Published online 18 September 2007Nucleic Acids Research, 2007, Vol. 35, No. 19 6339–6349
Termination of DNA synthesis by N6-alkylated, not
3’-O-alkylated, photocleavable 2’-deoxyadenosine
Weidong Wu1, Brian P. Stupi1, Vladislav A. Litosh1, Dena Mansouri2,
Demetra Farley3, Sidney Morris1,3, Sherry Metzker1and Michael L. Metzker1,2,3,*
1LaserGen, Inc., Houston, TX 77054,2Department of Molecular & Human Genetics and3Human Genome
Sequencing Center, Baylor College of Medicine, Houston, TX 77030, USA
Received July 25, 2007; Revised August 17, 2007; Accepted August 21, 2007
The Human Genome Project has facilitated the
sequencing of many species, yet the current
Sanger method is too expensive, labor intensive
and time consuming to accomplish medical rese-
quencing of human genomes en masse. Of the
termination (CRT) is a promising method with the
goal of producing accurate sequence information at
a fraction of the cost and effort. The foundation of
this approach is the reversible terminator (RT), its
chemical and biological properties of which directly
impact the performance of the sequencing technol-
ogy. Here, we have discovered a novel paradigm in
RT chemistry, the attachment of a photocleavable,
2-nitrobenzyl group to the N6-position of 2’-deoxy-
adenosine triphosphate (dATP), which, upon incor-
poration, terminates DNA synthesis. The 3’-OH
unblocked, providing favorable incorporation and
termination properties for several commercially
available DNA polymerases while maintaining good
discrimination against mismatch incorporations.
Upon removal of the 2-nitrobenzyl group with UV
light, the natural nucleotide is restored without
molecular scarring. A five-base experiment, illus-
trating the exquisite, stepwise addition through a
homopolymer repeat, demonstrates the applicability
of the N6-(2-nitrobenzyl)-dATP as an ideal RT for
Next-generation technologies are being developed to
advance sequencing to the $100000, and eventually the
$1000 genome. A number of strategies, albeit at different
stages of development, have been proposed including
reversible termination (CRT), real-time sequencing and
nanopore sequencing (1–6). CRT is a promising approach,
which is comprised of a three-step process of incorporat-
deprotecting after which the cycle begins again (5,6).
CRT reactions can be performed in a high-density format
using single-molecule arrays (3) or oligonucleotide arrays
(5,7), eliminating the requirement for gel electrophoresis
while significantly increasing sequence throughput. At the
center of this technology is the reversible terminator (RT),
whereby DNA polymerases exhibit specific and efficient
incorporation of the modified nucleotide into the growing
primer strand, with deprotection chemistries resulting in
the efficient removal of the terminating group.
To date, known RTs have contained labile blocking
groups at the 30-OH of the ribose sugar resulting
in termination of synthesis (7–11). In 1994, Metzker
et al. reported the synthesis of 30-O-(2-nitrobenzyl)-
20-deoxyadenosine and incorporation of its triphosphate
by several DNA polymerases (8). The 2-nitrobenzyl group
and its derivatives are widely used as photocleavable,
‘caging’ functionalities for altering normal biomolecular
processes (12). Recently, we discovered that the reported
synthesis of 30-O-(2-nitrobenzyl)-20-deoxyadenosine Ia
[designated as compound 7 in Metzker et al. (8)]
was incorrect, and that the actual product obtained from
the reaction of 20-deoxyadenosine with 2-nitrobenzyl
bromide using NaH in DMF is N6,N6-bis-(2-nitro-
benzyl)-20-deoxyadenosine IIa (Scheme 1). To investigate
whether the 30-O-alkylated compound could act as a
terminator of DNA synthesis and confirm the identity of
the active triphosphate reported by Metzker et al. (8), we
synthesized and characterized
20-deoxyadenosine analogs Ia–Ic and N6,N6-bis-(2-nitro-
benzyl)-20-deoxyadenosine analogs IIa–IIc (Figure 1),
both of which Ic and IIc triphosphates proved inactive
*To whom correspondence should be addressed. Tel: +17137987565; Fax: +17137985741; Email: firstname.lastname@example.org
? 2007 The Author(s)
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/
by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
by DNA polymerase incorporation assays. This finding
led to the synthesis and characterization of a third set of
N6-(2-nitrobenzyl)-20-deoxyadenosine analogs IIIa–IIIc
(Figure 1), of which compound IIIb was originally
identified by us as an intermediate during the ultraviolet
(UV) light-induced deprotection of compound IIb to its
corresponding parent analog. Here, we report that the
active triphosphate described by Metzker et al. (8) is
actually the N6-(2-nitrobenzyl)-dATP IIIc, representing a
novel, 30-unblocked terminator of DNA synthesis and an
ideal candidate as a RT for the CRT approach.
MATERIALS AND METHODS
Chemical reagents and solvents were purchased from Alfa
Aesar, Sigma-Aldrich, or EM Sciences. Oligonucleotides
were purchased from Integrated DNA Technologies.
20-Deoxyadenosine triphosphate (dATP), 20,30-dideoxy-
adenosine triphosphate (ddATP) and Q Sepharose Fast
Flow anion-exchange resin were purchased from GE
Healthcare Life Sciences. All DNA polymerases and
apyrase were purchased from New England Biolabs,
with the exception of AmpliTaqFS being purchased
from Applied Biosystems (AB). Analytical silica gel 60
F254 TLC plates were purchased from Whatman, and
silica gel 60 (230–400 mesh) was purchased from EM
Nucleosides andnucleotides synthesis
Complete experimental procedures describing synthesis of
the compounds used in this work are available in the
UV deprotection studies forcompounds Ib,IIb andIIIb
50-O-tert-butyldimethylsilyl (TBS) derivatives Ib, IIb and
IIIb in methanol (0.2ml, 1mM) were transferred to a
Wheaton scintillation vial and irradiated with a 0.5mW
transilluminator light source at either 302 or 365nm.
Aliquots of the irradiated solution were taken at different
time intervals and analyzed for the loss of starting
material and appearance of the deprotected product by
reverse-phase (RP) HPLC. Deprotection half-life times
(DT1/2) were determined from kinetic plots at which 50%
of the compound was deprotected (i.e. loss of the
2-nitrobenzyl group). UV deprotection experiments were
performed in triplicate for each compound.
correct structure for
compound 7 in reference (8)
reported structure for
compound 7 in reference (8)
Scheme 1. Correct structure of the product obtained from the reaction of 20-deoxyadenosine with 2-nitrobenzyl bromide using NaH/DMF
conditions. The reported compound 7 in reference (8) was misassigned as the structure Ia.
Ia R = H
Ib R = TBS
Ic R = P3O9
IIc R = P3O9
IIa R = H
IIb R = TBS
IIIc R = P3O9
IIIa R = H
IIIb R = TBS
Figure 1. Structures of 30-O-alkylated and N6-alkylated 20-deoxyadenosine analogs.
Nucleic Acids Research, 2007, Vol. 35, No. 19
As described for the polymerase end-point (PEP) assays,
compounds Ic, IIc and IIIc were tested for incorpo-
ration with Bst DNA polymerase at concentrations
of 200nM and 100mM using the BODIPY-R6G labeled
primer-1 (50-TTGTAAAACGACGGCCAGT) (13) and
CGTCGTTTTACA, interrogation base is underlined
and bolded) complex. Reactions were quenched with
10ml of stop solution and analyzed using an AB model 377
All DNA polymerases (see Supplementary Data for
definitions) were assayed in 1? ThermoPol buffer
(20mM Tris–HCl, pH 8.8; 10mM (NH4)2SO4; 10mM
KCl; 2mM MgSO4; 0.1% Triton X-100, New England
BioLabs). We found that the addition of Triton X-100
stimulated the activity of many of the enzymes tested (data
not shown), which is consistent with other reports (14,15).
For all polymerases evaluated in this study, 5nM
BODIPY-FL labeled primer-1 was annealed with 40nM
GTCGTTTTACA, interrogation base is underlined and
bolded) in 1? ThermoPol buffer at 808C for 30s, 578C for
30s and then cooled to 48C. The primer/template complex
was then diluted in half (i.e. its final concentration was
2.5nM in a volume of 10ml) by the addition of DNA
polymerase, nucleotide analog and ThermoPol buffer.
This defines the lower limit of the IC50value for nucleotide
titrations to 1.25nM (i.e. [primer]=[primer plus incor-
porated nucleotide]). Polymerase reactions were incubated
at their appropriate temperature (Supplementary Table 1)
for 10min, then cooled to 48C and quenched with 10ml
of stop solution (98% deionized formamide; 10mM
Na2EDTA, pH 8.0; 25mg/ml Blue Dextran, MW
2000000). Stopped reactions were heated to 908C for
30s and then placed on ice. The extension products were
analyzed on a 10% Long Ranger (Cambrex) polyacryl-
amide gel using an AB model 377 DNA sequencer, the
quantitative data of which are displayed as a linear–log
plot of product formation versus compound concentra-
tion. PEP assays were performed in triplicate, for each
DNA polymerase/nucleotide analog combination, to
calculate the average IC50?1 SD.
Compound IIIc and ddATP were then titrated using the
PEP assay with the eight DNA polymerases (unit activities
defined in Supplementary Table 1) in the concentration
range of either 0.1nM to 100nM, 1nM to 1mM, 10nM to
10mM or 100nM to 100mM. Average IC50?1 SD values
were calculated for compound IIIc and ddATP using
oligoTemplate-2 as described above.
PEP termination assays forcompound IIIc
Biotin, interrogation bases are underlined and bolded),
annealed to 5nM BODIPY-FL labeled primer-1 and
assayed as described above. The PEP titrations were then
performed at 1?, 5? and 25? IC50values for compound
IIIc, using the eight DNA polymerases, and reported as %
primer product, % first-base product and % second-base
PEPmismatch assays forcompound IIIc
OligoTemplate-2 was substituted with either 40nM of
GTTTTACA), 40nM of oligoTemplate-5 (50-TACG
GAGCAGCACTGGCCGTCGTTTTACA) or 40nM of
CGTTTTACA, interrogation bases are underlined and
bolded), annealed to 5nM BODIPY-FL labeled primer-1,
and assayed as described above. Compound IIIc and
dATP were assayed in the concentration range of 100nM
to 100mM, and ddATP in the range of 500nM to 500mM.
Average IC50?1 SD values were calculated for dATP and
compound IIIc using oligoTemplate-4, -5 and -6, as
UV deprotection studies forcompound IIIc incorporated
into theprimer/template complex
As described for the PEP assays, compound IIIc was
incorporated at a concentration of 100nM, using the
BODIPY-FL labeled primer-1/oligoTemplate-2 complex,
and quenched with 10ml of stop solution. The stopped
reactions were exposed to 365nm light for 0, 10, 20, 30, 45
or 60s, using our custom-designed UV deprotector
(Supplementary Figure 1), then analyzed using an AB
model 377 DNA sequencer. The quantitative data are
displayed as a linear–log plot of product formation versus
time. Deprotection assays were performed in triplicate to
calculate the average DT1/2?1 SD.
Five-basesequencing experiment using compoundIIIc as
An 80nM solution of oligoTemplate-3 in 1M NaCl and
1? ThermoPol buffer (final volume: 12.5ml) was incubated
for 15min at room temperature with 5ml of streptavidin-
coated M-270 magnetic Dynabeads (Invitrogen), which
had been previously washed three times with 5ml 1?
ThermoPol buffer. The oligoTemplate-3 bound beads
were then washed an additional three times with 5ml 1?
GCTGTAAAACGACGGCCAGT) in 1? ThermoPol
buffer at 808C for 30s, 578C for 30s and then cooled to
48C. The beads were then washed twice with 5ml 1?
Incorporation step. The BODIPY-FL labeled primer-2/
oligoTemplate-3 complex bound beads were incubated
with four units of Bst DNA polymerase and 250nM of
compound IIIc in 1? ThermoPol buffer (reaction volume:
20ml) at 658C for 6min, then placed on ice. Compound
IIIc-incorporated beads were washed four times with 50ml
W10 washing solution (10mM Tris–HCl, pH 8.0; 10mM
Na2EDTA; 0.1% Triton X-100), then once with 20ml W10
Nucleic Acids Research, 2007, Vol. 35, No. 196341
Deprotection step. The beads were resuspended in 20ml
deprotection solution (20% aqueous deionized forma-
mide; 10mM Na2EDTA, pH 8.0; 16.6mg/ml Blue
Dextran, MW 2000000), exposed to 365nm light for
9min (i.e. 3?3min exposures interrupted with a 15-s
mixing step to ensure good resuspension of the beads)
using the customized UV deprotector (Supplementary
Figure 1), then washed four times with 50ml 1? apyrase
buffer (100U/l apyrase in 1? ThermoPol), three times
with 50ml W10 washing solution, twice with 50ml 1?
ThermoPol buffer and then once with 20ml 1? ThermoPol
The entire cycle was then repeated from the incorpora-
tion step. Final reactions were washed twice with 50ml
W10 washing solution, once with 20ml W10 washing
solution, quenched with 10ml of stop solution, heated to
508C for 30s and placed on ice. The extension products
were analyzed on a 10% Long Ranger polyacrylamide gel
using an AB model 377 DNA sequencer.
Adenine–thymine and N6-(2-nitrobenzyl)-adenine–thymine
base pairs were created with the nucleobases planar to
each other, using Watson–Crick hydrogen-bond distances
of 2.82A˚ (N1...H–N3) and 2.91A˚
software packages (CambridgeSoft). A series of 3D,
N6-(2-nitrobenzyl)-adenine–thymine base pairs were then
created by rotating the 2-nitrobenzyl group, pivoted on
the N6-position of adenine, 3608 at 58 intervals, with the
nitro group at 08, 308, 458, 608 and 908 intervals relative to
the plane of the phenyl group. Chem3D structures were
then further optimized using the GAMESS program (17).
Restricted Hartree–Fock (RHF) energy calculations were
initially determined using the STO-3G atomic orbital/shell
data set, and each nitro group conformation (i.e. 08, 308,
458, 608 or 908 intervals) was plotted as RHF energy
versus degrees of rotation of the 2-nitrobenzyl group. Our
calculations revealed that a 458 rotation of the nitro
group, relative to the plane of the phenyl group, gave the
lowest RHF energy calculations. The N6-(2-nitrobenzyl)-
adenine–thymine base-pair conformations were further
characterized using the more stringent 6-31G?atomic
orbital/shell calculations and plotted as RHF energy
versus degrees of rotation of the 2-nitrobenzyl group
(data not shown). To evaluate the efficacy of our existing
software tools, we modeled the natural adenine–thymine
base pair and compared our results to those reported by
Sˇponer et al. (18,19). Here, we used the 6-31G??atomic
orbital/shell set, with the X, Y, Z coordinates described
in Ref. (19) to perform the calculations. Showing
good agreement with these reports (Supplementary
Table 3), this served as an independent validation of our
Nucleosides and nucleotides synthesis
To synthesize the 30-O-(2-nitrobenzyl)-20-deoxyadenosine
analog Ia, the 50-hydroxyl and 6-amino groups of
20-deoxyadenosine were protected with tert-butyldimethyl-
silyl (TBS) and tert-butyloxycarbonyl (Boc) groups,
respectively, to yield intermediate 1, according to Furrer
and Giese (20). Transformation of this precursor into
intermediate 2 occurred via deprotection and selective
Alkylation of intermediate 2 with 2-nitrobenzyl bromide,
using phase transfer catalysis under basic conditions, gave
the desired 30-O alkylated intermediate 3 in 91% yield.
The bis-Boc groups were removed by heating on silica gel
under vacuum (20) to give compound Ib in 91% yield,
followed by the removal of the 50-O-TBS group with
tetra(n-butyl)ammonium fluoride to give compound Ia in
23% yield. Synthesis of the triphosphate was performed
using the ‘one-pot’ procedure described by Ludwig (21),
followed by purification using Q Sepharose FF anion-
exchange chromatography to yield compound Ic as an
Treatment of 20-deoxyadenosine with NaH in DMF
at 08C followed by 2-nitrobenzyl bromide gave bis-N6,
N6-(2-nitrobenzyl)-20-deoxyadenosine IIa in 22% yield
(Scheme 3). The assignment of the IIa structure was
based on the
showed a total of eight aromatic hydrogens derived from
the two 2-nitrobenzyl moieties and two D2O-exchangeable
20-deoxyribose. Selective 50-O-TBS protection gave com-
pound IIb in 57% yield. Phosphorylation of compound
IIa was performed using the same procedure described for
compound Ia, with the exception that purification of
triphosphate IIc was achieved by preparative HPLC.
N6-(2-Nitrobenzyl)-20-deoxyadenosine IIIa was pre-
pared based on the work of Wan et al. (22). Treatment
of 20-deoxyinosine with 2-nitrobenzylamine in the pre-
sence of 1-H-benzotriazol-1-yloxy-tris(dimethylamino)
N-diisopropylethylamine (DIPEA) in anhydrous DMF
gave compound IIIa in 98% yield. Selective 50-O-TBS
protection gave compound IIIb in 63% yield (Scheme 4).
Triphosphate IIIc was prepared from compound IIIa
using the one pot procedure (21) and purified in a manner
similar to that for triphosphate Ic.
Triphosphates Ic, IIc and IIIc (Figure 1) were further
purified by preparative RP-HPLC, without UV detection,
to provide modified triphosphates free from contamina-
tion by dATP, resulting from the deprotection of the
2-nitrobenzyl group under ordinary laboratory light
conditions during synthesis and purification processes.
in Scheme 2.
1H NMR spectra (in DMSO-d6), which
Identification ofactive nucleoside triphosphate IIIc
Triphosphates Ic and IIc were initially tested for base-
specific termination of DNA synthesis using a fluorescent-
based oligonucleotide template assay. All compounds
were handled in low light conditions to minimize
2-nitrobenzyl deprotection. A five-base poly-thymidine
template was employed to test for specific incorporation of
natural and modified dATP analogs. As shown in
Figure 2A, compounds Ic and IIc did not show significant
incorporation using Bst DNA polymerase, even at
high concentrations (100mM), although compound Ic
did exhibit natural nucleotide contamination, even after
Nucleic Acids Research, 2007, Vol. 35, No. 19
a second round of RP-HPLC purification was performed.
The dATP contamination could, however, be substantially
reduced using the Mop-Up assay (23). Next, we examined
the rates of UV deprotection for 50-O-TBS analogs Ib
and IIb, conducted at wavelengths of 302 and 365nm.
Compound Ib exhibited the expected first-order profile
with deprotection half-life times (DT1/2) of 60 and 152s at
302 and 365nm, respectively. The deprotection rate for
compound IIb was approximately 3-fold faster than that
of compound Ib at both deprotection wavelengths
(Figure 2B). UV deprotection of compound IIb, however,
revealed a transient intermediate before the appearance of
the 50-O-TBS-20-deoxyadenosine product (Figure 2C). We
suspected, and later confirmed, the identity of the
intermediate to be the mono N6-(2-nitrobenzyl)-20-deoxy-
Following synthesis of N6-(2-nitrobenzyl)-dATP IIIc,
we examined its incorporation using Bst DNA polymerase,
which showed efficient, base-specific termination of DNA
synthesis at a final concentration of 200nM (Figure 2A).
The 2-nitrobenzyl group was efficiently removed from the
DNA duplex using a custom built UV deprotector
(Supplementary Figure 1), evidenced by the band shift
of the extended dye-primer to the termination position of
the first thymidine of the oligonucleotide template.
Examination of the UV deprotection data for compound
IIIb revealed a first-order reaction with DT1/2values of
46 and 144s at 302 and 365nm, respectively (Figure 2B).
The UV deprotection data suggest that the attachment of
a single 2-nitrobenzyl group to either the 30-O or the
N6-aromatic amine position does not significantly alter the
rate of reaction for UV light-induced cleavage.
These data help to explain the structure misassignment
and positive incorporation data presented in the Metzker
et al. paper (8). Alkylation of adenosine with 2-nitrobenzyl
bromide using NaH in DMF has been reported to occur
on either the 20- or 30-hydroxyl group of the ribose ring
(24), but not on the exo-cyclic amino group of the adenine
base. Alkylation of 20-deoxyadenosine under the same
conditions, however, exclusively gave bis-N6,N6-alkylated
Scheme 2. Synthesis of 30-O-(2-nitrobenzyl)-20-deoxyadenosine analogs Ia–Ic. (i) TBSCl, imidazole, DMF, room temperature, overnight; Boc2O,
DMAP, DMF, room temperature, overnight, 83%; (ii) n-Bu4NF, THF, 08C, then gradually warmed to room temperature, 96%; (iii) TBSCl,
imidazole, DMF, 83%; (iv) n-Bu4NOH, NaI, NaOH, CH2Cl2/H2O, 2-nitrobenzyl bromide, room temperature, 91%; (v) SiO2, high vacuum, 70–808C,
24h, 91%; (vi) n-Bu4NF, THF, 08C, then gradually warmed to room temperature, 23%; (vii) POCl3, (MeO)3PO, minus 208C; (n-Bu3NH)2H2P2O7,
n-Bu3N, DMF; 1M HNEt3HCO3, 31%.
Nucleic Acids Research, 2007, Vol. 35, No. 196343
compound IIa, albeit in lower yield. Based on the data
presented here, we now conclude that the structure of
the alkylation product reported by Metzker et al. (8) is
the bis-N6,N6-(2-nitrobenzyl)-20-deoxyadenosine analog
(Scheme 1). The reported termination of the correspond-
ing triphosphate, occurring at a concentration of 250mM
[Figure 5 in Metzker et al. (8)], is most likely the result of
the contamination of minute quantities of triphosphate
IIIc, derived from triphosphate IIc during its handling
under ordinary laboratory light conditions. One of the
minor termination bands observed for compound IIc
incorporation with Bst DNA polymerase (Figure 2A,
100mM lane) reveals the presence of IIIc triphosphate.
Our data in Figure 2C further support our supposition
that during UV light-directed deprotection, compound IIb
is transformed into intermediate IIIb before undergoing
loss of the second 2-nitrobenzyl group to yield the natural
nucleotide (Figure 2D). From this investigation, we have
discovered that the N6-(2-nitrobenzyl)-dATP IIIc is the
active species of the three triphosphates examined here.
Compound IIIc isactive with avarietyof DNA polymerases
Numerous groups have employed qualitative, Sanger-
based assaystoestimate incorporation efficiencies,
although these methods are not feasible for assaying
modified analogs in the absence of natural nucleotides
(8,25,26). This led us to develop a quantitative, PEP assay,
which could be utilized for high-throughput screening of
modified nucleotides against commercially available DNA
polymerases. The PEP assay is designed with a polymerase
concentration in excess of the primer/template complex,
thereby limiting the reaction to nucleotide binding and
coupling steps. The desired nucleotide is then titrated
across the appropriate concentration range to observe
extension of a dye-primer by gel electrophoresis. The end-
point concentration, or IC50value, is the point at which
the number of moles of substrate equals that of the
product. The primer/template complex concentration
tide titrations to 1.25nM (i.e. [primer]=[primer plus
incorporated nucleotide]). The number of activity units
for eight commercially available DNA polymerases was
determined by titration with dATP (concentration range
from 0.1 to 100nM), with the goal of reaching the PEP
IC50limit of 1.25nM. In general, increasing the number
the exceptions being Therminator and Therminator II
(see Supplementary Data for polymerase definitions).
IIa R = H
IIb R = TBS
Scheme 3. Synthesis of N6,N6-bis-(2-nitrobenzyl)-20-deoxyadenosine analogs IIa–IIc. (i) NaH, 2-nitrobenzyl bromide, DMF, 08C, 22%; (ii) TBSCl,
imidazole, DMF, 57%; (iii) POCl3, (MeO)3PO, minus 208C; (n-Bu3NH)2H2P2O7, n-Bu3N, DMF; 1M HNEt3HCO3, yield not determined.
IIIa R = H
IIIb R = TBS
Scheme 4. Synthesis of N6-(2-nitrobenzyl)-20-deoxyadenosine analogs IIIa–IIIc. (i) BOP, DIPEA, 2-nitrobenzyl amine, DMF, room temperature,
20h, then 508C 3h, 98%; (ii) TBSCl, imidazole, DMF, 63%; (iii) POCl3, (MeO)3PO, minus 208C; (n-Bu3NH)2H2P2O7, n-Bu3N, DMF; 1M
Nucleic Acids Research, 2007, Vol. 35, No. 19
For these enzymes, increasing the number of units
yielded an increase in IC50 values for dATP. This
observation was not investigated further. In these
cases, the number of units used for subsequent PEP
assays were those yielding the lowest IC50 value for
dATP (Supplementary Table 1, highlighted blue boxes).
N6-(2-Nitrobenzyl)-dATP IIIc was tested for base-
specific incorporation using eight polymerases with the
PEP assay and compared with assay data for dATP and
ddATP. Compound IIIc was incorporated by all poly-
merases examined, with IC50values ranging from 2.1nM
to 2.1mM (Table 1). Five of the eight polymerases revealed
a less than 4-fold preference for dATP over compound
IIIc, with Therminator and Vent(exo?) showing the least
bias of 1:1.3, providing evidence that compound IIIc is
incorporated almost as efficiently as dATP itself. In all
cases except AmpliTaqFS, compound IIIc was preferred
over ddATP with an incorporation bias range of
3.8?10?3:1 to 0.32:1, respectively. The F667Y mutation
in AmpliTaqFS has been shown to prefer ddATP over
dATP, with a ratio of 0.59:1 (25), which is in good
agreement with the ratio of 0.62:1 reported in Table 1.
Next, we examined the ability of N6-(2-nitrobenzyl)-
dATP IIIc to terminate DNA synthesis using an
oligonucleotide template containing a stretch of 10
thymidine bases. As expected, Bst DNA polymerase
extended the growing primer utilizing dATP as a substrate
in a concentration-dependent manner. At a concentration
of 25? its IC50value, the enzyme completely extended the
10 thymidine template and partially misincorporated
cpd IIIc (UV)
cpd Ic (Mop)
200 nM100 µM
IIb IIIb 5′-TBS-dA
* Removal of first 2-nitrobenzyl group only
60 ± 8 152 ± 13
17 ± 6 49 ± 13
46 ± 4 144 ± 11
DT½ at 302 nm
DT½ at 365 nm
Figure 2. Identification of N6-(2-nitrobenzyl)-20-deoxyadenosine triphosphate IIIc as the active analogue. (A) OligoTemplate assay for compounds Ic,
IIc and IIIc, analyzed on a 377 DNA sequencer. The final concentration for ddATP was 50mM. (B) Summary of the UV deprotection experiments
for compounds Ib, IIb and IIIb. (C) Time plot of the UV light-induced deprotection of compound IIb at 302nm. (D) Proposed stepwise
photocleavage of N6,N6-bis-(2-nitrobenzyl)-20-deoxyadenosine IIb.
Nucleic Acids Research, 2007, Vol. 35, No. 19 6345
dATP against a ‘G’ template base at the 11th base position
(Figure 3). In contrast, the N6-(2-nitrobenzyl)-dATP IIIc
efficiently incorporated and terminated Bst DNA synth-
esis at the first-base position up to 25? its IC50value. The
difference in electrophoretic mobility of the first-base
product for dATP and that of compound IIIc is due to the
N6-attached 2-nitrobenzyl group.
IC50values for compound IIIc were retitrated using the
poly(dT) template. In some cases, IC50 values differed
from those in Table 1, reflecting DNA sequence context
effects. PEP termination assays were performed for the
eight polymerases at 1?, 5? and 25? the IC50 values
(Table 2). Three of the four Family A DNA polymerases
resulted in efficient termination of DNA synthesis at the
first-base position using compound IIIc, while all Family
B DNA polymerases showed significant, but variable,
levels of second-base product. Therminator provided
the most extreme example with ?98.5% of the growing
primer extended as a second-base product. The majority
of polymerases extended the primer with an efficiency of
?99%. Based upon the desired properties of termination
at the first-base position and efficient primer extension, Bst
emerged as a promising CRT polymerase in combination
with compound IIIc.
Compound IIIc shows good mismatch discrimination with
Bst DNA polymerase
PEP discrimination assays were then performed to
evaluate the specificity of N6-(2-nitrobenzyl)-dATP IIIc
against mismatched template bases (i.e. A, C or G).
Comparing IC50 values of mismatched versus matched
bases for compound IIIc revealed nucleotide discrimina-
tion of greater than two orders of magnitude (Figure 4A).
Surprisingly, the cytosine–adenine mismatch revealed the
highest discrimination ratio of 1100-fold over the comple-
ment thymidine base, and only slightly less than that for
dATP (Supplementary Table 2). The ddATP analog did
not show mismatch incorporation at a final concentration
of 500mM (data not shown). These data suggest that
compound IIIc incorporates as a base-specific terminator
and reveals nucleotide characteristics more similar to
dATP than those of ddATP, which we attribute to the
presence of the 30-OH group.
The N6-proton of 2-nitrobenzyl-adenine base, still
capable of base pairing, may aid in the specificity observed
in Figure 4A. To examine this further, we performed
ab initio calculations for a thymine–N6-(2-nitrobenzyl)-
adenine base pair, using the Hartree–Fock method coupled
with the 6-31G?atomic shell set (17). The optimal
Table 2. PEP termination results for compound IIIc
% Primer % 1st Base % 2nd Base
Table 1. PEP assay results for dATP, compound IIIc and ddATP
DNA polymerase Average IC50?1 SD Incorporation bias
dATPCpd IIIcddATP Cpd IIIc/dATPCpd IIIc/ddATP ddATP/dATP
1x 5x 25xP 1x2x25x
dATP cpd IIIc
IC50: 10 nM
IC50: 1.2 nM
1st Base Cpd IIIc
Figure 3. PEP termination assay comparing dATP with compound IIIc
using Bst DNA polymerase.
Nucleic Acids Research, 2007, Vol. 35, No. 19
molecular structure was a nitro group positioned 458 and
a 2-nitrobenzyl group positioned 808 counterclockwise,
relative to the aromatic amino group (Figure 4B).
Hydrogen bond distances were determined to be 2.93A˚
(N1...H–N3) and 2.97A˚(N6–H...O4), which are longer
than those reported by Watson and Crick (16). A ?E of
–2.68kcal/mol for the modified base pair, however,
suggests that hydrogen bonding is favorable. Active site
tightness (27,28) of compound IIIc, involving hydro-
phobic interactions of the 2-nitrobenzyl with key amino
acids, may also contribute to the observed enzymatic
may also be involved with misalignment of the 30-OH
group, preventing nucleophilic attack by the incoming
nucleotide, thus terminating DNA synthesis.
Five-baseCRT sequencing withcompound IIIc
To test N6-(2-nitrobenzyl)-dATP IIIc as a RT in CRT
sequencing, a five-base experiment was performed using
a biotinylated template containing a poly(dT) stretch
(Figure 5A). For the first cycle, incorporation ‘a’ and
deprotection ‘b’ products are shown in the gel, with
subsequent cycles showing only incorporation products.
The data illustrate an advantage over the pyrosequencing
method (29), namely the stepwise addition through a
homopolymer repeat. The gel image in Figure 5A was
analyzed further by quantitating the fluorescent bands at
different CRT cycles (Figure 5B). During the first cycle,
the product of incorporation efficiency (I: 98.6%) and
deprotection efficiency (D: 94.0%) resulted in a cycle
Figure 4. (A) PEP discrimination curves for compound IIIc. (B) Proposed thymine–N6-(2-nitrobenzyl)-adenine base pair.
P 1a 2a3a 5a 1b4a
P 1a 1b 2a3a 4a5a
Figure 5. (A) Five-base CRT sequencing with N6-(2-nitrobenzyl)-dATP IIIc. Lanes: ‘P’ represents the dye-primer, ‘1a–5a’ represents final
incorporations of compound IIIc at different cycles, and ‘1b’ represents the first base addition followed by UV deprotection. (B) Histogram plot
of the gel image in Figure 5A represents the signal intensities of band products. ‘T’ is the percentage of total signal to that of the dye-primer signal
(lane P), ‘I’ is the incorporation efficiency, ‘D’ is the deprotection efficiency and ‘C’ is the percentage of correct +1 product to that of the total signal.
Nucleic Acids Research, 2007, Vol. 35, No. 19 6347
efficiency (Ceff) of 92.7% on a solid support. The estimated
signal ‘S’, calculated from the equation [S=(Ceff)RL] (5),
is 68.4% for the five-base read. At cycle five, the signal for
the correct +1 product is 40% (i.e. T?C), the difference
of which we attribute to photobleaching of the dye-primer
with increasing exposure to the UV light (data not shown).
Signal loss is also due to dephasing, observed as n?1
(incomplete deprotection) and n+1 products (natural
nucleotide carryover from the previous cycle). While
ongoing efforts are focused on reducing dephasing
products, ?80% of the total signal is derived from the
correct +1 product, with base-calling easily performed
from the primary data in Figure 5A and B.
We have discovered that the attachment of a small,
photocleavable 2-nitrobenzyl group to the N6-position
of 20-deoxyadenosine results in this triphosphate acting
as a RT ‘without blocking the 30-end’. The novel RT,
N6-(2-nitrobenzyl)-dATP IIIc, provides favorable enzy-
matic properties with a variety of wild-type and mutant
DNA polymerases. This is unlike the situation for
30-modified nucleotides, which typically act as poor
substrates for DNA polymerases. For example, screening
30-O-allyl-dATP with eight different DNA polymerases
revealed limited activity at high micromolar concentra-
tions with only Vent(exo?) DNA polymerase (8). The
highly related 98N(exo?) DNA polymerase (30), con-
taining the A485L and Y409V amino acid variants
requires up to 50min per single base addition, highlighting
the difficulty of incorporating these analogs (7,10). The
A485L and Y409V mutations are analogous to those
described for Vent(exo?) DNA polymerase (26), with the
Y409V residue acting as a ‘steric’ gate for incorporation of
ribonucleotides (26,31–33). Little is known regarding the
mechanism by which a 20-steric gate residue alters the
incorporation of 30-O-allyl terminators.
Research efforts have been focused on optimizing the
cycle efficiency and time, which determine read-length and
throughput, respectively. Targeting a 50% loss in signal as
an end-point (5), the cycle efficiency must be ?97.3% to
achieve a 25 base read-length. Although we show a slightly
smaller cycle efficiency of 92.7% with compound IIIc,
primarily due to its deprotection efficiency of 94%, work
is ongoing to improve this by substitution of the
2-nitrobenzyl group. We also anticipate further improve-
ments in cycle efficiency with development of instrumenta-
tion supportingthe CRT
integration and automation of the incorporation, imaging,
deprotection and washing steps.
Although the cycle efficiency is primarily the product of
incorporation and deprotection efficiencies of the RT,
other factors including dephasing (natural nucleotide
carryover) and accumulating ‘molecular scars’ (residual
linker structures left over after deprotection) can also
influence the efficiency in an adverse manner. The
N6-attachment to the adenine nucleobase is also unique,
differing from that of the traditional 7-deaza position of
BigDye terminators (34) and 30-O-allyl terminators (7,10).
Upon chemical deprotection of the 30-O-allyl terminators,
a residual propargyl amino group remains on the
nucleobase, resulting in an accumulating molecular scar
with subsequent CRT cycles. The primary advantage of
N6-alkylation is that, upon directed photocleavage with
365nm UV light, the modified nucleotide is transformed
back into its natural state without molecular scarring
(Figure 2D). The enhanced enzymatic properties and the
N6-cleavage site, transforming the efficiently incorporated
RT back into natural DNA, are anticipated to improve
the cycle efficiency and read-length of the CRT method.
These observations suggest that the N6-(2-nitrobenzyl)-
dATP IIIc represents an ideal candidate as a RT for CRT
sequencing, illustrated by stepwise, single base addition
through a homopolymer repeat. Of the polymerases tested
here, however, not all exhibited this property, with the
Family B polymerases revealing incorporation of a second
modified nucleotide, albeit at varying levels. Recently, we
have discovered that substitution of the 2-nitrobenzyl
group can also ‘tune’ the termination properties of
these polymerasesto give
products (unpublished data). To exploit the application
of 30-unblocked RTs in CRT sequencing, efforts are
underway to create fluorescently labeled analogs of
compound IIIc, and extrapolate this nucleotide model to
the remaining nucleobases for production of a novel, four-
color RT set. We note that the adenine example as a RT
may not be directly applicable to the remaining nucleo-
bases, which present their own unique challenges.
Nonetheless, a base modification strategy can still be
employed with careful selection of the attachment site of
the 2-nitrobenzyl group on each nucleobase structure, the
triphosphates of which exhibit similar enzymatic proper-
ties described in this report and transform into its natural
nucleobase structure upon UV deprotection (manuscript
We anticipate that 30-unblocked terminators will have
utility beyond the application of CRT sequencing. For
example, a complete set of non-fluorescent RTs could be
used in pyrosequencing, with the advantage of improving
accuracy through homopolymer repeat stretches. Reduced
incorporation biases of 30-unblocked terminators over
natural nucleotides, exhibited by several polymerases, may
also prove useful for more accurate heterozygote analysis
in Sanger sequencing. With other applications envisioned,
30-unblocked terminators may well find their way into the
general arsenal of molecular biology tools used in genomic
Supplementary Data are available at NAR Online.
National Institutes of Health (R01 HG003573, R41
HG003072, and R43 HG003443). Funding to pay the
Nucleic Acids Research, 2007, Vol. 35, No. 19
Open Access publication charges for this article was Download full-text
provided by R01 HG003573.
Conflict of interest statement. We declare that LaserGen
plans on commercializing this compound, along with its
derivatives. No other conflicts have been declared.
1. Shendure,J., Mitra,R.D., Varma,C. and Church,G.M. (2004)
Advanced sequencing technologies: methods and goals. Nat. Rev.
Genet., 5, 335–344.
2. Bai,X., Edwards,J. and Ju,J. (2005) Molecular engineering
approaches for DNA sequencing and analysis. Expert Rev. Mol.
Diagn., 5, 797–808.
3. Bennett,S.T., Barnes,C., Cox,A., Davies,L. and Brown,C. (2005)
Toward the $1000 human genome. Pharmacogenomics, 6, 373–382.
4. Chan,E.Y. (2005) Advances in sequencing technology. Mutat. Res.,
5. Metzker,M.L. (2005) Emerging technologies in DNA sequencing.
Genome Res., 15, 1767–1776.
6. Fan,J.-B., Chee,M.S. and Gunderson,K.L. (2006) Highly parallel
genomic assays. Nat. Rev. Genet., 7, 632–644.
7. Ju,J., Kim,D.H., Bi,L., Meng,Q., Bai,X., Li,Z., Li,X., Marma,M.S.,
Shi,S. et al. (2006) Four-color DNA sequencing by synthesis using
cleavable fluorescent nucleotide reversible terminators. Proc. Natl
Acad. Sci. USA, 103, 19635–19640.
8. Metzker,M.L., Raghavachari,R., Richards,S., Jacutin,S.E.,
Civitello,A., Burgess,K. and Gibbs,R.A. (1994) Termination of
DNA synthesis by novel 30-modified deoxyribonucleoside tripho-
sphates. Nucleic Acids Res., 22, 4259–4267.
9. Canard,B. and Sarfati,R. (1994) DNA polymerase fluorescent
substrates with reversible 30-tags. Gene, 148, 1–6.
10. Ruparel,H., Bi,L., Li,Z., Bai,X., Kim,D.H., Turro,N.J. and Ju,J.
(2005) Design and synthesis of a 30-O-allyl photocleavable
fluorescent nucleotide as a reversible terminator for DNA sequen-
cing by synthesis. Proc. Natl Acad. Sci. USA, 102, 5932–5937.
11. Barnes,C., Balasubramanian,S., Liu,X., Swerdlow,H. and Milton,J.
(2006) US patent 7,057,206 B2.
12. Corrie,J.E.T. (2005) Photoremovable protecting groups used for
caging of biomolecules. In Goeldner,M. and Givens,R. (eds),
Dynamics Studies in Biology. Wiley-VCH Verlag GmbH & Co.
KGaA, Weinheim, pp. 1–28.
13. Metzker,M.L., Lu,J. and Gibbs,R.A. (1996) Electrophoretically
uniform fluorescent dyes for automated DNA sequencing. Science,
14. Innis,M.A., Myambo,K.B., Gelfand,D.H. and Brow,M.A.D. (1988)
DNA sequencing with Thermus aquaticus DNA polymerase and
direct sequencing of polymerase chain reaction-amplified DNA.
Proc. Natl Acad. Sci. USA, 85, 9436–9440.
15. Kong,H., Kucera,R.B. and Jack,W.E. (1993) Characterization of a
DNA polymerase from the hyperthermophile archaea Thermococcus
litoralis. J. Biol. Chem., 268, 1965–1975.
16. Watson,J.D. and Crick,F.H. (1953) Molecular structure of nucleic
acids; a structure for deoxyribose nucleic acid. Nature, 171,
17. Schmidt,M.W., Baldridge,K.K., Boatz,J.A., Elbert,S.T.,
Gordon,M.S., Jensen,J.H., Koseki,S., Matsunaga,N.,
Nguyen,K.A. et al. (1993) General atomic and molecular electronic
structure system. J. Comput. Chem., 14, 1347–1363.
18. Sˇponer,J., Leszczynski,J. and Hobza,P. (1996) Structures and
energies of hydrogen-bonded DNA base pairs. A nonempirical
study with inclusion of electron correlation. J. Phys. Chem., 100,
19. Sˇponer,J., Jurec ˇ ka,P. and Hobza,P. (2004) Accurate interaction
energies of hydrogen-bonded nucleic acid base pairs. J. Am. Chem.
Soc., 126, 10142–10151.
20. Furrer,E. and Giese,B. (2003) On the distance-independent hole
transfer over long (A?T)n-sequences in DNA. Helv. Chim. Acta, 86,
21. Ludwig,J. (1981) A new route to nucleoside 50-triphosphates. Acta
Biochem. Biophys. Acad. Sci. Hung., 16, 131–133.
22. Wan,Z.K., Binnun,E., Wilson,D.P. and Lee,J. (2005) A highly
facile and efficient one-step synthesis of N6-adenosine and
N6-20-deoxyadenosine derivatives. Org. Lett., 7, 5877–5880.
23. Metzker,M.L., Raghavachari,R., Burgess,K. and Gibbs,R.A. (1998)
Elimination of residual natural nucleotides from 30-O-modified-
dNTP syntheses by enzymatic Mop-Up. BioTechniques, 25,
24. Ohtsuka,E., Tanaka,S. and Ikehara,M. (1977) Studies on transfer
ribonucleic acids and related compounds. XVI. Synthesis of
ribooligonucleotides using a photosensitive o-nitrobenzyl
protection for the 20-hydroxyl group. Chem. Pharm. Bull., 25,
25. Tabor,S. and Richardson,C.C. (1995) A single residue in DNA
polymerases of the Escherichia coli DNA polymerase I family is
critical for distinguishing between deoxy- and dideoxyribonucleo-
tides. Proc. Natl Acad. Sci. USA, 92, 6339–6343.
26. Gardner,A.F. and Jack,W.E. (1999) Determinants of nucleotide
sugar recognition in an archaeon DNA polymerase. Nucleic Acids
Res., 27, 2545–2553.
27. Kool,E.T. (2002) Active site tightness and substrate fit in DNA
replication. Annu. Rev. Biochem., 71, 191–219.
28. Henry,A.A. and Romesberg,F.E. (2003) Beyond A, C, G and T:
augmenting nature’s alphabet. Curr. Opin. Chem. Biol., 7,
29. Ronaghi,M., Uhlen,M. and Nyren,P. (1998) A sequencing method
based on real-time pyrophosphate. Science, 281, 363–365.
30. Southworth,M.W., Kong,H., Kucera,R.B., Jannasch,H.W. and
Perler,F.B. (1996) Cloning of thermostable DNA polymerases from
hyperthermophilic marine Archaea with emphasis on Thermococcus
sp. 9 degrees N-7 and mutations affecting 30-50exonuclease activity.
Proc. Natl Acad. Sci. USA, 93, 5281–5285.
31. Gao,G., Orlova,M., Georgiadis,M.M., Hendrickson,W.A. and
Goff,S.P. (1997) Conferring RNA polymerase activity to a
DNA polymerase: a single residue in reverse transcriptase
controls substrate selection. Proc. Natl Acad. Sci. USA, 94,
32. Astatke,M., Ng,K., Grindley,N.D.F. and Joyce,C.M. (1998) A
single side chain prevents Escherichia coli DNA polymerase I
(Klenow fragment) from incorporating ribonucleotides. Proc. Natl
Acad. Sci. USA, 85, 3402–3407.
33. Fa,M., Radeghieri,A., Henry,A.A. and Romesberg,F.E. (2004)
Expanding the substrate repertoire of a DNA polymerase by
directed evolution. J. Am. Chem. Soc., 126, 1748–1754.
34. Rosenblum,B., Lee,L., Spurgeon,S., Khan,S., Menchen,S.,
Heiner,C. and Chen,S. (1997) New dye-labeled terminators for
improved DNA sequencing patterns. Nucleic Acids Res., 25,
Nucleic Acids Research, 2007, Vol. 35, No. 19 6349