Synthesis, anti-HIV activity, and resistance profiles of ribose modified nucleoside phosphonates

Article abstract of DOI:10.1016/j.bmcl.2007.10.038

A series of nucleoside phosphonate reverse transcriptase (RT) inhibitors have been synthesized and their anti-HIV activity and resistance profiles evaluated. The most potent analog [5-(6-amino-purin-9-yl)-2,5-dihydro-furan-2-yloxymethyl]-phosphonic acid (d4AP) demonstrated a HIV EC₅₀ = 2.1 μM, and the most favorable resistance profile against HIV-1 variants with K65R, M184V or multiple thymidine analog mutations in RT.

Full text of DOI:10.1016/j.bmcl.2007.10.038

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 17 (2007) 6785–6789
Synthesis, anti-HIV activity, and resistance profiles of
ribose modified nucleoside phosphonates
Richard L. Mackman,* Constantine G. Boojamra, Vidya Prasad, Lijun Zhang,
Kuei-Ying Lin, Oleg Petrakovsky, Darius Babusis, James Chen, Janet Douglas,
Deborah Grant, Hon C. Hui, Choung U. Kim, David Y. Markevitch, Jennifer Vela,
Adrian Ray and Tomas Cihlar
Gilead Sciences, Inc., 333 Lakeside Drive, Foster City, CA 94404, USA
Received 27 August 2007; revised 11 October 2007; accepted 11 October 2007
Available online 17 October 2007
Abstract—A series of nucleoside phosphonate reverse transcriptase (RT) inhibitors have been synthesized and their anti-HIV activ-
ity and resistance profiles evaluated. The most potent analog [5-(6-amino-purin-9-yl)-2,5-dihydro-furan-2-yloxymethyl]-phosphonic
acid (d4AP) demonstrated a HIV EC50 = 2.1 lM, and the most favorable resistance profile against HIV-1 variants with K65R,
M184V or multiple thymidine analog mutations in RT.
Ó 2007 Elsevier Ltd. All rights reserved.
The recommended first line therapy for treatment-na¨ıve
HIV-infected patients includes a combination of two
nucleos(t)ide reverse transcriptase inhibitors (N(t)RTIs)
and one non-nucleoside RT or protease inhibitor. There
are 8 licensed N(t)RTIs for the treatment of HIV, with
most of the drugs exhibiting some limitations due to
clinical resistance, toxicities or drug–drug interactions.1
Consequently, there is continued interest in developing
new N(t)RTIs with improved cross-resistance and toxi-
cological profiles.
diacid is delivered into cells, via a prodrug, the high
polarity and reduced permeability now favor intracellu-
lar accumulation of the diacid and its metabolites. The
bis-isopropyloxymethylcarbonyl (bisPOC) prodrug is
used to deliver tenofovir orally and more recently an
amidate prodrug (GS-7340) has been demonstrated to
be an effective alternative enhancing the nucleotide
delivery into lymphocytes.4,5 Only a handful of reports
have described antiviral activity of phosphonomethoxy
nucleoside analogs on cyclic 5-membered ring ribose
scaffolds, for example, 2–7 (Fig. 1).6–9 Therefore, we
have synthesized a variety of novel NtRTI inhibitors
and compared their anti-HIV activity and resistance
profiles toward several key clinical HIV RT mutants.
Nucleotide, 9-R-(2-phosphonomethoxypropyl)adenine
(tenofovir) 1,2 the phosphonate component of tenofovir
disoproxil fumarate (tenofovir DF, VireadÒ), stands out
amongst the N(t)RTI for two reasons (Fig. 1). First, it is
a nucleoside phosphonate and second, it does not have a
formal 5-membered ring ribose core. The phosphono-
methoxy group is a bioisostere of the a-phosphate group
of nucleotide monophosphates but differs in that it is
more enzymatically stable than the a-phosphate group.3
A limiting property of the phosphonate diacids is that
they are charged at physiological pH and therefore exhi-
bit poor membrane permeability. However, once the
Phosphonomethoxy analogs of 20,30-didehydro-20,30-
dideoxy (d4) adenosine 2 (d4AP), and 20,30-dideoxy
(dd) adenosine, 4 (ddAP), were prepared following the
procedures of Kim et al.6 Thymidine analogs 3 (d4TP)
and 5 (ddTP), were prepared according to reported
procedures.6,7 Uridine analogs, 12 (d4UP) and 13 (5-
F-d4UP) were prepared from commercially available
20-deoxyuridine 8a and 5-fluoro-20-deoxyuridine 8b,
respectively (Scheme 1). Initial oxidation using platinum
oxide10 was followed by decarboxylative dehydration to
the intermediate glycal.11 Subsequent activation of the
glycal with PhSeCl on the least hindered a-face of the ri-
bose ring directed introduction of the dimethyl hydrox-
ymethylphosphonate ester to the b-face.6 Olefin
Keywords: Anti-HIV; Phosphonates; Resistance; Reverse transcrip-
tase.
*
Corresponding author. Tel.: +1 650 522 5258; fax: +1 650 522
5899; e-mail: rmackman@gilead.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.10.038

6786
R. L. Mackman et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6785–6789
treatment with NCS or NBS, respectively, followed by
NH2
N
TMSBr mediated ester hydrolysis.12 Carbocyclic analog
6 was prepared according to modified literature proce-
dures.13 The dd analog, 16 (ddUP), was prepared from
9, by hydrogen reduction over 10% Pd/C prior to re-
moval of the phosphonate ester groups. Cytidine analog
17 (d4CP) was prepared from uridine analog 9.14 The
procedure for the preparation of 18 (iso-ddAP) has been
briefly described and will be reported in full elsewhere.15
The active diphospho-phosphonate (triphosphate equiv-
alents) metabolites of the phosphonate diacids 2–5, 12,
16, and 18 were prepared by treatment of the phospho-
nates with CDI followed by pyrophosphate.7 Antiviral
testing methods and resistance profiling have been re-
ported previously.7
N
N
N
H2O3P
O
1, PMPA
O
H2O3P
O
Base
O
H2O3P
O
Base
4 Base = A, ddAP
5 Base = T, ddTP
2 Base = A, d4AP
3 Base = T, d4TP
NH2
NH2
N
N
N
N
N
N
O
N
The wild-type (WT) antiviral activity of the unsaturated
d4 phosphonate analogs is dependent on the nucleobase
(Table 1, entries 2–3, 12–15, and 17). Adenine analog, 2
(d4AP), has the most potent antiviral (EC50) activity,
12-fold more potent than thymidine 3, and significantly
more potent than the uridine analogs 12 and 13. 5-Hal-
ogenated uridines, 14 (5-Cl) and 15 (5-Br), have compa-
rable antiviral activity to thymidine analog 3, but
increased cellular toxicity. The saturated dd analogs, 4,
5, and 16, have weaker antiviral activity than their
corresponding d4 analogs (2, 3, and 12, respectively).
Antiviral activity is dependent on the ability of the phos-
phonate diacid to penetrate cells, become metabolized to
the active diphospho-phosphonate (triphosphate equiv-
alent) species, and inhibit HIV RT. Diphospho-phos-
phonate concentrations, measured in the CEM-CCRF
H2O3P
O
N
H2O3P
O
7, PMDTA
6, 4'C-d4AP
Figure 1. Structures of tenofovir and ribose modified phosphonates.
The nomenclature describing the phosphonate analogs is derived by
adding the ‘P’ to the typically accepted nucleoside notation, for
example, the equivalent phosphonate analog of nucleoside d4T is
referred to as d4TP.
formation was effected via hydrogen peroxide-mediated
oxidation of selenium and subsequent elimination.
TMSBr hydrolysis of the phosphonate ester groups pro-
vided compounds 12 and 13.6 Further halogenated uri-
dine analogs, 14 and 15, were prepared from 9 by
O
O
O
X
X
X
NH
NH
(i)-(iv)
NH
(v)
O
O
H2O3P
O
N
O
N
O
O
Me2O3P
O
N
O
HO
HO
12 X = H
13 X = F
3 X = Me
X = H
(vi), (v)
9
X = H,
8a X = H
8b X = F
8c X = Me
10 X = F,
11 X = Me
O
N
X
NH
O
O
H2O3P
O
X = H
X = H, Me
(vii), (v)
(viii), (v)
14 X = Cl
15 X = Br
O
NH2
N
NH2
N
X
NH
N
N
O
O
N
O
H2O3P
O
H2O3P
O
N
O
H2O3P
O
N
O
18
17
16 X = H
5 X = Me
Scheme 1. Reagents and conditions: (i) Pt/C, O2, 70 °C, H2O, NaHCO3, 3 days, 45%; (ii) DMF–dineopentylacetal, DMF, 80 °C, 75%; (iii) PhSeCl,
ꢀ70 °C, CH2Cl2 then HOCH2PO(OMe)2, AgClO4 71%; (iv) H2O2, H2O, dioxane, 85%; (v) TMSBr, DMF, 0 °C then NH4OH; (vi) NBS (or NCS),
pyr, DMF, 90 °C, 50–60%; (vii) Pd/C, H2, EtOH, 75%; (viii) p-ClPhOP(@O)Cl2, triazole, pyridine, then NH4OH, 55%.

R. L. Mackman et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6785–6789
Table 1. Anti-HIV activity, RT potency, and antiviral resistance profile
6787
Compound
Code
WT HIVa
MT-2
CC50/lM
WT RT
IC50/lM
Nucleosidec WT
M184Vb
Fold res.
6TAMsb
Fold res.
K65Rb
Fold res.
EC50/lM
RT IC50/lM
2
d4AP
d4TP
2.1 (1.0)
26 (8.8)
>200
>1000
>1000
—
0.60 (0.16)
0.39 (0.08)
—
0.14 (0.07)
0.06 (0.03)
0.9 (0.6)
1.1
—
2.9 (1.0)
12.7 (3.9)
—
2.9 (1.1)
5.2 (0.3)
—
3
17
12
13
14
15
6
d4CP
d4UP
—
0.55 (0.15)
>200
>200
>1000
—
5.7
—
—
—
—
—
—
—
5-F-d4UP
5-Cl-d4UP
5-Br-d4UP
40C-d4AP
ddAP
—
—
28.5 (16.2)
15
413 (174)
935
—
—
—
0.8
1.3
>10
>10
6.7
7.5
—
—
51.7 (13.4)
12 (2.5)
>200
—
—
1.1
—
—
4
>1000
—
—
4.1 (1.8)
5.4 (2.1)
31
0.26 (0.14)
0.12 (0.02)
0.83 (0.28)
—
3.4 (0.3)
—
—
4.9 (1.4)
—
—
5
ddTP
ddUP
—
—
16
18
1
>200
iso-ddAP
Tenofovir
Abacavir
9.5 (5.6)
3.6 (1.5)
0.32 (0.15)
863 (238)
>1000
190 (45)
0.32 (0.14)
0.38 (0.20)
0.13 (0.07)
11.0 (3.0)
0.7 (0.2)
7.1 (4.4)
3.2 (0.6)
8.8 (3.7)
5.7 (3.1)
2.2 (1.1)
4.3 (1.5)
3.8 (1.8)
—
—
—
a Values are results of at least 2 experiments, standard deviation is given in parentheses.
b Resistance determined in MT-2 cell lines.
c Data for the equivalent nucleoside triphosphate analog.
T-cell line, were shown to be higher for d4 analogs, 2
and 3, compared to dd analogs, 5 and 16 (data not
shown), consistent with the more potent antiviral activ-
ity of the d4 analogs.16
carbon yielded analog 6 (40C-d4AP).9 However, this
conversion to the carbocyclic ring led to a 25-fold
drop in the antiviral activity. Nevertheless, the corre-
sponding diphosphate of 6 is reported9 to be a potent
inhibitor of HIV RT, suggesting that, once again the
replacement of the 40-O with carbon reduces
phosphorylation.
Examination of the RT inhibition trends between
unsaturated d4 and saturated dd phosphonates shows
that the d4 analogs are consistently more potent
inhibitors of RT than the saturated dd analogs. This
trend is also found in the d4 and dd nucleosides
suggesting that the phosphonate bioisostere of the
nucleoside a-phosphate does not lead to grossly differ-
ent binding modes to HIV RT. However, despite this
similar trend, all the phosphonate analogs are weaker
RT inhibitors than their nucleoside counterparts, espe-
cially the dd phosphonate analogs. For example, the
dd phosphonate analogs are 16-fold (A analog 4),
45-fold (T analog 5), and 37-fold (U analog 16) less
potent than the equivalent dd nucleosides (Table 1,
column 6). In contrast, the d4 phosphonates are all
within 10-fold of the d4 nucleosides with respect to
RT inhibition. This more substantial drop in activity
for the dd phosphonates has been noted previously
and hypothesized to be the result of introducing a sec-
ond anomeric center that significantly perturbs the
conformation of the dd phosphonate analogs
compared to their nucleoside dd counterparts.7
Conformational analysis using NMR has shown that
the second anomeric center is more powerful than
the nucleobase anomeric center and drives the confor-
mation to the 30-exo, 20-endo (Southern) conforma-
tion. This perturbation is minimized in unsaturated
d4 analogs due to the more rigid and planar geometry
of the ribose ring. We therefore sought to evaluate the
effects of reducing the number of anomeric centers by
transposing the 40-O around the ring, for example,
iso-analog 18. Inhibition of RT improved by >10-fold
resulting in the most potent phosphonate RT inhibitor
of the series. However, antiviral activity was weaker
than d4 analog 2 due to decreased phosphorylation
relative to 2 (Table 1).16 Replacement of the 40-O with
Antiviral resistance profiles against HIV variants with
the major N(t)RTI mutations M184V, K65R, and
6TAMs were determined for the most active analogs
and compared to tenofovir 1. As anticipated, the d4
and dd phosphonomethoxy analogs, 2–4 and 14–15,
all retain their full activity against HIV strain contain-
ing M184V. In contrast, iso-ddAP 18 suffered a 11-
fold drop in antiviral activity similar to that observed
for abacavir (Table 1). The common feature between
these 2 inhibitors is the replacement of the 40-O with
carbon. Recent studies have determined that the
M184V mutation reduces the incorporation efficiency
of abacavir more than the corresponding 40-O ribose
analog, d4G triphosphate.17 Thus, the 40-O of the ri-
bose ring is critical for optimal binding and incorpora-
tion kinetics toward the M184V mutation. Modeling
of iso-ddAP 18 in HIV RT indicates that the ring
pucker changes to alleviate interactions between the
adenine C-8 hydrogen and 40-CH2, thus placing the
40-CH2 in closer proximity to residue M184 (Fig. 2).
This explains the susceptibility to the M184V muta-
tion that increases the steric interactions in this part
of the active site.
Pyrimidine analogs, 3, 14, and 15 showed the greatest
loss in susceptibility (>10-fold) due to multiple thymi-
dine analog mutations (6TAMs) whereas purines 2, 4,
and 18 were all superior to tenofovir, 1. The TAM
mutations promote primer unblocking by excision of
the nucleotide chain terminator prior to translocation
from the Nucleotide-site (N-site) to the Primer-site (P-
site).18 The increased effect of 6TAMs on the activity
of thymidine analog 3 compared to adenine 2 clearly

6788
R. L. Mackman et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6785–6789
Figure 2. (a) Compound 2 and (b) compound 18 bound in WT HIV RT. M184 interacts more closely to the 40-CH2 in (b) due to the different
puckering of the isomeric ribose ring compared to compound 2.
indicates that the nucleobase has an important role in
reducing the propensity for excision. Comparing the
TAMs profiles for four adenine analogs, acyclic phos-
phonate tenofovir 1 (8.8-fold resistance), d4 analog 2
(2.9-fold resistance), dd analog 4 (3.4-fold resistance),
and iso-analog 18 (3.2-fold resistance), indicates how
differences in the ribose portion of the molecule affect
the antiviral resistance due to TAMs. Further studies
that examined the in vitro removal of the four adenine
analogs from the 30-terminated DNA primer by RT
indicated that the cyclic analogs displayed reduced
efficiency of excision compared to tenofovir (data
not shown). These data corroborate the differences
in the TAMs susceptibility observed in the antiviral
assays. Taken together, cyclic ribose phosphonates
containing adenine base exhibit the most favorable
TAMs resistance profile among all tested nucleotide
analogs.
d4AP 2 has a superior resistance profile to both iso-
ddAP 18 and tenofovir 1, the only FDA-approved
NtRTI.
References and notes
1. Barreiro, P.; Jimenez-Nacher, I.; Garcia, B.; Garcia-
Benayas, T.; Rivas, P.; Soriano, V. Curr. Opin. Invest.
Drugs 2005, 6, 812.
2. Balzarini, J.; Holy, A.; Jindrich, J.; Naesens, L.; Snoeck,
R.; Schols, D.; De Clercq, E. Antimicrob. Agents Chemo-
ther. 1993, 37, 332.
3. Engel, R. Chem. Rev. 1977, 77, 349.
4. Robbins, B. L.; Srinivas, R. V.; Kim, C.; Bischofberger,
N.; Fridland, A. Antimicrob. Agents Chemother. 1998, 42,
612.
5. Lee, W. A.; He, G.-X.; Eisenberg, E.; Cihlar, T.; Swami-
nathan, S.; Mulato, M.; Cundy, K. C. Antimicrob. Agents
Chemother. 2005, 49, 1898.
6. Kim, C. U.; Luh, B. Y.; Martin, J. C. J. Org. Chem. 1991,
56, 2642.
Finally, the lysine 65 residue is located in the finger sub-
domain of RT and donates a hydrogen bond to the c-
phosphate oxygen of the dNTP upon its binding in the
enzyme active site.19 Mutation to arginine at position
65 introduces additional hydrogen bonds that reduce
the conformational mobility of the RT fingers domain
resulting in diminished incorporation and excision of
nucleotide analogs. Compared to tenofovir 1, the purine
analogs, d4AP 2 and iso-ddAP 18, have marginally re-
duced susceptibility due to K65R whereas the pyrimi-
dines 3, 14–15 tended to show more significant loss of
activity. However, the small differences are not substan-
tial enough to draw any firm structural or mechanistic
conclusions.
7. Mackman, R. L.; Zhang, L.; Prasad, V.; Boojamra, C.
G.; Douglas, J.; Grant, D.; Hui, H.; Kim, C. U.;
Laflamme, G.; Parrish, J.; Stoycheva, A. D.; Swamina-
than, S.; Wang, K.; Cihlar, T. Bioorg. Med. Chem. 2007,
15, 5519.
8. Wu, T.; Froeyen, M.; Kempeneers, V.; Pannecouque, C.;
Wang, J.; Busson, R.; De Clercq, E.; Herdewijn, P. J. Am.
Chem. Soc. 2005, 127, 5056.
9. Khandazhinskaya, A. L.; Shirokova, E. A.; Skoblov, Y.
S.; Victorova, L. S.; Goryunova, L. Y.; Beabealashvilli, R.
S.; Pronyaeva, T. R.; Fedyuk, N. V.; Zolin, V. V.;
Pokrovsky, A. G.; Kukhanova, M. K. J. Med. Chem.
2002, 45, 1284.
10. Moss, G. P.; Reese, C. B.; Schofield, K.; Shapiro, R.;
Lord; Todd J. Chem. Soc. 1963, 1, 1149.
A series of nucleoside phosphonates that combine differ-
ent ribose modifications and base changes led to the
identification of 2 potent RT inhibitors, d4AP 2 and
iso-ddAP 18. Resistance profiling against HIV strains
with clinically relevant RT mutations determined that
11. Zemlicka, J.; Gasser, R.; Freisler, J. V.; Horwitz, J. P.
J. Am. Chem. Soc. 1972, 9, 3213.
12. Beres, J.; Bentrude, W. G.; Kalman, A.; Parkanyi, L.;
Balzarini, J.; De Clercq, E. J. Med. Chem. 1986, 29,
488.

R. L. Mackman et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6785–6789
6789
13. (a) Dyatkina, N. B.; Theil, F.; Janta-Lipinski, M. V.
Tetrahedron 1995, 51, 761; (b) Boojamra, C. G. et al.
Manuscript in preparation.
14. Wang, P.; Gullen, B.; Newton, M. G.; Cheng, Y.-C.;
Schinazi, R. F.; Chu, C. K. J. Med. Chem. 1999, 42, 3399.
15. Mackman, R. L. Presented at the 19th International
Conference on Antiviral Research, San Juan, Puerto Rico,
May 2006; paper 37.
nate determined by LC MS/MS by comparing the signal to
that of a standard curve prepared in cellular extract.
17. Ray, A. S.; Yang, Z.; Shi, J.; Hobbs, A.; Schinazi, R.
F.; Chu, C. K.; Anderson, K. S. Biochemistry 2002, 41,
5150.
18. White, K. L.; Chen, J. M.; Feng, J. Y.; Margot, N. A.; Ly,
J. K.; Ray, A. S.; MacArthur, H. L.; McDermott, M. J.;
Swaminathan, S.; Miller, M. D. Antiviral Ther. 2006, 11,
155.
16. CEM-CCRF cells were incubated in media containing
10 lM of each tested phosphonate diacid. After 24 h cells
were washed, lysed, and the level of diphospho-phospho-
19. Boyer, P. L.; Sarafianos, S. G.; Arnold, E.; Hughes, S. H.
J. Virol. 2001, 75, 4832.