Efficient preparation of organic substrate-RNA conjugates via in vitro transcription

Article abstract of DOI:10.1021/ja051179m

A concise synthetic way has been developed for the preparation of guanosine monophosphate derivatives carrying a decaethylene glycol spacer at their 5′-oxygen to which are attached a range of organic substrates. The four different compounds, prepared via a convergent synthetic strategy, carry a tethered benzylallyl ether residue (1a), an anthracene (1b), a benzyl carbamate residue (1c), or a primary amino group (1d), respectively. All four compounds have been successfully incorporated at the 5′-end of a 25-mer long RNA transcript via T7 RNA polymerase, and no inhibition of chain elongation could be observed. Under proper conditions, 1a and 1b can be incorporated up to 90-95% and 1c up to 68%. The amino-terminated initiator 1d is incorporated less efficiently although still up to 49%. These results show that the more hydrophobic the guanosine monophosphate derivative is, the higher is its enzymatic incorporation.

Full text of DOI:10.1021/ja051179m

Published on Web 06/03/2005
Efficient Preparation of Organic Substrate-RNA Conjugates
via in Vitro Transcription
Roberto Fiammengo, Kamil Mus´ılek, and Andres Ja¨schke*
Contribution from the Institute of Pharmacy and Molecular Biotechnology, UniVersity of
Heidelberg, Im Neuenheimer Feld 364, 69120 Heidelberg, Germany
Received February 24, 2005; E-mail: jaeschke@uni-hd.de
Abstract: A concise synthetic way has been developed for the preparation of guanosine monophosphate
derivatives carrying a decaethylene glycol spacer at their 5′-oxygen to which are attached a range of organic
substrates. The four different compounds, prepared via a convergent synthetic strategy, carry a tethered
benzylallyl ether residue (1a), an anthracene (1b), a benzyl carbamate residue (1c), or a primary amino
group (1d), respectively. All four compounds have been successfully incorporated at the 5′-end of a 25-
mer long RNA transcript via T7 RNA polymerase, and no inhibition of chain elongation could be observed.
Under proper conditions, 1a and 1b can be incorporated up to 90-95% and 1c up to 68%. The amino-
terminated initiator 1d is incorporated less efficiently although still up to 49%. These results show that the
more hydrophobic the guanosine monophosphate derivative is, the higher is its enzymatic incorporation.
Introduction
during in vitro selection, as they are either incompatible with
the enzymatic steps or not selective or reliable enough for
In vitro selection and in vitro evolution techniques1,2 have
been used to isolate a large number of RNA catalysts3-5 and
RNA aptamers from combinatorial libraries (pools).6-8 If new
RNA catalysts are to be discovered via in vitro selection, a
substrate must be covalently attached to every sequence of the
RNA pool. This is especially true when attempting the direct
selection of ribozymes that catalyze a reaction between small
organic substrates.9,10 In these cases, the biopolymeric part is
catalyzing the reaction between a covalently linked reactant and
a free reactant without being itself modified. The occurrence
of the desired reaction should nevertheless bring about a change
which allows preparative isolation of the active RNA sequences,
i.e. those that catalyze the reaction. For instance, the reaction
of the RNA-attached substrate with a second substrate could
allow tagging of the active sequences with biotin.10-12 The
biotinylated sequences can then be isolated from the rest of the
pool, enzymatically copied and amplified, and used as input
for the next round of the iterative in vitro selection process.
While chemical methods are known for the covalent attachment
of small organic moieties to RNA, most of them cannot be used
derivatizing complex mixtures such as combinatorial pools of
nucleic acids.
In vitro transcription from DNA templates by DNA-dependent
T7 RNA polymerase, on the other hand, is a simple and reliable
enzymatic method allowing the synthesis of virtually any RNA
sequence.13,14 Using this method guanosine monophosphate
derivatives can be incorporated selectively at the 5′-end of a
transcribed oligoribonucleotide.15-17 Transcription reactions in
the presence of guanosine derivatives carrying covalently
appended an organic moiety (the reaction substrate during the
selection process) result therefore in the site-specific covalent
attachment of this moiety to the RNA. Most importantly, the
conjugation occurs under very mild conditions, which are in
principle compatible with a wide variety of substrates. Although
very powerful, this method has so far found only limited
application,10,18-20 probably because of the laborious multistep
synthesis required for the preparation of these guanosine
monophosphate derivatives, often named initiator nucleotides.
We report here a more versatile and concise synthetic strategy
for the preparation of initiator nucleotides bearing the potential
substrate attached to the end of a decaethylene glycol spacer
(1a-d). The adopted synthetic scheme is highly convergent and
(1) Bittker, J. A.; Phillips, K. J.; Liu, D. R. Curr. Opin. Chem. Biol. 2002, 6,
367-374.
(2) Breaker, R. R. Chem. ReV. 1997, 97, 371-390.
(3) Wilson, D. S.; Szostak, J. W. Annu. ReV. Biochem. 1999, 68, 611-647.
(4) Pan, T. Curr. Opin. Chem. Biol. 1997, 1, 17-25.
(5) Williams, K. P.; Bartel, D. P. In Catalytic RNA; Eckstein, F., Lilley, D.
M. J., Eds.; Springer: Berlin, 1996; Vol. 10, pp 367-381.
(6) Murphy, M. B.; Fuller, S. T.; Richardson, P. M.; Doyle, S. A. Nucleic
Acids Res. 2003, 31, e110.
(13) Milligan, J. F.; Groebe, D. R.; Witherell, G. W.; Uhlenbeck, O. C. Nucleic
Acids Res. 1987, 15, 8783-8798.
(14) Milligan, J. F.; Uhlenbeck, O. C. Methods Enzymol. 1989, 180, 51-62.
(15) Martin, C. T.; Muller, D. K.; Coleman, J. E. Biochemistry 1988, 27, 3966-
3974.
(7) Famulok, M.; Mayer, G.; Blind, M. Acc. Chem. Res. 2000, 33, 591-599.
(8) Famulok, M. Curr. Opin. Struct. Biol. 1999, 9, 324-329.
(9) Keiper, S.; Bebenroth, D.; Seelig, B.; Westhof, E.; Ja¨schke, A. Chem. Biol.
2004, 11, 1217-1227.
(16) Contreras, R.; Cheroutre, H.; Degrave, W.; Fiers, W. Nucleic Acids Res.
1982, 10, 6353-6362.
(17) Konarska, M. M.; Padgett, R. A.; Sharp, P. A. Cell 1984, 38, 731-736.
(18) Eisenfuhr, A.; Arora, P. S.; Sengle, G.; Takaoka, L. R.; Nowick, J. S.;
Famulok, M. Bioorg. Med. Chem. 2003, 11, 235-249.
(19) Zhang, L.; Sun, L.; Cui, Z.; Gottlieb, R. L.; Zhang, B. Bioconjugate Chem.
2001, 12, 939-948.
(10) Sengle, G.; Eisenfuhr, A.; Arora, P. S.; Nowick, J. S.; Famulok, M. Chem.
Biol. 2001, 8, 459-473.
(11) Jadhav, V. R.; Yarus, M. Biochemistry 2002, 41, 723-729.
(12) Seelig, B.; Ja¨schke, A. Chem. Biol. 1999, 6, 167-176.
(20) Seelig, B.; Ja¨schke, A. Bioconjugate Chem. 1999, 10, 371-378.
9
10.1021/ja051179m CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 9271-9276
9271

A R T I C L E S
Fiammengo et al.
Scheme 1. Guanosine Monophosphate Derivatives Used as Initiator Nucleotides in T7 RNAP Transcriptions (N, Nontemplated Nucleotide
Added at the 3′-end of the Transcripts)
Scheme 2. Synthesis of Monoprotected Decaethylenethylene
relies on two common intermediates, viz., monoprotected
decaethylene glycol 2 and base-unprotected guanosine 5′-
phosphoramidite 3. Furthermore screening of the transcription
reaction conditions has allowed an unprecedented efficient
incorporation of all synthesized compounds at the 5′-end of a
25-mer RNA transcript. Thus up to 90-95% of the transcribed
RNA molecules carry a benzylallyl ether moiety or an an-
thracene residue when the transcription mixtures contain 1a and
1b, respectively. The incorporation of initiator nucleotides 1c
and 1d are up to 68 and 49%, suggesting that more hydrophobic
substrates are incorporated more efficiently, all the other
parameters remaining unchanged (spacer nature and length,
transcription conditions).
Glycol 2a
a Reagents and conditions: (a) DMTCl (0.23 equiv), Et3N (0.37 equiv),
DMAP (0.01 equiv), CH2Cl2, rt, 16 h, 79%; (b) TsCl (1.1 equiv), Et3N
(3.0 equiv), DMAP (0.05 equiv), CH2Cl2, rt, 12 h, 99%; (c) hexaethylene
glycol (3.0 equiv), NaH (2.0 equiv), DMF, rt, 13 h, 73%. DMT ) 4,4′-
dimethoxytriphenylmethyl, DMAP ) 4-(dimethylamino)pyridine.
Results and Discussion
Recent developments in ribozyme catalysis require chemoen-
zymatic access to complex RNA conjugates. The incorporation
of guanosine monophosphate derivatives at the 5′-end of RNA
transcripts has been used as a valuable tool for attaching
potential substrates to each individual sequence of a RNA pool
during in vitro selection of ribozymes.12,18 It is therefore
interesting to understand how versatile this method could be
and under which conditions best results concerning conjugation
could be achieved. To this end, four guanosine monophosphate
derivatives 1a-d have been prepared displaying different
functional groups at the end of a decaethylene glycol spacer
(Scheme 1). These functional groups are benzylallyl ether,
anthracene, benzyl carbamate, and a primary amino group for
1a, 1b, 1c, and 1d, respectively.
The novel synthetic strategy here reported allows the prepara-
tion of these four compounds using two common intermediates,
namely, mono-DMT-decaethylene glycol 2 and 2′,3′-bisTBDMS
protected guanosine 5′-O-phosphoramidite 3. The potential
substrates, corresponding to the R groups of compounds 1a-
1c, are linked to the terminal hydroxyl group of the ethylene
glycol spacer via alkylation reactions, while 1d was obtained
by catalytic hydrogenolysis of the Cbz protected precursor 1c.
Synthesis of Common Intermediates. The synthesis of 2
started from inexpensive and commercially available tetraeth-
ylene glycol which was mono-DMT protected in 79% yield.
Slow addition over 16 h of dimethoxytrityl chloride (0.23 equiv)
to the glycol (4) gave substantially higher yields of mono- DMT-
tetraethylene glycol 5 than the 63% so far reported.21 Tosylation
(giving 6) followed by monoalkylation of commercially avail-
able hexaethylene glycol afforded 2 in 73% yield under
optimized conditions. Multigram amounts of this internediate
are therefore readily available involving only two chromato-
graphic steps for the purification of 5 and 2, respectively
(Scheme 2). This recursive synthesis of oligoethylene glycols
should allow the simple preparation of monodisperse tethers of
virtually any length just by iterating the tosylation and alkylation
step.
2′,3′-bisTBDMS-protected guanosine 722,23 is readily available
from unprotected guanosine and was easily converted to the
corresponding 5′-O-phosphoramidite 3 in 98% yield using
2-cyanoethyl-N,N-diisopropylchlorophosphoramidite as phos-
phitylating agent. Notably no protecting groups have been used
on the nucleobase in this reaction (Scheme 3).
Preparation of Substrate-Carrying Decaethylene Glycol
Derivatives. One of the key synthetic steps allowing the
preparation of initiator nucleotides with different potential
substrates appended at the end of a long tether is the alkylation
(21) Rumney, S. I. V.; Kool, E. T. J. Am. Chem. Soc. 1995, 117, 5635-5646.
(22) Kotch, F. W.; Sidorov, V.; Lam, Y.-F.; Kayser, K. J.; Li, H.; Kaucher, M.
S.; Davis, J. T. J. Am. Chem. Soc. 2003, 125, 15140-15150.
(23) Zhu, X.-F.; Williams, H. J.; Scott, A. I. J. Chem. Soc., Perkin Trans. 1
2000, 2305-2306.
9
9272 J. AM. CHEM. SOC. VOL. 127, NO. 25, 2005

Efficient Preparation of Organic Substrate
−RNA
A R T I C L E S
Scheme 3. Synthesis of Guanosine Phosphoramidite 3a
Scheme 4. Synthesis of Substrate-Decaethylene Glycol
Derivativesa
a Reagents and conditions: (a) 2-cyanoethyl-N,N-diisopropylchlorophos-
phoramidite (1.05 equiv), DIPEA (3.2 equiv), CH2Cl2, 0 °Cfrt, 1 h, 98%.
DIPEA ) ethyldiisopropylamine. Compound 3 is a mixture of two
diastereoisomers (ca. 1:1).
of 2 with the desired substrate in the form of an alkylating agent.
Compounds 11 and 14 were therefore prepared in a few-step
synthesis starting from 4-chloromethyl benzyl alcohol and 1,4-
diaminobutane (Scheme 4).
More in detail, oxidation of 4-chloromethylbenzyl alcohol 8
to the corresponding aldehyde 924 followed by reaction with
vinylmagnesium bromide gave benzyl allyl alcohol 10. The
hydroxyl group was then protected with TBDMSCl affording
11. 1,4-Diaminobutane 12 was monoprotected with N-(benzyl-
oxycarbonyloxy)succinimide in 44% yield and successively
reacted with bromoacetyl bromide under Schotten-Baumann
conditions affording 14 in 75% yield.
Alkylation of 2 with 9-chloromethylanthracene, 11, and 14
in acetonitrile in the presence of NaH gave the desired adduct
15, 17, and 21, in 73, 90, and 62% yield, respectively (Scheme
4). NaI was used as catalyst only when the leaving group was
a chloride. Further elaboration on the protecting groups was
then required in order to obtain the hydroxyl-terminated
derivatives to be used in the phoshoramidite coupling reaction.
Therefore the TBDMS group of 17 was removed and replaced
by a benzyl to give 19 before removal of the DMT group
affording the desired alcohol 20. Removal of the DMT group
from 15 and 21 afforded directly alcohols 16 and 22.
Synthesis of Guanosine Monophosphate Derivatives. Ac-
cording to several literature reports even relatively long oligo-
nucleotides (e.g. 60-mers) were synthesized via the phosphor-
amidite method without nucleoside base protection.25,26 Moreover,
it has been observed that guanosine does not undergo N-
phosphitylation during phosphoramidite coupling in the presence
of activators such as imidazolium triflate, benzimidazolium
triflate, 1-H-tetrazole, or 5-(p-nitrophenyl)-1H-tetrazole.25 To
keep the synthetic effort to a minimum, we decided therefore
to perform our coupling reactions also in the absence of
additional protecting groups on the nucleoside base (guanine).
Guanosine monophosphate derivatives were obtained by cou-
pling 3 with hydroxyl-terminated decaethylene glycol building
blocks (16, 20, and 22) carrying the substrate at the other end
of the tether in the presence of activators such as 4,5-
dicyanoimidazole, 5-benzylthiotetrazole, or imidazolium triflate
(Schemes 5 and 6).
a Reagents and conditions: (a) PCC (1.2 equiv), CH2Cl2, rt, 2 h, 94%;
(b) vinylmagnesium bromide (1.05 equiv), THF, -15 °C, 30 min, 83%;
(c) TBDMSCl (1.3 equiv), imidazole (3.3 equiv), DMF, rt, 1 h, 99%; (d)
Z-OSu (0.25 equiv), DMAP (0.005 equiv), THF, rt, 23 h, 44%; (e)
bromoacetylbromide (1.5 equiv), aq Na2CO3/CH2Cl2 1:1, 0 °C, 1 h, 75%;
(f) 2 (1.0 equiv), 9-chloromethylanthracene (1.1 equiv), NaH (1.6 equiv),
NaI (1.2 equiv), ACN, 40 °C, 0.5 h, 73%; (g) 3% TCA in DCE (excess),
rt, 5 min, 88%; (h) 2 (1.0 equiv), 11 (1.5 equiv), NaH (1.6 equiv), NaI (1.1
equiv), ACN, 40 °C, 1.5 h, 90%; (i) TBAF (2.0 equiv), THF, rt, 1 h, 90%;
(j) BnBr (3.0 equiv), NaH (1.5 equiv), ACN, rt, 1h, 91%; (k) 3% TCA in
DCE (excess), rt, 5 min, 89%; (l) 2 (1.0 equiv), 14 (1.5 equiv), NaH (1.6
equiv), ACN, 40 °C, 1.5 h, 62%; (m) 3% TCA in DCE (excess), rt, 5 min,
93%. PCC ) pyridinium chlorochromate, Z-OSu ) N-(benzyloxycarbo-
nyloxy)succinimide, ACN ) acetonitrile, TBAF ) tetra-n-butylammonium
fluoride, Bn ) benzyl, TCA ) trichloroacetic acid, DCE ) 1,2-dichloro-
ethane. Compound 20 and its precursors are racemic.
mixture of diastereoisomers due to the stereogenic phosphor
atom. However, the phosphodiester group in the final products
1 (after removal of the 2-cyanoethyl protecting group) is not
stereogenic. It proved therefore more convenient to directly
treat the phosphotriester intermediates with TBAF to simulta-
neously remove all protecting groups (TBDMS and 2-cyano-
ethyl). The crude products were then purified by RP-chroma-
tography giving 1a, 1d, and 1c in 42, 31, and 35% yield,
respectively.
The coupling reactions were quenched by addition of tert-
butylhydroperoxide, giving a phosphotriester intermediate as a
(24) Pierce, M. E.; Harris, G. D.; Islam, Q.; Radesca, L. A.; Storace, L.;
Waltermire, R. E.; Wat, E.; Jadhav, P. K.; Emmett, G. C. J. Org. Chem.
1996, 61, 444-450.
(25) Hayakawa, Y.; Kataoka, M. J. Am. Chem. Soc. 1998, 120, 12395-12401.
(26) Ohkubo, A.; Ezawa, Y.; Seio, K.; Sekine, M. J. Am. Chem. Soc. 2004,
126, 10884-10896.
9
J. AM. CHEM. SOC. VOL. 127, NO. 25, 2005 9273

A R T I C L E S
Fiammengo et al.
Scheme 5. Synthesis of Guanosine Phosphate Derivatives 1a, 1c, and 1da
a Reagents and conditions: (a) 20 (1.0 equiv), 3 (1.4 equiv), IMT (0.56 equiv), ACN/THF 1:1, rt, 1 h, then TBHP (10 equiv), rt, 10 min; volatiles
removal, and then TBAF (10 equiv), THF, rt, 1h, 41% over two steps; (b) 22 (1.0 equiv), 3 (1.2 equiv), DCI (3.0 equiv), THF, rt, 1 h, then TBHP (10 equiv),
rt, 10 min; volatiles removal, and then TBAF (10 equiv), THF, rt, 3h, 36% over two steps; (c) H2, Pd/C, MeOH, rt, 2 h, 94%. IMT ) imidazolium triflate,
TBHP ) tert-butylhydroperoxide, DCI ) 4,5-dicyanoimidazole.
Scheme 6. Two Synthetic Strategies for
Finally phosphate 1d, bearing a primary amino group at the
end of the decaethylene glycol spacer, was obtained removing
Anthracene-Decaethylene Glycol Guanosine Phosphate 1ba
the Cbz group from 1c via catalytic hydrogenation.
Incorporation of Compounds 1a-d at the 5′-end of RNA
Transcripts. Bacteriophage RNA polymerases (RNAP) have
been shown to be highly active for in vitro transcriptions from
synthetic DNA templates.13,27 In fact, T7 RNAP transcriptions
can be used for the synthesis of sizable amounts of RNA (up
to several milligram) starting from synthetic DNA templates.14
It is also known that a number of modifications can be
introduced at the 5′-end of RNA transcripts using bacteriophage
RNAP. Thus, “capped” RNA transcripts are obtained simply
by adding to the mixture of nucleotide triphosphates (required
for polymerization) an appropriate guanosine monophosphate
derivative.16,17 Even dinucleotides and nucleotides with modified
bases can be used as initiators. The incorporation occurs
selectively at the 5′-end since during the elongation step only
triphosphates can be used. On the basis of previous data from
this laboratory,20,28 and from others19 it was known that
guanosine monophosphate derivatives carrying long, flexible
a Reagents and conditions: (a) 2-cyanoethyl-N,N-diisopropylchlorophos-
phoramidite (1.2 equiv), DIPEA (3.0 equiv), CH2Cl2, 0 °Cfrt, 1 h, 98%;
poly(ethylene glycol) spacers were accepted by T7 RNAP during
in vitro transcriptions with moderate incorporation ratios.
However, only very limited information exists regarding the
influence of various parameters (like concentration of initiator
nucleotide, of GTP, and of the other NTPs) on the incorporation
efficiency and the total transcription yields.18-20 Moreover, the
influence of the chemical nature of the residue attached at the
end of the spacer has not been explored at all.
(b) 3 (1.2 equiv), DCI (1.0 equiv), THF, rt, 20 min, then TBHP (10 equiv),
rt, 10 min; volatiles removal, and then TBAF (6.0 equiv), THF, rt, 1h, 31%
over two steps; (c) 7 (1.2 equiv), BTT (10 equiv), ACN, rt, 9 h and then
TBHP (10 equiv), rt, 10 min; (d) TBAF (5.0 equiv), THF, rt, 1h, 20% over
two steps. BTT ) 5-benzylthiotetrazole.
An alternative strategy was also considered for the preparation
of guanosine monophosphate derivatives, namely, the coupling
between 7 and phosphoramidites prepared from the decaethylene
glycol derivatives (e.g. 23). One disadvantage of this approach
is that one different phosphoramidite should be synthesized for
each of the different chains to be coupled. Nonetheless, the
synthesis of 1b (Scheme 6) was carried out following this second
strategy (23 + 7, 20% yield). These experiments showed two
major problems, namely, the very poor solubility of 7 in all the
solvents typically used for coupling (ACN, THF, CH2Cl2) and
a considerable more difficult purification of the desired product.
In fact, several anthracene-carrying byproducts with polarities
similar to that of the desired product were observed when
phosphoramidite 23 was used, probably originated from side
reactions at the phosphorus. In contrast the purification of 1b,
prepared according to the first strategy (16 + 3, 31% yield),
was considerably simpler. Only two species containing an-
thracene were then present in the reaction mixture, i.e. the
desired 1b and unreacted 16, and they could be easily separated
by RP chromatography.
We have systematically investigated the incorporation condi-
tions for compounds 1a,b,d using a double-stranded DNA
template encoding for a 25-nucleotide long transcript. Then,
under optimized conditions, all four initiator nucleotides 1a-d
have been incorporated in high yield.
The first set of experiments was performed keeping constant
the concentrations of ATP, CTP, and UTP (each 1 mM) and
the amounts of 1a,b,d (4 mM), and varying the amount of GTP
(which competes with the initiator nucleotide in the initiation
step but is essential for elongation) between 0.2 and 2 mM (ratio
1:GTP from 20:1 to 2:1). Gel electrophoresis analysis of the
transcription reactions showed for 1a,b,d a similar trend in total
yield of RNA transcripts, i.e. the sum of both the desired
(27) Melton, D. A.; Krieg, P. A.; Rebagliati, M. R.; Maniatis, T.; Zinn, K.;
Green, M. R. Nucleic Acids Res. 1984, 12, 7035-7056.
(28) Seelig, B.; Ja¨schke, A. Tetrahedron Lett. 1997, 38, 7729-7732.
9
9274 J. AM. CHEM. SOC. VOL. 127, NO. 25, 2005

Efficient Preparation of Organic Substrate
−RNA
A R T I C L E S
Figure 1. Enzymatic incorporation of initiator nucleotides 1a ([), 1b (b), and 1d (2) (4 mM) via T7 RNAP transcription (2 h reaction time) of a 44-mer
dsDNA template in the presence of ATP, CTP, and UTP (each 1 mM) and varying concentration of GTP. (A) Phosphorimages of the PAGE analyses. Lanes
1-4: GTP 0.2, 0.5, 1.0, and 2.0 mM. Transcription products a and b: 25- and 26-mer nonmodified transcripts. c and d: 25- and 26-mer modified transcripts.
26-mers result from nontemplated addition of one nucleotide to the 3′-end of transcribed RNA. (B) The total yields (a + b + c + d) increase with increasing
GTP concentration, while the incorporation ratios (c + d divided by the total yield) decrease (C).
initiated transcript and the normal one with a G at its 5′-end,
under identical conditions (Figure 1).
In fact, the total yield of RNA is increased by increasing the
amount of GTP showing almost no saturation behavior at the
used concentrations, as expected from previous reports.13
However, it is interesting to note that the total yields are
influenced by the nature of the different initiator nucleotides:
systematically higher yields are obtained when compound 1b
was present in the reaction mixture. The incorporation yield,
i.e. the amount of initiated transcripts divided by the total amount
of transcripts, increased by lowering the amount of GTP
(increasing the 1:GTP ratio), but the effect was notably different
for the various initiator nucleotides. At 2 mM GTP, compounds
1a and 1d are incorporated only modestly (29 and 9%), while
1b is incorporated much more efficiently (64%). Decreasing
the GTP concentration from 2 to 0.2 mM resulted in ∼3 and
∼5 times better incorporation yields (90 and 49%) for 1a and
1d, respectively. For 1b the observed increment was ∼1.5 times,
affording 97% of the initiated transcript.
It is also known that an increase in the concentration of
nucleotide triphosphates (NTPs) up to 4 mM gives a substantial
increment in the total transcription yields.13 We therefore carried
Figure 2. Enzymatic incorporation of initiator nucleotides 1a ([), 1b (b),
and 1d (2) (4mM) via T7 RNAP transcription (2 h reaction time) of a
44-mer dsDNA template: GTP, 1 mM; concentrations of ATP, CTP, and
out a set of experiments keeping constant the ratio between
1a,b,d and GTP (4 and 1 mM) and varying the concentrations
of the other NTPs. The results showed indeed that the total yields
increased increasing NTP concentrations. Unexpectedly also the
incorporation ratio slightly increased (5-10%) (Figure 2). This
result indicates that the conditions favoring higher transcription
yields plays in the same direction as those favoring higher
incorporation. Thus, it should be possible to increase the total
transcription yields while still keeping high incorporation yields
by working at higher NTP concentration (>1 mM) and
increasing the GTP concentration (>0.2 mM). The next set of
experiments confirmed our expectations (Figure 3). By raising
the GTP concentration to 0.35 mM and the ATP, CTP, and UTP
UTP (in equimolar amount), varying between 1 and 4 mM. Transcription
yields (top) and the incorporation ratios (bottom) increase with increasing
ATP, CTP, and UTP concentrations.
to 4 mM, initiator nucleotides 1a-c were incorporated in 81,
91 and 68%, respectively, and the total yields for 1a and 1b
increased by 70 and 80%, respectively. These increments
compare well with the ones observed in previous experiments
(Figures 1 and 2). Initiator 1d was incorporated in a more
modest 41% (the highest incorporation observed in the previous
set of experiments was 49%), but the total yield increased almost
150%. It is also interesting to note that under these conditions
9
J. AM. CHEM. SOC. VOL. 127, NO. 25, 2005 9275

A R T I C L E S
Fiammengo et al.
for in vitro selection of novel ribozymes. The synthetic effort
to prepare such molecules has been minimized by developing
a versatile path which involves the two common intermediates
2 and 3. It is likely that a wide variety of small molecule
functionalities attached via a PEG-type spacer to the phosphate
moiety of GMP can be prepared by making use of these
intermediates. Intermediate 21 is expected to be of particular
synthetic value, since it could be deprotected and reacted
selectively at the amino terminus with active ester derivatives.
Therefore even delicate substrates could then be attached to the
decaethylene glycol spacer, without involving them directly
during the harsh alkylation step. Efforts in this direction are
currently underway in our laboratories. The data here presented
also show that initiator nucleotides 1a-d are well-accepted as
substrates by the enzyme T7 RNAP in the runoff transcription
reactions of synthetic dsDNA templates, without giving rise to
any detectable inhibition. Under appropriate conditions, like
GTP e 0.35 mM, the majority of the full length transcript is
initiated by the initiator nucleotide (up to 95%), the only
exception being the initiator nucleotide carrying an unprotected
amino group. Current work in our laboratory is now focusing
on the development of a novel ribozyme which catalyses allylic
substitution reactions in combination with transition metal
complexes making use of initiator nucleotide 1a.
Figure 3. Optimized transcriptions initiation reactions (2 h reaction time)
using a 44-mer dsDNA template. Initiator nucleotides 1a (lane 2), 1b (lane
3), 1c (lane 5), and 1d (lane 4), 4 mM; GTP, 0.35 mM; ATP, CTP, and
UTP, each 4 mM. a and b: 25- and 26-mer nonmodified transcripts. c and
d: 25- and 26-mer modified transcripts. In lane 1 the transcription reaction
is run under identical conditions except for the absence of any initiator
nucleotide.
the total transcription yields remained substantially constant for
all initiator nucleotides investigated (Figure 3) and were
comparable with the transcription yield obtained in the absence
of any initiator nucleotide.
The reasons for the lower incorporation of 1d are at present
not fully understood. Most probably the primary amino group
at the end of the decaethylene glycol chain is protonated at the
pH used in the transcription experiments, significantly affecting
the recognition of this substrate by T7-RNAP.
Acknowledgment. This work was supported by the Deutsche
Forschungsgemeinschaft (Grant SFB 623) and the Fonds der
Chemischen Industrie. The authors thank Mihaela Caprioara for
performing the MALDI TOF measurements and Prof. R. Kra¨mer
and Dr. A. Mokhir for helpful discussion and technical support.
The authors also thank Nancy Sauer for technical assistance.
Supporting Information Available: Full Experimental Sec-
tion including procedures and spectroscopic data for all new
compounds and in vitro transcription procedures and MALDI
TOF analysis of the transcripts. This material is available free
of charge via the Internet at http://pubs.acs.org.
Conclusion
Incorporation of guanosine monophosphate derivatives at the
5′-end of RNA transcripts is an efficient, reliable method to
produce RNA conjugates bearing covalently attached substrates
JA051179M
9
9276 J. AM. CHEM. SOC. VOL. 127, NO. 25, 2005