4066
J. Kim et al. / Bioorg. Med. Chem. Lett. 22 (2012) 4064–4067
Table 2
triphosphate 3. This increase in efficiency was imparted by a six-
fold increase in the kpol as well as the threefold tighter binding
affinity. A phosphonate group has several advantages over a
phosphate. The phosphorous–carbon bond is not susceptible to
hydrolysis by phosphatases. Also, once the prodrug form of a
nucleoside has entered a cell, the presence of the phosphonate
group bypasses the first phosphorylation step, which is considered
to be the rate limiting step in the conversion of the triphosphate
form of a nucleoside by cellular kinases. A recent review discusses
these advantages and gives examples of other antiviral phospho-
nate candidates.8 Interestingly, compound 5 was 4.4-fold more
efficiently incorporated by RTWT compared to the tenofovir diphos-
phate 2 (Fig. 1, solid bars). Compound 5 had a twofold faster kpol
and an almost threefold tighter binding affinity compared to the
FDA approved nucleotide analog. Compound 5 was able to act as
a chain terminator by inhibiting incorporation of the next correct
nucleotide, as shown in Supplementary Figure 1.
Pre-steady state kinetic data for incorporation of deoxyadenosine triphosphate
(dATP), tenofovir diphosphate (TFV-DP) and cyclobutyl-adenosine-phosphonyl-
diphosphate (5) using HIV-1 RTK65R
Compound kpol (sꢀ1
)
Kd
(
l
M)
Efficiency
(sꢀ1 Mꢀ1
Discrimination
l
)
dATP
TFV-DP
5
0.4 0.3
0.15 0.01
0.07 0.004
6.6 1.2 0.61
29.1 6.0 0.005
4.7 1.0 0.015
122
41
and a twofold decrease in the binding affinity. The RTK65R was able
to incorporate the cyclobutyl compound 5 more efficiently com-
pared to the tenofovir diphosphate 2 (Fig. 1, cross hatch bars) by
threefold. The discrimination of the two nucleotide analogs was
also determined. The RTK65R mutant was threefold less discrimina-
tive against compound 5 versus compound 2.
Recent work has been performed to solve the crystal structures
of wild type and the K65R mutant of RT in the presence of a pri-
mer-template and dATP or TFV-DP.9,10 Three major residues are in-
volved in the interactions with the incoming dNTP in both
polymerases—K(R)65, R72, and Y115. R72 interacts with the base,
Another cyclobutyl containing compound 6, was synthesized as
an adenosine analog to base-pair with dTTP. This compound was
not a good substrate for incorporation by RT, and was almost two-
fold less efficient (0.007 sꢀ1 Mꢀ1) compared to the parent com-
l
pound 3. This difference was due to the almost twofold decrease
sugar, and
a-phosphate of the incoming nucleotide. When K65 is
in the kpol
.
mutated to an R, guanidinium stacking occurs with this amino acid
and R72, causing steric restriction on R72 and therefore affecting
the incorporation rate by RT. The amino acid Y115 stacks against
the ribose ring of the natural substrate, which is partially lost with
TFV-DP due to its acyclic structure as shown in Figure 2a. In order
to analyze potential binding interactions of 5, the compound was
docked into the RT active site bound to primer-template (Tuske
et al. Ref. 10, PDB code 1T05). The cyclobutyl ring of compound 5
may be able to stack with Y115, or form favorable Van der Waals
contacts, which may explain its better binding affinity compared
to TFV-DP as shown in Figure 2a and b. Similar to the natural sub-
strate and TFV-DP, the K65R mutation greatly decreases the rate of
incorporation perhaps in the same manner, through the restriction
of the R72 and its interaction with the phosphonate group present
in the cyclobutyl compound. It is interesting to note that from the
wild type RT structure with TFV-DP, the side chains of Lys65 and
The HIV-1 RT’s ability to discriminate a nucleotide analog was
determined, which is the ratio of the efficiency of incorporation
for the natural substrate 1 to the efficiency of incorporation for
the analog. The RTWT was fourfold less discriminative against com-
pound 5 versus compound 2. The enzyme was 4- to 7-fold more
discriminative against the other cyclobutyl analogs (3, 4, 6) versus
compound 2, and 19- to 40-fold more discriminative versus com-
pound 5.
Tenofovir-diphosphate 2, is an acyclic adenosine analog that
also contains a phosphonate group, similarly to compound 5. Upon
administration of tenofovir to infected persons, a drug resistant
form of HIV-1 RT may arise containing a K65R mutation. We deter-
mined the pre-steady state kinetic parameters for the natural sub-
strate dATP, 1, tenofovir diphosphate 2, and the most promising
cyclobutyl compound, 5 using the K65R mutant form of HIV-1
RT, as shown in Table 2.
Lys219 are positioned to make potential salt bridges with the
a-
The RTK65R was fourfold less efficient in incorporating the natu-
ral substrate 1, and 14-fold less efficient in incorporating the cyclo-
butyl compound 5. The large decrease in incorporation efficiency
phosphate of the phosphonate. Based on the binding predictions
from our model, potential salt bridges may be formed between
compound 5, Lys65, and Lys219 (Fig. 2b). This differs from dNTP
in which only Lys65 is positioned to make a potential salt bridge
with the a-phosphate. The additional salt bridge formed with the
phosphonate may give rise to potency of the compound. The dock-
using the mutant RT was observed through a decrease in the kpol
,
compared to the wild type enzyme. Compound 2 was ninefold less
efficiently incorporated by the mutant enzyme compared to the
wild type which was imparted by a fivefold decrease in the kpol
,
ing model and structural data may also help explain the increased
Figure 1. Incorporation efficiency for deoxyadenosine triphosphate (dATP), tenofovir diphosphate (TFV-DP), cyclobutyl adenosine phosphonyl disphosphate (5) for wild type
(solid black) and K65R (cross hatch) HIV-1 RT. The natural substrate, dATP, is the most efficiently incorporated. The cyclobutyl analog, 5, is more efficiently incorporated by
both forms of RT compared to the FDA approved TFV-DP