Enzymatic phosphorylation of unnatural nucleosides

Article abstract of DOI:10.1021/ja028050m

In an effort to expand the genetic alphabet, a number of unnatural, predominantly hydrophobic, nucleoside analogues have been developed which pair selectively in duplex DNA and during enzymatic synthesis. Significant progress has been made toward the efficient in vitro replication of DNA containing these base pairs. However, the in vivo expansion of the genetic alphabet will require that the unnatural nucleoside triphosphates be available within the cell at sufficient concentrations for DNA replication. We report our initial efforts toward the development of an unnatural in vivo nucleoside phosphorylation pathway that is based on nucleoside salvage enzymes. The first step of this pathway is catalyzed by the D. melanogaster nucleoside kinase, which catalyzes the phosphorylation of nucleosides to the corresponding monophosphates. We demonstrate that each unnatural nucleoside is phosphorylated with a rate that should be sufficient for the in vivo replication of DNA.

Full text of DOI:10.1021/ja028050m

Published on Web 11/14/2002
Enzymatic Phosphorylation of Unnatural Nucleosides
Yiqin Wu,† Ming Fa, Eunju Lee Tae, Peter G. Schultz, and Floyd E. Romesberg*
Contribution from the Department of Chemistry, The Scripps Research Institute,
10550 N. Torrey Pines Road, La Jolla, California 92037
Received August 7, 2002
Abstract: In an effort to expand the genetic alphabet, a number of unnatural, predominantly hydrophobic,
nucleoside analogues have been developed which pair selectively in duplex DNA and during enzymatic
synthesis. Significant progress has been made toward the efficient in vitro replication of DNA containing
these base pairs. However, the in vivo expansion of the genetic alphabet will require that the unnatural
nucleoside triphosphates be available within the cell at sufficient concentrations for DNA replication. We
report our initial efforts toward the development of an unnatural in vivo nucleoside phosphorylation pathway
that is based on nucleoside salvage enzymes. The first step of this pathway is catalyzed by the D.
melanogaster nucleoside kinase, which catalyzes the phosphorylation of nucleosides to the corresponding
monophosphates. We demonstrate that each unnatural nucleoside is phosphorylated with a rate that should
be sufficient for the in vivo replication of DNA.
Introduction
activation may also aid in the design of pharmacologically active
nucleoside analogue drugs that also rely on cellular activation.
In an effort to expand the genetic alphabet by supplementing
the natural base pairs dA:dT and dG:dC with an unnatural base
pair, we have synthesized and characterized a variety of
predominantly hydrophobic nucleobases analogues.1-6 Several
of the unnatural base pairs formed between these hydrophobic
nucleobases are stable in duplex DNA and are also inserted,
proofread, and extended efficiently and selectively by DNA
polymerases in vitro.7 On the basis of these results, as well as
pioneering studies from the Kool lab,8,9 it is apparent that the
requirements for duplex stability and replication do not limit
the genetic code to hydrogen bonded (H-bonded) base pairs,
and hydrophobic interactions may be sufficient to control
information storage and retrieval. The in vivo replication of
DNA containing these unnatural base pairs requires that the
unnatural nucleoside triphosphates be available intracellularly
at sufficient concentrations for DNA synthesis. Thus, we have
begun to examine different pathways for intracellular activation
of unnatural nucleosides. These studies of unnatural nucleoside
The simplest strategy to supply E. coli with unnatural
nucleoside triphosphates is to supplement the growth media with
the unnatural nucleosides. It is possible that these small,
hydrophobic molecules will either passively diffuse or be
actively transferred across the lipophilic cell membrane and
become trapped inside the cell, provided that they are phos-
phorylated by cellular kinases of the nucleoside salvage
pathway.10 In this pathway nucleosides are converted to the
corresponding triphosphates by the successive action of nucleo-
side, monophosphate, and diphosphate kinases. This pathway
produces nucleoside triphosphates in sufficient concentrations
for a variety of cellular functions. For example, in mammalian
cells, DNA repair, mitochondrial DNA synthesis in G1 phase
cells, and the activation of antiviral and cytostatic nucleoside
analogues,11,12 all rely on nucleoside triphosphates from the
salvage pathway.
The first step in the synthesis of nucleoside triphosphates is
the nucleoside kinase catalyzed transfer of the γ-phosphate from
a donor molecule (ATP) to the nucleoside C5′-OH acceptor
to yield the nucleoside monophosphate. In some cases, this is
the rate-limiting step in the pharmacological activation of
nucleoside drug analogues.13 Thus, a variety of antiviral
therapies are based on the ability of a virus-encoded kinase to
phosphorylate a broader range of substrates than the host cellular
kinases. For example, acyclovir is selectively activated by herpes
simplex virus type-1 thymidine kinase (HSV-1 TK),14 which
* To whom correspondence should be addressed. E-mail: floyd@
scripps.edu.
† Present address: Syrrx, Inc. 10450 Science Center Drive, Suite 100,
San Diego, CA 92121.
(1) McMinn, D. L.; Ogawa, A. K.; Wu, Y.; Liu, J.; Schultz, P. G.; Romesberg,
F. E. J. Am. Chem. Soc. 1999, 121, 11 585-11 586.
(2) Ogawa, A. K.; Wu, Y.; McMinn, D. L.; Liu, J.; Schultz, P. G.; Romesberg,
F. E. J. Am. Chem. Soc. 2000, 122, 3274-3287.
(3) Ogawa, A. K.; Wu, Y.; Berger, M.; Schultz, P. G.; Romesberg, F. E. J.
Am. Chem. Soc. 2000, 122, 8803-8804.
(4) Wu, Y.; Ogawa, A. K.; Berger, M.; McMinn, D. L.; Schultz, P. G.;
Romesberg, F. E. J. Am. Chem. Soc. 2000, 122, 7621-7632.
(5) Berger, M.; Ogawa, A. K.; McMinn, D. L.; Wu, Y.; Schultz, P. G.;
Romesberg, F. E. Angew. Chem., Int. Ed. 2000, 39, 2940-2942.
(6) Berger, M.; Luzzi, S. D.; Henry, A. A.; Romesberg, F. E. J. Am. Chem.
Soc. 2002, 124, 1222-1226.
(10) Kornberg, A.; Baker, T. A. DNA Replication, 2nd ed.; W. H. Freeman and
Company: New York, 1992.
(11) Plunkett, W.; Saunders: P. P. Pharm. Theraput. 1991, 49, 239-268.
(12) Arner, E. S. J.; Eriksson, S. Pharm. Theraput. 1995, 67, 155-186.
(13) Johansson, M.; Van Rompay, A. R.; Degre`ve, B.; Balzarini, J.; Karlsson,
A. J. Biol. Chem. 1999, 274, 23 814-23 819.
(14) Black, M. E.; Newcomb, T. G.; Wilson, H.-M. P.; Loeb, L. A. Proc. Natl.
Acad. Sci. U.S.A. 1996, 93, 3525-3529.
(7) Tae, E. L.; Wu, Y.; Xia, G.; Schultz, P. G.; Romesberg, F. E. J. Am. Chem.
Soc. 2001, 123, 7439-7440.
(8) Matray, T. J.; Kool, E. T. J. Am. Chem. Soc. 1998, 120, 6191-6192.
(9) Morales, J. C.; Kool, E. T. Nature Struct. Biol. 1998, 5, 950-954.
9
14626
J. AM. CHEM. SOC. 2002, 124, 14626-14630
10.1021/ja028050m CCC: $22.00 © 2002 American Chemical Society

Enzymatic Phosphorylation of Unnatural Nucleosides
A R T I C L E S
Scheme 1. Unnatural Nucleotides.
Kinase Expression. A plasmid containing the Dm-dNK gene was
kindly provided by Prof. Magnus Johansson (Karolinska Institute,
Sweden). The gene was subcloned into the pET-15b expression vector,
and confirmed by sequencing. This expression plasmid results in fusion
of the carboxy terminus of the protein to a (His)6 tag. Following
transformation of BL21(DE3) pLysS E. coli (Stratagene), protein was
overexpressed by IPTG induction and purified by affinity chromatog-
raphy with a Ni-NTA affinity column (Novagen). The protein was
dialyzed extensively against 50 mM Tris (pH 8.0), 1 mM DTT, 50
mM NaCl buffer. The size and purity of the recombinant protein was
determined by SDS-PAGE. The protein concentration was determined
using the Bio-Rad protein assay with BSA as a standard. Approximately
2 mg of pure protein was obtained from a 1 L culture. All nucleoside
analogues were synthesized as reported previously,1-5 with the excep-
tions of dT, dA, AZT, and acyclovir which were obtained from Sigma.
Kinetic Assay. Nucleoside monophosphorylation was assayed with
standard literature protocols based on the phophorylation of radiolabeled
nucleosides.18,20 However, to rapidly screen many different unnatural
nucleosides an alternative assay was developed that is based on
supplementing the ATP phosphate donor pool with a small amount of
γ-32P labeled ATP. Product formation was monitored by chromato-
graphic separation and subsequent quantitation of the 32P labeled
monphosphorylation product. This assay worked well for Dm-dNK,
which has a high affinity for ATP (KM(ATP) ) 2 µM),15 allowing assays
to be run under saturating conditions with only 50 µM of total ATP.
The assay was less reproducible for the HSV-1 TK enzyme, which
has a higher KM(ATP) (100 µM)21,22 (necessitating higher concentrations
of ATP and complicating the quantification of the product relative to
ATP).
The kinetic assays were performed with 1-5 nM kinase in 50 mM
Tris (pH 8.0), 5 mM MgCl2, 1 mM DTT, and 50 µg/mL BSA. This
reaction buffer was supplemented with 50 µM of ATP that was spiked
with 0.1% γ-32P labeled ATP. Reactions were performed with varying
concentrations of nucleoside substrates. Twenty microliter reactions
were incubated (37 °C for 1-10 min) and terminated by heating (95
°C for 5 min). One microliter of the reaction was then spotted on
polyethyleneimine cellulose F thin-layer chromatography sheets (J. T.
Baker), and the nucleosides and product monophosphates were
separated using 66% isobutyric acid, 33% water, and 1% NH4OH.
Radioactivity was quantitated using a PhosphorImager (Molecular
Dynamics), with overnight exposures and ImageQuant software (Mo-
lecular Dynamics). The steady-state kinetic data were evaluated by
plotting the initial rates against nucleoside concentration and fitting
the data to the Michaelis-Menten equation. Data presented are the
average of three independent experiments.
also phosphorylates a broad range of nucleoside analogues with
modifications in both the furanose ring and the nucleobase.
Another deoxynucleoside kinase with a broad substrate tolerance
is the kinase from D. melanogaster (Dm-dNK).13,15 Dm-dNK
is 14% identical to HSV-1 TK and is the only known kinase in
Drosophila.13,15 Dm-dNK was the first kinase shown to phos-
phorylate all four natural deoxyribonucleosides with reasonable
efficiency, although it is most active with pyrimidine nucleo-
sides.
The replication of DNA containing an unnatural base pair in
E. coli might be based on phosphorylation of the unnatural
nucleoside precursors by endogenous salvage enzymes. How-
ever, preliminary examination of the endogenous nucleoside
kinase activities in E. coli whole cell lysates was not promising.
Furthermore, unlike HSV-1 TK and Dm-dNK, there is little
evidence that E. coli nucleoside kinases are capable of phos-
phorylating a broad range of substrates.16-19 Therefore, instead
of focusing on bacterial kinases, we examined the activity of
HSV-1 TK and Dm-dNK toward the unnatural nucleosides
shown in Scheme 1. Preliminary kinetic characterization of both
HSV-1 TK and Dm-dNK demonstrated that the unnatural
nucleosides were more efficiently phosphorylated by Dm-dNK.
Thus, we initiated a more complete characterization of the
specificity and catalytic efficiency of Dm-dNK toward the
unnatural nucleosides. A comparison of steady-state kinetic rate
data obtained for a variety of unnatural nucleosides allowed
the effects of base structure and hydrophobicity to be examined.
The results suggest that E. coli transformed with a plasmid
encoding Dm-dNK should be capable of providing unnatural
monophosphates at physiologically relevant concentrations.
Results and Discussion
Phosphorylation of Unnatural Nucleosides. Steady-state
kinetic data for the Dm-dNK phosphorylation of deoxythymidine
(dT), deoxyadenosine (dA), and thirteen unnatural nucleosides
are reported in Table 1. The unnatural nucleosides tested
included AZT and acyclovir, both of which have modified
furanose rings, and eleven unnatural nucleosides with hydro-
phobic “bases” (Scheme 1). Eight of the unnatural hydrophobic
nucleosides are based on the isocarbostyryl (ICS) or 7-azaindole
(7AI) rings, or derivatives thereof (dICS, dPICS, dMICS, d7AI,
dPP and d3MN). The incorporation of each nucleoside into
DNA has been characterized and a number of them form base
pairs that are promising candidates for the replication of
Experimental Section
Abbreviations. HSV-1 TK, herpes simplex virus type-1 thymidine
kinase; Dm-dNK, kinase from Drosophila melanogaster; AZT, 3′-azido-
3′-deoxythymidine; dA, deoxyadenosine; dT, deoxythymidine; dG,
deoxyguanosine; dC, deoxycytidine; KF, Klenow fragment of DNA
polymerase I; pol â, DNA polymerase â.
(20) Prota, A.; Vogt, J.; PIlger, B.; Perozzo, R.; Wurth, C.; Marquez, V. E.;
Russ, P.; Schulz, G. E.; Folkers, G.; Scapozza, L. Biochemistry 2000, 39,
9597-9603.
(15) Munch-Petersen, B.; Piskur, J.; Søndergaard, L. J. Biol. Chem. 1998, 273,
3926-3931.
(16) Kaneko, T.; Yamamoto, H. Nucleosides Nucleotides 1997, 16, 2111-2122.
(17) Valentin-Hansen, P. Methods Enzymol. 1978, 51, 308-321.
(18) Chen, M. S.; Prusoff, W. H. Methods Enzymol. 1978, 51, 354-360.
(19) Oeschger, M. P. Methods Enzymol. 1978, 51, 473-482.
(21) Wild, K.; Bohner, T.; Folkers, G.; Schulz, G. E. Prot. Sci. 1997, 6, 2097-
2106.
(22) Wild, K.; Bohner, T.; Aubry, A.; Folkers, G.; Schulz, G. E. FEBS Lett.
1995, 368, 289-292.
9
J. AM. CHEM. SOC. VOL. 124, NO. 49, 2002 14627

A R T I C L E S
Wu et al.
Table 1. Kinetic Parameters for Dm-Dnk toward Different
The enzymatic phosphorylation of ICS analogues shown in
Scheme 1 was then evaluated. Overall, the ICS analogue
nucleosides were found to be good substrates for Dm-dNK
(Table 1). dICS is phosphorylated much more efficiently than
either AZT or acyclovir; the rate is actually equivalent to that
of dA (kcat/KM ) 2.0 × 104 M-1s-1). Like dA, the decreased
rate of phosphorylation of dICS, relative to dT (kcat/KM ) 3.0
× 106 M-1s-1) results from weaker binding of the nucleoside
(KM ) 152 µM and 1.6 µM for dICS and dT, respectively).
Substitution of the ICS ring was found to have small effects
on the phosphorylation efficiency. C3-methyl substitution
(dMICS), C7-propynyl substitution (dPICS), and both C3 and
C7 substitutions (dPIM) of the ICS ring were examined. The
increased bulk of C3-methyl substitution of dMICS resulted in
a 2-fold increase in kcat/KM. However, while the additional steric
demands of the C7-propynyl group of dPICS resulted in a slight
decrease in KM, the decrease was more than offset by an almost
4-fold decrease in kcat. Taken together, these effects result in
an efficiency that is 2-fold decreased relative to dICS. dPIM,
containing both C2-methyl and C7-propynyl substitutions, was
converted to the monophosphate with the same efficiency as
dICS, presumably due to the effects of the two substitutions
on kcat. In all cases, the increased bulk at the ICS ring decreased
the KM of the enzyme for the substrate, but the effect on
kcatwas variable. Consistent with this notion, dPIM is the most
tightly bound of the hydrophobic base analogues examined
(KM ) 85 µM).
Substratesa
k
cat/KM
nucleoside
dT
dA
dICS
dPICS
dMICS
dPIM
d7AI
k
cat (s-1
)
KM (µM)
(M-1s-1
)
4.8 ( 1.0
5.2 (.0
1.6 ( 0.5
247 ( 50
152 ( 30
94 ( 10
92 ( 20
85 ( 10
686 ( 100
306 ( 50
510 ( 100
75 ( 30
682 ( 70
438 ( 123
105 ( 40
15 ( 2
3.0 × 106
2.1 × 104
2.0 × 104
8.5 × 103
4.0 × 104
1.8 × 104
1.2 × 103
4.6 × 103
1.4 × 103
3.1 × 104
2.3 × 102
4.6 × 102
1.2 × 104
1.8 × 103
e 1.0 × 102
3.0 ( 0.5
0.8 ( 0.2
3.7 ( 0.5
1.5 ( 0.5
0.8 ( 0.2
1.4 ( 0.4
0.7 ( 0.2
2.3 ( 0.2
0.16 ( 0.02
0.20 ( 0.03
1.3 ( 0.2
0.028 ( 0.005
n.d.b
dPP
dImPy
dP7AI
dTM
d3MN
d2MN
AZT
Acyclovir
n.d.b
a
See text for experimental details. b Not determined independently due
to the slow rate of reaction.
Scheme 2. Unnatural Base Pairs.
The phosphorylation of the purine-like unnatural nucleosides
d7AI, dPP, dImPy, and dP7AI were also characterized (Table
1). d7AI is not as efficiently phosphorylated by the nucleoside
kinase as is dICS. The d7AI phosphorylation efficiency (kcat/
KM) was 1.2 × 103 M-1s-1. This order of magnitude reduction
in rate, relative to phosphorylation of dA, resulted from a 6.5-
fold reduction and a 2.8-fold increase in kcat and KM, respec-
tively. Apparently, H-bond acceptors (N1, N3, N7, or the
exocyclic amine) or associated polarizability effects make dA
a better phosphorylation substrate than the azaindole ring of
d7AI. Structural studies (see below) with purine analogue
substrates have implicated that important nucleobase-protein
H-bonding interactions exist involving the N1 and exocyclic
C2 substituents, as well as the N7 atom. To assess the
importance of these interactions, we examined the effect of aza-
substitution of the indole ring at positions 4 (dPP) and 3
(dImPy), corresponding to the purine 6 and 7 positions. Aza-
substitution at C3 of the azaindole ring resulted in little change
in kcat/KM (1.2 × 103 M-1s-1 and 1.4 × 103 M-1s-1 for d7AI
and dImPy, respectively). Aza-substitution at C4 (dPP) resulted
in an almost 4-fold increase in kcat/KM (4.6 × 103 M-1s-1) due
to an increase in both kcat and a decrease in KM. It therefore
appears likely that H-bonding or electronic interactions between
the protein and the N1/exocyclic C2 substituent are moderately
important for activity.
unnatural DNA, including the 7AI:7AI,2,4,7 PICS:PICS,1,2 and
3MN:3MN3 self-pairs and the ICS:7AI2 and PP:MICS4 het-
eropairs (Scheme 2). The unnatural nucleosides dTM, d2MN,
dP7AI, dImPy, and dPIM were also studied as kinase substrates
in order to evaluate the effects of alternative base structures
and hydrophobicity.
A rapid kinase assay was developed based on the transfer of
γ-32P from ATP to the nucleoside substrate. The kinetic
constants obtained for substrates dT and dA using the assay
were compared to previously reported literature values.13 Native
substrate dT is phosphorylated most efficiently (kcat/KM ) 3.0
× 106 M-1 -1
s ), whereas phosphorylation of dA is approximately
100-fold less efficient (kcat/KM ) 2.1 × 104 M-1 -1
s ), predomi-
nantly due to an elevated KM (1.6 µM and 247 µM for dT and
dA, respectively). These data are in good agreement with
literature values.13
The enzymatic phosphorylation of the unnatural nucleosides
was then investigated. Dm-dNK phosphorylates AZT with
moderate efficiency (kcat/KM ) 1.8 × 103 M-1s-1). Although
the enzyme binds this nucleoside analogue with a reduced
affinity compared to dT (KM ) 15 µM and 1.6 µM for AZT
and dT, respectively), the decrease in efficiency is largely due
to a reduced rate of catalytic turnover (kcat ) 2.8 × 10-2 s-1
vs 4.8 s-1 for AZT and dT, respectively). Dm-dNK shows no
detectable activity toward the acyclovir nucleoside.
The addition of a C3 propynyl group to the azaindole ring
(dP7AI) had a significant positive effect on phosphorylation
efficiency. The added propynyl group resulted in a 3-fold
increase in kcat and a 9-fold decrease in KM, relative to d7AI,
leading to a kcat/KM for phosphorylation that is greater than that
for either dA or dICS. In fact, dP7AI is phosphorylated only
100-fold less efficiently than dT (3.1 × 104 M-1s-1 and 3.0 ×
106 M-1 -1
for dP7AI and dT, respectively), 20-fold faster than
s
9
14628 J. AM. CHEM. SOC. VOL. 124, NO. 49, 2002

Enzymatic Phosphorylation of Unnatural Nucleosides
A R T I C L E S
AZT, and at least 300-fold faster than acyclovir. Apparently,
the Dm-dNK active site is able to accommodate the increased
steric demands of the propynyl group of dP7AI (see below).
To further address the role of hydrophobicity, we examined
dTM, d2MN, and d3MN as substrates for Dm-dNK. The dTM
nucleoside is an approximate shape mimic of the enzymes
natural substrate, dT, but is only inefficiently phosphorylated
(2.3 × 102 M-1s-1) due to both a reduced kcat (nearly 20-fold),
as well as an increased KM (4.5-fold) relative to dICS. As
discussed below, Dm-dNK is expected to make numerous
H-bonding and hydrophobic contacts with the base analogue.
The absence of H-bond donor and acceptors, along with the
minimal aromatic surface area, is apparently sufficient to render
dTM a poor kinase substrate. The increased aromatic surface
area of d3MN does not result in increased phosphorylation
relative to dTM. However, the d2MN nucleoside, with the same
surface area as d3MN, is phosphorylated 50-fold faster than is
dTM, due to an increase in kcat (8-fold) as well as a decrease in
KM (6-fold). These data demonstrates that both nucleobase
H-bonding groups and shape play important roles in substrate
recognition.
Protein-Nucleobase Interactions. The crystal structure of
Dm-dNK bound to deoxycytidine was recently solved.23 The
enzyme has an R/â architecture with a central five-strand parallel
sheet. The substrate dC is located in a deep pocket at the
C-terminus of the â-sheet. The cytosine base is packed between
F111 on one face and W57 and F80 on the other face. Other
residues involved in hydrophobic packing are M69, Y70, V84,
M88, A110, and M118. The base also makes two hydrogen
bonds to Q81 via N3 and N4, whereas the 2-carbonyl oxygen
of dC is hydrogen bonded to two water molecules.
In addition to the Dm-dNK structure, several structures have
been reported for HSV-1 TK.21,22,24-26 With approximately 30%
of amino acid sequence similarity, the two enzymes show similar
structure with several critical residues conserved in the substrate
binding site.13 A structural basis for the HSV-1 TK polyspeci-
ficity has been well documented and involves the reorganization
of several amino acid side chains in the nucleobase binding site
that allows for efficient packing of different substrates. When
bound to dT, the thymine base makes three H-bonding contacts
with the protein, two between the nucleobase O4 and N3 atoms
and the donor and acceptor functionalities of Q125, respectively
(this residue is conserved in Dm-dNK as Q81; hereafter, the
corresponding Dm-dNK residues are given in parentheses for
comparison).13 The 2-carbonyl oxygen of dT also forms a water-
mediated hydrogen bond with R176 (S123). The thymine base
is hydrophobically packed by M128 (V84), I100 (deleted in
Dm-dNK), W88 (W57), Y132 (M88), R163 (R111) and Y172
(R119). HSV-1 TK is able to phosphorylate uracil derivatives
with bulky C5 substituents by readjustments of the hydrophobic
pocket, including W88 (W57), R163 (R111), A167 (V115), and
particularly Y132 (M88).27 Structural rearrangements of the
Dm-dNK residues may accommodate the propynyl group of
dP7AI and facilitate phosphorylation. To accommodate gua-
nosine analogues, HSV-1 TK is again able to conformationally
reorganize, especially at M128 (V84), I100 (deleted), Y172
(R119), and M231 (Y179) and efficiently pack the modified
nucleobases.24,26 The H-bonding of Q125 (Q81) to the substrate
is preserved in these complexes involving contacts with the
guanine N1 and O6 atoms. The 6-carbonyl group of the purine
also hydrogen bonds to the guanidinium group of R176 (S123).
It is apparent that H-bonding groups located in the N1/O6 region
of purine scaffolds mediate important interactions with HSV-1
TK residues Q125 (Q81) and R176 (S123). Similar interactions
between Dm-dNK residues S123 and Q81 may facilitate the
phosphorylation of dPP, relative to d7AI. Moreover, as dis-
cussed above, structural rearrangements involving W57, R111,
V115, M88, V84, R119, and Y179 may allow for favorable
interactions between the protein and the dP7AI propynyl group
that result in the efficient activation of this nucleoside.
Implications for Unnatural Nucleobase Design. Several
trends emerge from an analysis of Dm-dNK phosphorylation
rates that may facilitate the future design of nucleoside analogues
that are efficiently phosphorylated. Electronic effects resulting
from aza-substitution are apparent but generally small. Structural
effects are found to be more significant. The larger nucleosides
d2MN and d3MN have the same hydrophobic surface area
relative to dTM, but the additional aryl ring is oriented
differently with respect to the glycosidic bond. Although the
larger hydrophobic surface area of d3MN, relative to dTM,
affords no increase in the reaction rate, the same increase in
surface area of d2MN results in a 50-fold increase. The addition
of a propynyl group had a similarly variable effect. Addition
of a C7-propynyl group to the ICS ring had little effect, whereas
the addition of the same group to the C3 position of the 7AI
ring resulted in a 26-fold increased rate of phosphorylation. The
disparate effects that result from the same increase in hydro-
phobic surface area reflect the importance of substituent
orientation. It seems likely that a carefully designed increase
in hydrophobic surface area may generally facilitate unnatural
nucleoside monophosphorylation.
d7AI has emerged as a particularly interesting unnatural
nucleobase due to the recent demonstration that DNA containing
the unnatural 7AI:7AI self-pair may be efficiently replicated
in vitro.7 We have demonstrated that the d7AI nucleoside is
phosphorylated by Dm-dNK with a rate that should supply the
cell with sufficient concentrations of the monophosphate for
further processing. However, derivatives that maintain replica-
tion efficiency and fidelity, and also improve activation of the
nucleoside may be desirable. In this regard, dPP and dP7AI
are particularly interesting, and their behavior implies that further
derivatization may result in an unnatural nucleoside that is a
good substrate for both kinase and polymerase enzymes. For
example, the combination of C3 propynyl- and C4 aza-
substitution, which each individually increase phosphorylation
efficiency by Dm-dNK, might cooperatively increase the rate
of phosphorylation. Moreover, aza- and alkyl-substitution at
other indole positions might also result in more efficient
activation of the nucleoside. Further experiments are currently
in progress to evaluate these possibilities. Additional modifica-
tions of the ICS scaffold, by aza- and alkyl-substitution are also
being investigated.
(23) Johansson, K.; Ramaswamy, S.; Ljungcrantz, C.; Knecht, W.; Piskur, J.;
Munch-Petersen, B.; Eriksson, S.; Eklund, H. Nature Struct. Biol. 2001, 8,
616-620.
(24) Brown, D. G.; Visse, R.; Sandhu, G.; Davies, A.; Rizkallah, P. J.; Melitz,
C.; Summers, W. C.; Sanderson, M. R. Nature Struct. Biol. 1995, 2, 876-
881.
(25) Lavie, A.; Vetter, I. R.; Konrad, M.; Goody, R. S.; Reinstein, J.; Schlichting,
I. Nature Struct. Biol. 1997, 4, 601-604.
(26) Champness, J. N.; Bennett, M. S.; Wien, F.; Visse, R.; Summers, W. C.;
Herdewijn, P.; de Clercq, E.; Ostrowski, T.; Jarvest, R. L.; Sanderson, M.
R. Proteins: Struct. Funct. Genet. 1998, 32, 350-361.
(27) De Winter, H.; Herdewijn, P. J. Med. Chem. 1996, 39, 4727-4737.
9
J. AM. CHEM. SOC. VOL. 124, NO. 49, 2002 14629

A R T I C L E S
Wu et al.
Nature has evolved a system of information storage and
retrieval based on complementary nucleobase H-bonds. How-
ever, there is no reason H-bonding should be uniquely capable
of supporting the intermolecular interactions required for
information storage and replication. We have demonstrated that
hydrophobicity, known to be a dominant force in proteins,28 is
also capable of providing the intermolecular interactions required
for nucleic acid structure and function, including base pair
stability in duplex DNA, enzymatic triphosphate insertion,
proofreading, and extension. We have recently developed a
binary polymerase system, comprised of KF and pol â, that is
capable of the in vitro replication of DNA containing three base
pairs.7 It is now apparent that the unnatural nucleoside analogues
may also be efficiently phosphorylated to their corresponding
monophosphates by Dm-dNK. Therefore, the first component
of an unnatural phosphorylation pathway in E. coli could be
based on transforming the bacteria with a plasmid coding for
the Dm-dNK, and supplementing the growth media with the
desired unnatural nucleosides. Although several important
enzymes remain to be identified, including mono- and diphos-
phate kinases, Dm-dNK joins KF and pol â as part of an
evolving system for the in vivo replication of DNA containing
three base pairs.
Acknowledgment. The research described in this report was
supported by the National Institutes of Health (GM 60005 to
F.E.R.) and the Skaggs Institute for Chemical Biology (F.E.R.
and P.G.S.).
(28) Dill, K. A. Biochem. 1990, 29, 7133-7155.
JA028050M
9
14630 J. AM. CHEM. SOC. VOL. 124, NO. 49, 2002