Matching active site and substrate structures for an RNA editing reaction

Article abstract of DOI:10.1021/ja9034076

The RNA-editing adenosine deaminases (ADARs) catalyze deamination of adenosine to inosine in a double-stranded structure found in various RNA substrates, including mRNAs. Here we present recent efforts to define structure/activity relationships for the AD

Full text of DOI:10.1021/ja9034076

Published on Web 07/30/2009
Matching Active Site and Substrate Structures for an RNA
Editing Reaction
Subhash Pokharel,‡ Prasanna Jayalath,† Olena Maydanovych,‡ Rena A. Goodman,†
Selina C. Wang,† Dean J. Tantillo,† and Peter A. Beal*,†
Department of Chemistry, UniVersity of California, DaVis, California 95616, and Department of
Chemistry, UniVersity of Utah, 315 South 1400 East, Salt Lake City, Utah 84112-0850
Received April 27, 2009; E-mail: beal@chem.ucdavis.edu
Abstract: The RNA-editing adenosine deaminases (ADARs) catalyze deamination of adenosine to inosine
in a double-stranded structure found in various RNA substrates, including mRNAs. Here we present recent
efforts to define structure/activity relationships for the ADAR reaction. We describe the synthesis of new
phosphoramidites for the incorporation of 7-substituted-8-aza-7-deazaadenosine derivatives into RNA. These
reagents were used to introduce the analogues into mimics of the R/G-editing site found in the pre-mRNA
for the human glutamate receptor B subunit (GluR B). Analysis of the kinetics of the ADAR2 reaction with
analogue-containing RNAs indicated 8-aza-7-deazaadenosine is an excellent substrate for this enzyme
with a deamination rate eight times greater than that for adenosine. However, replacing the C7 hydrogen
in this analogue with bromine, iodine, or propargyl alcohol failed to increase the deamination rate further
but rather decreased the rate. Modeling of nucleotide binding in the enzyme active site suggested amino
acid residues that may be involved in nucleotide recognition. We carried out a functional screen of a library
of ADAR2 mutants expressed in S. cerevisiae that varied the identity of these residues to identify active
deaminases with altered active sites. One of these mutants (ADAR2 R455A) was able to substantially
overcome the inhibitory effect of the bulky C7 substituents (-Br, -I, propargyl alcohol). These results
advance our understanding of the importance of functional groups found in the edited nucleotide and the
role of specific active site residues of ADAR2.
an exon at a position in a codon that changes its meaning.5,6
These sites are often in mRNAs encoding proteins important
Introduction
Pathways for the regulation of gene expression that operate
via modification of mRNA (or pre-mRNA) are recognized for
their potential in the development of new therapeutics.1 For
instance, recent advances in the discovery of small molecules
capable of modulating alternative splicing have led to the
pharamcological control of disease-relevant pre-mRNAs.2 In
addition, antagonists of microRNA-programmed knockdown of
specific messages are showing therapeutic potential for hyper-
cholesterolemia and chronic hepatitis C infections.3 Adenosine
to inosine RNA editing catalyzed by RNA-editing adenosine
deaminase (ADAR) enzymes is another example of post-
transcriptional regulation of gene expression. Thousands of
human transcripts are targets for A to I RNA editing with the
majority of these reactions occurring in untranslated regions for
yet to be defined reasons.4 However, many examples are known
of recoding that occurs when the editing event takes place within
for nervous system function such as glutamate receptors,7
serotonin receptors,8 and voltage gated ion channels.9 ADAR
knockout animals often display developmental defects but also
have defects in their nervous systems arising from the lack of
these critical editing events.10-13 Importantly, a recent report
describes a link between high levels of RNA editing in the pre-
mRNA of the 2C subtype of the serotonin receptor and Prader-
Willi syndrome, a congenital human disease with both devel-
opmental and behavioral symptoms.14 In addition, Gallo and
co-workers have discovered a role for ADAR activity in
controlling cell proliferation and migration in pediatric
astrocytomas.15 Despite the clear importance of the A to I
editing process in controlling gene expression and in human
disease, no reasonably potent low molecular weight inhibitors
are known for ADARs. Nor have any high-resolution
structures of ADAR-substrate complexes been reported.
Efforts in these areas will benefit from a more complete
‡ University of Utah.
† University of California.
(1) Cooper, T. A.; Wan, L.; Dreyfuss, G. Cell 2009, 136, 777–793.
(2) Soret, J.; Bakkour, N.; Maire, S.; Durand, S. b.; Zekri, L.; Gabut, M.;
Fic, W.; Divita, G.; Rivalle, C.; Dauzonne, D.; Nguyen, C. H.; Jeanteur,
P.; Tazi, J. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 8764–8769.
(3) Elmen, J.; Lindow, M.; Schutz, S.; Lawrence, M.; Petri, A.; Obad,
S.; Lindholm, M.; Hedtjarn, M.; Hansen, H. F.; Berger, U.; Gullans,
S.; Kearney, P.; Sarnow, P.; Straarup, E. M.; Kauppinen, S. Nature
2008, 452, 896–899.
(5) Jepson, J. E. C.; Reenan, R. A. Biochim. Biophys. Acta, Gene Regul.
Mech. 2008, 1779, 459–470.
(6) Li, J. B.; Levanon, E. Y.; Yoon, J.-K.; Aach, J.; Xie, B.; LeProust,
E.; Zhang, K.; Gao, Y.; Church, G. M. Science 2009, 324, 1210–
1213.
(7) Higuchi, M.; Single, F. N.; Koehler, M.; Sommer, B.; Sprengel, R.;
Seeburg, P. H. Cell 1993, 75, 1361–70.
(8) Burns, C. M.; Chu, H.; Rueter, S. M.; Hutchinson, L. K.; Canton, H.;
Sanders-Bush, E.; Emeson, R. B. Nature 1997, 387, 303–308.
(9) Bhalla, T.; Rosenthal, J. J. C.; Holmgren, M.; Reenan, R. Nat. Struct.
Mol. Biol. 2004, 11, 950–956.
(4) Bhuiya, M. W.; Rittery, E. M.; Bernard A. Brown, I. In RNA and
DNA Editing; Harold, C. S., Ed.; Wiley-Interscience: Hoboken, NJ,
2008; pp 340-368.
9
11882 J. AM. CHEM. SOC. 2009, 131, 11882–11891
10.1021/ja9034076 CCC: $40.75  2009 American Chemical Society

RNA Editing with 7-Substituted-8-aza-7-deazadenosine
A R T I C L E S
Figure 1. Nucleoside analogues introduced into ADAR2 substrate RNAs. (A) 7-deazaadenosine,17 (B) 8-azaadenosine,23 (C) 8-aza-7-deazaadenosine and
7-substituted-8-aza-7-deazaadenosines (this work).
Figure 2. Proposed stepwise mechanism for adenosine deamination catalyzed by ADAR2.
understanding of the ADAR mechanism and its structure/
activity relationships.
and will lead to new approaches to targeting ADARs with
nucleoside/oligonucleotide ligands.
Biochemical studies and structures of ADAR fragments have
shed some light on how these enzymes recognize substrates and
catalyze adenosine deamination in RNA.16-32 In general,
members of the ADAR family contain multiple RNA-binding
motifs that bind double-stranded RNA substrates in addition to a
zinc-containing catalytic domain where deamination occurs.28,33,34
Interestingly, the ADAR catalytic domain shares sequence
similarity with cytidine deaminases (CDAs), enzymes that
catalyze deamination of cytidine to uridine.35-37 Indeed, the
crystal structure of the human ADAR2 deaminase domain
revealed a CDA-like active site with structural variations near
the periphery that are likely responsible for the adenosine
specificity associated with ADARs.28 However, the crystal
structure of this domain lacks any ligand in the active site. Our
understanding of adenosine recognition by ADARs relies on
models based on this structure and that of CDAs bound to
nucleoside inhibitors.28,38
Results
Our design of substrate analogues for ADAR2 is guided by
the proposed reaction mechanism shown in Figure 2. ADARs
deaminate adenosines within (or immediately adjacent to) a
duplex secondary structure, so the enzyme must first extrude
(10) Higuchi, M.; Maas, S.; Single, F. N.; Hartner, J.; Rozov, A.; Burnashev,
N.; Feldmeyer, D.; Sprengel, R.; Seeburg, P. H. Nature 2000, 406,
78–81.
(11) Palladino, M. J.; Keegan, L. P.; O’Connell, M. A.; Reenan, R. A.
Cell 2000, 102, 437–449.
(12) Tonkin, L. A.; Saccomanno, L.; Morse, D. P.; Brodigan, T.; Krause,
M.; Bass, B. L. EMBO J. 2002, 21, 6025–6035.
(13) Wang, Q.; Khillan, J.; Gadue, P.; Nishikura, K. Science 2000, 290,
1765–1768.
(14) Kishore, S.; Stamm, S. Science 2006, 311, 230–232.
(15) Cenci, C.; Barzotti, R.; Galeano, F.; Corbelli, S.; Rota, R.; Massimi,
L.; Di Rocco, C.; O’Connell, M. A.; Gallo, A. J. Biol. Chem. 2008,
283, 7251–7260.
(16) Yi-Brunozzi, H. Y.; Easterwood, L. M.; Kamilar, G. M.; Beal, P. A.
Nucleic Acids Res. 1999, 27, 2912–2917.
(17) Easterwood, L. M.; Veliz, E. A.; Beal, P. A. J. Am. Chem. Soc. 2000,
122, 11537–11538.
Our laboratory has used the reactivity of adenosine analogues
in RNA as a means of probing substrate recognition in the
ADAR reaction.17,23,25,32,39,40 For instance, we have shown that
7-deazaadenosine 1 in RNA is a good substrate for ADAR2,
indicating that N7 can be removed with little to no effect on
ADAR reactivity (Figure 1A).17 In contrast, introduction of
nitrogen into adenosine at position 8 (8-azaadenosine 2)
substantially enhances ADAR reactivity (Figure 1B).23 This
observation led us to prepare RNA containing 8-azanebularine,
which is capable of trapping ADAR2 in a high affinity complex,
likely via the covalent hydrate.25 Here we describe the effect
of combining the 7-deaza-and the 8-aza modifications of
adenosine (8-aza-7-deazaadenosine 3) (Figure 1C). In addition,
we describe new RNA editing substrates containing 7-substituted-
8-aza-7-deazaadenosines 4-6 bearing different C7 groups and
the role of active site structure in controlling ADAR2’s reaction
with these compounds. The results of these studies further define
the importance of specific functional groups in the reactive
nucleotide and their interaction with ADAR active site residues
(18) Stephens, O. M.; Yi-Brunozzi, H. Y.; Beal, P. A. Biochemistry 2000,
39, 12243–12251.
(19) Veliz, E. A.; Stephens, O. M.; Beal, P. A. Org. Lett. 2001, 3, 2969–
2972.
(20) Yi-Brunozzi, H. Y.; Stephens, O. M.; Beal, P. A. J. Biol. Chem. 2001,
276, 37827–37833.
(21) Cho, D.-S. C.; Yang, W.; Lee, J. T.; Shiekhattar, R.; Murray, J. M.;
Nishikura, K. J. Biol. Chem. 2003, 278, 17093–17102.
(22) Gallo, A.; Keegan, L. P.; Ring, G. M.; O’Connell, M. A. EMBO J.
2003, 22, 3421–3430.
(23) Veliz, E. A.; Easterwood, L. M.; Beal, P. A. J. Am. Chem. Soc. 2003,
125, 10867–10876.
(24) Dawson, T. R.; Sansam, C. L.; Emeson, R. B. J. Biol. Chem. 2004,
279, 4941–4951.
(25) Haudenschild, B. L.; Maydanovych, O.; Veliz, E. A.; Macbeth, M. R.;
Bass, B. L.; Beal, P. A. J. Am. Chem. Soc. 2004, 126, 11213–11219.
(26) Macbeth, M. R.; Lingam, A. T.; Bass, B. L. RNA 2004, 10, 1563–
1571.
(27) Stephens, O. M.; Haudenschild, B. L.; Beal, P. A. Chem. Biol. 2004,
11, 1239–1250.
(28) Macbeth, M. R.; Schubert, H. L.; VanDemark, A. P.; Lingam, A. T.;
Hill, C. P.; Bass, B. L. Science 2005, 309, 1534–1539.
(29) Maydanovych, O.; Beal, P. A. Chem. ReV. 2006, 106, 3397–3411.
(30) Pokharel, S.; Beal, P. A. ACS Chem. Biol. 2006, 1, 761–765.
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11883

A R T I C L E S
Pokharel et al.
Table 1. Computed Hydration Free Energy Relative to
analogues, allowing for an assessment of the effect of different
halogens. In addition, the 7-propargyl alcohol derivative was
analyzed. This analogue was chosen as an example of a
derivative with a more sterically demanding C7 substituent
linked via a carbon-carbon bond. Also, since a variety of
alkyne-containing substituents can be introduced via Sonogashira
couplings with a C7 halo precursor, it was important to evaluate
the effect of a simple C7-alkynyl modification to guide the
design of additional new analogues bearing this linkage.
The DFT calculations indicate that an 8-aza-7-deazapurine
is more readily hydrated than a simple purine. Furthermore, the
calculations suggest that the hydrates of the 7-substituted-8-
aza-7-deazapurines are more stable than that of the unsubstituted
analogue (Table 1) with the hydrate of the 7-bromo compound
the most stable of this group. Therefore, if only the intrinsic
reactivity of the base is considered, one would expect the
corresponding 8-aza-7-deazaadenosines to be competent sub-
strates for ADARs with the 7-bromo analogue reacting the
fastest. More difficult to predict is how these modifications might
affect the base-flipping step of the ADAR reaction. Furthermore,
this analysis does not consider steric effects and how they may
affect the fit of the nucleoside analogue within the enzyme active
site.
To test the reactivity of these analogues in an ADAR reaction,
we prepared phosphoramidites necessary to introduce them into
duplex RNA substrates for human ADAR2. Nucleosides 3-5
were known compounds prior to our work (Figure 1). The
propargyl alcohol derivative was prepared from 7-iodo-8-aza-
7-deazaadenosine 7 via a Sonogashira coupling to give 8 in good
yield (Scheme 1). Silyl protection of the propargyl alcohol
followed by removal of the ribose protecting groups gave
nucleoside 10.
Nucleosides 3-5 and 10 were each protected at the C6 amine
as the dimethylacetamidine (Scheme 2). Dimethoxytrityl protec-
tion of the 5′-hydroxyls along with 2′-O-TBDMS protection and
phosphitylation gave phosphoramidites 23-26 necessary for
generation of modified RNA (Scheme 2).
These phosphoramidites were used to prepare oligoribonucle-
otides corresponding to a sequence found near the R/G editing
site in human GluR B pre-mRNA (Table 2). We have used this
duplex substrate in the past for evaluating the effect of adenosine
modifications on the ADAR2 reaction.16,18,23 Each of the new
phosphoramidites coupled efficiently during automated RNA
synthesis and survived RNA deprotection conditions unaltered
as indicated by ESI-MS of the purified oligonucleotide products
(see Supporting Information).
The effect each modification had on the ADAR2 deamination
rate was evaluated under single turnover conditions at the
saturating concentration of the enzyme. These conditions were
chosen, since kinetic studies with ADAR2, where [substrate].
[enzyme], are complicated by long reaction times for this RNA
substrate.18 In addition, ADAR2 substrates that differ only in
structure at the reactive adenosine typically vary little in
measured binding affinity (Kd) due to the dominant role played
by the duplex RNA-binding domain in determining the affinity
for substrate (see below and Figure S.1).18,23 However, large
differences can be observed in the rate constant for deamination
measured under single turnover conditions at saturating enzyme
concentration for various adenosine analogues in this RNA.16,17,23
We found that 8-aza-7-deazaadenosine (3) at the R/G site
was deaminated to product by ADAR2 with a rate of 0.53 (
0.03 min.-1 (Table 2). This rate is nearly eight times faster than
that for adenosine at the same site in this substrate.16,23 The
9-Methylpurine (kcal/mol)
the nucleotide to be edited from the duplex and into the zinc-
containing active site (i.e., an initial base flipping step). Then
in an SNAr-type reaction, attack of a zinc-bound hydroxide ion
and N1 protonation generates the covalent hydrate of the adenine
ring (i.e., the Messenheimer intermediate). Proton transfer from
hydroxyl to the leaving amine is followed by departure of
ammonia and formation of the inosine product. Recent kinetic
isotope studies with the related tRNA modifying enzyme TadA
are consistent with this proposal.41
Considering this mechanism, one might expect the propensity
to form a covalent hydrate by the addition of water across the
C6-N1 double bond for various purine analogues to be a good
predictor of the C6 deamination rate for related substrates.
Indeed, the hydrate of 8-azapurine is significantly more stable
(based on computed hydration free energies) than that of purine,
while 8-azaadenosine in RNA is deaminated much more rapidly
by ADAR2 than is adenosine.23,42 This correlation between the
stability of covalent hydrates and rate of deamination for related
amino-substituted compounds has been observed for other
heterocycles.42-47 In an attempt to predict the ADAR reactivity
of 7-substituted-8-aza-7-deazaadenosines, we computed covalent
hydration free energies using density functional theory (DFT)
for a group of 8-aza-7-deazapurine analogues bearing a 9-methyl
group (see Experimental Section for computational details)
(Table 1). These compounds included 7-bromo and 7-iodo
(31) Stefl, R.; Xu, M.; Skrisovska, L.; Emeson, R. B.; Allain, F. H.-T.
Structure 2006, 14, 345–355.
(32) Maydanovych, O.; Easterwood, L. M.; Cui, T.; Veliz, E. A.; Pokharel,
S.; Beal, P. A. Methods Enzymol. 2007, 424, 369–86.
(33) Bass, B. L.; Weintraub, H. Cell 1987, 48, 607–13.
(34) Doyle, M.; Jantsch, M. F. J. Cell Biol. 2003, 161, 309–319.
(35) Melcher, T.; Maas, S.; Herb, A.; Sprengel, R.; Seeburg, P. H.; Higuchi,
M. Nature 1996, 379, 460–4.
(36) Kim, U.; Wang, Y.; Sanford, T.; Zeng, Y.; Nishikura, K. Proc. Natl.
Acad. Sci. U.S.A. 1994, 91, 11457–61.
(37) Carter, C. W., Jr. Biochimie 1995, 77, 92–8.
(38) Betts, L.; Xiang, S.; Short, S. A.; Wolfenden, R.; Carter, C. W., Jr. J.
Mol. Biol. 1994, 235, 635–56.
(39) Maydanovych, O.; Beal, P. A. Org. Lett. 2006, 8, 3753–3756.
(40) Jayalath, P.; Pokharel, S.; Veliz, E.; Beal, P. A. Nucleosides,
Nucleotides, Nucleic Acids 2009, 28, 78–88.
(41) Luo, M.; Schramm, V. L. J. Am. Chem. Soc. 2008, 130, 2649–2655.
(42) Erion, M. D.; Reddy, M. R. J. Am. Chem. Soc. 1998, 120, 3295–
3304.
(43) Evans, B.; Wolfenden, R. Biochemistry 1973, 12, 392–398.
(44) Agarwal, R. P.; Sagar, S. M.; Parks, R. E., Jr. Biochem. Pharmacol.
1975, 24, 693–701.
(45) Albert, A. AdV. Heterocycl. Chem. 1976, 20, 117–43.
(46) Shewach, D. S.; Krawczyk, S. H.; Acevedo, O. L.; Townsend, L. B.
Biochem. Pharmacol. 1992, 44, 1697–700.
(47) Hernandez, S.; Ford, H.; Marquez, V. E. Bioorg. Med. Chem. 2002,
10, 2723–2730.
9
11884 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

RNA Editing with 7-Substituted-8-aza-7-deazadenosine
A R T I C L E S
Scheme 1a
a Conditions: (a) Pd(PPh3)4, CuI, Et3N, CHCCH2OH, DMF, 86%; (b) TBDMSCl, Et3N, DCM, 79%; (c) NaOMe/MeOH 70%.
Scheme 2a
a Conditions: (a) CH3C(OCH3)2N(CH3)2, CH3OH; (b) DMTCl, pyridine, AgNO3, THF; (c) TBDMSCl, AgNO3, Et3N, THF; (d) ClP(OCH2CH2CN)(N(iPr)2),
DIPEA, THF.
large rate acceleration observed for 3 is similar to that found
for 8-azaadenosine (2), underscoring the importance of the 8-aza
modification.23 Also, this observation confirmed the 7-deaza
modification has minimal effect on the ADAR2 reaction.
Interestingly, under the same reaction conditions, 7-bromo-8-
aza-7-deazaadenosine (4) was a slower substrate than 3, with a
uents were expected to accelerate the rate of deamination based
on their intrinsic propensities to hydrate (see above). However,
analysis of the crystal structure of the catalytic domain of
ADAR2 with AMP modeled into the active site suggested a
possible clash between a bulky substituent at the purine
7-position and the side chain of R455 (Figure 3).28 We
hypothesized that such an unfavorable interaction within the
enzyme/substrate complex may cause a poor fit within the active
site and contribute to the lower than expected rate of deami-
nation for the 7-substituted compounds. To test this idea, we
used a screen developed previously in our laboratory to identify
functional mutants of ADAR2 with altered active site residues.30
The screen relies on the ability of active ADAR2 mutants to
deaminate within a stop codon in frame with the coding
sequence for the reporter enzyme R-galactosidase while both
enzyme and substrate RNA are expressed in the yeast S.
cereVisiae. The ADAR reaction converts the stop codon to one
for tryptophan, and expression of the reporter is observed as
colored yeast colonies. For this study, two amino acid positions
were chosen for randomization (375 and 455 in the human
rate of deamination similar to that for adenosine (0.08 min-1
)
(Table 2). In addition, the 7-iodo compound (5) and 7-propargyl
alcohol compound (6) each were converted to product substan-
tially more slowly than 3 (88-fold and 176-fold, respectively)
in this RNA. Indeed, the rate of deamination for each of these
two analogues was >10-fold lower than that observed for
adenosine (Table 2). As expected from our previous studies,
these effects are not due to differences in binding since all four
RNA substrates (where X ) H, Br, I, or propargyl alcohol
(Table 2)) are completely bound at this ADAR2 concentration
(Figure S.1).23
The observation that RNAs modified with 7-substituted 8-aza-
7-deazaadenosines 4-6 were slower substrates than 8-aza-7-
deazadenosine 3 was somewhat surprising, since these substit-
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11885

A R T I C L E S
Pokharel et al.
Table 3. Results of Screening ADAR2 T375X,R455X Library
Table 2. Single Turnover Rate Constants for Reaction of ADAR2
with RNAs Bearing 7-Substituted-8-aza-7-deazaadenosine
Analoguesa
no. of
clones
Amino acids
(375, 455)
codon(s)
AGT, GCT
AGT, AGC
ACC, CGT
TCG, GGC
AGT, ACA
AGT, AGA
GGG, TCG
TGC, AGA
TCA, TCA
AGT, AGT
AGT, AGA
AGC, GCG
TCC, GGT
ACA, GCA
1
2
2
2
1
2
3
2
1
1
2
1
1
1
Ser, Ala
Ser, Ser
Thr, Arg
Ser, Gly
Ser, Thr
Ser, Arg
Gly, Ser
Cys, Arg
Ser, Ser
Ser, Ser
Ser, Arg
Ser, Ala
Ser, Gly
Thr, Ala
b
c
krel
enzyme
substrate (N)
kobs, min-1
ADAR2
A
0.07 ( 0.03
0.53 ( 0.03
1
7.6
1.1
0.09
0.04
X ) H
Table 4. Single Turnover Rate Constants for Reaction of
ADAR2(R455A) with RNAs Bearing
X ) Br
0.08 ( 0.01
7-Substituted-8-aza-7-deazaadenosine Analoguesa
X ) I
0.006 ( 0.001
0.003 ( 0.0003
X ) CCCH2OH
c
d
krel
enzyme
substrate (N)b
kobs, min-1
a Reactions were carried out with 250 nM enzyme, 25 nM RNA
substrate. b Data were fitted to the equation: [P]t ) R[1 - exp(-kobs ·t)].
c krel ) kobs for analogue/kobs for adenosine.
ADAR2(R455A)
A
0.037 ( 0.002
0.13 ( 0.004
0.22 ( 0.01
0.05 ( 0.01
0.05 ( 0.004
1
X ) H
3.5
5.9
1.4
1.4
X ) Br
X ) I
X ) CCCH2OH
a Reactions were carried out with 250 nM enzyme, 25 nM RNA
substrate. b Substrate RNA as seen in Table 2. c Data were fitted to the
d
equation: [P]t ) R[1 - exp(-kobs ·t)]. krel ) kobs for analogue/kobs for
adenosine.
to 9 different protein sequences. A summary of the results of
the screen is found in Table 3. Interestingly, a large fraction
(8/9) of mutants identified had small hydrogen bond donors
(threonine, serine, or cysteine) at position 375. This is consistent
with the proposed role of this residue in a hydrogen-bonding
interaction with the 2′ hydroxyl of the reactive nucleotide.30
Importantly, several small residues were found in the active pool
at position 455 (alanine, serine, threonine, and glycine). Thus,
a critical interaction between the guanidinium group of R455
and the purine ring of the reactive nucleotide seems unlikely.
Furthermore, these smaller active site residues may allow
ADAR2 to process more sterically demanding substrates. Along
these lines, we chose to study further the R455A mutant. This
protein was overexpressed and purified as previously de-
scribed.26 RNA substrates containing adenosine and the 7-deaza-
8-azaadenosines 3-6 were then subjected to the reaction of the
ADAR2 R455A mutant. Importantly, this mutant is nearly as
active as wild type ADAR2 with adenosine at the R/G site,
exhibiting rate constants within a factor of 2 for the two proteins
(Tables 2, 4).
However, the reactivity profile for the different analogues is
markedly different with the R455A mutant compared to the wild
type enzyme. With this enzyme, all the 8-aza-7-deazaadenosine
analogues tested were at least as reactive as, if not superior to,
adenosine in this substrate context. RNA containing 7-bromo-
8-aza-7-deazaadenosine 4 was the best substrate, with a rate
nearly six times higher than that for adenosine (Table 4). 8-Aza-
7-deazadenosine also reacts faster than adenosine, but only by
3.5-fold (compared to 8-fold higher for wild type ADAR2).
RNAs containing both the 7-iodo compound 5 and 7-propargyl
alcohol derivative 6 were good substrates for the mutant with
rates slightly higher than that for adenosine. Importantly, each
of the analogues with “bulky” substitutents at C7 (4-6) had
faster rates with the R455A mutant than with the wild type
Figure 3. ADAR2 active site. Active site residue positions for human
ADAR2 as determined by X-ray crystallography for the catalytic domain.28
AMP (red) has been modeled into this active site based on structures of
the related CDAs bound to inhibitors.38
ADAR2 sequence) that are not conserved in the ADAR family
but are in locations near the entrance to the zinc-containing
active site and may be involved in nucleotide recognition (Figure
3).28 If functional mutants of ADAR2 could be discovered with
different active site recognition properties, nucleoside analogues,
such as the 7-substituted 8-aza-7-deazaadenosines described
here, may function as substrates (or inhibitors).48-52
Saturation mutagenesis was carried out at the codons for
residues 375 and 455 of human ADAR2. The resulting plasmid
library was then used to screen for mutants active for RNA
editing in our screen. A total of 22 active clones were sequenced
and 14 unique DNA sequences were identified corresponding
(48) Clemons, P. A.; Gladstone, B. G.; Seth, A.; Chao, E. D.; Foley, M. A.;
Schreiber, S. L. Chem. Biol. 2002, 9, 49–61.
(49) Kraybill, B. C.; Elkin, L. L.; Blethrow, J. D.; Morgan, D. O.; Shokat,
K. M. J. Am. Chem. Soc. 2002, 124, 12118–12128.
(50) Shah, K.; Shokat, K. M. Chem. Biol. 2002, 9, 35–47.
(51) Belshaw, P. J.; Schreiber, S. L. J. Am. Chem. Soc. 1997, 119, 1805–
1806.
(52) Lin, Q.; Jiang, F.; Schultz, P. G.; Gray, N. S. J. Am. Chem. Soc. 2001,
123, 11608–11613.
9
11886 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

RNA Editing with 7-Substituted-8-aza-7-deazadenosine
A R T I C L E S
in this substrate relative to adenosine, but it reacts 40 times
faster with the R455A mutant than it does with wild type
ADAR2.
Discussion
Tight binding nucleoside analogues are valuable in ADAR
structural studies and could function as inhibitors.53,54 We had
previously shown that 8-azapurines are well suited for binding
the ADAR2 active site.25 This is most likely due to the fact
that 8-azapurine is more susceptible than purine to covalent
hydration, a key step in the deamination mechanism (Figure
2).42 Thus, 8-azapurine analogues were promising as new
ligands for ADARs. However, derivatives with substituents at
C2 or C6 bound poorly to ADAR2.23,39 For these reasons, we
were interested in the ADAR reactivity of 8-aza-7-deazaad-
enosine and 7-substituted derivatives. If these were found to
be good ADAR substrates, one could imagine developing new
ligands with substituents at C7 capable of stabilizing contacts
with the ADAR active site (or that of a mutant). Furthermore,
these compounds could be used to test the existing substrate
recognition model.
Figure 4. R455A mutation of ADAR2 accelerates the deamination for
analogues with bulky substitutents at C7 of 8-aza-7-deazaadenosine.
Reactivity of Analogues and Mutants Support Nucleotide
Binding Model for ADAR2 Active Site. The model for nucle-
otide binding in the ADAR2 active site suggests a close
approach between the side chain of R455 and adenine N7
(Figure 3). Interestingly, the nucleoside-modifying enzyme
adenosine deaminase (ADA) binds its substrate via hydrogen
bonds to each of the purine nitrogens, including N7.55 An
aspartic acid in the ADA active site (D296 in mouse ADA)
hydrogen bonds to N7, and this residue is critical for activity
(0.001% of wild type kcat/Km for the D296A mutation). In
addition, removal of N7 from the substrate is highly inhibitory
to the ADA reaction, and 7-substituted 7-deaza compounds are
poor ADA substrates.56,57 In contrast, an important, direct
interaction between the R455 side chain of ADAR2 and
substrate N7 is unlikely, since 7-deazaadenosine and 8-aza-7-
deazaadenosine in RNA are both good substrates17 (Tables 2,
4) and mutants of ADAR2 that vary the position 455 residue
have near wild type activity, including R455A (Table 3). One
can understand the difference in importance of the interaction
between the enzyme and N7 of the adenine for the different
types of adenosine deaminases when one considers that ADARs
bind their substrates tightly via their RNA-binding domains and
are not as dependent on specific interactions to the reacting base
to hold the substrate in the active site (as is the case for the
nucleoside processing enzyme ADA). Indeed, dissociation
constants for RNA binding to the isolated RNA-binding domain
of ADAR2 are similar in magnitude to those for the full-length
enzyme.20,27
Figure 5. 31-mer R/G site RNA substrate used to assay wild type ADAR2
and ADAR2(R455A) editing activity. GluR-B R/G site indicated by N is
substituted with modified nucleotide analogue.
enzyme (R455A kobs/ADAR2 kobs > 1.0), whereas adenosine and
8-aza-7-deazadenosine reacted faster with wild type ADAR2
(R455A kobs/ADAR2 kobs < 1.0) (Figure 4).
The substrate RNA used for the studies described above bears
the reactive nucleotide at the 5′ end of the RNA duplex (Table
2). We use this test substrate design because it facilitates data
acquisition since simple end labeling of the top strand tags the
reactive nucleotide with a radioactive phosphate, necessary for
our TLC-based assay. However, editing site adenosines in
natural substrates are most often flanked on both sides by a
duplex structure. To determine the effect of a flanking duplex
structure on the reactivity of a 7-substituted-8-aza-7-deazaad-
enosine, we prepared a substrate using end labeling and splint
ligation as previously described by us to add RNA structure 5′
to a labeled editing site (Figure 5).18 As expected, this change
in RNA structure accelerates adenosine deamination for both
wild type ADAR2 and the R455A mutant (wild type ADAR2
rate ) 0.8 min-1, R455A rate ) 0.4 min-1) (data not shown).
However, with the 7-iodo derivative 5 at the editing site, the
wild type ADAR2 rate decreased 3-fold (to 2 ( 0.2 × 10-3
min-1) and the R455A rate increased slightly (to 0.08 ( 0.002
min-1) in comparison to the shorter substrate. Thus, the
reactivity of the 7-iodo compound is suppressed even further
Nevertheless, the R455A mutation in ADAR2 clearly com-
pensates for the poor reactivity of the wild type enzyme with
analogues that project sterically demanding substituents from
the C7 position of 8-aza-7-deazaadenosine (Br, I, propargyl
alcohol) (Figure 4). Thus, we conclude that substrate N7 and
the R455 side chain are in proximity at some point during the
reaction but their interaction is not necessary for editing. This
observation has important implications for future design of
nucleotide-based ligands for ADARs. Since both the C7 position
of a 7-deazapurine substrate and the R455 residue of ADAR2
(53) Verdine, G. L.; Norman, D. P. G. Annu. ReV. Biochem. 2003, 72, 337–
366.
(55) Wilson, D. K.; Rudolph, F. B.; Quiocho, F. A. Science 1991, 252,
1278–84.
(54) Jiang, Y. L.; Krosky, D. J.; Seiple, L.; Stivers, J. T. J. Am. Chem.
Soc. 2005, 127, 17412–17420.
(56) Frederiksen, S. Arch. Biochem. Biophys. 1966, 113, 383–388.
(57) Seela, F.; Xu, K. Org. Biomol. Chem. 2007, 5, 3034–3045.
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11887

A R T I C L E S
Pokharel et al.
can be altered in ways that maintain a productive interaction
between enzyme and substrate, these positions are good choices
for engineering in new interactions, such as salt bridges or
covalent cross-links. Stabilized complexes of this type have
aided structural studies of other protein-nucleic acid interac-
tions, and these efforts are currently underway in our
laboratory.40,53 Furthermore, sequence alignment of human
ADAR2 and ADAR1 indicate that ADAR1 has an alanine
residue (A970) at the position corresponding to R455 in
ADAR2.36 Therefore, it is possible that 7-substituted 8-aza-7-
deazaadenosine derivatives described here will be good sub-
strates for wild type ADAR1, although this remains to be tested
and will require development of a suitable in vitro substrate
for ADAR1. If so, 7-substutituted 8-aza-7-deazapurine analogues
may be capable of differentiating ADAR1 from ADAR2.
the editing sites reside in duplexes with a varying flanking
sequence, in internal loops, or at duplex sites with A•C or A•U
pairs.
Experimental Section
General Synthetic Procedures. Glassware for all reactions was
flame-dried or oven-dried at 175 °C overnight and cooled in a
desiccator prior to use. Reactions were carried out under an
atmosphere of dry nitrogen when anhydrous conditions were
necessary. All reagents were purchased from commercial sources
(Sigma/Aldrich or Fischer Scientific) and were used without further
purification unless otherwise stated. Tetrahydrofuran was either
distilled from sodium metal and benzophenone or passed through
a column of activated aluminum.66 TLC analysis of reaction
mixtures was performed on Merck silica gel 60 F254 precoated TLC
plates. Short and long wave visualization was performed with a
Mineralight multiband ultraviolet lamp at 254 and 365 nm,
respectively. Flash column chromatography was performed on
Are C7 Analogues Difficult to Flip? Within the group of
analogues with bulky C7 substituents, our DFT calculations on
the stabilities of covalent hydrates were useful in predicting the
relative reactivity of the substrate analogues (Br > I >
CCCH2OH) (Tables 1, 2, and 4). However, in contrast with the
predicted result, all three of these were less reactive than 8-aza-
7-deazaadenosine with wild type ADAR2. The increase in rate
realized when introducing the R455A mutation explains some
of this lower than expected reactivity as arising from a steric
clash. However, even for the R455A mutant, the 7-iodo and
7-propargyl alcohol compounds were less reactive than 8-aza-
7-deazaadenosine (Table 4). This inhibitory effect could arise
for several reasons. For instance, base modification can alter
the preferred sugar pucker in nucleosides, which has been shown
to be a factor in controlling rates of enzyme-catalyzed nucleoside
deamination.58-63 It is also true that halogens and alkynes at
the7-positionof7-deazapurinesstabilizenucleicacidduplexes.64,65
Since the edited nucleotide must flip out of the duplex and into
the ADAR active site, stabilization of a base-stacked conforma-
tion could inhibit this step and slow the reaction.18,20 Results
with extended duplex substrates are consistent with this latter
hypothesis (Figure 5). The rate of ADAR2-catalyzed adenosine
deamination is higher when this nucleotide is flanked by a
duplex on both sides, yet the 7-iodo analogue reacts more slowly
in this context compared to its reactivity at the duplex end. It
stands to reason that the reactive nucleotide would be less
conformationally flexible within the extended duplex and the
barrier for nucleotide flipping would be higher. Rates of flipping
for the 7-iodo analogue will be needed to confirm that this step
is the reason for the low reactivity observed. Once this is
established, the 7-substituted analogues could be used to probe
the role of base flipping in the various ADAR substrates where
1
Mallinckrodt Baker silica gel 150 (60-200 mesh). H, 13C, and
31P nuclear magnetic resonance spectra were recorded with Varian
VNMRS 600, Varian Mercury 300 spectrometers and referenced
to CDCl3 unless otherwise noted. The abbreviations such as s, t,
m, brs, dd, and d stand for singlet, triplet, multiplet, broad singlet,
doublet of doublets, and doublet. High-resolution mass spectra were
obtained at the Department of Chemistry, University of Utah or at
the University of California, Davis and the University of California,
San Diego Mass spectrometry facility.
4-Amino-3-(3′′-hydroxy-1′′-propynyl)-1-(2′,3′,5′-tri-O-benzoyl-ꢀ-
D-ribofuranosyl)-1H-pyrazolo[3,4-d]pyrimidine (8). A suspension
of 7 (518 mg 0.734 mmol), Pd(PPh3)4 (104 mg, 0.09 mmol), and
CuI (47.5 mg, 0.25 mmol) in anhydrous DMF (3 mL) was treated
with propargyl alcohol (290 µL, 3.67 mmol) followed by anhydrous
triethylamine (81 µL, 0.734 mmol). The mixture was stirred under
Ar at room temperature. After the reaction was complete (TLC),
the solvent was removed in vacuo, and the residue was chromato-
graphed on a flash silica gel column, eluting with CH2Cl2-CH3OH
(95:5) to give 8 (398 mg, 86%) as a white form. 1H NMR (CDCl3,
600 MHz): δ (ppm) 8.31 (s, 1H), 8.08-7.93 (m, 5H), 7.55-7.33
(m, 10H), 6.80 (d, J ) 3.0 Hz, 1H), 6.38 (dd, J ) 3.0, 5.4 Hz,
1H), 6.30 (dd, J ) 6.0, 6.6 Hz, 1H), 4.85 (dd, J ) 1.8, 4.2 Hz,
1H), 4.76 (dd, J ) 3.6, 12 Hz, 1H), 4.64 (dd, J ) 4.8, 12 Hz, 1H),
4.55 (s, 2H). 13C NMR (CDCl3, 150 MHz): δ (ppm) 166.6, 165.6,
165.5, 157.8, 157.1, 154.6, 133.9, 133.8, 133.3, 130.2, 130.1, 130.0,
129.8, 129.0, 128.9, 128.8, 128.7, 128.6, 128.5, 94.1, 87.2, 80.3,
77.1, 74.9, 72.1, 64.2, 51.4. ESIHRMS: calcd for C34H25N5O8 (M
+ H)+ 634.1938, obsd 634.1930.
4-Amino-3-(3′′-O-tert-butyldimethylsilyloxypropyne)-1-(2′,3′,5′-
tri-O-benzoyl-ꢀ-D-ribofuranosyl)-1H-pyrazolo[3,4-d]pyrimidine (9).
To a suspension of 8 (324 mg, 0.51 mmol) and DMAP (50 mg) in
CH2Cl2 (6 mL) was added TBDMSCl (116 mg, 0.77 mmol)
followed by anhydrous triethylamine (0.11 mL, 1.0 mmol). The
mixture was stirred under Ar at room temperature for 12 h and
purified by flash silica gel column chromatography with
CH2Cl2-CH3OH (99:1) to give 9 as a white foam (304 mg, 79%).
1H NMR (CDCl3, 600 MHz): δ (ppm) 8.18 (s, 1H), 7.29-7.91 (m,
2H), 7.81-7.76 (m, 4H), 7.39-7.33 (m, 3H), 7.24-7.16 (m, 6H),
6.64 (d, J ) 3.0 Hz, 1H), 6.20 (dd, J ) 3.6, 5.4 Hz, 1H), 6.13 (t,
J ) 6.6 Hz, 1H), 5.91 (brs, 2H), 4.66 (dd, J ) 4.8, 10.8 Hz, 1H),
4.57 (dd, J ) 4.2, 12.0 Hz, 1H), 4.46 (dd, J ) 4.8, 12.0 Hz, 1H),
4.44 (s, 2H), 0.77 (s, 9H), 0.01 (s, 6H). 13C NMR (CDCl3, 150
MHz): δ (ppm) 166.4, 165.4, 165.2, 157.9, 157.3, 154.7, 133.7,
133.6, 133.5, 132.3, 130.1, 130.9, 129.9, 129.8, 129.0, 128.9, 128.6,
128, 57, 128.50, 94.4, 87.2, 80.3, 76.6, 74.7, 72.0, 64.1, 52.3, 26.0,
25.9, 18.4, -5.02, -5.03. ESIHRMS: calcd for C40H42N5O8Si (M
+ H)+ 748.2803, obsd 748.2794.
(58) Luyten, I.; Thibaudeau, C.; Chattopadhyaya, J. J. Org. Chem. 1997,
62, 8800–8808.
(59) Jeong, L. S.; Buenger, G.; McCormack, J. J.; Cooney, D. A.; Hao,
Z.; Marquez, V. E. J. Med. Chem. 1998, 41, 2572–2578.
(60) Marquez, V. E.; Russ, P.; Alonso, R.; Siddiqui, M. A.; Hernandez,
S.; George, C.; Nicklaus, M. C.; Dai, F.; Ford, H., Jr. HelV. Chim.
Acta 1999, 82, 2119–2129.
(61) Ford, H., Jr.; Dai, F.; Mu, L.; Siddiqui, M. A.; Nicklaus, M. C.;
Anderson, L.; Marquez, V. E.; Barchi, J. J. Biochemistry 2000, 39,
2581–2592.
(62) Seela, F.; Zulauf, M.; Debelak, H. HelV. Chim. Acta 2000, 83, 1437–
1453.
(63) Hernandez, S.; Ford, H.; Marquez, V. E. Bioorg. Med. Chem. 2002,
10, 2723–2730.
(64) Seela, F.; Zulauf, M. J. Chem. Soc., Perkin Trans. 1 1999, 479–488.
(65) Peng, X.; Li, H.; Seela, F. Nucleic Acids Res. 2006, 34, 5987–6000.
(66) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518–20.
9
11888 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

RNA Editing with 7-Substituted-8-aza-7-deazadenosine
A R T I C L E S
4-Amino-3-(3′′-O-tert-butyldimethylsilyloxypropyne)-1-(ꢀ-D-ri-
bofuranosyl)-1H-pyrazolo[3,4-d]pyrimidine (10). To a solution of
compound 9 (304 mg, 0.406 mmol) in anhydrous MeOH (10 mL)
was added MeONa (25 mg, 0.45 mmol) until the pH reached 11.0.
The mixture was stirred at room temperature for 1 h and neutralized
with Amberlyst 15 ion-exchange resin. The resin was filterd and
washed with MeOH, and the solvent was removed under vacuum.
The residue was purified by flash silica gel column chromatography
with CH2Cl2/CH3OH (9:1f 6:1) to give 10 as a white solid (124
added to a solution of a N-(dimethylacetamidine)-protected deriva-
tive (1.08-3.96 mmol) in freshly distilled THF (5-15 mL). The
resulting reaction mixture was stirred at room temperature for 12 h.
It was then diluted with EtOAc (25 mL), filtered, and washed with
saturated aqueous NaHCO3 (1 × 40 mL). The organic portion was
dried (Na2SO4), filtered, and concentrated under reduced pressure.
Purification was carried out by flash column chromatography on a
silica gel using CH2Cl2/CH3OH/TEA 96:3:1.
4-{[(Dimethylamino)ethylidene]amino}-1-(ꢀ-D-ribofuranosyl)-5′-
O-(4,4′-dimethoxytriphenylmethyl)-1H-pyrazolo[3,4-d]pyrimi-
dine (15). A white foam (344.1 mg, 50%). 1H NMR (CD2Cl2, 300
MHz): δ (ppm) 8.44 (s, 1H), 7.91 (s, 1H), 7.42-7.15 (m, 9H),
6.84-6.72 (m, 4H), 6.40 (d, J ) 3.7 Hz, 1H), 4.93-4.91 (m, 1H),
4.6 (t, J ) 5.13 Hz, 1H), 4.23-4.18 (m, 1H), 3.74 (s, 6H),
3.31-3.12 (m, 8H), 2.17 (s, 3H). 13C NMR (CD2Cl2, 75 MHz): δ
(ppm) 163.1, 163.0, 159.0, 156.3, 155.5, 145.7, 136.6, 136.5, 134.6,
130.6, 128.7, 128.3, 128.2, 127.1, 113.6, 113.5, 109.3, 89.3, 86.5,
84.0, 74.3, 72.4, 64.9, 55.7, 39.0, 38.6, 30.2, 17.4. HRFABMS:
calcd for C35H39N6O6 (M + H)+ 639.2931, obsd 639.2925.
4-{[(Dimethylamino)ethylidene]amino}-3-bromo-1-(ꢀ-D-ribo-
furanosyl)-5′-O-(4,4′-dimethoxytriphenylmethyl)-1H-pyrazolo[3,4-
d]pyrimidine (16). A white foam (201 mg, 85%). 1H NMR (CD2Cl2,
600 MHz): δ (ppm) 8.36 (s, 1H), 7.45-7.18 (m, 9H), 6.8-6.78
(m, 4H), 6.33 (d, J ) 4.2 Hz, 1H), 4.90 (t, J ) 4.8 Hz, 1H), 4.48
(t, J ) 5.4 Hz, 1H), 4.20 (dd, J ) 4.2, 8.4 Hz, 1H), 3.76 (s, 3H),
3.33 (dd, J ) 4.2, 10.2 Hz, 1H), 3.24 (s, 3H), 3.20 (dd, J ) 5.4,
10.2 Hz, 1H), 3.16 (s, 3H), 2.07 (s, 3H). 13C NMR (CD2Cl2, 150
MHz): δ (ppm) 158.7, 155.25, 145.3, 136.3, 136.2, 130.3, 128.4,
127.9, 126.8, 122.6, 113.2, 106.8, 88.7, 86.3, 83.9, 73.8, 72.0, 64.3,
55.4, 46.5, 38.9, 38.8, 17.7, 8.6. ESIHRMS: calcd for C35H38BrN6O6
(M + H)+ 717.2036, obsd 717.2048.
4-{[(Dimethylamino)ethylidene]amino}-3-iodo-1-(ꢀ-D-ribofura-
nosyl)-5′-O-(4,4′-dimethoxytriphenylmethyl)-1H-pyrazolo[3,4-d]py-
rimidine (17). A white foam (352.8 mg, 70%). 1H NMR (CD2Cl2,
300 MHz): δ (ppm) 8.34 (s, 1H), 7.48-7.16 (m, 9H), 6.8-6.77
(m, 4H), 6.40 (d, J ) 4.9 Hz, 1H), 5.00 (t, J ) 4.9, 1H), 4.46-4.43
(m, 1H), 4.26-4.23 (m, 1H), 3.74 (s, 6H), 3.38-3.02 (m, 8H),
2.14 (s, 3H). 13C NMR (CD2Cl2, 75 MHz): δ (ppm) 162.7, 162.2,
159.0, 156.3, 155.2, 145.6, 136.7, 136.5, 130.6, 128.7, 128.3, 127.1,
113.6, 110.7, 92.6, 88.8, 86.6, 84.5, 74.0, 72.4, 64.8, 55.7, 39.0,
17.8. HRESIFTMS: calcd for C35H38IN6O6 (M + H)+ 765.1897,
obsd 765.1899.
1
mg, 70%). H NMR (CD3OD, 600 MHz): δ (ppm) 8.03 (s, 1H),
6.04 (d, J ) 4.8 Hz, 1H), 4.53 (t, J ) 4.8 Hz, 1H), 4.48 (s, 2H),
4.23 (t, J ) 4.8 Hz, 1H), 3.92 (dd, J ) 4.8, 8.4 Hz, 1H), 3.61 (dd,
J ) 3.6, 12.6 Hz, 1H), (dd, J ) 5.4, 12.6 Hz, 1H), 0.76 (s, 9H),
0.01 (s, 6H). 13C NMR (CD3OD, 150 MHz): δ (ppm) 166.3, 160.4,
157.8, 131.6, 105.6, 97.8, 94.3, 89.8, 79.6, 78.1, 75.3, 66.7, 55.7,
28.9, 21.8, -2.3. ESIHRMS: calcd for C19H30N5O5Si (M + H)+
436.2016, obsd. 436.2027.
General Procedure for the Preparation of N-(Dimethylaceta-
midine)-7-substituted-8-aza-7-deazaadenosine Analogues. N,N-
Dimethylacetamide dimethyl acetal (3 equiv) was added to a
solution of the ribonucleoside (0.25-1.02 mmol) in CH3OH (1.3-8
mL), and the resulting reaction mixture was stirred at room
temperature for 12 h. It was then concentrated and purified by flash
column chromatography on a silica gel using 4% CH3OH/CH2Cl2
followed by 6% CH3OH/CH2Cl2.
4-{[(Dimethylamino)ethylidene]amino}-1-(ꢀ-D-ribofuranosyl)-
1H-pyrazolo[3,4-d]pyrimidine (11). A white foam (45.4 mg, 54%).
1H NMR (CD3OD, 300 MHz): δ (ppm) 8.50 (s, 1H), 8.13 (s, 1H),
6.36 (d, J ) 4.7 Hz, 1H), 4.87-4.83 (m, 1H), 4.53-4.5 (m, 1H),
4.19 (q, J ) 4.4, 1H), 3.91-3.71 (m, 2H), 3.26 (s, 6H), 2.27 (s,
3H). 13C NMR (CD3OD, 75 MHz): δ (ppm) 164.7, 164.2, 156.7,
155.7, 135.7, 110.1, 91.6, 87.1, 75.5, 72.8, 64.2, 39.1, 38.8, 17.6.
HRFABMS: calcd for C14H21N6O4 (M + H)+ 337.1624, obsd
337.1631.
4-{[(Dimethylamino)ethylidene]amino}-3-bromo-1-(ꢀ-D-ribo-
furanosyl)-1H-pyrazolo[3,4-d]pyrimidine (12). (187 mg, 75%). 1H
NMR (CDCl3, 600 MHz): δ (ppm) 8.27 (s, 1H), 6.23 (d, J ) 6.0
Hz, 1H), 4.88 (dd, J ) 5.4, 6.0 Hz, 1H), 4.46 (dd, J ) 1.2, 5.4 Hz,
1H), 4.28 (d, J ) 1.2 Hz, 1H), 3.90 (dd, J ) 1.8, 12.6 Hz, 1H),
3.70 (dd, J ) 1.2, 12.6 Hz, 1H), 3.28 (s, 3H), 3.16 (s, 3H), 2.12 (s,
3H). 13C NMR (CD3OD, 150 MHz): δ (ppm) 162.6, 161.4, 156.6,
154.0, 122.2, 107.2, 92.2, 87.5, 74.8, 73.2, 63.8, 38.9, 17.6
ESIHRMS: calcd for C14H20BrN6O4 (M + H)+ 415.0729, obsd
415.0729.
4-{[(Dimethylamino)ethylidene]amino}-3-(3′′-O-tert-butyldimeth-
ylsilyloxypropyne)-1-(ꢀ-D-ribofuranosyl)-5′-O-(4,4′-dimethoxytriph-
enylmethyl)-1H-pyrazolo[3,4-d]pyrimidine (18). A white foam (186
4-{[(Dimethylamino)ethylidene]amino}-3-iodo-1-(ꢀ-D-ribofura-
nosyl)-1H-pyrazolo[3,4-d]pyrimidine (13). A yellow foam (305.3
1
mg, 68%). H NMR (CD2Cl2, 600 MHz): δ (ppm) 8.36 (s, 1H),
1
mg, 65%). H NMR (CD3OD, 300 MHz): δ (ppm) 8.41 (s, 1H),
7.39-7.37 (m, 2H), 7.28-7.23 (m, 4H), 7.21-7.18 (m, 2H),
7.15-7.14 (m, 1H), 6.74-6.72 (m, 4H), 6.32 (d, J ) 4.2 Hz, 1H),
4.88 (dd, J ) 3.6, 4.8 Hz, 1H), 4.52 (s, 2H), 4.51 (t, J ) 7.2 Hz,
1H), 4.15 (dd, J ) 5.4, 9.6 Hz, 1H), 3.72 (s, 6H), 3.29 (dd, J )
4.2, 10.2 Hz, 1H), 3.19 (dd, J ) 5.4, 10.2 Hz, 1H) 3.18 (brs, 3H),
3.11 (bs, 3H), 2.13 (s, 3H), 0.88 (s, 9H), 0.10 (s, 6H). 13C NMR
(CD2Cl2, 150 MHz): δ (ppm) 168.0, 167.4, 164.0, 161.8, 160.3,
150.6, 141.6, 141.5, 135.6, 135.5, 134.7, 133.6, 133.3, 132.1, 118.5,
114.4, 97.0, 94.5, 91.6, 88.9, 82.5, 79.1, 77.4, 69.7, 60.7, 57.6, 51.8,
44.0, 31.1, 23.7, 22.6, 14.0, 0.0. ESIHRMS: calcd for C44H55N6O7Si
(M + H)+ 807.3902, obsd 807.3886.
6.25 (d, J ) 4.7 Hz, 1H), 4.78-4.75 (m, 1H), 4.44 (t, J ) 4.7 Hz,
1H), 4.15-4.12 (m, 1H), 3.85-3.66 (m, 2H), 3.29 (s, 3H), 3.22
(s, 3H), 2.18 (s, 3H). 13C NMR (CD3OD, 75 MHz): δ (ppm) 163.7,
157.0, 155.5, 111.8, 93.6, 91.3, 87.1, 75.4, 72.6, 64.1, 39.1, 18.0.
HRESIFTMS: calcd for C14H20IN6O4 (M + H)+ 463.0591, obsd
463.0576.
4-{[(Dimethylamino)ethylidene]amino}-3-(3′′-O-tert-butyldimeth-
ylsilyloxypropyne)-1-(ꢀ-D-ribofuranosyl)-1H-pyrazolo[3,4-d]pyri-
midine (14). (167 mg, 86%). 1H NMR (CDCl3, 600 MHz): δ (ppm)
8.31 (s, 1H), 6.25 (d, J ) 6.0 Hz, 1H), 5.93 (t, J ) 5.4 Hz, 1H),
4.50 (s, 2H), 4.47 (dd, J ) 1.2, 4.8 Hz, 1H), 4.29 (d, J ) 1.8 Hz,
1H), 3.91 (dd, J ) 1.8, 12.6 Hz, 1H), 3.71 (dd, J ) 1.2, 12.6 Hz,
1H), 3.22 (s, 3H), 3.13 (s, 3H), 2.10 (s, 3H), 0.88 (s, 9H), 0.11 (s,
6H). 13C NMR (CDCl3, 150 MHz): δ (ppm) 162.2, 161.9, 156.0,
153.3, 128.9, 109.0, 92.8, 91.9, 87.4, 76.4, 74.6, 72.8, 63.6, 52.1
38.7, 25.8, 18.3, 17.5, -5.0, -5.1. ESIHRMS: calcd for
C23H37N6O5Si (M + H)+ 505.2595, obsd 505.2588.
General Procedure for the Preparation of N-(Dimethylaceta-
midine)-5′-O-(4,4′-dimethoxytrityl)-2′-O-(tert-butyldimethyl-silyl)-
7-substituted-8-aza-7-deazaadenosine Analogues. Triethylamine
(1.9 equiv) and tert-butylchlorodimethylsilane (1.1 equiv) were
consequently added to a solution of 5′-O-DMT protected derivative
(0.54-0.46 mmol) in freshly distilled THF (8 mL). AgNO3 (1.1
equiv) was added after stirring for 5 min. The resulting reaction
mixture was stirred at room temperature for 12 h. It was then diluted
with EtOAc (25 mL), filtered, and washed with 5% aqueous
NaHCO3 (1 × 30 mL). The organic portion was dried (Na2SO4),
filtered, and concentrated under reduced pressure. Purification was
General Procedure for the Preparation of N-(Dimethylaceta-
midine)-5′-O-(4,4′-dimethoxytrityl)-7-substituted-8-aza-7-deazaad-
enosine Analogues. Anhydrous pyridine (6.0 equiv), 4,4′-dimethoxy-
trityl chloride (1.1 equiv), and AgNO3 (1.1 equiv) were consequently
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11889

A R T I C L E S
Pokharel et al.
carried out by flash column chromatography on a silica gel by using
a solvent system listed for individual compounds. The identity of
2′-O-TBDMS isomer vs 3′-O-TBDMS isomer was confirmed by
2D-NMR (COSY).
General Procedure for the Preparation of N-(Dimethylaceta-
midine)-5′-O-(4,4′-dimethoxytrityl)-3′-O-[(2-cyanoethyl)(N,N-diiso-
propylamino)phosphino]-2′-O-(tert-butyldimethylsilyl)-7-substituted-
8-aza-7-deazaadenosine Analogues. N,N-Diisopropylethylamine (6
equiv) and 2-cyanoethyl-(N,N-diisopropylamino)chlorophosphite
(1.1 equiv) were consequently added to a solution of 5′-O-DMT,
2′-O-TBDMS protected derivative (0.18-0.11 mmol) in freshly
distilled THF (1 mL). The resulting reaction mixture was stirred at
room temperature for 12 h. It was then diluted with EtOAc (30
mL), filtered, and washed with 5% (w/v) aqueous NaHCO3 (2 ×
15 mL). The organic portion was dried (Na2SO4), filtered, and
concentrated under reduced pressure. Purification was carried out
by flash column chromatography on silica gel using the solvent
system listed for individual compounds.
4-{[(Dimethylamino)ethylidene]amino}-1-(ꢀ-D-ribofuranosyl)-5′-
O-(4,4′-dimethoxytriphenylmethyl)-3′-O-[(2-cyanoethyl)(N,N-diiso-
propylamino)phosphino]-2′-O-(tert-butyldimethylsilyl)-1H-pyra-
zolo[3,4-d]pyrimidine (23). Chromatography, EtOAc/hexanes/TEA
39:60:1. A white foam (137.9 mg, 78%). 31P NMR (CD2Cl2, 300
MHz): δ (ppm) 150.1, 148.5.
4-{[(Dimethylamino)ethylidene]amino}-3-bromo-1-(ꢀ-D-ribo-
furanosyl)-5′-O-(4,4′-dimethoxytriphenylmethyl)-3′-O-[(2-cyanoeth
yl)(N,N-diisopropylamino)phosphino]-2′-O-(tert-butyldimethylsilyl)-
1H-pyrazolo[3,4-d]pyrimidine (24). Chromatography, ethyl acetate/
TEA 99:1. A white foam (108 mg, 92%). 31P NMR (CD2Cl2, 300
MHz): δ (ppm) 151.5, 150.7, 149.8. ESIHRMS: calcd for
C50H69BrN8O7PSi (M + H)+ 1031.3978, obsd 1031.3979.
4-{[(Dimethylamino)ethylidene]amino}-3-iodo-1-(ꢀ-D-ribofura-
nosyl)-5′-O-(4,4′-dimethoxytriphenylmethyl)-3′-O-[(2-cyanoeth-
yl)(N,N-diisopropylamino)phosphino]-2′-O-(tert-butyldimethylsilyl)-
1H-pyrazolo[3,4-d]pyrimidine (25). Chromatography, EtOAc/TEA
99:1. A white foam (110.9 mg, 96%). 31P NMR (CD2Cl2, 300
4-{[(Dimethylamino)ethylidene]amino}-1-(ꢀ-D-ribofuranosyl)-5′-
O-(4,4′-dimethoxytriphenylmethyl)-2′-O-(tert-butyldimethylsilyl)-
1H-pyrazolo[3,4-d]pyrimidine (19). Chromatography, EtOAc/
hexanes 1:2. 2′-O-TBDMS isomer, a white foam (155 mg, 38%).
1H NMR (CD2Cl2, 300 MHz): δ (ppm) 8.53 (s, 1H), 7.99 (s, 1H),
7.49-7.42 (m, 9H), 6.78-6.76 (m, 4H), 6.35 (d, J ) 4.9 Hz, 1H),
5.18 (t, J ) 5.1 Hz, 1H), 4.35 (dd, J ) 4.9 Hz, 1H), 4.17-4.13
(m, 1H), 3.76 (s, 6H), 3.78-3.11 (m, 8H), 2.74 (d, J ) 4.9 Hz,
1H), 2.22 (s, 3H), 0.84 (s, 9H), 0.03 (s, 3H), -0.13 (s, 3H). 13C
NMR (CD2Cl2, 75 MHz): δ (ppm) 163.4, 162.8, 159.0, 156.7, 156.2,
145.8, 136.8, 136.6, 134.7, 130.7, 128.8, 128.2, 127.1, 113.5, 109.5,
88.8, 86.6, 84.3, 75.2, 72.5, 64.7, 55.7, 39.0, 38.5, 26.0, 18.4, 17.3,
-4.7, -4.9. HRFABMS: calcd for C41H53N6O6Si (M + H)+
753.3796, obsd 753.3771.
4-{[(Dimethylamino)ethylidene]amino}-3-bromo-1-(ꢀ-D-ribo-
furanosyl)-5′-O-(4,4′-dimethoxytriphenylmethyl)-2′-O-(tert-butyldi-
methylsilyl)-1H-pyrazolo[3,4-d]pyrimidine (20). Chromatography,
ethyl acetate/hexane 99:1. 2′-O-TBDMS, a white foam (135 mg,
55%). 1H NMR (CD2Cl2, 600 MHz): δ (ppm) 8.62 (s, 1H),
7.62-7.52 (m, 2H), 7.51-7.48 (m, 4H),7.38-7.36 (m, 2H),
7.31-7.28 (m, 1H), 6.93-6.90 (m, 4H), 6.42 (d, J ) 5.4 Hz, 1H),
5.28 (t, J ) 5.4, 1H), 4.36 (brs, 1H), 4.26 (dd, J ) 3.6, 7.8 Hz,
1H), 3.87 (s, 6H), 3.49 (dd, J ) 3.6, 10.8 Hz, 1H), 3.37 (s, 3H),
3.27 (s, 3H), 3.22 (dd, J ) 4.2, 10.2 Hz, 1H), 2.82 (brs, 1H), 2.36
(s, 3H), 0.96 (s, 9H), 0.14 (s, 3H), -0.0 (s, 3H). 13C NMR (CD2Cl2,
150 MHz): δ (ppm) 167.6, 167.5, 163.8, 162.3, 161.6, 150.6, 141.6,
141.3, 135.5, 133.6, 133.1, 131.9, 127.5, 118.42, 118.41, 112.8,
93.3, 91.6, 89.3, 79.6, 77.3, 69.4, 60.5, 43.9, 43.6, 30.8, 23.2, 22.4,
0.2, 0.0. ESIHRMS: calcd for C41H52BrN6O6Si (M + H)+ 832.8777,
obsd 832.8779.
MHz):
δ (ppm) 151.6, 149.7. HRESIFTMS: calcd for
C50H69IN8O7PSi (M + H)+ 1079.3841, obsd 1079.3838.
4-{[(Dimethylamino)ethylidene]amino}-3-(3′′-O-tert-butyldimeth-
ylsilyloxypropyne)-1-(ꢀ-D-ribofuranosyl)-5′-O-(4,4′-dimethoxytriph-
enylmethyl)-3′-O-[(2-cyanoethyl)(N,N-diisopropylamino)phosphino]-
2′-O-(tert-butyldimethylsilyl)-1H-pyrazolo[3,4-d]pyrimidine (26).
Purified by radial chromatography (1 mm plate), ethyl acetate/
hexane/TEA 64:35:1. A white foam (92 mg, 91%). 31P NMR
(CD2Cl2, 300 MHz): δ (ppm) 151.5, 149.9. ESIHRMS: calcd for
C59H86N8O8PSi2 (M + H)+ 1121.5845, obsd 1121.5840.
General Biochemical Procedures. Distilled, deionized water
was used for all aqueous reactions and dilutions. Biochemical
reagents were purchased from Sigma/Aldrich unless otherwise
noted. Common enzymes were purchased from Roche, Promega,
or New England Biolabs. Phosphoramidites 23-26 were synthe-
sized as described above. [γ-32P]ATP (6000 Ci/mmol) was obtained
from Perkin-Elmer Life Sciences. Storage phosphor autoradio-
graphy was carried out using imaging plates from Eastman Kodak
Co and a Molecular Dynamics Typhoon 9400 or Typhoon Trio.
Protein Overexpression and Purification. Human ADAR2 (WT)
and ADAR2 (R455A) in yeast expression plasmid (YEpTOP2PGAL1)
were overexpressed in Saccharomyces cereVisiae and purified as
previously described.26
Description of T375X, R455X Library Generation and
Screening. A QuikChange XL site-directed mutagenesis kit (Strat-
agene) was used to generate plasmid libraries following the
manufacturer’s protocol. The two specified codons were randomized
sequentially. Codon 375 was randomized first. To generate this
library the following primers were used: 5′-GGTGATAAGT-
GTTTCTACAGGANNNAAATGTATTAATGGTGAATACA and
5′-TGTATTCACCATTAATACATTTNNNTCCTGTAGAAACA-
CTTATCACC, where N represents an equal mixture of A, C, G,
and T. The S. cereVisiae expression vector for human ADAR2
YEpTOP2PGAL1(T375Amber) was used as a PCR template.30 This
amber stop codon plasmid was used as a template to minimize the
4-{[(Dimethylamino)ethylidene]amino}-3-iodo-1-(ꢀ-D-ribofura-
nosyl)-5′-O-(4,4′-dimethoxytriphenylmethyl)-2′-O-(tert-butyldimeth-
ylsilyl)-1H-pyrazolo[3,4-d]pyrimidine (21). Chromatography, EtOAc.
1
2′-O-TBDMS, a white foam (123.5 mg, 30%). H NMR (CD2Cl2,
300 MHz): δ (ppm) 8.53 (s, 1H), 7.55-7.18 (m, 9H), 6.85-6.82
(m, 4H), 6.33 (d, J ) 5.6 Hz, 1H), 5.26 (t, J ) 5.3 Hz, 1H), 4.27
(brs, 1H), 4.2 (dd, J ) 3.2 Hz, 1H), 3.78 (s, 6H), 3.44-3.10 (m,
8H), 2.81 (brs, 1H), 2.26 (s, 3H), 0.85 (s, 9H), 0.05 (s, 3H), -0.1
(s, 3H). 13C NMR (CD2Cl2, 75 MHz): δ (ppm) 162.7, 162.6, 159.0,
157.0, 156.1, 145.7, 136.8, 136.5, 130.7, 128.8, 128.4, 127.1, 113.6,
111.1, 92.4, 88.5, 86.8, 84.5, 75.1, 72.6, 64.7, 55.7, 39.1, 38.9, 25.9,
18.4, 17.7, -4.6, -4.8. HRESIFTMS: calcd for C41H52IN6O6Si (M
+ H)+ 879.2762, obsd 879.2768.
4-{[(Dimethylamino)ethylidene]amino}-3-(3′′-O-tert-butyldimeth-
ylsilyloxypropyne)-1-(ꢀ-D-ribofuranosyl)-5′-O-(4,4′-dimethoxytriph-
enylmethyl)-2′-O-(tert-butyldimethylsilyl)-1H-pyrazolo[3,4-d]pyri-
midine (22). A white foam (83.3 mg, 52%). 1H NMR (CD2Cl2, 600
MHz): δ (ppm) 8.65 (s, 1H), 7.62-7.60 (m, 2H), 7.50-7.46 (m,
4H), 7.40-7.36 (m, 2H), 7.32-7.29 (m, 1H), 6.93-6.90 (m, 4H),
6.45 (d, J ) 4.8 Hz, 1H), 5.24 (t, J ) 4.8 Hz, 1H), 4.68 (s, 2H),
4.41 (dd, J ) 4.8, 9.6 Hz, 1H), 4.28 (dd, J ) 3.6, 8.4 Hz, 1H),
3.89 (s, 6H), 3.49 (dd, J ) 3.6, 10.2 Hz, 1H), 3.37 (brs, 3H), 3.31
(dd, J ) 4.8, 10.2 Hz, 1H) 3.28 (brs, 3H), 2.82 (d, J ) 5.4 Hz,
1H), 2.34 (s, 3H), 1.04 (s, 9H), 0.97(s, 9H), 0.24 (s, 6H), 0.15 (s,
3H), 0.00 (s, 3H). 13C NMR (CD2Cl2, 150 MHz): δ (ppm)168.3,
167.4, 164.0, 162.3, 162.2, 161.0 150.7, 141.6, 141.5, 135.74,
135.71, 135.6, 134.7, 133.8, 133.7, 133.3, 132.1, 118.7, 118.6,
118.5, 114.7, 96.9, 93.92, 93.9, 91,7, 89.5, 89.4, 82.6, 80.1, 80.0,
77.5, 77.4, 69.6, 69.5, 60.7, 60.6, 57.8, 57.7, 57.6, 31.2, 31.1, 31.0,
30.9, 30.8, 23.7, 23.4, 22.5, 22.4, 0.4, 0.2, 0.14, 0.1, 0.05, 0.0.
ESIHRMS: calcd for C50H69N6O7Si2 (M + H)+ 921.4766, obsd
921.4765.
9
11890 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

RNA Editing with 7-Substituted-8-aza-7-deazadenosine
A R T I C L E S
over-representation of wild type DNA in the library. XL 10 gold
E. coli cells were transformed with the resulting PCR products using
ampicillin selection. Approximately 500 bacterial colonies were
harvested, combined and grown in LB-ampicillin selection. Plasmid
DNA was isolated using a QIAprep Spin Miniprep kit (Qiagen).
This resulting library DNA was used as a template to randomize
the 455 codon, where the following two primers were used: 5′-
C C T C T C C C T G T G G A G A T G C C N N N A T C T -
TCTCACCACATGAGCC and 5′-GGCTCATGTGGTGAGAA-
GATNNNGGCATCTCCACAGGGAGAGG. The PCR reaction
was used to transform bacteria, and approximately 5000 colonies
were picked to isolate the T375X, R455X DNA library. Random-
ization at the corresponding sites was confirmed by DNA sequencing.
change.32 A 10-fold excess of modified unlabeled strand was added,
and the resulting mixture was hybridized to the unlabeled comple-
ment strand.
Deamination Assay. The deamination assay and data quantifica-
tion were carried out as previously described with the slight
change.32 Single turnover editing reactions are carried out with 250
nM ADAR2 or mutant ADAR2(R455A) and 25 nM labeled RNA
duplexes.
Oligos were incubated in deamination buffer containing no
enzyme for 90 min at 30 °C to determine if any degradation
products arise that comigrate with the deamination product during
TLC. No observable degradation products with similar RF values
to those of the deaminated nucleotide analogue were observed.
Computational Methods. Hydration free energies (purine
analogue + water f C6-hydrated purine analogue) in water were
calculated using the CPCM-B3LYP/6-31+G(d,p) method.67-71 For
calculations on the iodine-containing system, a split basis set was
used: 6-31+G(d,p) was used for carbon, hydrogen, nitrogen, and
oxygen, and the LANL2DZ basis set,72 with pseudopotential, was
used for iodine. Geometries of all structures (purines with methyl
groups in place of the sugar moiety, hydrated versions thereof, and
water) were fully optimized in solvent, and UA0 radii73 and a
dielectric constant of 78.39 were used in all calculations. The
GAUSSIAN program suite was used for all calculations.74 Coor-
dinates and energies can be found in the Supporting Information.
Additional calculations using different methods and including
different heterocycles will be reported separately.
Synthesis, Purification, and Mass Spectrometric Analysis
of RNA. RNA oligonucleotides were synthesized on an ABI 394
synthesizer (DNA/Peptide Core Facility, University of Utah) using
5′-DMT, 2′-O-TBDMS protected ꢀ-cyanoethyl phosphoramidites
on a 1.0 µmol scale with coupling times of 25 min for more efficient
coupling. Oligonucleotide deprotection and purification were carried
out as previously described with the slight change for the depro-
tection of the 7-substituted-8-aza-7-deazaadenosine CPG-bound
oligonucleotide.32 These RNA oligonucleotides were incubated with
ammonium hydroxide at room temperature for 24 h prior to removal
of TBDMS with neat triethylamine trihydrofluoride. The RNA
oligonucleotides after polyacrylamide gel electrophoresis purifica-
tion were submitted to mass spectrometric analysis at the Mass
Spectrometry and Proteomics Core Facility, University of Utah.
Oligonucleotides were analyzed in an acetonitrile solution at pH
) 11 by electrospray ionization on a Quattro-II mass spectrometer
(Micromass, Inc.). The solution was introduced at 3 µL/min with
ESI conditions of 2.7 kV capillary voltage, 35 eV cone voltage,
and nitrogen nebulization. The instrument was scanned from 700
to 1400 m/z, and scans were combined for approximately 1 min.
The multiply charged molecular ions were deconvoluted into a
molecular mass spectrum using MaxEnt (Micromass, Inc.) software.
The Quattro-II instrument was operated using Masslynx software
version 3.4 (Micromass, Inc.). ESI mass spectrometry analysis of
26-mer RNA oligonucleotide containing compound 3: calcd: 8455.1;
obsd: 8456.1. ESI mass spectrometry analysis of RNA oligonucle-
otide containing compound 4: calcd: 8535.06; obsd: 8534.0. ESI
mass spectrometry analysis of RNA oligonucleotide containing
compound 5: calcd: 8582.1; obsd: 8581.8. ESI mass spectrometry
analysis of RNA oligonucleotide containing compound 6: calcd:
8510.18; obsd: 8510.1. To ensure the C7 substituent remains
unaltered during the ADAR2 reaction, RNA containing 7-bromo-
8-aza-7-deazaA was subjected to the deamination reaction, purified,
and analyzed by mass spectrometry. ESI mass spectrometry analysis
of RNA oligonucleotide containing compound 4: calcd: 8535.06;
obsd: 8534.0. Observed mass after deamination 8535.2.
Acknowledgment. P.A.B. acknowledges the National Institutes
of Health for financial support in the form of Grant GM061115.
We would like to acknowledge the assistance of mass spectrometry
facilities at the University of Utah, the University of California,
Davis, and the University of California, San Diego.
Supporting Information Available: NMR and mass spectra
for new compounds, ESI mass spectra for modified oligoribo-
nucleotides, deamination kinetic data, computed geometries and
hydration free energies for 8-aza-7-deaza-purine analogues, and
full ref 74. This material is available free of charge via the
Internet at http://pubs.acs.org.
JA9034076
(67) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(68) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B: Solid Sate 1988, 37,
785–789.
(69) Stephens, P. J.; Devlin, F. J.; Cabalowski, C. F.; Frisch, M. J. J. Phys.
Chem. 1994, 98, 11623–11627.
(70) Barone, V.; Cossi, M. J. J. Phys. Chem. A 1998, 102, 1995–2001.
(71) Barone, V.; Cossi, M.; Tomasi, J. J. Comput. Chem. 1998, 19, 404–
417.
Preparation of RNA Duplexes. The formation of labeled RNA
duplexes with 8-aza-7-deazaA and 7-substituted-8-aza-7-deazaA
modifications was carried out as previously described with the slight
(72) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270–283.
(73) Takano, Y.; Houk, K. N. J. Chem. Theory Comput. 2005, 1, 70–77.
(74) Frisch, M. J.; et al. Gaussian03; Gaussian, Inc.
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11891