI-1640 + 10% of heat-inactivated fetal calf serum; DMEM + 10% of heat-
inactivated fetal calf serum), and in cell extracts (CEM; Huh7) in order to mimic their behavior under the
experimental conditions of the in vitro assays. All compounds were also incubated in water, in acidic
buffer (pH 1.2), in RPMI-1640 and in DMEM medium to differentiate between chemical and enzyme-
1
1

ACCEPTED MANUSCRIPT
mediated hydrolysis. Additionally, compounds 8, 9, 11, 12, 13, 14, 15 and 16 were incubated in buffer
with or without addition of purified pig liver carboxyesterases (PLE) and/or with snake venom
phosphodiesterases (SVP). PLE is a widely accepted model of intracellular esterase activity [53-55].
During incubation, samples were removed and analyzed using an “on-line cleaning” method [56]which
allows the direct HPLC analysis of compounds and metabolites in biological media without any pre-
treatment.
For each of the conditions studied, the whole sample was incubated at 37 °C and at each chosen time
interval, 50 µL were sampled and directly injected onto an analytical HPLC column where the analytical
conditions allowed visualising and properly separating the starting compound from the corresponding
metabolites. All compounds could accurately be UV-detected (260 nm) and quantified. Identification of
the metabolites was confirmed by comparative injection of authentic samples. For example, after
incubation of compound 14 in the various investigated media, aliquots were HPLC-analyzed and several
signals could be observed in resulting chromatograms: bis-POC-S-PMPA 14, mono-POC-S-PMPA 16
(two diastereoisomers), S-PMPA 9 and PMPA. All signals were unambiguously assigned by co-injection
with authentic samples (retention time, UV spectra).
Having assigned the various HPLC signals, the kinetic data could be treated according to a general
“
consecutive-competitive pseudo-first order” model. A proposed decomposition pattern is shown in
Scheme 2, in which we established that the overall metabolic sequence implies two steps and the
formation of an intermediate mono-POC or mono-POM during the conversion process. Each step of
decomposition could be characterized by half-lives (t
½ 1 2
) and rate constants (k and k ). Moreover, the
observed rate constant k is the sum of the rate constants of the chemical-mediated hydrolysis (k1c and k2c
)
and the enzyme-mediated hydrolysis (k1e and k2e) in the serum-containing media and cell extracts (table 4).
In RPMI-1640 or DMEM, the rate constants obtained reflected the chemical hydrolysis kinetics. In both
conditions, the rate constants k
higher (5- to 9-fold) than the rate constants of the thiophosphonates diesters 13 and 14, indicating that
thiophosphonates were more resistant to chemical hydrolysis than the corresponding phosphonates (t1/2
1
for the first step of conversion of phosphonates diesters 11 and 12 were
≈
1
2

ACCEPTED MANUSCRIPT
1
2
to 2 h). For all tested compounds, the second conversion (k ) was very slow and half-lives were higher
than 72 or 48 h.
When tested in FCS-containing culture medium (RPMI-1640 + 10% FCS), the rate of decomposition (k
resulted predominantly from the chemical degradation but slightly decomposition of compounds 11, 12,
3 and 14 occurred, due to enzymatic degradation. k1e was slow and equivalent for thiophosphonates
1
)
1
-3
-1
-3
-1
(
1.34 and 1.06x10 min ) and phosphonates derivatives (1.58 and 1.1x10 min ). In contrast, in DMEM
+
10% FCS, the rate constant k1e expressing the enzymatic degradation was higher especially for
-3
-1
phosphonate derivatives (6.93 and 6.41x10 min ). Such behaviour was correlated to the presence of
heat-inactivated serum in the medium and demonstrated a higher remaining enzymatic activity in DMEM
+
10% FCS than in RPMI-1640 + 10% FCS. For both series, the k
were higher than 72h or 48h for compounds 13 and 14 and were 55h/31h and 33h/> 48h for compounds
1 and 12.
2
values were very low and half-lives
1
In cell extracts (CEM and Huh7), we observed a difference of behaviour between the POM and POC
prodrug series, in addition to the impact of the thio modification. Indeed, bis-POM-S-PMEA 13 and bis-
-3
POM-PMEA 11 were rapidly deprotected in both cell extracts with equivalent k1e values of 2.99x10
-1
-3
-1
-3
-1
-3
-1
min and 3.36x10 min for compound 13 and 7.35x10 min and 6.93x10 min for compound 11. As
expected, thiophosphonate 13 was less efficiently deprotected than phosphonate 11. However and
2
unfortunately, the k values for compound 13 were very low and half-lives were higher than 72h or 48h,
meaning that the parent compound 8 was not efficiently released in cell extracts. In comparison, PMEA
was released in cell extracts with a half-live of 11 h, being more compatible with the experimental
conditions of the in vitro assays.
The combination of POC protection and thio modification was worse in terms of parent compound
release. Indeed, the removal of the first POC protecting group in cell extracts (CEM and Huh7), was less
efficient in comparison to the POM-protected (thio)phosphonates. Bis-POC-S-PMPA 14 was slowly
-3
-1
-3
-1
hydrolysed in both cell extracts with k1e values of 0.77x10 min and 0.11x10 min . Moreover, the k
2
1
3

ACCEPTED MANUSCRIPT
values for compound 14 were very low and half-lives were higher than 72h or 48h, meaning that the
parent compound 9 was not efficiently released in the cell extracts.
However, caution must be taken in extrapolating the kinetic findings to intact living cells: the
concentration of protein in the cell lysate used in these experiments is certainly lower than that in intact
cells; it is also possible that some of the enzymes responsible for metabolism in intact cells are bound to
membranes or organelles and are removed during centrifugation or that enzyme activity is suboptimal
when present in a crude extract compared to the (compartmentalized) intact cells.
Nevertheless, the foregoing studies demonstrated that bis-POM-S-PMEA 13 and bis-POC-S-PMPA 14
were prodrugs of S-PMEA 8 and S-PMPA 9, but the kinetics of release was rather slow. The studies also
revealed that mono-POM-S-PMEA and mono-POC-S-PMPA were intermediates in the conversion
process and led to the establishment of a coherent kinetic scheme for the overall metabolic sequence.
We decided to investigate more carefully the metabolism of the thiophosphonate prodrugs to better
understand the decomposition pathway. Therefore, compounds 8, 9, 11, 12, 13, 14, 15 and 16 were
incubated in buffer with or without addition of purified pig liver carboxyesterases (PLE) and/or snake
venom phosphodiesterases (SVP).
As described previously, each step of decomposition could be characterized by half-lives (t
constants (k and k ). Moreover, the observed rate constant k is the sum of the rate constants of the
chemical-mediated hydrolysis (k1c and k2c) and the enzyme-mediated hydrolysis (k1e and k2e).
½
) and rate
1
2
The half-lives were deduced from the experimental data and the calculated rate constants of each
decomposition step in the different media are reported in Table 5.
Pig liver carboxyesterase (PLE) was used as a model for hydrolysis of the bis-POC and bis-POM
prodrugs by intracellular esterase(s) [53-55]. At a concentration of 0.5 U/mL of carboxyesterase, the
formation of the mono-POC or mono-POM intermediates was fast at 37°C, with k1e values from 60 to
-3
-1
1
30x10 min but at this enzyme concentration, none of the parent compound formation was observed
after several hours of incubation. The decomposition of thiophosphonate diesters 13 and 14 into
monoesters was PLE-catalyzed, hence k and k1e values were equivalent meaning that the chemical
1
1
4

ACCEPTED MANUSCRIPT
hydrolysis was negligible. On the contrary, the decomposition of phosphonate diesters 11 and 12 into
-3
-1
1
monoesters was predominantly chemical-mediated with a k value (230x10 min ) markedly higher
-3
-1
than the k1e value (60x10 min ).
A relatively high concentration (5 U/mL) of enzyme was required to observe the release of parent
-3
-1
thiophosphonates S-PMEA and S-PMPA, with k
2
values reaching 1.78 to 3.38x10 min . Under these
conditions, phosphonate monoesters were really poor substrates for carboxyesterase with a t1/2 higher than
-3
-1
7
2
hours (k < 1.44x10 min ).
These model experiments have indicated that conversion of thiophosphonate diesters 13 and 14 into
monoesters catalyzed by intracellular esterase(s) was efficient and thiophosphonate diesters 13 and 14
were better substrates for intracellular esterase(s) than the corresponding phosphonate diesters 11 and 12.
Esterase-catalyzed hydrolysis of thiophosphonate monoesters into parental thiophosphonates S-PMEA 8
and S-PMPA 9 was plausible but unfavorable. It has been shown that a carboxyesterase partially purified
from rat liver was able to hydrolyse only one of the two carbethoxy groups of malathion and could attack
only non-ionized substrates [57].
Snake venom phosphodiesterase (SVP) was used as a model for hydrolysis of the bis-POC and bis-POM
prodrugs by intracellular phosphodiesterase(s). The incubation of thiophosphonate diesters 13 and 14
-3
-1
with a concentration of 0.5 U/mL SVP, led to formation of monoesters with k
1
values of 4.8x10 min
-3
-1
and 4.62x10 min . In comparison, phosphonate diesters 11 and 12 were very rapidly hydrolyzed with k
1
-3
-1
-3
-1
values of 170x10 min and 140x10 min . However, this hydrolysis rate corresponded to a chemical
degradation in the buffer employed and therefore, compounds 11-14 were found not to be significant
substrates of SVP. On the contrary, the phosphonate monoesters were good substrates of SVP and were
-
3
-1
rapidly hydrolyzed into parent compounds PMEA and PMPA with k
2
values of 140x10 min and
20x10 min . Unfortunately, thiophosphonate monoesters were very poor substrates of SVP and k
values could not be accurately determined.
-
3
-1
1
2
To confirm this difference of behaviour between phosphonate monoester and thiophosphonate monoester,
mono-POC-PMPA 15 and mono-POC-S-PMPA 16 were incubated with 0.5 U/mL SVP. It should be
1
5

ACCEPTED MANUSCRIPT
noted that mono-POC-S-PMPA 16 was synthesized as a diastereoisomeric mixture. Diastereoisomers
t
were separated by HPLC analysis, the more polar isomer was designated “ fast „ for fast-eluting ( R 29.1
min in the S3 system analysis) and the more hydrophobic isomer was designated „ Slow “ for slow-
t
eluting ( R 29.3 min in the S3 system analysis).
Following the phosphodiesterase-catalyzed hydrolysis of thiophosphonate monoester 16 into parent
compound S-PMPA 9 by HPLC, we observed that only the „Slow“ diastereoisomer was smoothly
-3
-1
hydrolyzed by SVP (k
2
value of 3.85x10 min ). Indeed, the observed signal area of the „Fast“ isomer
remained constant, meaning that this isomer was not substrate of SVP. In comparison, mono-POC-PMPA
-3
-1
was a really good substrate of SVP (k value of 230x10 min ). This result may indicate that
2
thiophosphonate monoesters were still poor substrates of SVP and that the SVP showed clear
stereopreference.
To propose a coherent decomposition scheme, compounds 11-14 were incubated with 0.5 U/mL of PLE
and SVP as a mixture of enzymes. As we hypothesized, PLE should be involved in the first
decomposition step and SVP should be involved in the second decomposition step.
As observed previously, the decomposition of thiophosphonate diesters 13 and 14 into monoesters was
PLE-catalyzed and the decomposition of phosphonate diesters 11 and 12 into monoesters was
predominantly chemical-mediated (k
1
and k1e values). The phosphonate monoesters were good substrates
values of
40x10 min and 60x10 min . On the contrary, thiophosphonate monoesters were really poor
of SVP and were rapidly hydrolyzed into the parent compounds PMEA and PMPA with k
2
-3
-1
-3
-1
1
-3
-1
-3
-
substrates of SVP and k values were low (k values of 3.65x10 min and 5.63x10 min ).
2
2
In summary, this study allowed to propose a reliable decomposition scheme for thiophosphonate prodrugs
2 and 14 in comparison to the phosphonate prodrugs 11 and 13, as reported in Scheme 3.
1
In cell extracts, bis-POM and bis-POC-thiophosphonates were readily transformed to mono-POM and
mono-POC-thiophosphonates. Very likely, this transformation occurred upon carboxyesterase activation
of the thiophosphonate diester to the corresponding thiophosphonate monoester, which proved stable and
was easily detected. The intermediate monoester was a poor substrate for cellular carboxyesterase(s) and
1
6

ACCEPTED MANUSCRIPT
for cellular phosphodiesterase(s). However, the parent compounds S-PMEA and S-PMPA were still
efficiently released in cell extracts. In comparison, phosphonate diesters were hydrolyzed chemically
predominantly into monoesters, which were then good substrates for intracellular phosphodiesterase(s)
In culture media, the same degradation pathway was observed, but the carboxyesterase-controlled step
was significantly slower.
We observed also a minor degradation of parent compound S-PMEA and S-PMPA into PMEA and
PMPA. This desulfuration may be mediated by phosphodiesterases [58]. It has been shown that
phosphoramidases were also able to transform nucleoside 5’-O-phosphorothioates to nucleoside 5’-0-
phosphates [59]. Additional studies concerning the elucidation of the mechanism of the enzyme-catalyzed
desulfuration process should be performed.
3
. Conclusion
In conclusion, we have investigated a novel practical synthetic procedure to efficiently obtain
thiophosphonates S-PMEA 8 and S-PMPA 9 and their prodrug forms bis-POM-S-PMEA 13 and bis-
POC-S-PMPA 14 using a common strategy, thereby limiting the number of synthetic steps. Compound 13
bis-POM-S-PMEA displayed interesting activity against HIV, HBV, HSV-1 and HSV-2 and compound
1
4 bis-POC-S-PMPA displayed potent activity against HIV-1, HIV-2 and HBV, but not HSV. Generally,
the prodrug strategy led to an important increase of antiviral efficacy, especially against HIV for 13 and
4 and HSV for 13. Sometimes (in case of PMEA) the increased antiviral activity was accompanied by an
1
increase of the cytotoxicity. The thiophosphonate prodrugs 13 and 14 displayed the same antiviral
specificity as the corresponding phosphonate prodrugs 11 and 12.
Bis-POM-S-PMEA 13 and bis-POC-S-PMPA 14 behaved as prodrugs of S-PMEA and S-PMPA. Mono-
POM-S-PMEA and mono-POC-S-PMPA 16 were intermediates in the conversion process and permitted
the proposition of a coherent kinetic scheme for the overall metabolic conversion sequence.
Thiophosphonate diesters were good substrates of carboxyesterase(s) leading to the formation of
1
7

ACCEPTED MANUSCRIPT
thiophosphonate monoesters, which were then rather poor substrates of phosphodiesterase(s).
Consequently, the parent compounds S-PMEA 8 and S-PMPA 9 were released smoothly in the cell
extracts. Furthermore, the release of S-PMEA or S-PMPA in the cell extracts may explain the unexpected
antiviral activity of this kind of bis-POC / bis-POM prodrugs in infected cell cultures. Moreover, the
metabolism in cells of prodrug forms to release parent compounds is sufficiently different between
thiophosphonate and phosphonate series to induce important differences in antiviral efficacy. Because,
thiosphosphonate prodrugs possess very long half-life in cell extracts, parent compounds are not
effectively released and accumulated in cells and this is the limiting factor for antiviral efficacy. On the
contrary, phosphonate prodrugs are quickly deprotected in cells, parent compounds are released
efficiently, phosphorylated and could act effectively as chain terminators.
Finally, with this understanding of the conversion pathway and kinetic degradation of bis-POM-S-PMEA
1
3 and bis-POC-S-PMPA 14 in various biological media, it should be possible to rationally design other
prodrugs to identify compounds with better metabolic and pharmacological characteristics for clinical use.
In particular, the slow removal of the two protecting groups represents a first target for improvement.
Optimizing the delivery of these active nucleoside thiophosphonates remains our major goal for the
successful treatment of viral infections with these type of agents.
4
. Experimental
4
.1 Chemistry
1
13
31
The H NMR, C NMR and P NMR (Nuclear Magnetic Resonance) spectra were determined with a
BRUKER AMX 250 MHz, a BRUKER Avance DPX 200 MHz and a BRUKER Avance III 600 MHz.
Chemical shifts were expressed in ppm and coupling constants (J) are in hertz (s = singlet, d = doublet, dd
=
double doublet, ddd = double-double doublet, t = triplet, dt = double triplet, m = multiplet). FAB High
Resolution Mass Spectra (HRMS) were recorded in positive-ion mode on a JEOL SX 102 mass
spectrometer using a cesium ion source and a glycerol/thioglycerol matrix; ESI High Resolution Mass
Spectra were recorded in the negative-ion mode on a Micromass Q-TOF Waters (Laboratoire de Mesures
1
8

ACCEPTED MANUSCRIPT
Physiques RMN, USTL, Montpellier, France). Flash column chromatographies were performed on a
®
®
MPLC SP4 system (Biotage ) with Biotage SNAP cartridge HP-Sil columns. Analytical thin layer
chromatographies (TLC) were performed on silica gel 60F 254 aluminium plates (Merck) of 0.2 mm
thickness. The spots were examined with a UV light and Molybdenum blue spray. HPLC was performed
on a Waters 600E controller system equipped with a 2998 photodiode array detector (detection 260 nm),
auto-injector 717 and on-line degazer. HPLC-MS Coupling analyses were performed on a Thermofisher
Accela 600 HPLC system (detection 260nm) and a Finnigan survery MSQ (MS) with elecrospray
ionization mode (ESI +/-). The UV-detector was connected to the mass spectrometer, and the whole
eluting flow was both UV and mass analyzed. Analyses were performed on a column Hypersil Gold C18
(2.1 x 50mm, 1.9 ꢀm) and samples were eluted at a flow of 550 ꢀL/min and detection at 260 nm. Eluent
A: 0.1% of Formic acid (Sigma Aldrich) in water; eluant B: 0.1% of Formic acid in acetonitrile. The
gradient starts at 98% eluent A with an increase to 10% eluent B after 5 min, to 50% eluent B after 9 min
and continues at 50% eluent B until 10 min, to 100% eluent B after 11 min and continues at 100% eluent
B until 12 min.
Buffer solutions: HPLC, MPLC and HPLC-MS buffers are prepared daily. TEAB buffer: A 1M
triethylamine solution in water was prepared mixing 140 mL of a triethylamine solution with 860 mL of
deionized water. A stock solution of triethylammonium bicarbonate buffer TEAB (pH 7.5) 1M was
prepared by addition of dry-ice in a 1M triethylamine solution in water to reach pH 7.5 and filtered with
membrane 0.22 µM GV-type (Millipore).
Analytical reverse phase (RP) chromatography was carried out on a column X-Bridge (3.5 ꢀm, 4.6 x150
mm) + precolumn X-Bridge C18 (5 ꢀM, 100 Å, 10 x10 mm) and samples were eluted at a flow rate of 1
mL/min and detection at 260 nm. Eluent A: 0.05 M triethylammonium bicarbonate buffer (pH 7.5); eluent
B: solution A containing 50% of acetonitrile. S1: the gradient starts at 100% eluent A with an increase to
1
1
0% eluent B after 10 min, to 20% eluent B after 25 min, to 100% eluent B after 35 min and continues at
00% eluent B until 50 min.
1
9

ACCEPTED MANUSCRIPT
Preparative purifications were achieved on HPLC system on a preparative column X-Bridge (5 ꢀm, 10 x
2
50 mm) + precolumn X-Bridge C18 (5 ꢀM, 100 Å, 10 x10 mm) and the samples were eluted at a flow
rate of 4.5 mL/min and detection at 260 nm Eluent A: 0.05 M triethylammonium bicarbonate buffer (pH
.5); eluent B: solution A containing 50% of acetonitrile. (S2: .the gradient starts after 5 min at 100%
7
eluent A with an increase to 20% eluent B after 35 min, to 100% eluent B after 40 min continues at 100%
eluent B until 50 min).
General procedure for the synthesis of 9-[2-(diethylphosphonomethoxy)ethyl]adenine (4) and (R)-9-[2-
(diethylphosphonomethoxy)propyl]adenine (5).
A solution of 1 or 2 (1 equiv, 50 mmol) and magnesium di-tert-butoxide (1 equiv, 50 mmol) in anhydrous
N,N-dimethylformamide (440 mL) was stirred at room temperature under argon for 1h. A solution of
diethyl p-toluenesulfonyloxymethanephosphonate 3 (1 equiv, 50 mmol) in anhydrous N,N-
dimethylformamide (33 mL) was added dropwise. The mixture was stirred for 14h at 35°C. Analytical
samples (1 ꢀL) were collected, quenched with 250 ꢀL of triethylammonium bicarbonate 0.05 M and
analysed by HPLC (S1). Solution of 3 (0.5 equiv, 25 mmol) in anhydrous N,N-dimethylformamide (100
mL) and magnesium di-tert-butoxide (0.5 equiv, 25 mmol) were added under argon and stirred for
another 6h. The reaction was quenched then with acetic acid 80% (10 mL) and concentrated under
reduced pressure and co-evaporated with toluene. The residue dissolved in brine (500 mL) was extracted
twice with hot chloroform (500 mL). The organic extracts were dried over magnesium sulfate, filtered,
and concentrated under reduced pressure. The residue was purified by silica gel flash chromatography
[
eluent, stepwise gradient of methanol (0-10%) in dichloromethane] to give pure compound as a white
solid.
9
-[2-(diethylphosphonomethoxy)ethyl]adenine (4). Chemical formula : C H N O P, MW = 329.29
12 20 5 4
t
f
g/mol. Yield = 52%. According to the literature [17, 27]. R value 0.37 (DCM/MeOH, 9:1). HPLC R
2
0

ACCEPTED MANUSCRIPT
1
=
36.4 min (S1), purity > 95%. H NMR (CD OD) δ: 8.10 (s, 1H, H-2), 8.03 (s, 1H, H-8), 4.35 (t, J = 5.0
3
Hz, 2H, CH
2
N), 3.93 (dq, J = 8.0 and J = 7.0 Hz, 4H, OCH
d, JPH = 8.5 Hz, 2H, CH P), 1.11 (t, J = 7.0, 6H, OCH CH
50.67, 143.34, 119.34, 72.15 (d, J = 11.9 Hz), 66.63 (d, JCP = 166 Hz), 63.99 (d, J = 6.6 Hz), 44.54,
2
CH
3
), 3.86 (t, J = 5.0 Hz, 2H, CH
2
O), 3.76
13
(
2
2
). C NMR (CD OD) δ: 157.29, 153.71,
3 3
1
1
3
31
+
+
+
6.62 (d, J = 5.8 Hz). P NMR (CD OD) δ : 21.56. MS (GT, FAB ): 136 (Adenine+1H) , 330 (M+1H) ,
3
+
52 (M+Na) .
(R)-9-[2-(diethylphosphonomethoxy)propyl]adenine (5). Chemical formula : C13
H
22
N
5
O
4
P, MW = 343.32
t
g/mol. Yield = 53%. According to the literature [17, 27]. R
f
value 0.44 (DCM/MeOH, 9:1). HPLC R
1
=
37.2 min (S1), purity > 97%. H NMR : (CD OD) δ: 8.23 (s, 1H, H-2), 8.15 (s, 1H, H-8), 4.41 (dd, J =
3
1
4.5 Hz and J = 3.2 Hz, 1H, CH
a b
N), 4.22 (dd, J = 14.5 Hz and J = 7.7 Hz, 1H, CH N), 4.10 (m, 1H,
CHO), 4.03-3.69 (m, 6H, OCH CH and CH P), 1.33 (t, J = 7.1 Hz, 6H, OCH CH ), 1.27 (d, J = 6.2 Hz,
2
3
2
2
3
13
3
H, CH
3 3
). C RMN (CD OD) δ: 157.30, 153.70, 150.89, 143.63, 119.70, 77.65 (d, J = 12.3 Hz), 64.32
(
(
(
d, JCP = 167 Hz), 64.08 (d, J = 6.6 Hz), 63.95 (d, J = 6.6 Hz), 49.18, 16.73 (d, J = 5.8 Hz), 16.71, 16.70
31
+
+
+
3
d, J = 5.8 Hz). P RMN (CD OD) δ: 21.56. MS (GT, FAB ) : 136 (Adenine+1H) , 344 (M+1H) , 366
+
+
M+Na) , 687 (2M+1H) .
General procedure for the synthesis of 9-[2-(diethylthiophosphonomethoxy)ethyl]adenine (6) and (R)-9-
2-(diethylthiophosphonomethoxy)propyl]adenine (7).
[
A solution of 4 or 5 (1 equiv, 10 mmol) and 2,4-Bis(methylthio)-1,3,2,4-dithiadiphosphetane-2,4-
disulfide (1 equiv, 10 mmol) in anhydrous benzene (450 mL) was stirred at 80°C for 2h. The mixture was
then stirred at 0°C for 2h. After a filtration through a celite pad, benzene was concentrated under reduced
pressure and the residue was purified by silica gel flash chromatography [eluent, stepwise gradient of
methanol (0-5%) in dichloromethane] to afford the desired product as a white powder.
2
1

ACCEPTED MANUSCRIPT
9
-[2-(diethylthiophosphonomethoxy)ethyl]adenine (6). Chemical formula : C12
H
20
N
5
O
3
PS, MW =
t
1
3
45.36 g/mol. Yield = 43%. R
NMR : (CD
H, OCH CH
NMR : (CD OD) δ: 157.27, 153.63, 150.69, 143.55, 119.88, 72.37 (d, J = 132.9 Hz), 71.87 (d, J = 8.6
f
value 0.42 (DCM/MeOH, 9:1). HPLC R 40.9 min (S1), purity > 85%. H
3
OD) δ: 8.11 (s, 1H, H-2), 8.06 (s, 1H, H-8), 4.32 (t, J = 4.7 Hz, 2H, CH
2
N), 3.98 – 3.83 (m,
13
6
2
3
and CH O), 3.76 (d, JPH = 4.2 Hz, 2H, CH P), 1.09 (t, J = 7.1 Hz, 6H, OCH CH ). C
2
2
2 3
3
CP
31
+
Hz), 63.74 (d, J = 6.9 Hz), 44.68, 16.49 (d, J = 6.5 Hz). P NMR : (CD
3
OD) δ: 88.02. MS (TOF, ESI ) :
+
+
+
3
46 (M+1H) . HRMS (TOF, ESI ) calcd for C H N O PS (M+H) 346.1103, found 346.1105.
1
2
21
5
3
(R)-9-[2-(diethylthiophosphonomethoxy)propyl]adenine (7). Chemical formula : C13
H
22
N
5
O
3
PS, MW =
t
1
3
59.38 g/mol. Yield = 44%. R
f
value 0.49 (DCM/MeOH, 9:1). HPLC R 41.7 min (S1), purity > 85%. H
NMR : (CD
3
a
OD) δ 8.11 (s, 1H, H-2), 8.05 (s, 1H, H-8), 4.26 (dd, J = 14.3 Hz and J = 2.8 Hz, 1H, CH N),
4
.08 (dd, J = 14.3 Hz and J = 8.1 Hz, 1H, CH N), 4.00 – 3.73 (m, 6H, OCH CH and CH P), 4.03-3.69
b
2
3
2
1
3
(m, 6H, CHO), 1.33 (t, J = 7.1 Hz, 6H, OCH
2
3 3 3
CH ), 1.27 (d, J = 6.2 Hz, 3H, CH ). C NMR (CD OD) δ
1
57.60, 153.60, 150.91, 143.89, 119.72, 77.35 (d, J = 9.1 Hz), 70.06 (d, JCP = 134.6 Hz), 63.73 (d, J =
31
7
.3 Hz), 63.62 (d, J = 7.3 Hz), 16.81, 16.55(d, J = 1.3 Hz), 16.45 (d, J = 1.1 Hz). P NMR : (CD
3
OD) δ :
+
+
+
+
8
8.407. MS (TOF, ESI ) : 360 (M+1H) . HRMS (TOF, ESI ) calcd for C H N O PS (M+H) 360.1260,
1
3
23
5
3
found 360.1254.
General procedure for the synthesis of 9-[2-(thiophosphonomethoxy)ethyl]adenine (8) and (R)-9-[2-
thiophosphonomethoxy)propyl]adenine (9).
(
A solution of 6 or 7 (1 equiv, 3 mmol) and sodium bromide (40 equiv, 160 mmol) in anhydrous N-
methylpyrrolidinone (20 mL) was stirred vigouresly under argon. Trimethylsilyl chloride (40 equiv, 160
mmol) was added dropwise and the mixture was heated at 90°C for 12h. The reaction was quenched with
5
mL of triethylammonium bicarbonate 0.05 M and stirred for 1h. After solvents were concentrated in
vacuo and freezed-drying, the residue was purified by reversed-phase column chromatography (S2).
2
2

ACCEPTED MANUSCRIPT
Product fractions were collected, evaporated to dryness and lyophilized. Excess of triethylammonium
bicarbonate was removed by repeated freezed-drying with deionized water. The residue was dissolved in
+
water and eluted on a Dowex 50WX2 column (Na exchange) to give the pure compound as disodium salt
as a white powder after freeze-drying.
9
-[2-(thiophosphonomethoxy)ethyl]adenine (8). Chemical formula : C
8
H
10
N
5
Na
O) δ 8.13 (s, 1H, H-8), 8.01 (s,
H, H-2), 4.28 (t, J = 4.8 Hz, 2H, CH N), 3.90 (t, J = 4.8 Hz, 2H, CH O), 3.62 (d, J = 5.2 Hz, 2H, CH P).
2 3
O PS, MW = 333.22
t
1
g/mol. Yield = 40%. HPLC R 19.5 min (S1), purity > 98%. H NMR : (D
2
1
2
2
2
1
3
2
C NMR : (D O) δ 152.23, 149.12, 146.98, 141.98, 116.38, 73.01 (d, JCP = 120 Hz), 68.57 (d, JCP = 10.6
3
1
+
+
Hz), 42.27. P NMR : (D O) δ 61.76 (t, J = 5.2 Hz). MS (TOF, ESI ) : 290 (M+1H) . HRMS (TOF,
2
+
+
8 13 5 3
ESI ) calcd for C H N O PS (M+H) 290.0477, found 290.0576.
(
R)-9-[2-(thiophosphonomethoxy)propyl]adenine (9). Chemical formula : C
9
H
12
N
5
Na
2 3
O PS, MW =
t
1
3
7
1
47.24 g/mol. Yield = 48%. HPLC R 21.7 min (S1), purity > 99%. H NMR (D O) δ 8.12 (s, 1H, H-8),
2
.98 (s, 1H, H-2), 4.22 (dd, J = 3.1 Hz and J = 14.6 Hz, 1H, CH
a
N), 4.05 (dd, J = 6.7 Hz and J = 14.6 Hz,
H, CH N), 3.91 (m, 1H, CHO), 3.62 (dd, J = 5.6 Hz and J = 12.7 Hz, 1H, CH
b
a
P), 3.46 (dd, J = 6.1 Hz
13
and J = 12.8 Hz, 1H, CH
b
P), 1.05 (d, J = 6.2 Hz, 3H, CH
3
). C NMR (D
2
O)
δ
152.65, 149.12, 147.25,
3
1
1
42.25, 116.26, 74.46 (d, JCP = 10.7 Hz), 70.89 (d, JCP = 121.1 Hz), 46.39, 14.73. P NMR (D O) δ 61.11
2
+
+
+
+
(
9 15 5 3
t, J = 5.6 Hz). MS (TOF, ESI ) : 304 (M+1H) . HRMS (TOF, ESI ) calcd for C H N O PS (M+H)
-
3
04.0633, found 304.0619 (M) .
General
procedure
for
the
synthesis
of
9-{2-[O,O’-
(R)-9-{2-[O,O’-
bis(pivaloyloxymethyl)phosphonomethoxy]ethyl}adenine
(11)
and
bis(isopropoxycarbonyloxymethyl)phosphonomethoxy]propyl}adenine (12).
2
3

ACCEPTED MANUSCRIPT
A solution of 4 or 5 (1 equiv, 10 mmol) and sodium bromide (2.1 equiv, 21 mmol) in anhydrous N-
methylpyrrolidinone (10 mL) was stirred vigouresly under argon. Trimethylsilyl chloride (4 equiv, 40
mmol) was added dropwise and the mixture was heated at 60°C for 12h. The reaction was monitored by
HPLC until complete disparition of the starting material (S1). When completion was observed, the
reaction mixture was diluted with anhydrous ethyl acetate (16 mL) and stirred for 5 min. The ethyl acetate
and remaining trimethylsilyl chloride were then concentrated in vacuo at 60°C. Anhydrous isopropyl
alcohol (1.5 mL) was then added to the reaction mixture and stirred for 15 min. The mixture was diluted
with anhydrous ethyl acetate (16 mL) and solvents were evaporated under reduced pressure at 60°C.
Chloromethyl pivalate (POMCl, 5 equiv, 50 mmol) for the synthesis of 11 or chloromethyl
isopropylcarbonate 10 [50] (POCCl, 5equiv, 50 mmol) for the synthesis of 12 and triethylamine (3 equiv,
3
0 mmol) were added to the mixture and the reaction mixture was stirred at 60°C for 6h. The reaction was
diluted in anhydrous ethyl acetate (16 mL), stirred for 5 min, and the ethyl acetate and remaining
triethylamine were concentrated in vacuo at 60°C. The reaction mixture was diluted in ethyl acetate (160
mL) and extracted with water (60 mL), dried over magnesium sulfate, filtered, and concentrated to
dryness under reduced pressure. The resulting oil was purified by silica gel flash chromatography [eluent,
stepwise gradient of methanol (0-5%) in dichloromethane] to afford the desired product as a white
powder.
9
-{2-[O,O’-Bis(pivaloyloxymethyl)phosphonomethoxy]ethyl}adenine (11). Chemical formula
P, MW = 501.47 g/mol. Yield = 27%. According to the literature [33]. R value 0.32
DCM/MeOH, 9.5:0.5). HPLC R 42.5 min (S1), purity > 90%. H NMR (CDCl ) δ 8.30 (s, 1H, H-8),
:
C
20
H
32
N
5
O
8
f
t
1
(
3
7
.90 (s, 1H, H-2), 6.16 (s, 2H, NH
2
), 5.69-5.54 (m, 4H, OCH2O), 4.37 (t, J = 4.9 Hz, 2H, CH
2
N), 3.92 (t,
13
J = 4.9 Hz, 2H, CH
2
O), 3.83 (d, J = 7.6 Hz, 2H, CH
2
P), 1.18 (s, 18H, CH
3
(t-Bu)). C NMR (CDCl ) δ
3
1
3
55.76, 153.00, 149.99, 141.43, 119.52, 81.76 (d, J = 6.3 Hz), 71.48, 65.63 (d, J = 166.1 Hz), 43.43,
31
+
+
+
8.79, 26.89. P NMR : (CDCl ) δ 20.56. MS (TOF, ESI ) : 502 (M+1H) . HRMS (TOF, ESI ) calcd for
3
+
C
20
H
33
N
5
O
8
P (M+H) 502.2067, found 502.2057.
2
4

ACCEPTED MANUSCRIPT
(R)-9-{2-[O,O’-Bis(isopropoxycarbonyloxymethyl)phosphonomethoxy]propyl}adenine (12). Chemical
formula : C H N O P, MW = 519.44 g/mol. Yield = 33%. According to the literature [51]. R value
19
30
5
10
f
t
1
0
8
.32 (DCM/MeOH, 9.5:0.5). HPLC R 42.6 min (S1), purity > 90%. H NMR (CDCl
3
) δ 8.31 (s, 1H, H-8),
.02 (s, 1H,H-2), 6.49 (s, 2H, NH ), 5.70 – 5.57 (m, 4H, OCH
2
2
O), 4.96 - 4.87 (m, 2H, CH(i-Pr)), 4.37 (dd,
P), 3.72
J = 14.5 and 3.1 Hz, 1H, CHaN), 4.15 (dd, J = 14.5 and 7.3 Hz, 1H, CHbN), 3.94 (m, 2H, OCH
2
1
3
(
dd, J = 13.9 and 9.0 Hz, 1H, CHO), 1.32 - 1.29 (m, 12H, CH
3
(i-Pr)), 1.22 (d, J = 6.2 Hz, 3H, CH
3
). C
NMR : (CDCl
3
) δ 155.29, 153.24, 153.21, 142.29, 84.44, 77.37, 77.16, 76.94, 76.71, 73.59, 63.76, 62.64,
31
+
4
8.38, 32.07, 29.84, 22.83, 21.77, 16.53, 14.24, 1.15. P NMR : (CDCl ) δ 20.99. MS (TOF, ESI ) : 520
3
+
+
+
(
31 5
M+1H) . HRMS (TOF, ESI ) calcd for C19H N O10P (M) 520.1809, found 520.1802.
General
bis(pivaloyloxymethyl)thiophosphonomethoxy]ethyl}adenine
bis(isopropoxycarbonyloxymethyl)thiophosphonomethoxy]propyl}adenine (14).
procedure
for
the
synthesis
of
9-{2-[O,O’-
(13)
(R)-9-{2-[O,O’-
A solution of 11 or 12 (1 equiv, 2.5 mmol) and 2,4-Bis(methylthio)-1,3,2,4-dithiadiphosphetane-2,4-
disulfide (1 equiv, 2.5 mmol) in anhydrous benzene (40 mL) was stirred at 80°C for 30min. The mixture
was then stirred at 0°C for 2h. After filtration through a celite pad, benzene was concentrated in vacuo
and the residue purified by silica gel flash chromatography [eluent, stepwise gradient of methanol (0-5%)
in dichloromethane] to afford the desired product as a white powder.
9
-{2-[O,O’-Bis(pivaloyloxymethyl)thiophosphonomethoxy]ethyl}adenine (13). Chemical formula :
t
C
20
H
32
N
5
O
7
PS, MW = 517.54 g/mol. Yield = 29%. R
f
value 0.39 (DCM/MeOH, 9.5:0.5). HPLC R 46.1
1
min (S1), purity > 97%. H NMR : (CDCl
3
) δ: 8.73 (s, 1H, H-8), 8.11 (s, 1H, H-2), 7.26 (s, 2H, NH
2
),
5
5
.69-5.60 (m, 4H, OCH O), 4.40 (t, J = 4.9 Hz, 2H, CH N), 3.84 (t, J = 4.9 Hz, 2H, CH O), 3.96 (d, J =
2
2
2
13
Hz, 2H, CH
2
P), 1.21 (s, 18H, CH
3
(t-Bu)). C NMR (CDCl
3
) δ 155.64, 152.68, 149.63, 143.84, 122.7,
2
5

ACCEPTED MANUSCRIPT
3
1
8
1.83 (d, J = 6.2 Hz), 71.23, 65.17 (d, J = 166.1 Hz), 48.66, 38.86, 26.95. P RMN (CDCl ) δ 88.89.
3
+
+
+
+
MS (TOF, ESI ) : 518 (M+1H) . HRMS (TOF, ESI ) calcd for C20
33 5 7
H N O PS (M+H) 518.1838,found
5
18.1851.
(R)-9-{2-[O,O’-Bis(isopropoxycarbonyloxymethyl)thiophosphonomethoxy]propyl}adenine
(14).
Chemical formula : C19
H
30
N
5
O
9
PS, MW = 535.51 g/mol. Yield = 32%. R value 0.38 (DCM/MeOH,
f
t
1
9
6
.5:0.5). HPLC R 46.3 min (S1), purity > 98%. H NMR : (CDCl
3
) δ 8.31 (s, 1H, H-8), 8.23 (s, 1H, H-2),
2 2
.49 (s, 2H, NH ), 5.72-5.54 (m, 4H, OCH O), 4.91 (tq, J = 12.3 and J = 6.2 Hz, 2H, CH(i-Pr)), 4.43 (d,
J = 13.7, 1H, CHaN), 4.16 (dd, J = 14.3 and J = 7.6 Hz, 1H, CHbN), 4.09-3.95 (m, 2H, OCH
J = 13.6 and J = 4.3 Hz, 1H, CHO), 1.32-1.29 (m, 12H, CH (i-Pr))1.31 (d, J = 6.1, 3H, CH
CDCl
8.90, 46.02, 29.82, 16.56, 8.77. P NMR (CDCl
2
P), 3.75 (dd,
1
3
3
3
). C NMR
(
3
) δ 153.24, 153.21, 151.74, 144.66, 118.45, 76.30 (d, J = 8.5 Hz), 73.58, 69,27 (d, J = 137 Hz),
31
+
+
4
3
) δ 88.82. MS (TOF, ESI ) : 536 (M+1H) . HRMS :
+
+
(
TOF, ESI ) calcd for C19H N O PS (M+H) 536.1580, found 536.1572.
31 5 9
General
procedure
for
the
synthesis
of
(R)-9-{2-[O-
(R)-9-{2-[O-
mono(isopropoxycarbonyloxymethyl)phosphonomethoxy]propyl}adenine
(15)
mono(isopropoxycarbonyloxymethyl)thiophosphonomethoxy]propyl}adenine (16).
A solution of 12 or 14 (1 equiv, 0.05 mmol) in water (5 mL) was stirred vigorously at 60°C. The reaction
was maintained at 60°C and monitored by HPLC-MS for complete conversion to 15 or 16. Typical
t
t
retention times : R =0.49 min, PMPA, ESI-MS [M+H] m/z= 288.0; R =0.48 min, 9, ESI-MS [M+H]
t
t
m/z= 304.10; R =4.78 min, 15, ESI-MS [M+H] m/z= 404.09; R =6.36 and 6.50 min (2 diastereoisomers),
t
t
1
6, ESI-MS [M+H] m/z= 420.09; R =8.04 min, 12, ESI-MS [M+H] m/z= 520.19; R =8.92 min, 14, ESI-
MS [M+H] m/z= 536.10. When completion was observed, the reaction mixture was extracted with ethyl
acetate (5 mL) and the aqueous layer concentrated in vacuo and lyophilized. The crude product was
®
purified on reverse phase C18 (biotage SNAP cartridge KP-C18-HS 33g, flow : 10 mL/min, UV-
2
6

ACCEPTED MANUSCRIPT
detector : 260 nm, Eluant A : TEAB 0.05M, eluant B : solution A containing 50% of acetonitrile, gradient
starts at 100% eluent A with an increase to 10% eluent B after 66 min, to 20% eluent B after 83 min and
to 100% eluent B after 116 min). Each fraction was analysed by HPLC (S1), pure fractions were collected
+
and eluted on a Dowex 50WX2 column (Na exchange) to give the pure compound as sodium salt as a
white powder after freeze-drying.
(R)-9-{2-[O-Mono(isopropoxycarbonyloxymethyl)phosphonomethoxy]propyl}adenine (15). Chemical
t
1
formula : C H N NaO P, MW = 425.31 g/mol. Yield = 38 %. HPLC R 37.7 min (S1), purity > 98%. H
14
21
5
7
NMR : (D
2
O) δ: 8.26 (s, 1H, H-8), 8.23 (s, 1H, H-2), 5.29 (m, 2H, OCH
.3 Hz, 1H, CH(i-Pr)), 4.40 (dd, J = 14.8 and J = 2.7 Hz, 1H, CHaN), 4.24 (dd, J = 14.8 and J = 7.7 Hz,
H, CHbN), 3.78 (dd, J = 13.1 and J = 8.7 Hz, 2H, OCH P), 3.46 (dd, J = 13.2 and J = 10.5 Hz, 1H,
), 1.18 (d, J = 6.2 Hz, 6H, CH
2
O), 4.70 (dd, J = 12.6 and J =
6
1
2
31
CHO), 1.23 (d, J = 6.3 Hz, 3H, CH
3
3
(i-Pr)). P RMN (D
2
O) δ 17.10. MS
+
+
+
(
ESI ) : 404.09 (M+H) ; 807.23 (2M+H) .
(R)-9-{2-[O-Mono(isopropoxycarbonyloxymethyl)thiophosphonomethoxy]propyl}adenine
(16).
t
Chemical formula : C14
H
21
N
5
NaO
6
PS, MW = 441.38 g/mol. Yield = 30 %. HPLC R 38.7 (fast-eluting
1
isomer) and 38.9 (slow-eluting isomer) min (S1), purity > 98%. H NMR (peaks of both diastereoisomers
A and B are described) : (D
14.9 and J = 5.6 Hz, 2H, OCH
Pr) of A and B), 4.46 – 4.41 (m, 2H, CHaN of A and B), 4.30 – 4.22 (m, 2H, CHbN of A and B), 4.09 –
.03 (m, 2H, CHO of A and B), 3.91 (dd, J = 13.2 and J = 5.1 Hz, 1H, OCH P of A), 3.83 (dd, J = 13.0
2
O) δ 8.29 (s, 4H, H-8 of A and B and H-2 of A and B), 5.50 (ddd, J = 21.0, J
=
2
O of A), 5.39 – 5.34 (m, 2H, OCH O of B), 4.83 – 4.79 (m, 2H, CH(i-
2
4
a
and J = 3.1 Hz, 1H, OCH
a
P of B), 3.64 (dd, J = 13.0 and J = 8.0 Hz, 1H, OCH
b
P of A), 3.57 (dd, J = 13.2
31
and J = 5.0 Hz, 1H, OCH
b
P of B), 1.29 – 1.21 (m, 18H, CH (i-Pr) and CH of A and B). P NMR (D O)
3
3
2
+
+
δ 72.91, 72.26 (peaks of both diastereoisomers are described). MS (ESI ) : 420.09 (M+H) ; 839.18
+
(
2M+H) .
2
7