Synthesis of a nucleobase-modified ProTide library

Article abstract of DOI:10.1021/acs.orglett.6b02764

A new method for the construction of (aryloxy)phosphoramidate nucleoside prodrugs is presented. An (aryloxy)phosphoramidate ribose derivative as key building block was used for coupling with a number of nucleobases under Vorbrüggen reaction conditions yie

Full text of DOI:10.1021/acs.orglett.6b02764

Letter
pubs.acs.org/OrgLett
Synthesis of a Nucleobase-Modified ProTide Library
Ling-Jie Gao, Steven De Jonghe, and Piet Herdewijn*
KU Leuven, Rega Institute for Medical Research, Medicinal Chemistry, Minderbroedersstraat 10, 3000 Leuven, Belgium
S
* Supporting Information
ABSTRACT: A new method for the construction of (aryloxy)phosphoramidate nucleoside prodrugs is presented. An
(aryloxy)phosphoramidate ribose derivative as key building block was used for coupling with a number of nucleobases under
Vorbruggen reaction conditions yielding the protected ProTides in excellent yields. Selective hydrolysis of the acetoxy groups on
̈
the sugar moiety afforded a series of the desired ProTides. The advantage of this approach, when compared to classical
procedures, is the greater flexibility for achieving structural variety of the nucleobase moiety.
ucleoside analogues have widespread utility as antiviral and
1
Nantitumor agents. Their antiviral or antitumor effects
depend on the sequential intracellular conversion of the parent
nucleoside into the monophosphate, the diphosphate, and then
the triphosphate metabolite. In general, for antiviral nucleosides
the nucleoside−triphosphate acts as a chain terminator for the
viral DNA or RNA polymerase. For anticancer nucleosides,
different molecular mechanisms are known, such as inhibition of
the human DNA polymerase by the nucleoside triphosphate,
inhibition of the ribonucleotide reductase by the nucleoside
diphosphate, or inhibition of thymidylate synthase by nucleoside
monophosphate.2 Notably, all these mechanisms require
phosphorylation of the nucleoside in the target cells to at least
the monophosphate level.3 One of the most successful
approaches to overcome the dependence of bioactive nucleoside
Figure 1. Synthesis of ProTides.
analogues on kinase-mediated activation is the ProTide method-
ology, based on aryloxyphosphoramidate chemistry.4 It allows
one to bypass the first (and often rate limiting) step of kinase-
mediated activation of nucleosides. This method has been
extensively applied to anti-HIV,5 anti-HCV,6 and antitumoral
nucleosides.7 (Aryloxy)phosphoramidate nucleoside prodrugs
are generally prepared by coupling of the desired nucleosides
with phosphorochloridate by either activation as an imidazolium
intermediate by treatment with N-Me-imidazole (NMI) or by
deprotonation of the primary hydroxyl group of the nucleoside
with t-BuMgCl and subsequent substitution with the chlor-
ophosphoramidate (Figure 1).8 Numerous nucleoside prodrugs
have been synthesized using one of these procedures. These
methodologies present some limitations, mainly in the formation
of byproducts, which arise from phosphorylation of unprotected
hydroxyl groups on the sugar backbone. In addition, the nature of
the nucleobase plays an important role in the reaction behavior.
Whereas uracil and adenine as bases are not usually
phosphorylated, the preparation of cytosine, guanine, diamino-
purine, and hypoxanthine analogues is more problematic due to
competitive phosphorylation of the nucleobase and/or poor
solubility of the substrate in the reaction solvents. The formation
of byproducts has been reduced by appropriate protection of the
sugar hydroxyl groups and the competitive sites on the
nucleobase moiety.
An additional drawback of the current approaches, which has
not received much attention in literature, is the fact that these
methods require the preparation of a (protected) nucleoside
derivative, from which the corresponding phosphoramidate
prodrug is synthesized. If necessary, the last step involves
removal of the protecting groups on the sugar and/or nucleobase
moiety. This is not optimal from a medicinal chemistry
standpoint. It is not amenable to high-throughput chemistry.
For each new ProTide, a novel nucleoside derivative has to be
synthesized, and in the last step, this is converted to its
(aryloxy)phosphoramidate prodrug. In the early phases of drug
discovery, medicinal chemists prefer synthetic sequences in
which structural variety is introduced at a late stage of the
synthesis. This allows for large-scale synthesis of a key
Received: September 13, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.6b02764
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
intermediate from which, by a number of simple transformations,
structural variety can be introduced. With these limitations in
mind, and as part of our ongoing interest in the synthesis of novel
ProTides,9 we sought to develop a strategy that utilizes a key
intermediate from which in a single step a variety of nucleobases
can be introduced and hence can be used for the assembly of a
nucleobase-modified ProTide library. This requires the avail-
ability of a protected ribose derivative in which the primary
hydroxyl group is derivatized as an (aryloxy)phosphoramidate
moiety. The protected D-ribose derivative 2 was prepared from D-
ribose 1, with a slightly modified literature procedure.10
Protection of the primary hydroxyl group as a monomethoxy-
trityl group was followed by acetylation of the remaining
hydroxyl groups. Acidic deprotection affords the triacetylated D-
ribose derivative 2 in 80% yield over three steps. It was feasible to
separate the mixture of α- and β-anomers of compound 2 by flash
chromatography on silica gel. However, since the configuration
of the anomeric hydroxyl group has no influence on the
diastereoselective introduction of the base moiety, the mixture 2
was used as such in the next step. Commercially available phenyl
phosphorodichloridate was reacted with the diisoamyl esterified
L-aspartic acid hydrochloride salt 39a to afford phosphorochlor-
idate 4, which was allowed to react in situ with compound 2 in the
presence of NMI in anhydrous dichloromethane. The key
intermediate 5 was isolated in 85% yield after flash chromatog-
raphy on silica gel as a mixture of four isomers due to the chirality
of the phosphorus atom and the anomeric center of the sugar
moiety (Scheme 1). This mixture moves as a single, inseparable
However, depending on the nucleobase and sugar, regioselec-
tivity in the glycosylation can be problematic, leading to
regioisomeric mixtures.
Uracil was selected as a prototype nucleobase. Two sets of
reactions conditions were explored (Table 1). N,O-bis-
(trimethylsilyl)acetamide (BSA) was selected as the silylating
agent, trimethylsilyltriflate (TMSOTf) as the Lewis acid, and
Table 1. Coupling of ProTide Sugar 5 with Nucleobases
Scheme 1. Synthesis of ProTide Sugar 5
spot on TLC. We gave this sugar derivative 5 the name ProTide
sugar. The choice of L-aspartic acid diisoamyl ester as the amino
acid motif was dictated by the fact that this often imparts higher
antiviral activity to nucleoside analogues, when compared to the
classical L-alanine.9a
Having key intermediate 5 in hand, we started to explore the
reaction conditions for its coupling with a number of
nucleobases. In nucleoside chemistry, the Vorbruggen type of
̈
coupling is the method of choice for construction of the
glycosidic bond.11 It is based on the coupling of trimethylsily-
lated heterocyclic bases with protected sugar derivatives having
an acetoxy group at the anomeric center, and it is catalyzed by a
Lewis acid. Due to anchimeric assistance from a 2′-acyloxy group,
this reaction proceeds diastereoselectively to afford exclusively
nucleosides with a β-configuration at the anomeric center.
B
DOI: 10.1021/acs.orglett.6b02764
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Table 2. Selective O-Deacetylation Conditions
entry
conditions
result
a
1 N HCl or 1 N K2CO3
rt: recovered starting material; reflux: complex reaction mixture
b
c
1 N HCl or 1 N K2CO3 in CH3OH
7 N NH3 in MeOH/CH2Cl2, rt
2 N NH3 in iPrOH, 48 h, rt
25% aq NH3, iPrOH, 2 h, rt
hydrolysis of acetyl and isoamyl esters
hydrolysis of acetyl esters and formation of L-aspartic acid dimethyl ester
hydrolysis of acetyl esters
d
e
hydrolysis of acetyl esters
acetonitrile as the reaction solvent (method A). Because of the
reflux temperature, only the N-9 glycosylated purine derivatives
6i and 6j were isolated. An independent proof of regioselectivity
was obtained by resynthesis of 6i and 6j by a reported procedure
(method C).12 The chlorine-substituted purines are ideal
substrates for further transformation into adenine and 2,6-
diaminopurine by substitution of the chloride with sodium azide,
followed by catalytic hydrogenation yielding an amino group, as
reported previously.13
In order to demonstrate the general applicability of this
approach, a number of modified nucleobases were selected. The
synthesis of the pteridine-based nucleoside 6k proceeded
smoothly and in excellent yield (entry i). Glycosylation of
methyl 1H-1,2,4-triazole-3-carboxylate (entry j) gave the desired
ProTide 6l in excellent yield, whereas glycosylation of pyridin-
2(1H)-one (entry k) gave a lower yield. On the other hand, 4-
chloropyrrolo[2,3-d]pyrimidine (also called 6-chloro-7-deaza-
purine) did not undergo coupling with ProTide sugar 5 under
presumed lability of the phosphoramidate part, the Vorbruggen
̈
glycosylation reaction was initially run at room temperature. The
desired ProTide 6a was isolated in good yield after a reaction
time of 24−48 h. The reaction time could be dramatically
reduced, without a significant drop in reaction yield, by running
the same reaction at reflux temperature. Alternatively,
hexamethyldisilazane (HMDS) could be used as the silylating
agent with tin(IV) chloride as a Lewis acid (method B).
According to this procedure, the nucleobase was first treated with
HMDS in the presence of a catalytic amount of ammonium
sulfate, after which the silylated heterocycle was coupled with
ProTide sugar 5 in 1,2-dichloroethane or acetonitrile in the
presence of SnCl4 to yield ProTide 6a in excellent yield. In the
literature, most of the SnCl4-mediated glycosylation reactions are
quenched with a saturated NaHCO3/KHCO3 solution. This very
often results in emulsions that are problematic to the workup
procedure. However, quenching the reaction mixture with a
saturated NH4Cl solution led to a good separation between the
organic and aqueous phases, making extractions much easier.
Because of the convenience of this procedure, method B was
applied for most of the subsequent compounds. Given the good
results obtained with uracil, a set of other pyrimidines was
selected (Table 1). The reactions with 6-azauridine (entry b) and
5-F-cytosine (entry c) proceeded regioselectively, and only the
N-1 regioisomers were isolated in excellent yields (using both
methods A and B). In contrast, when 5-F-uracil was used for the
different Vorbruggen reaction conditions (data not shown). As it
̈
is well known that pyrrolo[2,3-d]pyrimidine nucleosides are
harder to prepare than the corresponding purine nucleosides,
due to reactivity differences of the pyrrole versus the imidazole
system, this is not surprising.14 Having the desired protected
ProTides 6a−m in hand, we turned our attention toward the
selective O-deacetylation without affecting the diisoamyl ester
and phenoxy moieties of the prodrug part. The uridine-
containing (aryloxy)phosphoramidate prodrug 6a was studied
as a prototype example (Table 2). Initial attempts focused on
mild, water-based, hydrolysis conditions (such as 1 N HCl or 1 N
K2CO3) at room temperature (entry a). The results were
disappointing due to the poor solubility of the substrates in
water. Heating the same reaction resulted in complex reaction
mixtures. Switching from water to methanol as the reaction
solvent (1 N HCl or 1 N K2CO3 in methanol) led to concomitant
hydrolysis of the acetyl groups on the ribose part and the
isoamylester groups in the prodrug moiety (entry b).
Encouraging results were obtained when the reactants were
subjected to a 7 N ammonia solution in methanol at room
temperature using dichloromethane as reaction solvent (entry c).
Careful monitoring of the reaction time and concentration of
ammonia used was critical for selective deacetylation. Ultimately,
employing a 7 N ammonia in methanol solution and dichloro-
methane in a 1:1 ratio for 12 h at room temperature was found to
be optimal for selective hydrolysis of the acetyl ester groups.
However, the desired product was always accompanied by the
corresponding L-aspartic acid dimethyl ester derivative, formed
by transesterification reaction with methanol. Due to their very
Vorbruggen condensation, a significant amount of the N-3
̈
glycosylated product was isolated (entry d), which could be easily
separated from the N-1 isomer by flash chromatography on silica
gel. When 5-aminouracil (entry e) was selected as the
nucleobase, only the N-3 glycosylated isomer 6f was isolated.
When the experimental reaction conditions of method A were
applied for the coupling of 2-thiouracil with ProTide sugar 5
(entry g), the desired ProTide was not obtained. At room
temperature, no reaction took place and only starting material
was recovered, whereas at reflux temperature, a complex reaction
mixture was formed. When the experimental conditions of
procedure B were applied using 1,2-dichloroethane as solvent,
this resulted in a 1:2 mixture of the N-1/N-3 regioisomers (82%
combined yield).
Besides pyrimidine-like bases, a number of purine nucleobases
were also introduced. In the Vorbruggen coupling of ProTide
̈
sugar 5 with 6-chloropurine (entry g) and 2-amino-6-
chloropurine (entry h), a mixture of two regioisomers, due to
N-7 and N-9 glycosylation, was isolated if the reaction was run at
room temperature. When the same reaction was performed at
C
DOI: 10.1021/acs.orglett.6b02764
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
similar polarity, separation of both compounds by silica gel flash
chromatography was tedious and difficult.
intact. This methodology is useful for the construction of a
ProTide library with structural variation of the nucleobase
moiety.
In an attempt to avoid the formation of this side product, we
ran the same reaction in a solvent mixture of 2 N ammonia in 2-
propanol and dichloromethane and found that the reaction
proceeded sluggishly. However, running the reaction in 2 N
ammonia in 2-propanol (without dichloromethane) for 48 h at
room temperature (entry d) led to full deprotection of the acetyl
groups without formation of transesterification-based products.
In an effort to shorten the reaction time, a mixture of 25%
ammonia in water and 2-propanol (1:2 volume ratio) was
evaluated (entry e). The reaction was run at room temperature,
and the acetyl groups were completely deprotected within 2 h to
yield the desired product in good yield. These optimized reaction
conditions were applied to the protected ProTides 6b−m (Table
3). The procedure works very well for most of the substrates with
isolated yields of 80% or more.
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.orglett.6b02764.
Experimental procedures and spectral data of compounds
2, 5, 6a−m, and 7a−l (PDF)
AUTHOR INFORMATION
Corresponding Author
■
*E-mail: piet.herdewijn@kuleuven.be. Tel: +32 16 32 26 57.
Notes
The authors declare no competing financial interest.
Table 3. Selective O-Deacetylation of ProTides 6b−m
REFERENCES
■
(1) Jordheim, L. P.; Durantel, D.; Zoulim, F.; Dumontet, C. Nat. Rev.
Drug Discovery 2013, 12, 447−464.
(2) Parker, W. B. Chem. Rev. 2009, 109, 2880−2893.
(3) Thornton, P.; Kadri, H.; Miccoli, A.; Mehellou, Y. J. Med. Chem.
2016, DOI: 10.1021/acs.jmedchem.6b00523.
(4) Mehellou, Y.; Balzarini, J.; McGuigan, C. ChemMedChem 2009, 4,
1779−1791.
(5) (a) Ray, A. S.; Fordyce, M. W.; Hitchcock, M. J. M. Antiviral Res.
2016, 125, 63−70. (b) Saboulard, D.; Naesens, L.; Cahard, D.; Salgado,
A.; Pathirana, R.; Velazquez, S.; McGuigan, C.; De Clercq, E.; Balzarini,
J. Mol. Pharmacol. 1999, 56, 693−704.
yield
entry substrate
B
product
(%)
a
b
c
d
e
f
6b
6c
6d
6e
6f
6-azauridine (N-1 isomer)
7b
7c
7d
7e
7f
83
85
80
86
90
82
86
81
80
87
5-fluorocytosine (N-1 isomer)
5-fluorouracil (N-1 isomer)
5-fluorouracil (N-3 isomer)
5-aminouridine (N-3 isomer)
2-thiouridine (N-1 isomer)
2-thiouridine (N-3 isomer)
6-chloropurine (N-9 isomer)
6-chloro-2-aminopurine (N-9 isomer)
4-amino-6-methyl-7-oxopteridine
(6) Sofia, M. J.; Bao, D.; Chang, W.; Du, J.; Nagarathnam, D.;
Rachakonda, S.; Reddy, G.; Ross, B. S.; Wang, P.; Zhang, H.; Bansal, S.;
Espiritu, C.; Keilman, M.; Lam, A. M.; Micolochick Steuer, H. M.; Niu,
C.; Otto, M. J.; Furman, P. A. J. Med. Chem. 2010, 53, 7202−7218.
(7) (a) Slusarczyk, M.; Lopez, M. H.; Balzarini, J.; Mason, M.; Jiang, W.
G.; Blagden, S.; Thompson, E.; Ghazaly, E.; McGuigan, C. J. Med. Chem.
2014, 57, 1531−1542. (b) McGuigan, C.; Murziani, P.; Slusarczyk, M.;
Gonczy, B.; Vande Voorde, J.; Liekens, S.; Balzarini, J. J. Med. Chem.
2011, 54, 7247−7258.
6g
6h
6i
7g
7h
7i
g
h
i
6j
7j
j
6k
6l
7k
7l
(8) Pradere, U.; Garnier-Amblard, E. C.; Coats, S. J.; Amblard, F.;
Schinazi, R. F. Chem. Rev. 2014, 114, 9154−9218.
a
k
methyl 1H-1,2,4-triazole-3-
85
carboxylate
(9) (a) Maiti, M.; Maiti, M.; Rozenski, J.; De Jonghe, S.; Herdewijn, P.
Org. Biomol. Chem. 2015, 13, 5158−5174. (b) Gao, L. J.; De Jonghe, S.;
Daelemans, D.; Herdewijn, P. Bioorg. Med. Chem. Lett. 2016, 26, 2142−
2146. (c) Maiti, M.; Gao, L. J.; Huang, C.; Ptak, R. G.; Murray, M. G.; De
Jonghe, S.; Herdewijn, P. Org. Biomol. Chem. 2016, 14, 8743−8757.
(10) Beigelman, L.; Mikhailov, S. N. Carbohydr. Res. 1990, 203, 324−
329.
(11) Vorbruggen, H.; Ruh-Pohlenz, C. Org. React. 1999, 1−630.
(12) (a) Moreau, C.; Kirchberger, T.; Zhang, B.; Thomas, M. P.;
Weber, K.; Guse, A. H.; Potter, B. V. L. J. Med. Chem. 2012, 55, 1478−
1489. (b) Moreau, C.; Kirchberger, T.; Swarbrick, J. M.; Bartlett, S. J.;
Fliegert, R.; Yorgan, T.; Bauche, A.; Harneit, A.; Guse, A. H.; Potter, B.
V. L. J. Med. Chem. 2013, 56, 10079−10102.
(13) Schinazi, R. F.; Cho, J. H.; Zhou, L.; Zhang, H.; Pradere, U.; Coats,
S. J. Patent WO2012/158811, 2012.
(14) Bio, M. M.; Xu, F.; Waters, M.; Williams, J. M.; Savary, K. A.;
Cowden, C. J.; Yang, C.; Buck, E.; Song, Z. J.; Tschaen, D. M.; Volante,
R. P.; Reamer, R. A.; Grabowski, E. J. J. J. Org. Chem. 2004, 69, 6257−
6266.
l
6m
3-cyano-2-oxopyridine
complex reaction
mixture
a
Yield of product 7l where B = 1,2,4-triazole-3-carboxamide.
In some cases, the formation of overreacted products (due to
hydrolysis of the isoamyl esters) was observed on TLC. This was,
however, not a major issue since they can be often easily
separated from the desired compound by flash chromatography
on silica gel. It is noteworthy that the chlorine at position 6 of the
purine moiety of compounds 6i and 6j is stable under these
reaction conditions, affording ProTides 7i and 7j, whereas the
methyl ester on the imidazole ring of 6j was transformed into the
corresponding carboxamide. On the other hand, for the pyridine
analogue 6m, no desired product was isolated due to the
decomposition of the substrate under the reaction conditions.
In summary, a conceptually new method for the construction
of (aryloxy)phosphoramidate nucleoside prodrugs is presented.
It involves the synthesis of ProTide sugar 5 as a key building
block that can be used for coupling with a number of
nucleobases, affording exclusively the β-nucleosides, although
mixtures of regioisomers were formed in some cases. Finally, a
method has been elaborated for the selective hydrolysis of the
acetoxy groups on the sugar moiety, leaving the prodrug moiety
D
DOI: 10.1021/acs.orglett.6b02764
Org. Lett. XXXX, XXX, XXX−XXX