Site-specific labeling of DNA and RNA using an efficiently replicated and transcribed class of unnatural base pairs

Article abstract of DOI:10.1021/ja207907d

Site-specific labeling of enzymatically synthesized DNA or RNA has many potential uses in basic and applied research, ranging from facilitating biophysical studies to the in vitro evolution of functional nucleic acids and the construction of various nanom

Full text of DOI:10.1021/ja207907d

ARTICLE
pubs.acs.org/JACS
Site-Specific Labeling of DNA and RNA Using an Efficiently Replicated
and Transcribed Class of Unnatural Base Pairs
Young Jun Seo,† Denis A. Malyshev,† Thomas Lavergne,† Phillip Ordoukhanian, and Floyd E. Romesberg*
Department of Chemistry and Center for Protein and Nucleic Acid Research, The Scripps Research Institute,
10550 N. Torrey Pines Road, La Jolla, California 92037, United States
S Supporting Information
b
ABSTRACT: Site-specific labeling of enzymatically synthesized DNA or RNA has many
potential uses in basic and applied research, ranging from facilitating biophysical studies to the
in vitro evolution of functional nucleic acids and the construction of various nanomaterials and
biosensors. As part of our efforts to expand the genetic alphabet, we have developed a class of
unnatural base pairs, exemplified by d5SICS-dMMO2 and d5SICS-dNaM, which are efficiently
replicated and transcribed, and which may be ideal for the site-specific labeling of DNA and RNA.
Here, we report the synthesis and analysis of the ribo- and deoxyribo-variants, (d)5SICS and
(d)MMO2, modified with free or protected propargylamine linkers that allow for the site-specific
modification of DNA or RNA during or after enzymatic synthesis. We also synthesized and
evaluated the α-phosphorothioate variant of d5SICSTP, which provides a route to backbone
thiolation and an additional strategy for the postamplification site-specific labeling of DNA. The
deoxynucleotides were characterized via steady-state kinetics and PCR, while the ribonucleosides
were characterized by the transcription of both a short, model RNA as well as full length tRNA.
The data reveal that while there are interesting nucleotide and polymerase-specific sensitivities to linker attachment, both
(d)MMO2 and (d)5SICS may be used to produce DNA or RNA site-specifically modified with multiple, different functional groups
with sufficient efficiency and fidelity for practical applications.
DNA, the efficiency of modified triphosphate incorporation is
generally reduced compared to their natural counterparts, result-
ing in lower yields of amplified product as well as strong sequence
dependences.14,25 Foroligonucleotide evolution experiments, this
sequence-dependent efficiency is especially problematic because it
decouples amplification level from copy number, and thus from
fitness, during the amplification step required after or during
selection. Moreover, high functional group density is unavoidable
in any nucleotide-specific labeling strategy (i.e., all or none of a given
nucleotide must be modified), but the majority of functional groups
are unlikely to contribute to any desired property while still
contributing to biases in replication. Finally, it is note-
worthy that proteins rarely (if ever) employ such a high density
of functional groups but instead typically have only one or a few that
are embedded in a controlled environment where their activity (i.e.,
pKa, nucleophilicity, redox potential, etc.) can be manipulated. Thus,
methods to site-specifically label nucleic acids with one or a few
functional groups attached to the fifth and sixth nucleotides of an
expanded genetic alphabet would enable less biased selections and
potentially the construction of more protein-like environments.
The development of an efficiently replicated and transcribed
unnatural base pair would make possible the site-specific mod-
ification of DNA and RNA. While the four-letter genetic alphabet
based on complementary hydrogen-bonding (H-bonding)
is conserved throughout nature, and orthogonal H-bonding
1. INTRODUCTION
The sequence-specific amplification of DNA and RNA has
revolutionized biotechnological and biomedical research, making
techniques such as cloning and sequencing routine and also en-
abling a variety of emerging techniques with important applica-
tions in diverse fields ranging from synthetic biology1 to nano-
materials2,3 and medical diagnostics.4,5 In addition, the use of
polymerases to amplify DNA molecules has made it possible to
evolve functional DNAs or RNAs with desired ligand recognition or
catalytic properties.6,7 However, the potential physical and functional
properties of nucleic acids are restricted by the limited diversity
of their constituent nucleotides. While the potential diversity
of DNA or RNA may be expanded via chemically synthesizing the
oligonuclotides,8ꢀ10 this approach is incompatible with enzymatic
amplification. Thus, efforts have been directed toward the modifica-
tion of one or more of the natural triphosphates with linkers that
enable the attachment of different functional groups without ablating
polymerase recognition.11ꢀ21 The best sites of linker attachment
have proven to be the nucleobases: specifically, the C5 position
of the pyrimidines and the C7 position of the 7-deaza purine
scaffold. With the pyrimidine scaffold, the detailed structure of
the linker is critical,13,15 with rigid alkynes or trans-alkenes being
optimal.13,19,22,23 At least with the propargylamide-like linkers and
A-family polymerases (such as Taq), this appears to result at least in
part from favorable hydrogen-bonding interactions between the
polymerase and the amide carbonyl group of the linker.24
While the linker modification of one or more natural dNTPs
allows for the polymerase-mediated synthesis of highly functionalized
Received: August 21, 2011
Published: October 08, 2011
r
2011 American Chemical Society
19878
dx.doi.org/10.1021/ja207907d J. Am. Chem. Soc. 2011, 133, 19878–19888
|

Journal of the American Chemical Society
ARTICLE
Figure 1. (A) Parental unnatural base pairs, (B) linker-modified analogues, and (C) α-phosphorothioate variant of d5SICSTP. Sugar and phosphate
backbone are omitted for clarity in A and B.
patterns have been used to design unnatural base pairs,26ꢀ29
we30ꢀ41 and others43ꢀ47 have demonstrated the potential of
hydrophobic and packing forces to control the efficient and
selective replication of unnatural base pairs. However, as with the
natural nucleotides, the general use of an unnatural base pair for
the site-specific labeling of DNA or RNA requires that the
unnatural nucleotides bear linkers or other functionalities that
allow different groups to be attached without ablating polymerase
recognition.
Two of the most promising unnatural base pairs that we have
developed are those formed between d5SICS and dMMO2
(d5SICS-dMMO2) or dNaM (d5SICS-dNaM) (Figure 1A).
Both unnatural base pairs can be amplified by PCR with good
to excellent efficiency and fidelity using a variety of DNA poly-
merases, even when embedded within difficult to replicate
sequences.30,34,38,42 While d5SICS-dNaM is replicated and tran-
scribed with greater efficiency, our efforts to develop an un-
natural base pair for in vitro applications include both d5SICS-
dNaM and d5SICS-dMMO2, because unlike the annular ring
scaffold of dNaM, the dMMO2 scaffold may be modified with
linkers at a position that is analogous to the C5 position of the
natural pyrimidines.
to site-specifically introduce a thiolate for functional group
attachment. Overall, we demonstrate that several of the modified
nucleotides are sufficiently well recognized by polymerases to
enable practical applications based on the site-specific labeling of
nucleic acids with multiple different functionalities.
2. RESULTS
2.1. Synthesis of Modified Unnatural Nucleotides. With
natural triphosphates, propargylamide linkers are relatively well
tolerated by different polymerases.13,23 To explore the use of
similar linkers with d5SICS-dNaM and d5SICS-dMMO2, we
synthesized (d)5SICSPATP, (d)5SICSATP, d5SICSBIOTP,
d5SICSSSBIOTP, (d)MMO2PATP, (d)MMO2ATP, and (d)-
MMO2BIOTP (Figure 1B), as described in detail in the Support-
ing Information. Briefly, d5SICSPA was generated from the
isosteryl compound, which was iodinated and then sulfonylated.
The modified base was coupled to (2R,5R)-5-chloro-2-(((4-
methylbenzoyl)oxy)methyl)tetrahydrofuran-3-yl4-methylbenzoate,
resulting in anomeric mixtures of nucleosides, with the pure
β-anomer obtained by column chromatography. After removal of
the toluoyl group, the iodo functionality was used to attach the
dichloro acetyl protected propargylamine via Sonogashira cou-
pling. Phosphorylation under Ludwig conditions49 provided
d5SICSPATP, which was purified by anion exchange chroma-
tography followed by HPLC, and then deprotected to provide
d5SICSATP. The biotinylated analogues, d5SICSBIOTP and
d5SICSSSBIOTP, were synthesized by coupling d5SICSATP to
NHS-PEG4-biotin or NHS-SS-PEG4-biotin, respectively. To
explore the use of a site-specifically incorporated backbone thiolate,
d5SICSαSTP (Figure 1C) was synthesized via the phosphoryla-
tion of the corresponding nucleoside in the presence of thiopho-
sphoryl chloride.50 Note that this usually results in a mixture of
Sp and Rp stereoisomers and that only the Sp is recognized by
DNA polymerases.51 For dMMO2 derivatives, we iodinated the
free nucleoside, which was synthesized as described previously,40
and then attached the dichloro acetyl protected propargylamine
to provide the desired protected nucleoside.52 The nucleo-
side was then phosphorylated to provide dMMO2PATP, which
was subsequently deprotected to provide dMMO2ATP. The
Here, we report the synthesis and evaluation of protected and
free propargylamine linker-derivatized variants of both the deoxy-
and ribonucleosides of MMO2 and 5SICS ((d)MMO2PA, (d)-
MMO2A, (d)5SICSPA, and (d)5SICSA), as well as the biotiny-
lated deoxynucleotides dMMO2SSBIO, d5SICSBIO, andd5SICSSSBIO
(Figure 1B). The deoxynucleotides were characterized via both
steady-state kinetics and PCR, and the ribonucleosides were
characterized by the transcription of short model RNAs as well as
Tyr
a full length amber suppressor tyrosyl tRNA (tRNA ) from
CUA
Methanocaldococcus jannaschii.48 Both the protected and free
amino variants were evaluated to fully explore double labeling
strategies, including those involving postsynthetic modification.
The data reveals that the modifications of the unnatural nucleo-
tides have varying effects on the different steps of replication and
transcription and that the linkers are generally less perturbative
within the (d)MMO2 scaffold. We also report the synthesis and
evaluation of the α-phosphorothioate variant of d5SICSTP
(d5SICSαSTP, Figure 1C) and show that it is an efficient route
19879
dx.doi.org/10.1021/ja207907d |J. Am. Chem. Soc. 2011, 133, 19878–19888

Journal of the American Chemical Society
ARTICLE
Table 1. Steady-State Kinetic Data for Kf-Mediated Insertion of Modified Triphosphates (dYTP) Opposite Their Cognate
Nucleotide in the Template (X)
50-d(TAATACGACTCACTATAGGGAGA)
30-d(ATTATGCTGAGTGATATCCCTCTXGCTAGGTTACGGCAGGATCGC)
X
Y
k
cat (minꢀ1
)
KM (μM)
kcat/KM ꢁ 105 (Mꢀ1 minꢀ1
)
T
Aa
NaMa
4.1 ( 0.3
14.6 ( 0.8
14 ( 2
0.0053 ( 0.0004
0.25 ( 0.04
35.8 ( 0.6
15 ( 2
7700
580
4.0
5SICS
MMO2a
MMO2PA
MMO2A
MMO2SSBIO
5SICSa
5SICSPA
5SICSA
11 ( 2
7.2
6.7 ( 0.3
1.5 ( 0.3
8.3 ( 1.1
8.1 ( 0.9
69 ( 8
0.97
0.34
2100
83
45 ( 10
NaM
0.039 ( 0.004
1.0 ( 0.1
19 ( 4
0.085 ( 0.05
6.5 ( 1.5
0.045
160
0.18
0.45
b
5SICSαS
5SICSBIO
5SICSSSBIO
0.42 ( 0.16
31 ( 12
0.55 ( 0.13
1.3 ( 0.3
29 ( 9
a Ref 42. b Mixture of Sp and Rp diastereomers.
cleavable biotinylated analogue, dMMO2SSBIOTP, was synthe-
sized via coupling of dMMO2ATP with NHS-SS-PEG4-biotin.
For the 5SICS ribonucleoside derivatives, 5-methyl isocarbos-
tyril was first coupled with (2R)-2-((benzoyloxy)methyl)-5-
oxotetrahydrofuran-3,4-diyl dibenzoate. The pure β anomer was
purified from an anomeric mixture by column chromatography
to provide the desired benzoyl protected nucleoside. Iodination,
sulfonylation, and benzoyl deprotection, followed by coupling
with the dichloro acetyl propargylamine, then provided
5SICSPA; phosphorylation provided 5SICSPATP, and deprotec-
tion provided 5SICSATP. For the MMO2 derivatives, the free
nucleoside was synthesized as reported previously,39 iodinated,
and then finally coupled to the dichloro acetyl protected pro-
pargylamine to produce MMO2PA. Phosphorylation and depro-
tection proceeded as with d5SICSA to provide MMO2PATP and
MMO2ATP.
2.2. Steady-State Kinetic Analysis of Linker-Modified
Unnatural Base Pair Synthesis. To characterize polymerase
recognition of the modified unnatural base pairs, we first explored
the efficiency with which the exonuclease deficient Klenow
fragment of E. coli DNA polymerase I (Kf) inserts the linker-
modified derivatives of d5SICSTP opposite dNaM (Table 1).
For comparison, Kf inserts dATP opposite dT and d5SICSTP
opposite dNaM with an efficiency of 7.7 ꢁ 108 Mꢀ1 minꢀ1 and
2.1 ꢁ 108 Mꢀ1 minꢀ1, respectively.42 We found that Kf inserts
d5SICSPATP with an efficiency of 8.3 ꢁ 106 Mꢀ1 minꢀ1. The
25-fold decreased insertion efficiency relative to d5SICSTP is
due entirely to an elevated apparent KM. d5SICSBIOTP and
d5SICSSSBIOTP were inserted less efficiently, with second-
order rate constants of 1.8 ꢁ 104 Mꢀ1 minꢀ1 and 4.5 ꢁ 104
Mꢀ1 minꢀ1, respectively, again, largely because of elevated
apparent KM values. The insertion of d5SICSATP opposite
dNaM by Kf is even less efficient, proceeding with a second-
order rate constant of only 4.5 ꢁ 103 Mꢀ1 minꢀ1. In this case, the
reduced insertion efficiency relative to d5SICSTP is due to both
a decreased apparent kcat and an increased apparent KM.
We then characterized the efficiency with which Kf inserts
the linker-modified derivatives of dMMO2TP opposite d5SICS
(Table 1). For comparison, Kf inserts dMMO2TP opposite
d5SICS with an efficiency of 4 ꢁ 105 Mꢀ1 minꢀ1.31 Interestingly,
we found that dMMO2PATP is inserted more efficiently (kcat/
KM = 7.2 ꢁ 105 Mꢀ1 minꢀ1) than dMMO2TP, because of a
2-fold decrease in the apparent KM, while dMMO2ATP is
inserted only slightly less efficiently (kcat/KM = 9.7 ꢁ 104 Mꢀ1
minꢀ1), because of small changes in both the apparent kcat and
KM. Because d5SICSSSBIOTP was more efficiently recognized
than d5SICSBIOTP, we also characterized dMMO2SSBIOTP and
found that it is inserted by Kf opposite d5SICS with an efficiency
of 3.4 ꢁ 104 Mꢀ1 minꢀ1, because of a 9-fold decrease in the
apparent kcat.
2.3. PCR Amplification of Modified DNA. To characterize
the effect of the linkers on the PCR amplification of DNA
containing the unnatural base pair, we incorporated dMMO2-
d5SICS into the middle of a 149-mer DNA duplex. The duplex
DNA was then PCR amplified using DeepVent (exo+) DNA
polymerase and a wide variety of different combinations of the
modified triphosphates. The amplification level was determined
for each combination of triphosphates, and then the amplicons
were sequenced to determine replication fidelity (Table 2, and
Figure S1, Supporting Information). In most cases, the amplifica-
tion efficiency approaches that of the fully natural control (996-
fold amplification), and the fidelity as measured by sequencing
is in excess of 99% per doubling. The single exceptions are the
amplifications with d5SICSATP, where the amplification level
was only ∼200-fold and the fidelity only ∼90% per round. This
agrees well with the steady-state kinetic data, which demon-
strated that d5SICSATP insertion is inefficient. However,
despite the slight reduction in steady-state insertion efficiency
of d5SICSPATP and d5SICSαSTP, relative to d5SICSTP, no
significant or systematic differences were apparent in amplifica-
tion efficiency or fidelity with these three triphosphates. Fidelities
of PCR amplifications involving biotin-modified triphosphates
were also determined by streptavidin gel shift (Figure 2). These
fidelities paralleled those determined by sequencing but were gene-
rally somewhat reduced. We attribute the differences to incomplete
binding of streptavidin to the biotin tag due to its location in the
middle of the large duplex where it is more obscured than when
present at the terminus, as has been more commonly examined.
19880
dx.doi.org/10.1021/ja207907d |J. Am. Chem. Soc. 2011, 133, 19878–19888

Journal of the American Chemical Society
ARTICLE
Table 2. PCR Fidelities and Amplification Efficienciesa
duplexes modified. We next attempted to immobilize the biotin-
labeled DNA to streptavidin solid support; however, unlike con-
ventional end-labeling via biotinylated primers,53 we found that
the nature and length of the spacer arm used to attach the
biotin is critical. Virtually all of the biotinylated DNA was bound
to the solid support when the linker was conjugated to biotin
via a hydrophilic PEG4 spacer (29 Å spacer arm) or an SS-PEG4
spacer (38 Å spacer arm including a cleavable disulfide bridge,
Figure 4A, B) coupled to dMMO2A, but only small amounts of
the DNA were bound when shorter and more hydrophobic
spacers were used (for example, after conjugation to EZ-Link
sulfo-NHS-SS-biotin which has a 24 Å spacer arm, data not
shown). In the case of SS-PEG4-immobilized DNA, after washing
to remove any contaminating natural DNA and cleaving the
disulfide with DTT, the double-stranded DNA was efficiently
released and then reconjugated to iodoacetyl-PEG2-biotin. Gel-
shift assays before and after the final conjugation revealed that
greater than 50% of the duplexes were labeled (Figure 4C).
To explore the postamplification labeling of DNA with two
different groups, we pursued two strategies. First, we examined
the amplification of DNA containing either d5SICSPA-dMMO2A
or d5SICSA-dMMO2PA (Figure 5A,B). After amplification, the
free amino groups were labeled with NHS-SS-PEG4-biotin and
characterized by gel mobility. After propargylamine deprotec-
tion, the newly liberated amine was labeled with sulfo-NHS-biotin,
with the labeling efficiency again characterized via gel mobility.
Labeling efficiencies were 60ꢀ85% for each step of labeling via
d5SICSPA-dMMO2A and d5SICSA-dMMO2PA (Figure 5D). A
second strategy for double labeling was based on the amplifica-
tion of DNA using d5SICSATP and dMMO2SSBIOTP (Figure 5C).
The level of biotinylation via direct dMMO2SSBIO incorporation
was characterized via streptavidin gel shift, and a second biotin
was then attached via coupling (DTT resistant) NHS-PEG4-
biotin to d5SICSA. After treatment with DTT to selectively
remove the biotin from dMMO2 (quantitative removal based on
control reactions, Figure 5E), the level of biotinylation at d5SICS
was characterized. Labeling efficiencies were 85% and 68%, respec-
tively, for the first and second steps via d5SICSA-dMMO2SSBIO
(Figure 5E).
fidelity
fidelity
dXTPs incorporated
amplification (sequencing)b (gel shift)c
dNaM and d5SICS
735
609
662
489
528
888
160
960
279
624
167
378
351
690
624
164e
217e
>99.7
99.6
ꢀ
ꢀ
dMMO2 and d5SICS
d
dNaM and d5SICSαS
99.6
ꢀ
dMMO2PA and d5SICS
dMMO2A, d5SICS
dNaM and d5SICSPA
dNaM and d5SICSA
dMMO2PA and d5SICSPA
dMMO2A and d5SICSA
dMMO2A and d5SICSPA
dMMO2PA and d5SICSA
>99.7
99.5
ꢀ
ꢀ
>99.7
91.1
ꢀ
ꢀ
>99.7
90.9
ꢀ
ꢀ
98.7
ꢀ
91.4
ꢀ
dMMO2A and d5SICSαS
99.4
ꢀ
d
dMMO2SSBIO and d5SICS
dNaM and d5SICSSSBIO
dNaM and d5SICSBIO
dNaM only
99.2
95.0
95.4
96.5
ꢀ
>99.7
>99.7
e
ꢀ
e
d5SICS only
ꢀ
ꢀ
a See Materials and Methods for experimental details. b Calculated as
average fidelities in both directions per doubling (see Materials and
Methods for details). c Calculated from gel mobility assay (see text).
d d5SICSαS was used as a mixture of Sp and Rp diastereomers.
e Unnatural base pair lost during amplification.
2.5. Thiolation of DNA as a Route to Site-Specific Labeling.
We also explored the use of d5SICSαSTP to site-specifically in-
corporate a reactive centerinto the backboneofDNA (Figure3C).
As expected, steady-state kinetics revealed that the α-phos-
phorothioate is only slightly perturbative, with Kf inserting
d5SICSαSTP opposite dNaM with an efficiency of 1.6 ꢁ
107 Mꢀ1 minꢀ1 (Table 1). To explore the use of labeling with
d5SICSαS for postamplification modification, we amplified DNA
with dNaMTP and d5SICSαSTP. As expected, d5SICSαS-dNaM
amplified with excellent efficiency and fidelity (Table 2). To
explore postsynthetic labeling, the resulting duplex DNA was
conjugated to EZ-Link iodoacetyl-PEG2-biotin. Gel-shift assays
with streptavidin demonstrated that 70% of the duplexes were
labeled (Figure 3D), in good agreement with values reported in
the literature for chemically synthesized α-phosphorothioate
DNA.54,55
2.6. Site-Specifically Modified RNA. To explore the site-
specific modification of RNA, we first characterized the ability of
the RNA polymerase from T7 bacteriophage (T7 RNAP) to
incorporate the linker-modified unnatural ribonucleotide tripho-
sphates into 17 nt transcripts (Figure 6A). We first examined the
transcription of a template containing d5SICS with the natural
ribotriphosphates and either MMO2PATP or MMO2ATP.
Figure 2. PCR fidelity determination via gel mobility. Based on
the intensity of the DNA band that shifts in the presence of
added streptavidin (SA), the biotin incorporation level is 65% for
d5SICS-dMMO2SSBIO, 64% for d5SICSSSBIO-dNaM, and 72% for
d5SICSBIO-dNaM. (Note that the slightly slower migrating band in lane
1 corresponds to unbiotinylated single-stranded DNA resulting from
incomplete annealing after PCR.) The overall incorporation levels are
converted to fidelities (Table 2) by normalizing by the number of
doublings. A 50 bp DNA ladder is loaded in the rightmost lane.
This conclusion is consistent with the immobilization studies
described below. However, we cannot exclude the possibility that
some of the biotin may have been lost by disulfide cleavage or
exchange during PCR.
2.4. Postamplification Modification of DNA. To examine
the postamplification labeling of DNA with a single functional
group, we first explored strategies based on the PCR amplifica-
tion of DNA with d5SICSATP and dNaMTP or d5SICSTP and
dMMO2ATP (Figure 3A,B). The resulting duplexes were bioti-
nylated using sulfo-NHS-SS-biotin and the labeling efficiency
was determined by streptavidin gel shift (Figure 3D). We found
that the reactions were complete after 1 h, with 60ꢀ70% of the
19881
dx.doi.org/10.1021/ja207907d |J. Am. Chem. Soc. 2011, 133, 19878–19888

Journal of the American Chemical Society
ARTICLE
Figure 3. Postamplification DNA labeling with single functional groups, using (A) d5SICSA, (B) dMMO2A, and (C) d5SICSαS. (D) Determination of
labeling efficiency via streptavidin gel shift. The biotin incorporation level is 55% for d5SICS-dMMO2A, 70% for d5SICSA-dNaM, and 70% for
d5SICSαS-dNaM. A 50 bp DNA ladder is included in the rightmost lane of the gel. The faster migrating, strong band corresponds to dsDNA, while the
slower migrating band corresponds to the 1:1 complex between dsDNA and streptavidin. The faint and most slowly migrating band in lane 6
corresponds to the 2:1 complex of dsDNA and streptavidin.
Figure 4. (A) Immobilization of biotinylated dsDNA on streptavidin affinity resin. (B) Gel mobility assay of PCR amplicons labeled with NHS-SS-
PEG4-biotin (P) compared to the unbound fraction remaining in the supernatant (S) after binding to the streptavidin solid support. Biotinylation levels
are 53% for d5SICS-dMMO2A and 70% for d5SICSA-dNaM. A 100 bp DNA ladder is loaded in the leftmost lane. (C) Conjugation of dsDNA to
iodoacetyl-PEG2-biotin after release from the streptavidin affinity resin via DTT treatment. Biotin incorporation levels are 89% for d5SICS-dMMO2A
and 53% for d5SICSA-dNaM. A 50 bp DNA ladder is loaded in the leftmost lane. In B and C, the faster migrating, strong band corresponds to dsDNA,
while the slower migrating band corresponds to the 1:1 complex between dsDNA and streptavidin.
Under the conditions employed, no full length product was
observed in the absence of an unnatural triphosphate or in the
presence of only a 5SICSTP derivative, and most of the
truncated product corresponded to the termination of transcrip-
tion immediately before d5SICS in the template (Figure 6B). In
contrast, when either MMO2PATP or MMO2ATP was present,
we observed efficient conversion to full-length product. We then
characterized transcription of the unnatural base pair in the
opposite strand context by examining the ability of dNaM to tem-
plate the transcription of RNA containing 5SICSPA or 5SICSA.
Again, in the absence of an unnatural triphosphate, virtually no
full-length product was observed, but addition of either cognate
unnatural triphosphate, 5SICSPATP or 5SICSATP, resulted
in the efficient production of the full-length transcription
product (Figure 6B). To characterize the fidelity of transcription,
the experiments were again run with [α-32P]ATP and the
19882
dx.doi.org/10.1021/ja207907d |J. Am. Chem. Soc. 2011, 133, 19878–19888

Journal of the American Chemical Society
ARTICLE
Table 3. Nucleotide Composition Analysis of T7 RNAP Transcription Productsa
normalized composition of nucleotides incorporated 50 to A during transcription of 17 nt RNA
template
YTP
Ap
Gp
Cp
Up
Yp
d5SICS
d5SICS
dNaM
dNaM
dMMO2A
dMMO2PA
d5SICSA
1.02 ( 0.01 [1]
0.99 ( 0.03 [1]
1.03 ( 0.02 [1]
1.07 ( 0.01 [1]
1.94 ( 0.03 [2]
1.99 ( 0.03 [2]
2.00 ( 0.01 [2]
1.95 ( 0.02 [2]
n.d. [0]b
n.d. [0]b
n.d. [0]b
n.d. [0]b
n.d. [0]b
n.d. [0]b
n.d. [0]b
n.d. [0]b
1.03 ( 0.02 [1]
1.01 ( 0.04 [1]
0.96 ( 0.02 [1]
0.95 ( 0.01 [1]
d5SICSPA
normalized composition of nucleotides incorporated 50 to A during transcription of tRNACTyUrA
template
YTP
Ap
Gp
Cp
Up
Yp
d5SICS
dNaM
MMO2A
5SICSA
4.06 ( 0.13 [4]
4.16 ( 0.16 [4]
3.03 ( 0.07 [3]
3.10 ( 0.04 [3]
5.91 ( 0.17 [6]
5.86 ( 0.06 [6]
0.98 ( 0.03 [1]
1.01 ( 0.02 [1]
1.02 ( 0.05 [1]
0.92 ( 0.04 [1]
a Values were determined by dividing the radioactivity observed for a particular monophosphate by the product of the total radioactivity and the expected
number of nucleotides. Predicted values assuming 100% fidelity are shown in brackets. The error reported is the standard deviation of at least three
independent determinations. b Below detection limit.
Figure 5. Postamplification DNA labeling with two different functional groups, using (A) d5SICSPA-dMMO2A, (B) d5SICSA-dMMO2PA, and
(C) d5SICSA-dMMO2SSBIO. (D) Efficiencies of first and second labeling of DNA from containing d5SICSPA-dMMO2A and d5SICSA-dMMO2PA
.
First labeling efficiencies are 61% for d5SICSPA-dMMO2A and 66% for d5SICSA-dMMO2PA, and second labeling efficiencies are 85% for d5SICSPA
-
dMMO2A and 78% for d5SICSA-dMMO2PA. A 100 bp ladder is loaded in the rightmost lane of each gel. (E) First and second labeling efficiencies of
DNA containing d5SICSA-dMMO2SSBIO. The first labeling efficiency is 85% in the absence of DTT and 0% in the presence of DTT (due to linker
cleavage). The second labeling efficiency is 68%. A 50 bp DNA ladder is loaded in the rightmost lane. In D and E, the faster migrating, strong band
corresponds to dsDNA, while the slower migrating band corresponds to the 1:1 complex between dsDNA and streptavidin. The faint and more slowly
migrating bands correspond to higher order complexes of dsDNA and streptavidin.
19883
dx.doi.org/10.1021/ja207907d |J. Am. Chem. Soc. 2011, 133, 19878–19888

Journal of the American Chemical Society
ARTICLE
Tyr
resulting transcripts were digested to nucleoside 30-phosphates,
which were separated by 2D-TLC (Figure S2, Supporting
Information), and the relative amount of each was quantified
via densitometry (Table 3). Nucleotide-composition analysis
confirmed that both unnatural nucleotides direct the transcrip-
tion of RNA containing a cognate unnatural nucleotide with
good selectivity. Interestingly, unlike with the replication of the
deoxy-variants discussed above, we observed little difference in the
efficiency or fidelity of 5SICSATP or 5SICSPATP incorporation.
To further explore the T7 RNAP-mediated transcription of
RNA for more practical applications, we synthesized a 96-nt
DNA fragment encoding the 77-nt M. jannaschii tRNA
with
CUA
either dNaM or d5SICS at the third position of the anticodon
(nucleotide 37). Full length product was efficiently produced in
the presence of each natural triphosphate and the cognate
unnatural triphosphate. However, unlike with the shorter model
template, we were unable to identify conditions where product
formation depended only upon the addition of the correct un-
natural ribotriphosphate, likely due to the higher concentrations
of natural NTPs required for efficient transcription of the longer
RNA. Thus, we characterized the fidelity of transcription of the
DNA template containing dNaM with 5SICSA and of the
template containing d5SICS with MMO2A by 2D-TLC and
densitometry (Table 3, and Figure S3, Supporting Information).
The incorporation of MMO2A opposite d5SICS proceeded with
excellent fidelity, while the incorporation of 5SICSA opposite
dNaM proceeded with a more modest fidelity of 92%, suggesting
that as with the deoxy variants and DNA polymerases, the free
amine is tolerated better in the MMO2 scaffold than in the
5SICS scaffold.
Finally, we explored the post-transcription site-specific
Tyr
labeling of RNA (Figure 7). Purified tRNA
product with
CUY
Y = MMO2A at the third position of the anticodon loop was first
unfolded by heating to 85 °C in the presence of 20% DMSO and
then reacted with NHS-PEG4-biotin at 37 °C for 1 h. Based on
streptavidin gel shift, the efficiency of site-specific biotin attach-
ment was greater than 75%.
3. DISCUSSION
The development of an unnatural base pair is the first step
toward creating semisynthetic organisms with increased poten-
tial for information storage and retrieval but would likely find
more immediate use in different applications based on the pro-
duction of site-specifically modified DNA or RNA. While func-
tional unnatural base pairs that make site-specific labeling pos-
sible have recently been identified,29,34,39,44,46 little is known
about their ability to accommodate the linkers required to attach
the various functional groups of interest. Presumably, just as
the development of the unnatural base pairs required extensive
optimization, so too will development of the linkers. Previously,
Figure 6. T7 RNAP transcription of site-selectively modified RNA. (A)
Sequences of DNA template and transcription product. (B) Denaturing
PAGE analysis with sequence indicated. Note that the bands for the
linker-modified full length transcripts are shifted slightly more than
those of the unlabeled full length product.
site-selectively modified with MMO2A. (B) PAGE analysis of the modified tRNA
containing
CUA
Tyr
CUA
Tyr
Figure 7. (A) Transcription of M. jannaschii tRNA
MMO2A before (left two lanes) and after (right two lanes) coupling to NHS-PEG4-biotin. Coupling efficiency calculated as greater than 75%. A 25 bp
DNA ladder is loaded in the rightmost lane.
19884
dx.doi.org/10.1021/ja207907d |J. Am. Chem. Soc. 2011, 133, 19878–19888

Journal of the American Chemical Society
ARTICLE
the Hirao group demonstrated the use of the same propargyl-
amine linker used here, along with an acetamidohexanamide
spacer, to attach a single fluorophore to a promising unnatural
base pair within DNA, as well as to directly attach a biotin tag to a
constituent nucleotide in RNA.44,46 To further explore the
potential of linker-modified unnatural base pairs to site-specifi-
cally label DNA or RNA with one or more different moieties,
either during or after enzymatic synthesis, we have explored the
replication and transcription of d5SICS-dMMO2/dNaM with
unnatural triphosphates bearing propargylamine-based linkers or
an α-phosphorothioate.
The steady-state kinetics with Kf polymerase revealed that
replication of the unnatural base pair is sensitive to the structure
of the linker, with the protected amino linker being the best
accommodated. However, the effects of linker attachment were
not the same for both nucleotides. While modification of
dMMO2TP with the free or the dithiol-linked biotin-modified
linker reduced the efficiency of insertion opposite d5SICS 4- to
10-fold, modification with the protected amino linker actually
increased insertion efficiency by 2-fold. In contrast, modifica-
tion of d5SICSTP with any of the linkers examined resulted
in reduced insertion efficiency opposite dNaM. Relative to
d5SICSTP, the insertion efficiency of d5SICSBIOTP and
d5SICSSSBIOTP is 3- to 4-orders of magnitude reduced, pre-
dominantly because of a reduction in the apparent binding of the
triphosphate. The decrease was even larger for d5SICSATP
because of effects on both the apparent binding and turnover.
In contrast, the insertion efficiency of d5SICSPATP opposite
dNaM was reduced only 25-fold (because of reduced apparent
binding).
Relative to the dMMO2, the greater sensitivity of d5SICSTP
insertion to linker modification may result from its being more
efficiently recognized, making it more susceptible to perturba-
tion. However, the increase in insertion efficiency observed with
dMMO2PA, and the fact that d5SICSATP is inserted less
efficiently opposite dNaM than dMMO2ATP is inserted oppo-
site d5SICS, suggests that the dMMO2 scaffold is more tolerant
of the modifications examined. Perhaps, hydrophobic and pack-
ing interactions between the π-rich triple bond of the linker and
the flanking nucleobases within the major groove are more
favorable when accessed from the dMMO2 scaffold than from
the d5SICS scaffold. This is consistent with the recently reported
structure of the ternary complex of the large fragment of Taq
DNA polymerase, which is homologous to Kf, bound to C5-
propargylamide-modified dTTP, which shows that the amide
moiety of the linker forms a stabilizing hydrogen-bonding inter-
action with the side chain of polymerase residue R660.24 This Arg
residue is conserved in Kf, and because the ring structure of
dMMO2 is similar in size and shape to a natural pyrimidine, it
likely engages in a similar stabilizing interaction with the pro-
pargylamide linker when presented within the dMMO2 scaffold.
Perhaps the extra aromatic ring of the d5SICS scaffold disrupts
this stabilizing interaction. The same model likely accounts for
the decreased tolerance of the free amine linker, as it is expected
to be protonated and thus to introduce electrostatic repulsion
with the positively charged side chain of R660.
straightforward to incorporate many different functional groups
into duplex DNA via amide linkage to either unnatural tripho-
sphate. Moreover, different combinations of unnatural tripho-
sphates bearing protected or free amines may also be incorpo-
rated into DNA via PCR and thereby provide a versatile route to
the postsynthetic labeling of DNA with two different functional
groups, although with d5SICS, the most efficient replication
requires that the amine is incorporated in a protected form. Thus,
the most simple route to the labeling of DNA with two different
functionalities would be either the use of triphosphates already
bearing the functional groups of interest or the amplification of
DNA containing d5SICSPA-dMMO2A, followed by modification
of dMMO2A, deprotection, and finally modification of d5SICSA.
The use of d5SICSSSBIO-dMMO2A also provides a convenient
route to purify DNA containing the unnatural base pair, and the
use of spacers that are longer and more hydrophilic to connect
the linkers to the functional groups of interest facilitates post-
amplification modification.
While the efficiency of d5SICSPA insertion is likely sufficient
for practical labeling applications, we nonetheless demonstrated
that the incorporation of d5SICSαSTP opposite dNaM, which
proceeds with a 2-fold greater efficiency, also provides a con-
venient route to site-specific labeling via the modification of the
α-phosphorothioate. While amplification with d5SICSPATP,
d5SICSαSTP, or d5SICSTP all proceeded with similar efficien-
cies and fidelities within the sequence context examined, it is pos-
sible that other sequences may be more sensitive to the specific
triphosphate used. Also note that the α-phosphorothioate- and
linker-based strategies are not mutually exclusive and when com-
bined should allow for a given site to be simultaneously modified
with up to three different functional groups, one attached to the
nucleobase of d5SICS, a second attached to the nucleobase of
dMMO2, and a third attached to the backbone immediately 50 to
an unnatural nucleotide (Figure 8).
As observed with the deoxy variants and Kf DNA polymerase,
the linker-modified unnatural ribotriphosphates were found to
be substrates for T7 RNAP. While we only characterized the
transcription of the protected- and free-amine modified ribotri-
phosphates, the data suggest that just as with their deoxy counter-
parts, the direct attachment of different functional groups to the
ribotriphosphates should be possible, providing a direct route to
the transcription of RNA labeled with one or two functional
groups. In addition, we demonstrated that the incorporation of
ribotriphosphates bearing free amine groups should provide a
general route to the post-transcription labeling of RNA with
different functionalities of interest (Figure 8). In contrast to the
Regardless of the detailed rates with which the different tri-
phosphates are inserted opposite their cognate unnatural nucleo-
tide in the template, PCR experiments with DeepVent clearly
demonstrate that different combinations of the propargylamine-
modified nucleotides may be efficiently and site-selectively incor-
porated into DNA with reasonable fidelity. Thus, it should be
Figure 8. General scheme for labeling DNA with three different
functional groups and RNA with two different functional groups. R1
and R2 may correspond to either amide-coupled functional groups of
interest or a combination of protected and free amine linkers for further
elaboration as described in the text.
19885
dx.doi.org/10.1021/ja207907d |J. Am. Chem. Soc. 2011, 133, 19878–19888

Journal of the American Chemical Society
ARTICLE
differences in replication observed with d5SICSATP and the
other d5SICSTP derivatives, the transcription of short RNAs
with 5SICS and 5SICSA appears similar, although fidelity was
somewhat reduced for the incorporation of 5SICSA into the
tRNA, which is longer and more challenging to transcribe.
Because alkylated α-phosphorothioates undergo strand cleavage
in RNA, this route for backbone labeling is not possible, and the
current methodology is limited to the site-specific labeling
of RNA with only two different functional groups. However,
the ability to induce site-specific RNA cleavage could also find
interesting applications.56,57
Our goal is to expand the potential physical and functional
properties of DNA and RNA without losing the inherent advan-
tages possessed by these biopolymers relative to other materials.
One property of DNA that has received much recent attention is
its ability to act as a scaffold for the display of different functio-
nalities with nanometer length-scale control for nanomaterial
development.2,3 The site-specific labeling approach described
here seems well suited to contribute to these efforts, as many of
the possible applications involve a level of distance control and
functional group isolation that would be challenging or impos-
sible to obtain with the higher functionalization density inherent
to nucleotide-specific labeling approaches. From the perspective
of creating (or evolving) controlled environments for molecular
recognition or catalysis,6,7 relative to the nucleotide-specific label-
ing approach, the site-specific approach is more analogous to that
employed by proteins, where the position of one or only a few
reactive centers are controlled within an environment provided
by the remainder of the oligonucleotide. Moreover, including
only one or a few modified nucleotides in different members
of an oligonucleotide library is less likely to incur unnecessary
replication biases than more heavily or fully functionalized DNA.
Efforts to deploy the linker-modified unnatural base pairs in
materials applications and for the selection of DNAs or RNAs
with expanded functional potential are currently underway.
4.3. Kinetic Assay. Primer oligonucleotides were 50-radiolabeled
with [γ-32P]ATP and T4 polynucleotide kinase and annealed to
template oligonucleotides by heating to 95 °C followed by slow cooling
to room temperature. Reactions were initiated by adding a solution
of 2ꢁ dNTP solution (5 μL) to a solution containing Kf polymerase
(0.10ꢀ1.23 nM) and primer template (40 nM) in 5 μL of Kf reaction
buffer (50 mM Tris-HCl, pH 7.5, 10 mM MgCl2, 1 mM DTT, and
50 μg/mL acetylated BSA). After incubation at 25 °C for 3 ꢀ 10 min, the
reactions were quenched with 20 μL of loading dye (95% formamide,
20 mM EDTA, and sufficient amounts of bromophenol blue and xylene
cyanol). Reaction products were resolved by 8 M urea 15% polyacry-
lamide gel electrophoresis, and gel band intensities corresponding to pri-
mer and extended primer were quantified by phosphorimaging (Storm
Imager, Molecular Dynamics) and Quantity One (BioRad) software.
Plots of kobs versus triphosphate concentration were fit to the Michaelisꢀ
Menten equation using the program Origin (Microcal Software) to
determine Vmax and KM. kcat was determined from Vmax by normalizing
by the total enzyme concentration. Each reaction was run in triplicate,
and standard deviations for both kinetic parameters were determined.
4.4. PCR Amplification. dsDNA template D1 was prepared as des-
cribed previously34 (see Supporting Information for details) and ampli-
fied by PCR under the following conditions: 1 ng of the template, 1ꢁ
ThermoPol reaction buffer (New England Biolabs), MgSO4 adjusted
to 6.0 mM, 0.6 mM of dNTP, 0.2 mM of each unnatural triphosphates,
1 μM of each primers, and 0.03 U/μL of DeepVent (exo+, New England
Biolabs) in an iCycler Thermal Cycler (Bio-Rad) with a total volume
of 25 μL under the following thermal cycling conditions: 94 °C, 30 s;
48 °C, 30 s; 65 °C, 8 min, 14 cycles. Sybr Green I (Invitrogen) was added
to the final concentration of 0.5ꢁ to monitor amplification by qPCR.
Upon completion, a 5 μL aliquot was analyzed by 2% agarose gel, and the
remaining material was purified utilizing the PureLink PCR Purification
Kit (Invitrogen). Amplification efficiency was quantified by fluorescent
dye binding (Quant-iT dsDNA HS Assay kit, Invitrogen), and fidelity
was quantified by either sequencing (3730 DNA Analyzer, Applied
Biosystems) (see Supporting Information and Malyshev et al.34) or gel
mobility (see below).
4.5. Gel Mobility Assays. DNA samples (10ꢀ50 ng) were mixed
with 1 μg of streptavidin (Promega) in phosphate labeling buffer
(50 mM sodium phosphate, pH 7.5, 150 mM NaCl, 1 mM EDTA),
incubated for 30 min at 37 °C, mixed with 5ꢁ nondenaturing loading
buffer (Qiagen), and loaded on 10% nondenaturing PAGE. The gel
was run at 180 V for 25ꢀ40 min, then soaked in 1ꢁ Sybr Gold Nucleic
Acid Stain (Invitrogen) for 30 min and visualized using a Molecular
Imager Gel Doc XR+ equipped with 520DF30 filter (Bio-Rad). Strong
bands corresponding to dsDNA (at ∼150 bp) and the 1:1 complex
between dsDNA and streptavidin (at ∼400 bp) were apparent. Faint
bands corresponding to higher order (slower migrating) complexes of
DNA and streptavidin or from unbiotinylated, single-stranded DNA
resulting from incomplete annealing after PCR in some cases were also
apparent.
4.6. General Procedures for Postamplification DNA Labeling.
For postenzymatic synthesis labeling, dsDNA with a free amino group
was incubated with 10 mM EZ-Link sulfo-NHS-SS-biotin or EZ-Link
NHS-PEG4-biotin (Thermo Scientific) for 1 h at rt in phosphate labeling
buffer and then purified using the Qiagen PCR purification kit. With
either dMMO2PA or d5SICSPA, the amine first required deprotection,
which was accomplished by overnight incubation in a concentrated
aqueous ammonia solution at rt. Ammonia was removed via a SpeedVac
concentrator (water aspirator followed by oil vacuum pump). To cleave
the disulfide-containing linkers(i.e., SS-biotinor SS-PEG4-biotin), dsDNA
was treated with DTT (final concentration of 30 mM) for 1 h at 37 °C.
For backbone labeling, dsDNA with a backbone phosphorothioate was
incubated with 25 mM EZ-Link iodoacetyl-PEG2-biotin (Thermo
Scientific) in phosphate labeling buffer overnight at 50 °C, and products
4. MATERIALS AND METHODS
4.1. General. All reactions were carried out in oven-dried glassware
under inert atmosphere, and all solvents were dried over 4 Å molecular
sieves with the exceptions of dichloromethane, which was distilled from
CaH2, and tetrahydrofuran, which was distilled from sodium metal. All
other reagents were purchased from Aldrich and Acros. 1H, 13C, and 31
P
NMR spectra were recorded on Varian Mercury 300, Varian Inova-400,
or Bruker AMX-400 spectrometers. High-resolution mass spectroscopic
data were obtained on an ESI-TOF mass spectrometer (Agilent 6200
Series) at the TSRI Open Access Mass Spectrometry Lab; MALDI-TOF
mass spectrometry (Applied Biosystems Voyager DE-PRO System
6008) was performed at the TSRI Center for Protein and Nucleic Acid
Research. Polynucleotide kinase and Kf were purchased from New
England Biolabs, T7 RNAP from Takara USA, and [α-32P]ATP and
[γ-32P]ATP from MP Biomedicals. DNA ladders were obtained from
Invitrogen.
4.2. Oligonucleotide Synthesis. Fully natural oligonucleotides
were purchased from Integrated DNA Technologies. Oligonucleotides
containing an unnatural nucleotide were prepared by the β-cyanoethyl
phosphoramidite method on controlled pore glass supports (1 μmol)
using an Applied Biosystems Inc. 392 DNA/RNA synthesizer. After
automated synthesis, the oligonucleotides were cleaved from the sup-
port and deprotected by heating in aqueous ammonia at 55 °C for 12 h.
After purification via 8 M urea 15% polyacrylamide gel electrophoresis
(PAGE), the oligonucleotides were electroeluted and ethanol precipi-
tated. Oligonucleotide concentration was determined by UV absorption.
19886
dx.doi.org/10.1021/ja207907d |J. Am. Chem. Soc. 2011, 133, 19878–19888

Journal of the American Chemical Society
ARTICLE
were purified with Qiagen PCR Purification Kit. All reactions manip-
ulating attached biotin moieties were quantified by streptavidin gel-shift
assays.
dryness with a SpeedVac concentrator. The transcripts were digested
with RNaseI (10 U, 1 μL, Epicenter, WI) in 2 μL of 10ꢁ RNaseI dilution
buffer in a final volume of 20 μL at 37 °C for 120 min. The digestion
products were analyzed by 2D TLC using high performance thin layer
chromatography (HPTLC) plates (100 ꢁ 100 mm) (Merck, Darmstadt,
Germany) with the following developing solvents: isobutyric acidꢀ
ammoniaꢀwater (66:1:33 v/v/v) for the first dimension, and sodium phos-
phate (0.1 M, pH 6.8), ammonium sulfate, n-propanol (100:60:2 v/w/v)
for the second dimension. The products on the gels and the TLC plates
were analyzed by phosphorimaging using ImageQuant (Quantity One).
For post-transcriptional labeling of tRNATCyUrA, the purified full length
RNA containing MMO2A (60 μL, 2.2 ng/μL) was mixed with 15 μL of
DMSO, heated at 85 °C for 5 min, and cooled immediately on ice. The
resulting solution was incubated with EZ-Link NHS-PEG4-biotin
(Thermo Scientific) (1.5 mg) in 30 μL of 4ꢁ phosphate labeling buffer
for 1 h at 37 °C. After ethanol precipitation, 10% of the resulting material
(3 μL) was mixed with 3 μL of 5ꢁ denaturing loading buffer, heated to
85 °C, cooled on ice, mixed with 3 μL of streptavidin (1 mg/mL), and
incubated for 30 min at 37 °C. The efficiency of labeling was determined
by streptavidin gel shift on a 7 M urea 15% polyacrylamide gel using 1ꢁ
TBE buffer (180 V for 25ꢀ40 min).
4.7. DNA Immobilization on Streptavidin Solid Support.
Streptavidin Sepharose High Performance affinity resin (GE Healthcare)
was washed twice with phosphate labeling buffer to remove ethanol.
Site-specifically biotinylated dsDNA was added to the prewashed resin
and incubated with occasional gentle mixing. After 1 h, the supernatant
was removed (and analyzed by gel shift mobility to confirm the absence
of biotinylated DNA), and the resin was washed three times with
phosphate labeling buffer to remove any unbound DNA. dsDNA was
recovered via DTT treatment (final concentration of 30 mM) for 1 h at
37 °C, followed by filtration and purification with the Qiagen PCR
Purification Kit.
4.8. Transcription Experiments. For the transcription of short,
17 nt RNAs, the 35 nt DNA template 50-CACTXCTCGGGATTCCC-
TATAGTGAGTCGTATTAT (X = d5SICS or dNaM) and 21 nt DNA
primer 50-ATAATACGACTCACTATAGGG were annealed in a 1:1
ratio in 10 mM Tris-HCl buffer (pH 7.6) containing 10 mM NaCl, by
heating at 95 °C for 3 min and slow cooling to 4 °C. Transcription was
carried out at 37 °C in 10 μL reactions containing 100 nM DNA primer-
template, 20 μM NTP, 0.25 mCi [α-32P]ATP, 50 U T7 RNAP (Takara),
and 10 μM MMO2PATP, MMO2ATP, 5SICSPATP, or 5SICSATP
in 1ꢁ Takara buffer (40 mM Tris-HCl, pH 8.0, 8 mM MgCl2, 2 mM
spermidine). All solutionswere preparedwithDEPC-treatedandnuclease-
free sterilized water (Fisher Bioreagents). After incubation for 2 h at
37 °C, the reactions were quenched by addition of gel loading dye solu-
tion (10 μL of 10 M urea, 0.05% bromophenol blue). This mixture was
heated at 75 °C for 3 min and then separated on a 7 M urea 20%
polyacrylamide gel using 1ꢁ TBE buffer. The gel was removed from the
apparatus, and radioactivity was quantified by phosphorimaging (overnight
’ ASSOCIATED CONTENT
S
Supporting Information. Nucleoside synthesis and char-
b
acterization as well as details of kinetic experiments, sequencing,
and transcription analysis.This material is available free of charge
via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
exposure) using ImageQuant (Quantity One).
Corresponding Author
floyd@scripps.edu
Tyr
For transcription of M. jannaschii tRNA , the template strand 50-
CUY
TGG TCC GGC GGG CCG GAT TTG AAC CAG CGC CAT GCG
GAT TXA GAG TCC GCC GTT CTG CCC TGC TGA ACT ACC
GCC GGT ATA GTG AGT CGT ATT ATC (X = d5SICS or dNaM)
and the nontemplate strand 50- GAT AAT ACG ACT CAC TAT ACC
GGC GGT AGT TCA GCA GGG CAG AAC GGC GGA CTC TAA
ATC CGC ATG GCG CTG GTT CAA ATC CGG CCC GCC GGA
CCA were annealed in Tris-HCl buffer (pH 7.6) containing 10 mM
NaCl, by heating at 95 °C for 3 min and cooling to 4 °C. Transcription
was carried out at 37 °C in 20 μL reactions containing 2 μM DNA
primer-template, 1 mM NTP, 0.25 mCi [α-32P]ATP (MP Biomedicals,
LLC, Solon, OH), 50 U T7 RNAP (Takara Bio Inc.), and 1 mM
MMO2ATP or 5SICSATP in 1ꢁ Takara buffer (40 mM Tris-HCl, pH
8.0, 8 mM MgCl2, 2 mM spermidine). All solutions were prepared with
DEPC treated and nuclease-free sterilized water (Fisher Bioreagents).
After incubation for 3 h at 37 °C, the reaction was quenched by the
addition of gel loading dye solution (20 μL of 10 M urea, 0.05% bromo-
phenol blue). This mixture was heated at 75 °C for 3 min and then
separated on a 7 M urea 12% polyacrylamide gel using 1ꢁ TBE buffer.
The gel was removed from the apparatus, and radioactivity was quan-
tified by phosphorimaging (overnight exposure) using ImageQuant
(Molecular Dynamics).
Transcription fidelity was characterized by 2D TLC analysis. Briefly,
gels containing transcription products were imaged (overnight exposure
with Kodak X-OmatAR5 film), and the image was used to identify the
radioactive spots corresponding to the full-length [32P]-RNA tran-
scripts. These portions of the gel were then excised and transferred to
2 mL microcentrifuge tubes for passive elution (6ꢀ12 h) at room tem-
perature with 400 μL of sterilized water. Eluted full-length products were
precipitated with cold ethanol (1.2 mL), CH3COONa (0.3 M, 20 μL),
and E. coli tRNA (1 mM, 1 μL, Sigma Aldrich) for 2 h. After centri-
fugation, the ethanol was removed and the product was evaporated to
Author Contributions
†These authors contributed equally.
’ ACKNOWLEDGMENT
Funding for this work was provided by the National Institute
for Health (GM060005).
’ REFERENCES
(1) Benner, S. A.; Sismour, A. M. Nat. Rev. Genet. 2005, 6, 533–543.
(2) Wang, H.; Yang, R.; Yang, L.; Tan, W. ACS Nano 2009, 3,
2451–2460.
(3) Chen, T.; Shukoor, M. I.; Chen, Y.; Yuan, Q.; Zhu, Z.; Zhao, Z.;
Gulbakan, B.; Tan, W. Nanoscale 2011, 3, 546–556.
(4) Ruigrok, V. J.; Levisson, M.; Eppink, M. H.; Smidt, H.; van der
Oost, J. Biochem. J. 2011, 436, 1–13.
(5) Heyduk, T. Biophys. Chem. 2010, 151, 91–95.
(6) Syed, M. A.; Pervaiz, S. Oligonucleotides 2010, 20, 215–224.
(7) Mayer, G. Angew. Chem., Int. Ed. 2009, 48, 2672–2689.
(8) Lonnberg, H. Bioconjugate Chem. 2009, 20, 1065–1094.
(9) Paredes, E.; Evans, M.; Das, S. R. Methods 2011, 54, 251–259.
(10) Lavergne, T.; Bertrand, J. R.; Vasseur, J. J.; Debart, F. Chem.—
Eur. J. 2008, 14, 9135–9138.
(11) Battersby, T. R.; Ang, D. N.; Burgstaller, P.; Jurczyk, S. C.;
Bowser, M. T.; Buchanan, D. D.; Kennedy, R. T.; Benner, S. A. J. Am.
Chem. Soc. 1999, 121, 9781–9789.
(12) Brakmann, S.; Lobermann, S. Angew. Chem., Int. Ed. 2001, 40,
1427–1429.
(13) Gourlain, T.; Sidorov, A.; Mignet, N.; Thorpe, S. J.; Lee, S. E.;
Grasby, J. A.; Williams, D. M. Nucleic Acids Res. 2001, 29, 1898–1905.
19887
dx.doi.org/10.1021/ja207907d |J. Am. Chem. Soc. 2011, 133, 19878–19888

Journal of the American Chemical Society
ARTICLE
(14) Jager, S.; Rasched, G.; Kornreich-Leshem, H.; Engeser, M.;
Thum, O.; Famulok, M. J. Am. Chem. Soc. 2005, 127, 15071–15082.
(15) Kuwahara, M.; Nagashima, J.; Hasegawa, M.; Tamura, T.;
Kitagata, R.; Hanawa, K.; Hososhima, S.; Kasamatsu, T.; Ozaki, H.;
Sawai, H. Nucleic Acids Res. 2006, 34, 5383–5394.
(16) Mehedi Masud, M.; Ozaki-Nakamura, A.; Kuwahara, M.; Ozaki,
H.; Sawai, H. ChemBioChem 2003, 4, 584–588.
(17) Perrin, D. M.; Garestier, T.; Helene, C. Nucleosides Nucleotides
1999, 18, 377–391.
(47) Mitsui, T.; Kitamura, A.; Kimoto, M.; To, T.; Sato, A.; Hirao, I.;
Yokoyama, S. J. Am. Chem. Soc. 2003, 125, 5298–5307.
(48) Liu, C. C.; Schultz, P. G. Annu. Rev. Biochem. 2010, 79,
413–444.
(49) Ludwig, J.; Eckstein, F. J. Org. Chem. 1989, 54, 631–635.
(50) Burgess, K.; Cook, D. Chem. Rev. 2000, 100, 2047–2060.
(51) Eckstein, F. Annu. Rev. Biochem. 1985, 54, 367–402.
(52) Kawai, R.; Kimoto, M.; Ikeda, S.; Mitsui, T.; Endo, M.;
Yokoyama, S.; Hirao, I. J. Am. Chem. Soc. 2005, 127, 17286–17295.
(53) Rapley, R. The Nucleic Acid Protocols Handbook; Humana Press:
New York, NY, 2000.
(18) Roychowdhury, A.; Illangkoon, H.; Hendrickson, C. L.; Benner,
S. A. Org. Lett. 2004, 6, 489–492.
(19) Sakthivel, K.; Barbas, C. F., III. Angew. Chem., Int. Ed. 1998,
37, 2872–2875.
(54) Fidanza, J. A.; Ozaki, H.; McLaughlin, L. W. Methods Mol. Biol.
1994, 26, 121–143.
(20) Tasara, T.; Angerer, B.; Damond, M.; Winter, H.; Dorhofer, S.;
Hubscher, U.; Amacker, M. Nucleic Acids Res. 2003, 31, 2636–2646.
(21) Hollenstein, M.; Hipolito, C. J.; Lam, C. H.; Perrin, D. M.
Nucleic Acids Res. 2009, 37, 1638–1649.
(55) Padmanabhan, S.; Coughlin, J. E.; Zhang, G.; Kirk, C. J.; Iyer,
R. P. Bioorg. Med. Chem. Lett. 2006, 16, 1491–1494.
(56) Gish, G.; Eckstein, F. Science 1988, 240, 1520–1522.
(57) Strobel, S. A.; Shetty, K. Proc. Natl. Acad. Sci. U.S.A. 1997, 94,
2903–2908.
(22) Held, H. A.; Benner, S. A. Nucleic Acids Res. 2002, 30, 3857–
3869.
(23) Lee, S. E.; Sidorov, A.; Gourlain, T.; Mignet, N.; Thorpe, S. J.;
Brazier, J. A.; Dickman, M. J.; Hornby, D. P.; Grasby, J. A.; Williams,
D. M. Nucleic Acids Res. 2001, 29, 1565–1573.
(24) Obeid, S.; Baccaro, A.; Welte, W.; Diederichs, K.; Marx, A. Proc.
Natl. Acad. Sci. U.S.A. 2010, 107, 21327–21331.
(25) Zhu, Z.; Waggoner, A. S. Cytometry 1997, 28, 206–211.
(26) Horlacher, J.; Hottiger, M.; Podust, V. N.; H€ubscher, U.;
Benner, S. A. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 6329–6333.
(27) Lutz, M. J.; Held, H. A.; Hottiger, M.; H€ubscher, U.; Benner,
S. A. Nucleic Acids Res. 1996, 24, 1308–1313.
(28) Piccirilli, J. A.; Krauch, T.; Moroney, S. E.; Benner, S. A. Nature
1990, 343, 33–37.
(29) Yang, Z.; Chen, F.; Chamberlin, S. G.; Benner, S. A. Angew.
Chem., Int. Ed. 2010, 49, 177–180.
(30) Hwang, G. T.; Romesberg, F. E. J. Am. Chem. Soc. 2008, 130,
14872–14882.
(31) Leconte, A. M.; Hwang, G. T.; Matsuda, S.; Capek, P.; Hari, Y.;
Romesberg, F. E. J. Am. Chem. Soc. 2008, 130, 2336–2343.
(32) Leconte, A. M.; Romesberg, F. E. In Protein Engineering;
RajBhandary, C. K. a. U. L., Ed.; Springer-Verlag: Berlin, 2009; pp
291ꢀ314.
(33) Malyshev, D. A.; Pfaff, D. A.; Ippoliti, S. I.; Hwang, G. T.;
Dwyer, T. J.; Romesberg, F. E. Chem.—Eur. J. 2010, 16, 12650–12659.
(34) Malyshev, D. A.; Seo, Y. J.; Ordoukhanian, P.; Romesberg, F. E.
J. Am. Chem. Soc. 2009, 131, 14620–14621.
(35) Matsuda, S.; Fillo, J. D.; Henry, A. A.; Rai, P.; Wilkens, S. J.;
Dwyer, T. J.; Geierstanger, B. H.; Wemmer, D. E.; Schultz, P. G.;
Spraggon, G.; Romesberg,F. E. J. Am. Chem. Soc. 2007, 129, 10466–10473.
(36) Matsuda, S.; Leconte, A. M.; Romesberg, F. E. J. Am. Chem. Soc.
2007, 129, 5551–5557.
(37) McMinn, D. L.; Ogawa, A. K.; Wu, Y.; Liu, J.; Schultz, P. G.;
Romesberg, F. E. J. Am. Chem. Soc. 1999, 121, 11585–11586.
(38) Seo, Y. J.; Hwang, G. T.; Ordoukhanian, P.; Romesberg, F. E.
J. Am. Chem. Soc. 2009, 131, 3246–3252.
(39) Seo, Y. J.; Matsuda, S.; Romesberg, F. E. J. Am. Chem. Soc. 2009,
131, 5046–5047.
(40) Seo, Y. J.; Romesberg, F. E. ChemBioChem 2009, 10, 2394–
2400.
(41) Yu, C.; Henry, A. A.; Romesberg, F. E.; Schultz, P. G. Angew.
Chem., Int. Ed. 2002, 41, 3841–3844.
(42) Lavergne, T.; Malyshev, D. A.; Romesberg, F. E. Chem. Eur. J.
2011, in press, DOI: 10.1002/chem.201102066.
(43) Hirao, I. Curr. Opin. Chem. Biol. 2006, 10, 622–627.
(44) Hirao, I.; Kimoto, M.; Mitsui, T.; Fujiwara, T.; Kawai, R.; Sato,
A.; Harada, Y.; Yokoyama, S. Nat. Methods 2006, 3, 729–735.
(45) Hirao, I.; Mitsui, T.; Kimoto, M.; Yokoyama, S. J. Am. Chem.
Soc. 2007, 129, 15549–15555.
(46) Kimoto, M.; Kawai, R.; Mitsui, T.; Yokoyama, S.; Hirao, I.
Nucleic Acids Res. 2009, 37, e14.
19888
dx.doi.org/10.1021/ja207907d |J. Am. Chem. Soc. 2011, 133, 19878–19888