Tritium labeling of full-length small interfering RNAs

Article abstract of DOI:10.1002/jlcr.2919

A simple procedure is described for full-length single internal [ ³H]-radiolabeling of oligonucleotides. Previous labeling strategies have been applied to large molecular weight compounds such as proteins and oligonucleotides, for example, iodination and ¹¹¹In labeling via covalently bounded chelators. However, a procedure has not yet been reported for single internal radiolabeling of oligonucleotides that preserves the molecular structure (³H replacing a ¹H). In following our strategy, the radiolabel can be strategically placed within a stable and predetermined internal position of the siRNA. This placement was accomplished by placing a 5-bromouridine or 5-bromo-2'-O-methyluridine phosphoramidite building block into the middle of the antisense strand using standard phosphoramidite chemistry. The deprotected full-length antisense strand was tritium labeled by bromine/tritium exchange, catalyzed by palladium on charcoal in the predetermined 5-position of either uridine or 2'-O-methyluridine. Internal placement of the building block within the oligonucleotide sequence and label placement at 5-position decreases the likelihood of the label to be readily cleaved from the oligonucleotide in vivo, and loss of the label by spontaneous tritium/hydrogen exchange. The tritiated single-stranded and double-stranded RNAs were also shown to be radio and chemically stable for at least 6 months at -80 °C. This allows more than sufficient time to conduct pharmaceutical formulation and pharmacokinetic studies. Copyright

Full text of DOI:10.1002/jlcr.2919

Research Article
Received 27 January 2012,
Revised 20 March 2012,
Accepted 20 March 2012
Published online 11 May 2012 in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/jlcr.2919
Tritium labeling of full-length small
interfering RNAs
Jesper Christensen,a François Natt,b Jürg Hunziker,b Joel Krauser,a
a
a
*
Hendrik Andres and Piet Swart
A simple procedure is described for full-length single internal [3H]-radiolabeling of oligonucleotides. Previous labeling stra-
tegies have been applied to large molecular weight compounds such as proteins and oligonucleotides, for example, iodina-
tion and 111In labeling via covalently bounded chelators. However, a procedure has not yet been reported for single internal
1
radiolabeling of oligonucleotides that preserves the molecular structure (3H replacing a H). In following our strategy, the
radiolabel can be strategically placed within a stable and predetermined internal position of the siRNA. This placement
was accomplished by placing a 5-bromouridine or 5-bromo-2′-O-methyluridine phosphoramidite building block into the
middle of the antisense strand using standard phosphoramidite chemistry. The deprotected full-length antisense strand
was tritium labeled by bromine/tritium exchange, catalyzed by palladium on charcoal in the predetermined 5-position of
either uridine or 2′-O-methyluridine. Internal placement of the building block within the oligonucleotide sequence and label
placement at 5-position decreases the likelihood of the label to be readily cleaved from the oligonucleotide in vivo, and loss
of the label by spontaneous tritium/hydrogen exchange. The tritiated single-stranded and double-stranded RNAs were also
shown to be radio and chemically stable for at least 6 months at ꢀ80 ꢁC. This allows more than sufficient time to conduct
pharmaceutical formulation and pharmacokinetic studies.
Keywords: oligonucleotides; siRNA; tritiation; halogen/tritium exchange; final step radiolabeling
make it challenging to interpret pharmacokinetic and meta-
Introduction
bolic studies.
As current pharmaceutical drug research still largely focuses on
End-labeling has typically been accomplished by addition of a
[32P]-phosphate group to the 5′-end of the oligonucleotide,2–7 or
interactions of small molecules with enzymes and proteins,
nucleic acid-derived molecules involved with RNA interference
(RNAi) offers a unique opportunity as novel drug candidates.
These siRNA drugs would be designed to trigger the inhibition
by using 14C-labeled uridines in the 3′-end region of the mole-
cule.8 This is a valid approach, but the position of the label still
may have some drawbacks for in vivo pharmacokinetics and
absorption, distribution, metabolism and elimination studies.
or breakdown of messenger RNA (mRNA).
The discovery of RNAi by Fire and Mello1 has been considered
a revolution in biotechnology, and the inventors were awarded
the 2006 Nobel Prize. The RNAi pathway is initiated by double-
stranded RNA (dsRNA) that, when introduced into cells from exo-
genous sources, are termed small interfering RNAs (siRNAs). In
principle, RNAi can silence or slow down the production of any
protein. In recent years, this role in cell regulation, which uses a
more specific and targeted approach, has led to a significant
interest in siRNA potential by the pharmaceutical industry. This
area of research offers a new way to treat human diseases that
may not be targetable with traditional approaches.
Uniform radiolabeling of phosphorothioate oligonucleotides9,10
with 35S is another option. However, this labeling technique
is restricted to oligonucleotides with a phosphorothioate
backbone. Facile removal of the radiolabel by exonucleases,
short half-lives and radiation protection measures against
strong beta radiations (32P) are some drawbacks using these
isotopes.11
Tritium radiolabeling procedures of oligonucleotides have also
been reported in the literature, for example, internal labeling of
deprotected phosphorothioate oligodeoxynucleotides randomly
labeled at the unstable C8-position of the purines by exchange
For the design and development of siRNAs throughout noncli-
nical and clinical pharmaceutical development, radiolabeled
molecules offer unique and beneficial advantages. Although
several oligonucleotide radiolabeling methods have been pre-
viously reported, limitations exist with respect to the choice of
radioisotope, position of the label, and chemical modifications
needed for coupling of the label.
aNovartis Pharma AG, Novartis Institutes for Biomedical Research, Drug Metabo-
lism and Pharmacokinetics, Basel, Switzerland
bNovartis Pharma AG, Novartis Institutes for Biomedical Research, Biologics
Center, Basel, Switzerland
Radiolabeled oligonucleotides labeled at chemically unstable
positions or at the end of the oligonucleotide might pose
the risk of the label being cleaved from the oligonucleotide
in vivo. The limitations of these labeling methods could
*Correspondence to: Piet Swart, Novartis Pharma AG, Novartis Institutes for Bio-
medical Research, Drug Metabolism and Pharmacokinetics, Fabrikstrasse 14,
1.02, CH-4002 Basel, Switzerland.
E-mail: piet.swart@novartis.com
J. Label Compd. Radiopharm 2012, 55 189–196
Copyright © 2012 John Wiley & Sons, Ltd.

J. Christensen et al.
with tritiated water as described by Graham et al.9 This procedure
has been adapted to siRNAs12,13 but has limitations in biological
assays because of radiolytic degradation and formation of
tritiated water. Sands et al.14 described an enzymatic procedure
for single internal final step labeling of deprotected oligodeoxynu-
cleotides. [3H]Methylation, which occurs at the C5-position of the
Experimental section
Materials and methods
All commercial reagents were used without further purification unless
otherwise noted. DNA/RNA oligonucleotides were synthesized in-house
(Novartis; Basel Switzerland). Deuterations and tritiations were performed
first deoxycytidine in the GCGC recognition site of the enzyme using a tritium manifold system from RC Tritec (external partner
company). All air and moisture-sensitive reactions were carried out in
argon-flushed, two-neck flasks sealed with rubber septa, and the
reagents were introduced by syringe and dried over molecular sieves
3 Å prior to use. Progress of reactions was monitored by thin layer chro-
matography on precoated Merck (Whitehouse Station, NJ, USA) silica gel
plates (60 F-254). Visualization was accomplished by UV light (254 nm).
Flash chromatography separations were performed on Merck silica gel
60 (0.040–0.063 mm). Melting points were recorded using Büchi 535 digi-
tal melting point apparatus (New Castle, DE, USA). Centrifugations were
carried out using Thermo Scientific Heraeus Multifuge 3SR + centrifuge
(HhaI methylase). The procedure requires synthesis of a comple-
mentary sense strand to prevent tritiation of the sense strand.
The sense strand is preferably made with 5-methyldeoxycytidine.
Single internal labeling of the stable C5′-position of the sugar
of oligodeoxynucleotide has been described by Tan et al.15
This labeling occurred during oligonucleotide synthesis by an
oxidation/reduction strategy. Le Doan et al.11 have reported a
procedure for final step labeling of deprotected oligodeoxynu-
cleotides containing a terminal amino group linker reacting
with [2,3-3H]succinimidyl propionate, and an example with (Waltham, MA, USA) or Mistral 1000 centrifuge. HPLC analysis and purifi-
cations were performed on Agilent 1100 or 1200 Series systems (Santa
Clara, CA, USA). Online radioactivity detection was performed in series
to UV detection by adding 4 ꢂ LC flow of Flo-Scint A scintillation cocktail
to the solvent stream. This mixture was subsequently investigated online
with a Berthold LB 513 system using a 200 or 100-ml Z-cell. Offline radio-
activity detection was performed by liquid scintillation counting (LSC)
using Perkin Elmer Wallac 1414 Liquid Scintillation Counter or Packard
2200CA TRI-CARB Liquid Scintillation Analyzer (Waltham, MA, USA).
NMR spectra were recorded using 400, 500, 600-MHz Bruker instruments.
Coupling constants (J) are reported in Hz and chemical shifts (d) in parts
the amino linker attached to the C5-position of a terminal
thymidine was also described.
Radioactive isotopes such as 111In have also been used for
radiolabeling. However, a major drawback of 111In is the short
half-life of the isotope (2.8 days), and the need of a cDTPA chela-
tor covalently attached to the molecule of interest. The chelator
as well as the bulky 111In group could significantly affect the
conformation or change the molecular properties of an oligonu-
cleotide drug. As a consequence, this may alter or confound its
pharmacodynamic and pharmacokinetic behavior. The resulting per million (ppm) relative to the residual proton signal of the deuterated
solvent used (CDCl3, DMSO-d6 or D2O, 7.26, 2.50 (13C 39.52) and 4.79,
data may not be well understood, straight forward, representa-
tive or even accurate. Other labeling techniques such as fluoro-
respectively). 3H-NMR spectra were calibrated indirectly via the deuter-
ated solvent in the 1H-NMR spectra. The following abbreviations are
phores or direct coupling of the radio isotope 125I have been
used to denote multiplicities: s, singlet; d, doublet; t, triplet; m, multi-
used for labeling siRNAs. However, these labeling approaches
plet; br, broad. MS (mass spectrometry) and high-resolution MS (HRMS)
may influence the pharmacokinetics and pharmacodynamics,
of monomers were performed by NIBR Analytical Sciences (Novartis) or
as highlighted already for modified proteins (Swart et al.16).
Linear oligonucleotides are prepared by solid phase synthesis in
recorded using Waters Micromass ZQ2000 or Thermo Finnigan LCQ
Advantage MAX spectrometers. MS and HRMS of oligonucleotides
a stepwise fashion, which suggests that the site-specific introduc-
tion of a radiolabel should be possible using correspondingly
labeled monomers. To this end, we (J. H., F. N., H. A., unpublished
results) originally explored the synthesis of a 5′-tritium labeled uridine
phosphoramidite building block (5′-O-[(4,4′-dimethoxyphenyl)
were recorded using Waters LCT Premier XE or Thermo LTQ-Orbitrap
XL mass spectrometers.
Synthesis of 5-bromo-2′O-methyluridine phosphoramidite
methyl]-2′-O-tert-butyldimethylsilyl-uridine-3′-O-[2-cyanoethyl Protected Nucleosides (1–3)
N,N-(diisopropylamino)phosphor amidite). Although this com-
pound could be used to incorporate a tritium label at low speci- 5-bromo-3′,5′-O-benzoyl-2′-O-methyluridine (1): To a suspension of 1,3,5-
tri-O-benzoyl-2-O-methyl-a/b-D-ribose (10.00 g, 21.0 mmol, 1.0 equiv)
and 5-bromouracil (4.01 g, 21.0 mmol, 1.0 equiv) in abs. acetonitrile
(80 ml) was added bis(trimethylsilyl)acetamide (10.41 ml, 8.54 g, 42.0 mmol,
2.0 equiv) under argon. The reaction mixture was heated to reflux upon
which a clear solution was formed. After 30 min, trifluoromethanesulfonic
acid trimethylsilylester (5.71 ml, 7.00 g, 31.5 mmol, 1.5 equiv) was added
and heating was continued for 2 h. The reaction mixture was allowed to cool
to room temperature (RT) and then diluted with chloroform (300 ml). The
solution was washed first with water (150 ml) and then with saturated
sodium bicarbonate (2 ꢂ 150 ml). The aqueous solutions were re-extracted
fic activity, any attempt to increase the tritium content in the
building block led to failure of the oligonucleotide coupling
reaction. The very same synthesis scheme, on the other hand,
could be used for the successful incorporation of deuterium
labels. We therefore presume a rapid, radioactivity-mediated
decay of the phosphorous (III) building block at high tritium
concentration.
Thus, a novel full-length labeling procedure was developed for
a single internal tritium labeling to avoid post labeling chemical
and conformational changes of the siRNA. This procedure with chloroform (150 ml). The combined organic layers were dried over
sodium sulfate, filtered and concentrated in vacuo. The a/b diastereomeric
ratio in the crude mixture was 1:1.8, the desired b-isomer was successfully
isolated by washing the impurities (including the a-diastereomer) away with
the following procedure: addition of heptane/EtOAc 1:1 (25 ml), vortex, cen-
trifugation (2000 rpm, 3 min, 4 ꢁC); removal of the supernatant, dissolving the
precipitate in EtOAc; and removal of the solvent in vacuo yielded 6.07 g of 1
(53%) as a white solid. MF: C24H21BrN2O8; MW: 545.3361; M.p. 180–181 ꢁC;
1H-NMR (400 MHz, CDCl3) d (ppm): 8.36 (br s, 1 H), 8.07 (dd, J= 12.6/8.3 Hz,
4 H), 7.80 (s, 1 H), 7.65–7.57 (m, 2H), 7.48 (t, J = 7.6Hz, 4H), 6.04
(d, J = 3.8Hz, 1 H ), 5.43 (t, J= 5.7Hz, 1H), 4.81–4.64 (m, 3 H), 4.19
(t, J= 4.4Hz, 1H), 3.48 (s, 3 H); Confirmation of the b-isomer by ROESY
depends upon the existence of a uridine or a 2′-O-methyluridine
residue in the sequence. Furthermore, the exact position of the
radio label is known. This oligonucleotide radiolabeling proce-
dure presents some advantages compared with previously
reported methods. For instance, there is no chemical or molecu-
lar modification (3H replacing a 1H), that is, minimal or no change
in the pharmacological activity, kinetics and metabolic behavior
of the molecule. In addition, an important prerequisite was
successfully demonstrated for the radiochemical stability of the
tritiated product.
www.jlcr.org
Copyright © 2012 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2012, 55 189–196

J. Christensen et al.
(500MHz, CDCl3): H1′ ↔ H4′, H1′ ↔ H2′-OMe, H2′ ↔ H6, H3′ ↔ H6′; 13C-NMR calculated for C40H48BrN4NaO9P [M + Na]+: m/z 861.2240; observed
(151 MHz, DMSO-d6) d (ppm): 165.52–164.87, 159.04, 149.69, 140.68,
133.70–133.50, 129.30–128.78, 96.77, 89.04, 80.02, 79.00, 71.03, 63.72, 58.38;
HRMS (positive ESI) calculated for C24H22BrN2O8 [M + H]+: m/z 545.0559;
observed 545.0557 (0.37 ppm).
861.2239 (0.12 ppm).
Labeling synthesis
[3H]-Radiolabeling of monomer 2 by bromine/tritium exchange
5-bromo-2′-O-methyluridine (2): 1 (5.06 g, 9.279 mmol, 1.0 equiv) was
dissolved in methanol (70 ml) and stirred under argon. 7 N NH3 in metha-
nol (22.5 ml) was added and the mixture was stirred at RT under argon for
4 days. Removal of the solvent in vacuo was performed in a cautious
manner. Flash chromatography (ethyl acetate/methanol) of the crude
product yielded 2.79 g of 2 (89%) as a white solid. MF: C10H13BrN2O6;
MW: 337.1240; M.p. 232–234 ꢁC, literature17 m.p. 235–237 ꢁC; 1H-NMR
(400 MHz, DMSO-d6) d (ppm): 11.83 (br s, 1 H), 8.53 (s, 1 H), 5.80
(d, J = 3.8 Hz, 1 H), 5.32 (t, J = 4.6 Hz, 1 H), 5.13 (d, J = 6.6 Hz, 1 H), 4.12
(q, J = 5.9Hz, 1 H), 3.89–3.83 (m, 1 H), 3.81 (t, J = 4.3Hz, 1 H), 3.75–3.54
(m, 2 H), 3.39 (s, 3 H), literature17 1H-NMR (DMSO-d6) d (ppm): 11.78 (s, 1 H),
8.50 (s, 1 H), 5.81 (d, J= 4 Hz, 1 H), 5.00–4.84 (m, 2 H), 3.41 (s, 3 H); 13C-NMR
(151 MHz, DMSO-d6) d (ppm): 159.19, 149.78, 140.13, 95.78, 86.76, 84.82,
83.08, 67.79, 59.67, 57.65; HRMS (negative ESI) calculated for C10H12BrN2O6
[MꢀH]ꢀ: m/z 334.9884; observed 334.9887 (0.90 ppm).
[5-3H]2′-O-methyluridine (5): 5-bromo-2′-O-methyluridine 2 (3.43 mg,
0.010 mmol, 1.0 equiv) was dissolved dimethylformamide (DMF) (0.5ml),
the catalyst (10% Pd/C, 2.67 mg, 0.25 equiv) was added, then N,N-
diisopropylethylamine (2 ml, 1.1 equiv). The reaction glass flask was
connected to a tritium gas line. The reaction flask was then frozen in
liquid nitrogen, evacuated, and filled with gaseous tritium (99 atom%
[3H]-H2, up to a pressure of 450 mbar at thawed state). After defrosting,
the Br/3H-exchange reaction was performed with vigorous stirring of
the reaction mixture at RT. The reaction was quenched after 2 h. The
reaction solvents were lyophilized, and labile tritium was removed by
lyophilizing twice with methanol (2 ꢂ ~1 ml). The dry crude product
was redissolved in 0.22 mm filtered Millipore H2O (~2 ml), filtered
(0.2 mm, 10 mm Whatman Anotop), and diluted with ethanol (25 ml,
8.64 GBq) and stored at ꢀ80 ꢁC until further use. Rotary evaporation,
then flash chromatography (ethyl acetate/ethanol) of the crude pro-
duct yielded 5, redissolved in ethanol (25 ml), and stored at ꢀ80 ꢁC
until further use (2.34 mg, 7.25 GBq, 88% chemical yield, 77% radioche-
mical yield). Yield was calculated via obtained radioactivity and 3H-
incorporation. MF: C10H14N2O6; MW: 258.0852; 1H-NMR (400 MHz,
DMSO-d6) d (ppm): 11.29 (br s, 1H), 7.91 (d, J = 8.5 Hz, 1H), 5.85
(d, J = 5.1 Hz, 1H), 5.63 (d, J = 8.0 Hz, 0.13% 1H ! 87% 3H-incorporated
(negative proof)), 5.17–5.05 (m, 2H), 4.11 (q, J = 5.1 Hz, 1H), 3.87–3.82
(m, 1H), 3.78 (t, J = 5.0 Hz, 1H), 3.67–3.52 (m, 2H), 3.35 (s, 3H); 3H-NMR
(400 MHz, DMSO-d6) d (ppm): 5.65 (d, J = 8.3 Hz). HRMS (positive ESI)
calculated for C10H15N2O6 [M + H]+: m/z 259.0925; observed 259.0929
(1.54 ppm). Specific radioactivity was determined by both MS and 1H-
NMR: 86.4% and 87% 3H-incorporation, respectively ! 919 MBq/mmol,
3.53 GBq/mg.
5-bromo-5′-[(4,4′-dimethoxyphenyl)methyl]-2′-O-methyluridine (3): 2 (2.79 g,
8.276 mmol, 1.0 equiv) was dissolved in pyridine (40 ml) under argon.
4,4′-dimethoxytrityl chloride (2.42 g, 7.143 mmol, 1.2 equiv) dissolved in
pyridine (30 ml) was added portion wise during 20 min. The mixture was
stirred at RT under argon for 24 h. Removal of the solvent in vacuo yielded
a yellow oil, which was dissolved in chloroform (200 ml). The solution was
first washed with saturated sodium bicarbonate (2 ꢂ 200 ml) and then
washed with brine (200 ml). The aqueous solutions were re-extracted with
chloroform (200 ml). The combined organic layers were dried over sodium
sulfate, filtered, and concentrated in vacuo by azeotropic removal of
pyridine with toluene (3 ꢂ 20 ml). Flash chromatography (heptane/ethyl
acetate + 0.1% triethylamine) of the crude product yielded 3.16 g of 3
(60%) as a yellow solid. MF: C31H31BrN2O8; MW: 639.4904; M.p. 106–107 ꢁC;
1H-NMR (400 MHz, CDCl3) d (ppm): 8.61 (br s, 1 H), 8.12 (s, 1 H), 7.43 (d,
J = 7.1 Hz, 2 H), 7.37–7.27 (m, 6 H), 7.25–7.20 (m, 1 H), 6.85 (d, J = 8.1 Hz,
4 H), 5.94 (d, J = 2.8 Hz, 1 H), 4.50–4.44 (m, 1 H), 4.09–4.05 (m, 1 H), 3.95–
3.91 (m, 1 H), 3.79 (s, 6 H), 3.64 (s, 3 H), 3.48 (d, J = 2.5 Hz, 2 H), 2.63
(d, J = 7.8 Hz, 1 H); 13C-NMR (151 MHz, DMSO-d6) d (ppm): 159.09,
158.14, 149.68, 144.69, 139.51, 135.45–135.25, 129.72, 127.93–127.65,
126.72, 113.29, 96.28, 87.57, 82.94, 82.29, 68.53, 62.91, 57.84, 55.04;
Labeling of oligonucleotides by bromine/deuterium exchange
Prior to the tritiation experiments, the reaction conditions were opti-
mized by means of deuterium labeling. Procedures as described for
optimized tritium labeling have the following exceptions: connected
to a deuterium gas line, reactions were quenched after maximum 8 h,
the reaction solvents were lyophilized, the dry crude product was redis-
solved in 0.22-mm filtered Millipore H2O, and filtered (0.2 mm, 25 mm
Whatman Anotop).
HRMS (negative ESI) calculated for
637.1191; observed 637.1196 (0.78 ppm).
C
31H30BrN2O8 [MꢀH]ꢀ: m/z
Phosphoramidite (4)
5-bromo-5′-[(4,4′-dimethoxyphenyl)methyl]-2′-O-methyluridine-3′-O-[2-
cyanoethyl N,N-(diisopropylamino)phosphoramidite] (4): 3(2.91 g, 4.551 mmol,
1.0 equiv) was dissolved in abs. dichloromethane (80 ml) under argon,
and diisopropylammonium-tetrazolide (1.17 g, 6.826 mmol, 1.5 equiv)
was added. Cyanoethyltetraisopropylphosphordiamidite (2.29 ml, 1.92 g,
6.371 mmol, 1.4 equiv) was added and the mixture was stirred at RT
under argon for 7 h. Chloroform (100 ml) was added to the mixture
and the solution was first washed with saturated sodium bicarbonate
(2 ꢂ 50 ml), and then with brine (50 ml). The aqueous solutions were
re-extracted with chloroform (50 ml). The combined organic layers are
dried over sodium sulfate, filtered, and concentrated in vacuo. Flash
chromatography (heptane/ethyl acetate + 0.1% triethylamine) of the
crude product yielded 3.51 g of 4 (92%) as a yellow solid (phosphor-
diastereomers). MF: C40H48BrN4O9P; MW: 839.7083; M.p. 85–86 ꢁC;
1H-NMR (400 MHz, CDCl3) d (ppm): 8.15 (br s, 1 H), 8.08 (s, 1 H), 7.43
(dd, J = 11.2/8.0 Hz, 2 H, Hc), 7.38–7.27 (m, 6 H), 7.26–7.20 (m, 1 H),
6.88–6.81 (m, 4 H), 6.01 and 5.94 (d, J = 4.8 & 3.3 Hz, 1 H), 4.62–4.46
(m, 1 H), 4.30–4.19 (m, 1 H), 4.08–4.03 (m, 1 H), 3.96–3.58 (m, 4 H), 3.80
(s, 6 H), 3–57/3.55 (3 H), 3.51–3.34 (m, 2 H), 2.65 (t, J = 6.2 Hz, 1 H), 2.41
(t, J = 5.7 Hz, 1 H), 1.23–1.02 (m, 12 H); 31P-NMR (400 MHz, DMSO-d6)
d (ppm): 149.6 and 149.3, no internal reference; HRMS (positive ESI)
Characterization of [2H]ssRNA-1
MF: C197H251N69O145P20 (unlabeled acid); EM: 6521.9140; MS (ESI-TOFMS
negative) calculated for C197H248N69O145P20 [MꢀH]3ꢀ
: m/z 2173.0;
observed 2173.0. Brominated starting material was not observed in the
MS spectrum of [2H]ssRNA-1.
Characterization of [2H]ssRNA-2
MF: C202H252N68O151P20 (unlabeled acid); EM: 6664.8883; MS (ESI-TOFMS
negative) calculated for C202H249N68O151P20 [MꢀH]3ꢀ
: m/z 2220.6;
observed 2220.7. Brominated starting material was not observed in the
MS spectrum of [2H]ssRNA-2.
Characterization of [2H]ssRNA-3
MF: C206H259N75O146P20 (unlabeled acid); EM: 6737.9900; MS (ESI-TOFMS
negative) calculated for C206H256N75O146P20 [MꢀH]3ꢀ
: m/z 2245.0;
observed 2245.1. Brominated starting material was not observed in the
MS spectrum of [2H]ssRNA-3.
J. Label Compd. Radiopharm 2012, 55 189–196
Copyright © 2012 John Wiley & Sons, Ltd.
www.jlcr.org

J. Christensen et al.
for 0′! 2.33% B/min ! 40% B after 20′! 400% B/min ! 80% B after
0.1′! 80% B for 2′! 750% B/min ! 5% B after 0.1′! 5% B for 2.8′; LC flow:
0.5 ml/min; l = 254 nm; T = 30 ꢁC; injection volume: 5 ml; RA-flow: 2.0 ml/min;
radiochemical purity = 73.3%; chemical purity = 83.6%, and 4.0% of [3H]
ssRNA-1 in, [3H]siRNA-1). Specific radioactivity (determined previously for
[3H]MRP4 AS) ! 70 MBq/mmol, 5.25 MBq/mg.
Labeling of oligonucleotides by bromine/tritium exchange
Optimized labeling procedure for [3H]ssRNA-1
The brominated starting material of ssRNA-1 was dissolved in 0.22-mm fil-
tered Millipore H2O (0.8 mmol, 5.5 mg), the catalyst (10% Pd/C, 75 equiv,
66.5 mg) was added, then N,N-diisopropylethylamine (147 equiv, 21 ml)
and DMF was added to a final volume of 0.5 ml (DMF/H2O ratio 80:20).
The reaction glass flask was connected to a tritium gas line. The reaction
flask was then frozen in liquid nitrogen, evacuated and filled with
gaseous tritium (99 atom% [3H]-H2, up to a pressure of 170–270 mbar
at frozen state ! 280–520 mbar at thawed state, 6-ml dead volume).
After defrosting, the Br/3H-exchange reaction was performed with vigor-
ous stirring of the reaction mixture at RT. The reaction was quenched
after 6 h. The reaction solvents were lyophilized, and labile tritium was
removed by lyophilizing twice with methanol (2 ꢂ ~1 ml). The dry crude
product was redissolved in 0.22-mm filtered Millipore H2O (~2 ml), filtered
(0.2 mm, 25 mm Whatman Anotop) into a centrifugal filter device (Amicon
Ultra – Millipore, 3 kDa cut-off) and centrifuged (4000 g, 60 min at RT) to
the supernatant volume was <250 ml. The supernatant was pipette off
and diluted to 5 ml (~85 MBq), and stored at ꢀ80 ꢁC until further use.
[3H]-radiolabeling, purification and annealing of [3H]siRNA-2
Procedures as described for [3H]siRNA-1 have the following exceptions: labeling
reaction time of 4 h, only prepurification performed, and annealing performed
in phosphate buffered saline (PBS). [3H]siRNA-2 was then stored at ꢀ80 ꢁC until
further use.
[3H]ssRNA-2: 2.0 mg, 9.41 MBq, 38% chemical yield, 1% radiochemical
yield. Quantity was measured by UV-absorption spectroscopy (60.31
OD, Epsilon 203980 [OD/mmol]). MF: C202H252N68O151P20 (unlabeled
acid); EM: 6664.8883; 3H-NMR (400 MHz, D2O)
d (ppm): 5.73 (s)
(decoupled spectrum to enhance the signals); HRMS (negative ESI) calcu-
lated for C202H242N68O151P20 [MꢀH]10ꢀ: m/z 665.4810; observed 665.4816
(0.89 ppm). Brominated starting material was not observed in the MS
spectrum of [3H]ssRNA-2. Purity of [3H]ssRNA-2 was analyzed by analyti-
cal IP-RP-HPLC-UV-RA (as described for ssRNA 1, except used the solvents
and gradient described for prepurification; LC flow: 0.9 ml/min; RA-flow:
3.6 ml/min); radiochemical purity = 60.9%; chemical purity = 86.5% (UV,
l = 260nm). Specific radioactivity was determined by LSC/HPLC linear
regression method (calibration curve constituted from eight dilutions
of ssRNA-2 reference substance ! R2 = 0.9605) ! 26 MBq/mmol, 3.9 MBq/mg.
[3H]siRNA-2: radiochemical purity = 75.1%; chemical purity = 92.0%,
and 4.8% of the sense strand in, [3H]siRNA-2 (UV, l = 254 nm). Specific
radioactivity (determined previously for [3H]ssRNA-2) ! 26 MBq/mmol,
1.9 MBq/mg.
Ion pair RP-HPLC purification of [3H]ssRNA-1
The crude ssRNA-1 was prepurified by semi-preparative IP-RP-HPLC-UV
(Waters XBridge Prep, RP C-18, 5 mm, 10 ꢂ 250 mm; eluent A: 50 mM
triethylammonium acetate (pH 7.0), eluent B: 30% A in acetonitrile; gradient:
5% B for 4′ ! 0.94% B/min ! 20% B after 16′! 75% B/min ! 95% B
after 1′ ! 95% B for 4′ ! -900% B/min ! 5% B after 0.1′! 5% B for 4.9′;
flow: 3.2ml/min; l = 260 nm; T = 30 ꢁC; injection volume: 500ml). In fol-
lowing the purification, the purity and integrity of the isolated material
were tested against the pure nonradiolabeled standard using IP-RP-HPLC-
UV-RA (Waters XBridge, RP C-18, 3.5mm, 3.0 ꢂ 150 mm; eluent A: 16.4-mM
triethylamine, 95-mM hexafluoroisopropanol (pH 8.4), eluent B: 16.4-mM
triethylamine, 95-mM hexafluoroisopropanol in 40% methanol; gradient:
Results and discussion
The [3H]-radiolabeling method of oligonucleotides is based on
catalytic halogen/tritium exchange, which is an established means
to introduce tritium within a molecule. Either 5-brominated
uridine or 2′-O-methyluridine (2) was used for the halogenated
precursors. The chemical modification 2′-O-methyl of RNA is an
existing modification.18 This modification was shown to be well
tolerated throughout the duplex.19 With the use of two precursors,
that is, uridine as unmodified and chemically modified 2′-O-
methyl, additional flexibility has been provided for siRNA labeling.
The likelihood that the siRNA contains at least one uridine or 2′-O-
methyluridine is significant. The brominated precursors were
built into the oligonucleotides by conversion to their respective
phosphoramidites via standard phosphoramidite chemistry. The
bromine/tritium exchange catalyzed by palladium on charcoal
under tritium gas occurs on the fully deprotected full-length oligo-
nucleotide strand at the predetermined 5-position of either the
uridine or 2′-O-methyluridine (Figure 1).
5% B for 0′ ! 3.75% B/min ! 80% B after 20′! 150%
B/min ! 95% B
after 0.1′! 95% B for 2′ ! -900% B/min ! 5% B after 0.1′ ! 5% B for 2.8′;
LC flow: 0.5ml/min; l = 260 nm; T = 30 ꢁC; injection volume: 5ml, RA-flow:
2.0ml/min). As deemed necessary, an additional purification was performed
using semi-preparative IP-RP-HPLC-UV (same conditions as for analytical IP-
RP-HPLC-UC-RA except; Waters XBridge Prep, RP C-18, 5mm, 10 ꢂ 250 mm;
flow: 2.0ml/min; injection volume: 500ml), followed by an analytical verifica-
tion of the purity by IP-RP-HPLC-UV-RA (as described previously), and the
corresponding fractions were combined, rotary evaporated, lyophilized
(white solid) and redissolved in 0.22-mm filtered Millipore H2O (1ml) to yield
[3H]ssRNA-1 (1.0mg, 11.96 MBq, 18% chemical yield, 1% radiochemical
yield). Quantity was measured by UV-absorption spectroscopy (34.42 OD,
Epsilon 186460 [OD/mmol]) and stored at ꢀ80 ꢁC until further use. MF:
C197H251N69O145P20 (unlabeled acid); EM: 6521.9140; HRMS (negative
ESI) calculated for C197H247N69O145P20 [MꢀH]4ꢀ: m/z 1629.4707;
observed 1629.4630 (4.39 ppm). Brominated starting material was
not observed in the MS spectrum of [3H]ssRNA-1. Purity of [3H]
ssRNA-1 analyzed by analytical IP-RP-HPLC-UV-RA (same conditions
as described above; radiochemical purity=90.6%; chemical purity=93.1%
(UV, l = 260 nm)). Specific radioactivity determined by LSC/HPLC
linear regression method (calibration curve constituted from seven
dilutions of ssRNA-1 reference substance! R2 = 0.9998) ! 70 MBq/mmol,
10.8 MBq/mg.
Synthesis of the brominated 2′-O-methyluridine
phosphoramidite
5-bromouridine phosphoramidite were obtained from Glen
Research. 5-Bromo-2′-O-methyluridine 2 has been described in
literature.17 The 5-bromo-2′-O-methyluridine phosphoramidite 4
was synthesized via a four-step synthesis route (Figure 2).
The benzoyl-protected sugar and the 5-bromouracil base in
the presence of bis(trimethylsilyl)acetamide and trimethylsilyl triflate
resulted in the protected nucleoside 1 and its a-diastereoisomer
(ratio 1.8:1). Separation of the two diastereoisomers was not success-
ful by flash chromatography. However, the diastereoisomers
were successfully separated via solvent extraction with a heptane /
Annealing procedure of [3H]siRNA-1
[3H]ssRNA-1 was lyophilized and redissolved in 0.9% NaCl solution (0.5 ml),
resulting in a 325nmol/ml solution. [3H]ssRNA-1 and the complementary
sense strand were mixed equimolarly, and the mixture was heated and sha-
ken for 3 min at 90 ꢁC and then for 1h at 37ꢁC in an Eppendorf thermomixer
comfort (650 rpm). [3H]siRNA-1 was then stored at ꢀ80 ꢁC until further use.
Purity of [3H]siRNA-1 was analyzed by analytical IP-RP-HPLC-UV-RA (Waters
XBridge, RP C-18, 3.5 mm, 3.0 ꢂ 150 mm; eluent A: 0.1 M NaCl, 16.4 mM triethy-
lamine, 95 mM hexafluoroisopropanol (pH 8.4), eluent B: 0.1 M NaCl, 16.4 mM
triethylamine, 95 mM hexafluoroisopropanol in 40% methanol; gradient: 5% B ethyl acetate mixture (1:1), which resulted in a 53% yield of the
www.jlcr.org
Copyright © 2012 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2012, 55 189–196

J. Christensen et al.
Figure 1. [3H]-Radiolabeling of oligonucleotides (ssRNA-1) by bromine/tritium exchange.
Figure 2. Synthesis route of 5-brominated-2′-O-methyluridine phosphoramidite (4). (a) bis(trimethylsilyl)acetamide (2.0 equiv), Trifluoromethanesulfonic acid methylsily-
lester (1.5 equiv), CH3CN, argon, reflux, 2 h, 53%; (b) 7N NH3 in MeOH, MeOH, argon, RT, 4 days, 89%; (c) DMT-Cl (1.2 equiv), pyridine, argon, RT, 24 h, 60%; d: Diisopropy-
lammonium-tetrazolide (1.5 equiv), cyanoethyltetraisopropylphosphoramidite (1.4 equiv), CH2Cl2, argon, RT, 7 h, 92%.
b-diastereoisomer. The protected nucleoside 1 was deprotected base was added to sequester tritium bromide (3HBr). The catalyst
with ammonia to give the deprotected 5-bromo-2′-O-methyluridine did not hydrogenate the double bond, and the reaction went to
2 in an 89% yield after purification. The 5-bromo-2′-O-methyluridine completion; thus, there was no need to evaluate alternative
2 was converted into its phosphoramidite 4 in two steps. First, catalysts for the siRNA models. For DNA models, one approach
the 5′-hydroxy group was protected with a DMT-group to afford 5- using the catalyst without the carrier (Pd black) and another with
bromo-5′-[(4,4′-dimethoxyphenyl)methyl]-2′-O-methyluridine 3 in Pd/BaSO4 were successful. However, these approaches were not
60% yield after purification. Subsequent phosphorylation of 3 with as efficient as the 10% Pd/C. When turning from the monomer to
cyanoethyltetraisopropylphosphordiamidite in the presence of the DNA oligonucleotide models (data not shown), we decided
diisopropylammonium-tetrazolide gave the desired 5-brominated- to proceed with a DMF/H2O solvent system in the ratios 80:20.
2′-O-methyluridine phosphoramidite 4 in a 92% yield.
Pure DMF would have been preferred, but we anticipated a 21-mer
RNA oligonucleotide would not be soluble under those terms.
Tritiation of 5-Br-2′-O-methyluridine
Stability of [5-3H]2′-O-methyluridine
Tritiation of 5-Br-2′-O-methyluridine 2 with 10% palladium on
charcoal in DMF in the presence of Hünig base was completed The chemical stability of the carbon–tritium bond in the C5-position
within 2 h. The obtained [5-3H]2′-O-methyluridine 5 had a high of the tritiated monomer [5-3 H]2′-O-methyluridine 5 was followed
specific radioactivity of 919 GBq/mmol (3.53 GBq/mg). Hünig for 17 days in phosphate buffer (PBS, pH 7.4) at 37 ꢁC, that is,
J. Label Compd. Radiopharm 2012, 55 189–196
Copyright © 2012 John Wiley & Sons, Ltd.
www.jlcr.org

J. Christensen et al.
mimicking physiological conditions. Two different analytical meth- brominated monomers. The deuteration and tritiation experiments
ods (LSC and 1H- & 3H-NMR) gave consistent results of less than 5% for the full-length ssRNA were performed in the low mg scale of
tritiated water observed after 17days. In addition, the compound approximately 5–6 mg, and were performed using three ssRNAs.
was stable after 1 year in EtOH at ꢀ80 ꢁC and gave results of less
During optimization of the ssRNA-1 tritiation, the quantity of
the catalyst and the base to obtain full conversion of the starting
material was found to be of major importance. A large amount of
catalyst and base equivalents (75 and 150, respectively) were
needed for a full conversion of the ssRNA within 4 h. Other
parameters such as temperature, and gas pressure during the
tritiation had no or limited impact on the conversion of the
starting material.
However, a deviation of more than ꢃ5% from the 80:20 ratio
of the DMF/H2O solvent system had a negative effect on the
conversion. The explorative studies on the tritiation conditions
showed a full conversion on the bromide into tritium/hydrogen
on the final conditions within a 6-h reaction time. The final
than a 5% loss of the tritium label.
Synthesis and tritiation of ssRNA
Three siRNAs (Table 1) with different sequences and two nucleo-
tide 3′-end overhangs were synthesized using standard phos-
phoramidite chemistry. siRNA-1 is a 21-mer siRNA, which has
the rat MRP4 as target and is a chemically unmodified 21-mer
oligonucleotide (van de Water et al.20). The 5-Br-uridine was incor-
porated at position 9 of the antisense strand. siRNA-2 sequence is
also known as SSB, and directs against the ubiquitous gene
Sjögren Syndrome antigen B. The antisense contained 5-Br-2′-
OMeU in position 7. siRNA-3 is a dummy sequence with a relative tritiated ssRNA-1 had a specific radioactivity of 70MBq/mmol
high ‘u’ density. This sequence was selected to study the effect of (10.8 MBq/mg). For further stability and kinetic studies, the ssRNA
was annealed equimolarly with the complementary sense strand
to afford [3H]siRNA-1. Similar conditions were applied for the
tritiation for ssRNA-2, containing the 5-bromo-2′-O-methyluridine
precursor in position 7, instead of the 5-bromouridine, resulting
the oligonucleotide sequences on success of the labeling.
In following their synthesis and purification, the brominated ssRNAs
were analyzed by liquid chromatography–mass spectrometry. For
all three siRNAs, we could confirm the incorporation of the
Table 1. 21-mer RNA oligonucleotides studied
Name
Sequence (Antisense and Sense)
Precursor
5-Br-U
siRNA-1
siRNA-2
siRNA-3
AS
S
AS
S
AS
S
5′-ACAGCUCCU9GACACCUCUCdTdT-3′
5′-GAGAGGUGUCAGGAGCUGUdTdT-3′
5′-UuAcAUu7AAAGUCUGUUGUuu-3′
5′-AcAAcAGAcuuuAAuGuAAuu-3′
5′-UGAACcuAGUu11AAcuUAAGuu-3′
5′- cuuAAGuuAAcuAGGuucAuu-3′
5-Br-2′-OMeU
5-Br-2′-OMeU
dX = DNA, X = RNA, and x = 2′-OMe RNA.
Figure 3. 3H-NMR spectrum of ssRNA-2.
www.jlcr.org
Copyright © 2012 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2012, 55 189–196

J. Christensen et al.
in full conversion of the starting material within 4 h, with a specific specific radioactivity, because of a more efficient exchange versus
radioactivity of 26 MBq/mmol (3.9 MBq/mg). A decoupled 3H-NMR bromine. Furthermore, the chemical stability of the carbon–tritium
spectrum of ssRNA-2 has been obtained (Figure 3), and the tritium bond in the C6-position of pyrimidines is known in the litera-
distribution observed is localized on one shift only (5.73 ppm), ture21,22 to be more stable than the C5-position. As the method is
which correspond to the 5-position of 2′-O-methyluridine 5. This dependent upon the existence of a uridine or a 2′-O-methyluridine
also indicates that no random incorporation (scrambling) of residue in the sequence, this method should also be transferable to
tritium into the molecule has occurred. The chemical yields of chemically modified as well as unmodified phosphodiester oligo-
the [3H]-radiolabeling reactions of ssRNA-1 and ssRNA-2 were nucleotides. This work provides a good indication that the method
18% and 38%, respectively, even though in both cases all starting might be universal for chemically modified as well as unmodified
material was converted. This was mostly because considerable phosphodiester oligonucleotides.
amounts of material were lost during work-up and purification;
the reaction scale (5–6 mg) itself might also have had an impact.
Several batches of [3H]ssRNA-1 and [3H]ssRNA-2 have been con-
ducted, with consistent results, and thereby the reproducibility
of the labeling procedure has been confirmed.
Acknowledgements
By testing the developed labeling method with the aforemen-
tioned ssRNAs, the conversion of the starting material did not lent MS and NMR support. A special acknowledgement to
We sincerely thank NIBR Analytical Sciences for their excel-
Dr. Christian Guenat for obtaining HRMS spectrum of the
5-Br-2′OMeU-CEPA, and to Lukas Oberer and Thomas Lochmann
for their scientific discussions, guidance and support with NMR
experiments (particularly 3H-NMR). We would also like to gratefully
acknowledge Ms. Farah Aiouna for her excellent technical
assistance, Julie Mabille for the synthesis of the non labeled
siRNA’s, Dr. Olivier Kretz for his careful review and comments
on the manuscript, and Dr. Tapan Ray for his expertise, guidance
and support.
seem to be affected much by the sequence or the position of
the brominated precursor or the choice of precursor. A lower
incorporation of tritium into ssRNA-2, than ssRNA-1 was
observed (26 and 70 MBq/mmol, respectively); this might be
due to the sequence, position of precursor, and/or choice of
precursor. However, the observed differences do not appear
significant. Attempts were made to characterize the impurities
with HRMS but were unsuccessful.
Stability of tritiated ssRNAs and dsRNAs
The stability of the [3H]-radiolabeled single strands (ssRNA-1 and Conflict of Interest
ssRNA-2) and duplexes (siRNA-1 and siRNA-2) have been studied
The authors did not report any conflict of interest.
by ion pair RP-HPLC with radio detection at +4 and ꢀ80 ꢁC. The
single strands were stored in water or in 10% EtOH. After 8-month
storage at +4 ꢁC, less than 10% tritiated water was observed and
less than 5% at ꢀ80 ꢁC for both solutions. The duplexes were
stored in PBS. After 30-month storage at +4 ꢁC and ꢀ80 ꢁC, less
than 5% tritiated water was observed, allowing batch wise synth-
esis of tritiated siRNA for multiple pharmaceutical formulation and
pharmacokinetic studies.
References
[1] A. Fire, S. Q. Xu, M. K. Montgomery, S. A. Kostas, S. E. Driver, C. C. Mello,
Nature 1998, 391(6669), 806–811.
[2] A. Malek, F. Czubayk, A. Aigner, J. Drug Target. 2008, 16(2), 124–139.
[3] S. Gao, F. Dagnaes-Hansen, E. J. B. Nielsen, J. Wengel, F. Besenbacher,
K. A. Howard, J. Kjems, Mol. Ther. 2009, 17(7), 1225–1233.
[4] A. Khan, M. Benboubetra, P. Z. Sayyed, K. W. Ng, S. Fox, G. Beck, I. F. Benter,
S. Akhtar, J. Drug Target. 2004, 12(6), 393–404.
Conclusions
The described final step labeling procedure for single internal
[3H]-radiolabeling of oligonucleotides, which exclusively depends
upon the existence of a uridine or a 2′-O-methyluridine residue
in the sequence is a very promising approach. The starting point
for development of the labeling method was the monomer. The
method was then optimized for DNA models with minor adjust-
ments and successfully transferred to the RNA oligonucleotides.
This radiolabeling procedure presents some definite advan-
tages over previously reported methods of labeling oligonucleo-
tides. This labeling did not alter the chemical or molecular
characteristics (3H replacing a 1H), that is, the reactivity and meta-
[5] S. Akhtar, S. Basu, E. Wickstrom, R. L. Juliano, Nucl. Acids Res. 1991, 19(20),
5551–5559.
[6] K. Yanagihara, M. Tashiro, Y. Fukuda, H. Ohno, Y. Higashiyama,
Y. Miyazaki, Y. Hirakata, K. Tomono, Y. Mizuta, K. Tsukamoto,
S. Kohno, J. Antimicrob. Chemother. 2006, 57(1), 122–126.
[7] G. Beale, A. J. Hollins, M. Benboubetra, M. Sohail, S. P. Fox, I. Benter,
S. Akhtar, J. Drug Target. 2003, 11(7), 449–456.
[8] D. Kenworthy, R. Chambers, D. Fraser, N. Patel, S. Hood, Tissue Distri-
bution, Metabolism and Excretion of [14C]GSK2402968 (PRO051) a
phosphorothioated exon skipping oliogonucleotide in CD-1 and
mdx mice. DMDG Open Meeting 2011.
[9] M. J. Graham, S. M. Freier, R. M. Crooke, D. J. Ecker, R. N. Maslova,
E. A. Lesnik, Nucl. Acids Res. 1993, 21(16), 3737–3743.
bolism of the molecule should remain unaffected compared with [10] S. Agrawal, J. Temsamani, J. Y. Tang, Proc. Natl. Acad. Sci. U.S.A. 1991,
88(17), 7595–7599.
[11] T. Ledoan, R. Auger, A. Benjahad, J. P. Tenu, Nucleosides Nucleotides
the unlabeled compound. The radiolabel is placed in a stable and
predetermined internal position, which decreases the likelihood of
Nucleic Acids 1999, 18(2), 277–289.
the label to be readily cleaved from the oligonucleotide in vivo.
[12] O. R. Mook, F. Baas, M. B. de Wissel, K. Fluiter, Mol. Cancer Ther. 2007,
Storage of either the radiolabeled single strands or the duplexes
does not seem to be an issue. The relatively low specific radioac-
tivity obtained for both ssRNAs was most likely due to the partially
aqueous solvent system. To address this, a third building block
with iodine in the C6-position of 2′-O-methyluridine is currently
in progress. This building block is anticipated to increase the
6(3), 833–843.
[13] O. R. F. Mook, J. Vreijling, S. Wengel, J. Wengel, C. Zhou, J. Chattopadhyaya,
F. Baas, K. Fluiter, Artificial DNA: PNA & XNA 2010, 1(1), 36–44.
[14] H. Sands, L. J. Gorey-Feret, A. J. Cocuzza, F. W. Hobbs, D. Chidester,
G. L. Trainor, Mol. Pharmacol. 1994, 45(5), 932–943.
[15] W. T. Tan, R. P. Iyer, D. Yu, S. Agrawal, Tetrahedron Lett. 1995, 36(21),
3631–3634.
J. Label Compd. Radiopharm 2012, 55 189–196
Copyright © 2012 John Wiley & Sons, Ltd.
www.jlcr.org

J. Christensen et al.
[16] P. J. Swart, E. Beljaars, C. Smit, A. Pasma, H. Schuitemaker, D. K. F. Meijer, [19] J. K. Watts, G. F. Deleavey, M. J. Damha, Drug Discov. Today 2008,
J. Drug Target. 1996, 4(2), 109–116.
13(19–20), 842–855.
[17] M. J. Robins, S. R. Naik, Biochemistry 1971, 10(19), 3591–3597.
[18] H. Grosjean, Nucleic acids are not boring long polymers of only four
types of nucleotides: a guided tour. In DNA and RNA Modification
Enzymes: Structure, Mechanism, Function and Evolution, (Ed.: H. Grosjean),
Landes Bioscience: Austin, USA, 2009, pp. 1–18.
[20] F. M. van de Water, O. C. Boerman, A. C. Wouterse, J. G. P. Peters,
F. G. M. Russel, R. Masereeuw, Drug Metab. Dispos. 2006, 34(8),
1393–1397.
[21] S. Noll, E. Mittag, React Kinet Catal L 1983, 22(1–2), 35–37.
[22] E. A. Evans, H. C. Sheppard, J. C. Turner, J Labelled Compd 1970, 6(1), 76–87.
www.jlcr.org
Copyright © 2012 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2012, 55 189–196