Synthesis, incorporation efficiency, and stability of disulfide bridged functional groups at RNA 5'-ends

Article abstract of DOI:10.1016/S0968-0896(00)00080-8

Modified guanosine monophosphates have been employed to introduce various functional groups onto RNA 5'-ends. Applications of modified RNA 5'-ends include the generation of functionalized RNA libraries for in vitro selection of catalytic RNAs, the attachm

Full text of DOI:10.1016/S0968-0896(00)00080-8

Bioorganic & Medicinal Chemistry 8 (2000) 1317±1329
Synthesis, Incorporation Eciency, and Stability of Disul®de
0
Bridged Functional Groups at RNA 5 -Ends
a
b
c
d
Gerhard Sengle, Andreas Jenne, Paramjit S. Arora, Burckhard Seelig,
c
d
a,
È
James S. Nowick, Andres Jaschke and Michael Famulok *
a
Kekule Institut fuÈr Organische Chemie und Biochemie, UniversitaÈt Bonn, Gerhard-Domagk-Straûe 1, 53121 Bonn, Germany
b
Institut fuÈr Biochemie-Genzentrum, Feodor-Lynen-Straûe 25, 81377 MuÈnchen, Germany
Department of Chemistry, 535B Rowland Hall, University of California at Irvine, Irvine, CA 92697-2025, USA
c
dInstitut fuÈr Chemie/Biochemie, Freie UniversitaÈt Berlin, Thielallee 63, D-14195 Berlin, Germany
Received 29 September 1999; accepted 27 January 2000
0
AbstractÐModi®ed guanosine monophosphates have been employed to introduce various functional groups onto RNA 5 -ends.
0
Applications of modi®ed RNA 5 -ends include the generation of functionalized RNA libraries for in vitro selection of catalytic
RNAs, the attachment of photoanity-tags for mapping RNA±protein interactions or active sites in catalytic RNAs, or the non-
radioactive labeling of RNA molecules with ¯uorescent groups. While in these and in similar applications a stable linkage is desired,
in selection experiments for generating novel catalytic RNAs it is often advantageous that a functional group is introduced rever-
sibly. Here we give a quantitative comparison of the di€erent strategies that can be applied to reversibly attach functional groups
0
via disul®de bonds to RNA 5 -ends. We report the preparation of functional groups with disul®de linkages, their incorporation
eciency into an RNA library, and their stability under various conditions. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
selection experiment, a pool of randomized RNA or
DNA molecules is subjected to iterative selection/
ampli®cation steps which allow the enrichment of active
nucleic acid catalysts until they dominate the library
even though they initially may have been represented in
a very low proportion.
The targeted modi®cation of RNA molecules with
arbitrary chemical functionalities at de®ned sites has the
potential for broad applications in molecular biology.1
One of the most straightforward methods to introduce a
particular functionality into an RNA molecule is the
,2
derivatization with guanosine monophosphorothioate
0
By far the most successful strategy for the isolation of
RNA- and DNA-based catalysts involves the direct
screening of nucleic acids libraries for catalytic activity.
In these direct selections, nucleic acids capable of cata-
lyzing a particular chemical transformation modify
themselves with an anity-tag or other characteristic
that allows its preferential enrichment over those mole-
cules which are catalytically inactive. This strategy also
allows for the screening of nucleic acids that catalyze
reactions between two small substrates if one of the
reactants is covalently attached to every individual
member of the starting RNA pool while the other reac-
tant in solution is carrying an anity tag. Consequently,
only those RNA species within the library that catalyze
bond formation between the two reactants become tag-
ged and therefore can be separated from non-active
(
GMPS) at the 5 -end and the subsequent chemical
modi®cation of the nucleophilic sulfur at the phos-
phothioate group. For example, Burgin and Pace have
0
for mapping the active site of RNaseP.
attached a photoanity tag to a 5 -thiophosphorylated
Phe
3
tRNA
Another potential application of this technology is the
mapping of RNA±protein interactions in large RNA/
protein complexes.
0
5
-Modi®ed RNA molecules have probably experienced
their widest application in the ®eld of in vitro selection
of novel catalytic RNAs. Various ribo- and deoxyribo-
zymes have been isolated by screening large combina-
torial nucleic acid libraries.4 In a typical in vitro
�6
7
molecules, e.g., by anity chromatography. If cleavable
linkers have been used to attach the functional group to
the RNA, active catalysts can be recovered speci®cally
*
3
Corresponding author. Tel.: +49-228-731-787; fax: +49-228-735-
88; e-mail: m.famulok@uni-bonn.de
0968-0896/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(00)00080-8

1
318
G. Sengle et al. / Bioorg. Med. Chem. 8 (2000) 1317±1329
after their immobilization. This has been achieved for
8
throughout the text, produces RNA transcripts which
0
example, by using a photosensitive group or a disul®de
bridge that can be easily cleaved to liberate the RNA
from the solid support.9
allow the modi®cation of RNA at the 5 -end through
phosphorothioate disul®de linkages. In studying these
strategies, we were concerned that the stability of the
phosphorothioate sul®de linkage might not be sucient
for in vitro selection experiments. In such experiments
the library is often incubated for several hours and
therefore the linkage has to be inert to decomposition,
otherwise the desired activity cannot be enriched. This
paper describes the implementation and evaluation of
these methods.
,10
In this paper, we report the preparation of functional
groups with disul®de linkages, their incorporation e-
ciency into an RNA library, and their stability under
various conditions. We have applied two di€erent stra-
tegies in this study: (a) The direct incorporation of
GMP-derived initiator nucleotides that already contain
a disul®de linkage into RNA by in vitro transcription,
0
and (b) the incorporation of GMPS (guanosine-5 -
3
monophosphorothioate) into RNA by GMPS-primed
Results and Discussion
transcription and subsequent coupling of the RNA with
a 2-thiopyridine-activated substrate to form a disul®de
linkage. A third strategy (c), involving the coupling of
GMPS-primed RNA by a sul®de bond to a substrate
containing a disul®de group, has been published pre-
We have established and analyzed two di€erent strate-
gies to attach certain functional units via cleavable di-
sul®de bridges onto RNA molecules (see Fig. 1). For
both methods, the eciency of disul®de-incorporation
and the stability of di€erent disul®de linkages at di€er-
ent bu€er conditions were investigated. For the GMPS-
coupling-method, a number of 2-thiopyridine activated
compounds (8, 10, 14) were synthesized. These com-
9
,10
viously
1
and was not investigated in this study. Figure
illustrates these strategies.
Strategy (a), designated as the incorporation-method
throughout the text, has recently been established to be
a useful method for the direct and site-speci®c incor-
poration of functional groups into RNA transcripts.1
Strategy (b), designated as the GMPS-coupling-method
0
pounds were linked to the 5 -phosphorothioate group of
an RNA library generated by T7 transcription in the
1,22,24
12
presence of GMPS. To investigate the stability of the
0
resulting 5 -RNA phosphorothioate sul®de linkage
0
Figure 1. There are di€erent strategies to introduce functional groups via cleavable disul®de bridges onto RNA 5 -ends. In this study, methods (a)
and (b) were investigated in detail. Various functionalized RNAs were prepared, and incorporation eciency and stability of the disul®de linkages
were analyzed. `X` stands for a functional group, `L` stands for a leaving group (e.g., a leaving group in an active ester).

G. Sengle et al. / Bioorg. Med. Chem. 8 (2000) 1317±1329
1319
(alkyl-S-S-PO -RNA), we synthesized the correspond-
ing model compounds 1a, 3a or directly analyzed 2b in
which the functional group is attached to the 5 -end of
RNA molecules (see Fig. 2). As a representative mole-
cule for related investigations into the second strategy,
the incorporation-method, we synthesized the initiator
nucleotide 4a, a molecule containing disul®de bridge
ester 7 with thiopyridine-derivatized cysteine 6 in
DMSO and triethylamine (TEA) a€orded thiopyridine-
activated 8 in 43% yield. The thiopyridine group of 8
was exchanged with GMPS to obtain the GMPS-
cysteine analogue 1a in 62% yield after puri®cation by
preparative HPLC. This reaction is conveniently
monitored by analytical HPLC, which indicates the
GMPS to be consumed and 1a to form within 10 min of
mixing of 7 equivalents of 8 with GMPS at 10 mM
thiopyridine-activated substrate concentrations.
3
0
¯
analyzed its incorporation by T7 RNA-polymerase
anked by alkyl groups (alkyl-S-S-alkyl-GMP) and
13,14
0
and the stability of the resulting 5 -modi®ed RNA.
Scheme 1 also illustrates the synthesis of the thiopyridine-
activated biotinylated cysteine 10, which was used to
a€ord biotin linked RNA 2b in subsequent experiments.
The biotin-derivative 10 was prepared from commercially
available NHS-activated biotin and thiopyridine-activated
Syntheses
The GMPS-linked benzoylcysteine 1a was prepared in two
steps from N-hydroxy succinimide (NHS)-activated ester
of benzoic acid (Scheme 1). Coupling of the NHS-activated
Figure 2. Compounds synthesized for and analyzed in this study. 1±3 were generated by reaction of thiopyridine activated compounds 8, 10, 14 with
GMPS or GMPS-primed RNA. 4b is the RNA conjugate prepared by transcription reaction in the presence of initiator nucleotide 4a. While com-
pounds 1, 2, and 3 contain a phosphorothioate sul®de linkage, compound 4 contains a `natural` alkyl disul®de bond.

1
320
G. Sengle et al. / Bioorg. Med. Chem. 8 (2000) 1317±1329
Scheme 1. Synthesis of compounds 1 and 10.
cysteine 6 in 80% yield. Incubation of compound 10
with GMPS-primed RNA a€orded biotin-linked RNA
The purity of 4a was quanti®ed by the HABA (4-
17
hydroxyazobenzene-2-carboxylic acid)-test which allows
quanti®cation of biotin contained in a solution. Using
the HABA-test, we calculated the yield of HPLC-pur-
i®ed 4a to be 53%.
2b.
Compound 14 was prepared from activated-citrulline 11
in three steps (Scheme 2). Coupling of 4-nitrophenyl
ester of N-Boc-protected citrulline15 11 with S-tritylcys-
teine 12 followed by deprotection of the trityl group
a€orded the citrullyl-cysteyl dipeptide 13 in 61% yield.
The activated 2-thiopyridyl compound 14 was synthe-
Incorporation studies
We performed a set of di€erent analyses to determine
the eciency of incorporation of the various functional
0
0
sized from 13 by reaction with 2,2 -dipyridyl disul®de in
79% yield. Compound 14 was subsequently condensed
groups into RNA 5 -ends by using both the incorpora-
tion-method and the GMPS-coupling-method (see Fig.
1). These analyses included anity chromatography and
gel-retardation assays of functionalized RNA.
with GMPS or GMPS-primed RNA, to a€ord 3a or 3b,
respectively. Completeness of the reaction was deter-
mined by quanti®cation the UV absorbance of the lib-
erated thiopyridone.
For the GMPS-coupling-method, we ®rst determined the
eciency of GMPS-incorporation by T7-transcription
of a random-dsDNA library in the presence of various
amounts of GMPS (Fig. 3). The percentage of GMPS-
The synthesis of 4a is shown in Scheme 3. Coupling of
protected guanosine 15 and the phosphoramidite 16
followed by iodine oxidation, and basic and acidic
primed RNA was determined by separation of the
32
0
deprotection steps, a€orded hexylamino guanosine-5 -
thiomodi®ed
P-labeled RNA from non-modi®ed
monophosphate 17. Subsequent coupling of sulfosucci-
nimidyl-2-(biotinamido)-ethyl-1,3-dithiopropionate 18
RNA by [(b-acryloylamino)phenyl]mercuric chloride
(APM polyacrylamide gel anity electrophoresis)
18,23
(
routinely used for the biotinylation of proteins such as
a water soluble N-hydroxysuccinimide ester which is
and subsequent quanti®cation of the retarded band on a
phosphor imager. Surprisingly, we observed that GMPS
had a stimulating e€ect on the transcription reaction:
the yields of RNA were generally higher when GMPS
1
6
IgG or cell surface proteins of lymphocytes) with 17 in
KH PO -bu€er at pH 8.0 for 16 h yielded 4a.
2
4

G. Sengle et al. / Bioorg. Med. Chem. 8 (2000) 1317±1329
1321
Scheme 2. Synthesis of 3b.
Scheme 3. Synthesis of 4a and precursors.
was present up to a 4-fold excess over the NTPs used in
the standard transcription (Fig. 3). At a GMPS-con-
centration of 5±15 mM and a concentration of 5 mM of
the other four NTPs the total yield of RNA comprised
equal molarity of each NTP, the transcription reaction
was enhanced by a factor of 3.5. When the mono-
nucleotide concentration exceeded 50±60 mM we
observed inhibition of transcription.
>
priming eciency have been obtained by other groups
using lower amounts of GMPS.
80% GMPS-primed transcripts. Similar results in
To compare the eciency of the second step of the
reaction, i.e., coupling of the thiopyridine-activated
compounds to the 5 -phosphorothioate group in GMPS
1
9,20
0
For comparison we also performed similar experiments
with the mononucleotides GMP and AMP. Whereas
AMP had no e€ect on the transcription yield at con-
centrations below 10 mM, the presence of GMP in
equal concentrations or moderate molar excess over the
four NTPs clearly led to an increase of RNA yields. At
a concentration of 5 mM GMP, corresponding to an
or GMPS-primed RNA, respectively, we used HPLC,
UV-spectroscopy or anity chromatography. The rate
of disul®de formation of the GMPS model-substrate
was determined by HPLC-quanti®cation of the 2-thio-
pyridone by-product formed during the coupling step.
At the reaction conditions described in the synthesis
section less than 5% of the initial GMPS concentration

1
322
G. Sengle et al. / Bioorg. Med. Chem. 8 (2000) 1317±1329
Figure 3. Dependence of transcription yields on the concentration of mononucleotides GMP, GMPS and AMP. Percentage values of the various
transcription reactions (NMP concentration: 4±80 mM) were normalized to a standard in vitro transcription containing each NTP at a concentra-
tion of 5.0 mM without any mononucleotide added.
remained in the reaction mixture after 5 min. We also
performed similar experiments with a 25- to 50-fold
excess of the activated dipeptide 14 using both GMPS-
primed RNA and GMPS alone. The analysis of the
coupling eciency of these reactions yielding com-
pounds 3a and 3b was performed by UV spectroscopy.
After a 5 min incubation time the UV±vis-spectra in the
range between 210 and 420 nm were recorded and sub-
tracted by a spectrum obtained with a reference sample
containing the dipeptide. From the absorption at 343
nm the amount of 2-thiopyridone was calculated pro-
viding an indirect but nevertheless very accurate mea-
sure of the coupling eciency. This analysis established
that the reaction was quantitative since the molar
amount of liberated thiopyridone was approximately
equivalent to the RNA concentration used (see meth-
ods). When using GMPS instead of GMPS-primed
RNA the formation of the phosphorothioate sul®de
bridge also appeared to be quantitative (data not
shown).
successive coupling step of at least 70%, the overall
yield of phosphorothioate sul®de derivatized RNA was
determined to be at least 60%. Taken together the
results obtained with the activated thiols 1±3 con®rm
that the coupling of GMPS/ GMPS-primed RNA with
thiopyridine-activated molecules is fast, selective, and
quantitative.
To compare the data obtained with the GMPS-coupling-
method we now analyzed the incorporation-method (see
Fig. 1a). It is known that T7 RNA polymerase tolerates
GMP,11 GMPS, a series of modi®ed nucleotides,
3
13,22
21
and various dinucleotides [NpG ] for initiation of
transcription. However, it has been also observed that
trimers or short oligonucleotides are only incorporated
with moderate success.1
4,28
To investigate whether the
initiator nucleotide 4a was accepted by T7 RNA poly-
merase, we added 4a to standard transcription reactions
and reduced the concentration of GTP accordingly. In
addition to the ratio of initiator nucleotide and GTP the
2
+
concentration of the DNA template and that of Mg
-
We then wanted to investigate the coupling eciency of
activated thiols to GMPS-primed RNA (see above for
ions are also important parameters for the general e-
ciency of transcription with T7 RNA polymerase.
Therefore, it was necessary to ®nd reaction conditions
that would lead to both a high ratio of initiator-nucleo-
tide-primed to unprimed RNA and a high overall tran-
scription yield.
1a, 3a) by analyzing the formation of compound 2b.
This compound contains a biotin moiety which enabled
32
the determination of streptavidin-bound
P-radio-
labeled RNA by scintillation counting. At 300-fold
excess of thiopyridine-activated biotinylated cysteine 10
the amount of phosphorothioate sul®de formation
appeared to be at least 70% which was around 25%
lower than for 2a. This diminished yield may result
For a randomized 213mer DNA-library we varied both
the template and Mg2+-concentrations and determined
the ratio for the concentrations of GTP and 4a where
highest absolute yields of primed RNA were obtained.
Highest yields of RNA were obtained at template con-
centrations between 4.0 and 5.0 mM and 12 mM MgCl2
(data not shown). Under these conditions we determined
the ratio of [GTP]/[4a] to be optimal at 8 mM/3 mM.
0
from inaccessibility of the 5 -phosphorothioate group
in the folded RNA for the reaction with 10 or non-
quantitative coupling of biotinylated RNA to the col-
umn resin. However, with a GMPS priming rate for the
RNA of more than 85% and a reaction eciency for the

G. Sengle et al. / Bioorg. Med. Chem. 8 (2000) 1317±1329
1323
The results of this analysis are shown in Figure 4. While
relatively high yields of total RNA are obtained at con-
centrations of 4a at or below 3 mM, the yield dramati-
cally dropped when the concentration of 4a was raised.
At a GTP/4a-ratio of 1:4, using 1.0 mM GTP and 4.0
mM 4a, the total RNA yield dropped to 10% of the
RNA yield obtained under standard conditions for T7
transcription, while with the same ratio at higher con-
centrations (2.5 mM GTP, 10 mM 4a), no transcribed
RNA was detected. This indicates that 4a strongly inhi-
bits transcription.
coupled to thiopyridine-activated sepharose) were incu-
bated at conditions typical for an in vitro selection
experiment. Aliquots were taken at 6±8 di€erent time
points and analyzed. The results are shown in Figure 5.
Compound 1a shows no detectable decomposition after
incubation for 72 h in bu€er (50 mM TRIS 7.4, 5 mM
MgCl , 500 mM NaCl; data not shown). In bu€ered
2
solution at a physiological pH there seems to be only a
slight di€erence in stability of a phosphorothioate sul-
®de linkage compared to the alkyl-¯anked disul®de-
bond in 4. The detected di€erences in stability for model
compounds 1a, 3a, 4a, compared to 2b, 4b, 4b*, most
likely are due to a certain extent of RNA degradation.
As similar curves were obtained for 4b and 4b* it is
obvious that the biotin-anchor-group in 4b had no
in¯uence on the stability of the disul®de-bond.
Stability studies
For many applications of disul®de-derivatized RNAs it
is critical that the modi®cation remains stably coupled
to the RNA. To analyze the stability of the phos-
phorothioate sul®de linkage resulting from the GMPS-
coupling-method we performed a time course experiment,
in which the decomposition of 1a was monitored by
HPLC. We also analyzed the more complex substrate 3a
The result obtained with 2b suggests that the stability
of the phosphorothioate sul®de-linkage is slightly
decreased compared to the alkyl-disul®de bridge.
Despite this slightly decreased stability it appears, how-
ever, that the stability of an RNA library modi®ed with
the disul®de linkage is sucient for an incubation period
of several hours in bu€ered solutions around a neutral
pH. With model compound 3a it could also be shown
that nucleophilic side chains which might attack the
linkage had no e€ect on the stability (see insert Fig. 5).
that includes a nucleophilic primary NH -group in close
2
proximity to the phosphorothioate sul®de, which we
assumed to a€ect the stability of this linkage. Further-
more, we investigated the phosphorothioate sul®de
linkage in the context of an RNA (2b). To compare the
stability of the phosphorothioate sul®de linkage with
that of a disul®de bridge ¯anked by short alkyl chains
we also analyzed compounds 4a and 4b by HPLC and
RNA anity chromatography, respectively.
To investigate, whether bu€er, temperature or salt con-
ditions a€ect the stability of the phosphorothioate sul-
®
de-linkage, we performed a time course experiment
32
Compounds 2b, 3a, 4a,b and 4b* (obtained from 4b
after disul®de reduction with 0.2 M DTT and then
in which GMPS-primed P-labeled RNA was coupled
to thiopyridine-activated sepharose and incubated the
Figure 4. Quantitative analysis of transcription yields and priming eciencies using 4a in the incorporation-method. The height of the bars indicate
the total yield of RNA after transcription in the presence of various ratios of GTP/4a (mM) at 2.4 mM DNA and 12 mM MgCl . The shaded
portions represent the fraction primed with 4a. Percentage values are normalized to an in vitro transcription reaction that contained each NTP at
.0 mM without 4a.
2
5

1
324
G. Sengle et al. / Bioorg. Med. Chem. 8 (2000) 1317±1329
Figure 5. Decomposition of RNA coupled substrates 2b, 4b and 4b* (obtained from 4b after disul®de-bond cleavage with 0.2 M DTT) in bu€ered
solution (50 mM K-MOPS, 200 mM NaCl, pH 7.4). Analysis was performed by RNA anity chromatography in duplicate (see Methods). The
compounds were analyzed for up to 30 h and the percentage of decomposition was plotted versus time. Inset: Decomposition of the model com-
pounds 3a, 4a under the same conditions was analyzed by reversed phase HPLC (see Methods). Almost no decomposition of 1a was detected after
7
bation at 95 C for 5 min in the same bu€er; only upon treatment with 2 M mercaptoethanol, 1a clearly reacted to a€ord N-Bz-cysteine and GMPS.
2 h incubation in aqueous bu€er (50 mM TRIS, 5 mM MgCl
ꢀ
2
, 500 mM NaCl, pH 7.4) (data not shown). Compound was also stable after incu-
immobilized RNA under various salt-, bu€er-, and
temperature-conditions. At various time points, unbound
RNA resulting from disul®de bond cleavage was eluted
and quanti®ed by scintillation counting. The results of
these experiments are shown in Figures 6 and 7.
alkyl-¯anked disul®de bond remains stable without any
detectable in¯uence of pH the phosphorothioate sul®de-
linkage clearly is much more susceptible to basic pH.
Most likely the alkyl-S-S-PO -RNA bridge is reduced-
3
to GMPS-primed RNA and the corresponding thiol as
the eluted RNA fractions could be almost quantitatively
re-immobilized on thiopropyl sepharose (data not
shown). Figure 7b con®rms that the high amount of
eluted RNA at high values of pH is not due to increased
RNA degradation. The RNA itself is equally stable
under the various elution conditions used.
Omission of any salts or bu€er results in a destabilized
linkage (19% decomposition after 30 h at 23 C, ®lled
ꢀ
rhombus) compared to a K-MOPS solution adjusted to
ꢀ
comparison of the reaction performed in plain water
pH 7.4 (9% after 30 h at 23 C, open square). The
2
+
(
19%, ®lled rhombus) with that in 100 mM Mg (1%,
2
+
®
linkage. The presence of 100 mM Mg
(
lled triangle) shows that Mg -ions are stabilizing the
2
+
in the bu€er
6%, open circle) had no in¯uence on stability com-
Conclusion
pared to the bu€ered solution without divalent metal
ions (9%, open square). Increasing the temperature has a
strong e€ect on the stability of this linkage. Cleavage in
It is critical for the success of an in vitro selection
experiment that (i) the substrate gets e€ectively attached
to RNA, (ii) the attached substrate remains bound to
the RNA during the course of the selection, and (iii) the
immobilized active RNA molecules can be easily
removed from the solid support.
2
+
ꢀ
bu€ered solution in the absence of Mg at 37 C reaches
about 25% after 30 h of incubation (open triangle) com-
ꢀ
pared to only 9% at 23 C under otherwise identical
conditions (open square). No degradation of the eluted
RNA molecules was detected as analyzed by polyacryl-
amide electrophoresis of eluted RNA (data not shown).
We have compared two di€erent strategies allowing
the incorporation of functional groups into RNA
molecules at the 5 -ends. The two strategies involve
0
Figure 7a shows a systematic investigation of the e€ect
of pH on the stability of the two linkages. Whereas the
direct priming of T7 transcripts with a highly functio-
nalized initiator nucleotide in a one-step protocol, the

G. Sengle et al. / Bioorg. Med. Chem. 8 (2000) 1317±1329
1325
Figure 6. Stability of the phosphorothioate sul®de-linkage under various conditions. Radiolabeled GMPS-primed RNA was immobilized on thio-
propyl-sepharose and incubated at room temperature. Aliquots were taken at various time points and analyzed. Percentage of decomposition as
ꢀ
detected by scintillation counting is plotted versus time. (!) 100 mM MgCl
2
in distilled water, (&) in 50 mM K-MOPS, pH 7.4, 200 mM NaCl, at 23 C,
ꢀ
(^) in destilled water, (~) in 50 mM K-MOPS, pH 7.4, 200 mM NaCl at 37 C, (*) in 50 mM K-MOPS, pH 7.4, 200 mM NaCl, 100 mM MgCl
2
.
incorporation-method, and a two-step strategy, the
GMPS-coupling-method, in which the RNA transcripts
are ®rst primed with GMPS and then functionalized at
the thiophosphate group with thiopyridine activated
substrates. We also determined the stability of the
it was shown that another initiator nucleotide, anthra-
cene-polyethylene glycol-GMP did not show any inhibi-
tory e€ect on transcription by T7 RNA polymerase.
1
3,22
The incorporation-method certainly requires time con-
suming optimization of conditions for incorpora-
0
resulting linkages and found that the 5 -modi®cations
tion.1
3,22
The higher yield obtained by the GMPS-
introduced via the GMPS-coupling-method in which a
phosphorothioate sul®de is obtained exhibited slightly
reduced stability compared to a linkage in which the
disul®de groups are ¯anked by alkyl chains. We also
found that several parameters, such as bu€er composi-
tion, temperature and pH play a very critical role for
stability of the various linkages studied. The phosphor-
othioate sul®de linkage is signi®cantly more sensitive
when exposed pH values above 7.5 compared to the
alkyl disul®de linkage which might argue for the incor-
poration-method. Interestingly, high concentrations of
the initiator nucleotide 4a in the T7 in vitro transcrip-
tion reaction resulted in a markedly reduced overall
yield of RNA. This suggests that 4a has a strong inhi-
bitory e€ect on T7 RNA polymerase at concentrations
exceeding 3 mM. The addition of GMP is known to
coupling-method, however, is somewhat compromised
by the slightly reduced stability observed for the phos-
phothioate-disul®de linkage in comparison to the more
inert character of the disul®de group embedded between
alkyl groups introduced into RNA by 4. Another
advantage of the incorporation-method is that it pro-
duces `ready to use` RNA conjugates directly after
transcription reaction.
Our results suggest that following the GMPS-incor-
poration-method might be advantageous if large
amounts of derivatized RNA are desired and if the
RNA is not exposed to values of pH exceeding 7.5,
ꢀ
temperatures higher than 23 C, and low-salt conditions.
In cases where maximum stability of the disul®de linkage
is desired but the adherence to these conditions cannot
be guaranteed it is advantageous to follow the incor-
poration-method or the two-step process (c) in Figure 1.
1
1
preferentially initiate transcription
which could
explain the improved yield- up to 3.5-fold over standard
in vitro transcription reactions.
In summary, the incorporation eciency of complex
initiator nucleotides such as 4a that are introduced by
the incorporation-method is markedly reduced when
compared to the GMPS-coupling-method that takes
advantage of the much more ecient incorporation of
GMPS followed by subsequent derivatization with acti-
vated thiols that occurs almost quantitatively. However,
this conclusion cannot be generalized. In previous studies
Experimental
General materials and methods
MALDI-TOF mass spectra were obtained on a Bruker
Re¯ex spectrometer. Electron ionization spray (ESI)
mass spectroscopy was performed on a Finnigan MAT

1
326
G. Sengle et al. / Bioorg. Med. Chem. 8 (2000) 1317±1329
Figure 7. (A) Decomposition of phosphorothioatesul®des and alkyldisul®des at various pHs. Percentage of cleavage from the matrix is plotted
versus pH. GMPS-primed RNA was coupled to thiopropyl-sepharose and incubated for 24 h at room temperature in 50 mM K-MOPS, 200 mM
NaCl, and the pH value given in the x-axis. 4b was immobilized on a streptavidin-agarose column and incubated at the same conditions. (B) PAGE
analysis of the eluted fractions at pH 9.0, 8.0, 7.4 obtained from the pH-studies shown in (7A) (double determinations). Lower bands correspond to
GMPS-primed 94mer RNA, whereas shifted bands correspond to oxidized 94mer (RNA-S-S-RNA).
7
M spectrometer. A Bruker AMX-400 spectrometer
counting was done in a Beckman LS 6000 Sc liquid
a
1
was used for 400 M Hz H NMR analyses. Polyacryl-
amide gel electrophoresis (PAGE) was carried out as
described unless otherwise stated. Radioactivity was
assayed on a Molecular Dynamics Phosphor Imager.
HPLC was carried out on a Millipore, Millennium
v2.10 or Beckman System Gold programmable solvent
module 126 chromatograph using analytical and pre-
parative C18 reversed-phase columns (Beckman).
dNTPs, NTPs, DNase I were purchased from Boehrin-
ger Mannheim, radiochemicals from NEN. Taq poly-
merase was from Stratagene, RNasin from Promega, T7
RNA polymerase from various sources as indicated.
Oligonucleotides were synthesized on a Millipore Expedite
oligonucleotide synthesizer using standard phosphor-
amidite chemistry unless otherwise stated. Scintillation
scintillation counter. N -Boc-citrulline-p-nitrophenyl-
ester 11 and S-Trityl-cysteine 12 were purchased from
Bachem, N-Monomethoxytrityl-aminohexyl-cyanoethyl-
N,N-diisopropyl phosphoramidite from PerSeptive Bio-
systems. [(N-acryloylamino)phenyl)]-mercuric chloride
(APM) was synthesized18 and mercuro-polyacrylamide
1
8
gels were prepared with this material as described.
Guanosine monophosphorothioate (GMPS) was syn-
thesized and puri®ed as described.12
Preparation of DNA oligonucleotides
The 222mer DNA library used for transcription studies
with GMPS, GMP and AMP to generate a 196mer
9
RNA-library was synthesized as described previously.

G. Sengle et al. / Bioorg. Med. Chem. 8 (2000) 1317±1329
1327
0
The 113mer DNA (5 -TCT AAT ACG ACT CAC TAT
each matrix, equilibrated in coupling bu€er. The matrix
was washed subsequently with 20 column volumes each
of washing bu€er (25 mM K-MOPS, 1 M NaCl, 5 mM
EDTA, pH 7.4), denaturating bu€er (4 M urea, 5 mM
EDTA, adjusted with K-MOPS to pH 7.4) and ®nally
with water. The RNA linked to the matrix was then
transferred to a spin ®lter tube, resuspended in the buf-
fer and incubated for the time given in the ®gure
legends. After the time of incubation, the solution was
centrifuged at 4900 g for 1 min. Each sample was addi-
tionally washed with 20 column volumes with incuba-
tion-solution and the amount of eluted RNA was
detected by scintillation counting.
AGG GAG AGA CAA GCT TGG GTC TGG TTT
GAA CGG GGG CGC ATC AGC CAT GCT ATA
AAC TCC AAA GAA GAG AAA GAG AAG TTA
ATT AAG GAT CCT CAG) was serving as a template
to produce the 94mer RNA used for 2b, 4b and 4b*. The
213mer DNA-library M213.1 was synthesized as
ssDNA the 111mer M111.1, 113mer M113.1 DNA-
2
5
libraries amplifying, ligating and ®nally amplifying
6
2
again as previously described.
T7 Transcriptions
T7 transcriptions were performed essentially as descri-
2
7
bed. DNA templates for T7 transcription were syn-
thesized by standard phosphoramidite chemistry. The
transcription reactions with GMPS, GMP and AMP
were performed in a typical 100 mL transcription reaction
mixture containing 40 mM TRIS±HCl, pH 8.0, 22 mM
Syntheses
BzCys-S-Py (8). A solution of benzoic acid (0.071 g,
0.585 mmol), N-hydroxysuccinimide (NHS, 0.067 g,
0.585 mmol), and 1,3-dicyclohexylcarbodiimide (0.145
g, 0.702 mmol) in CH Cl (5 mL) was stirred at room
MgCl , 6.8 mM spermidine, 30 mM DTT, 0.1% Triton
2
2
2
X 100, 5 mM of each NTP, 4±80 mM of NMP, 2.4 mM
dsDNA template, 20U ribonuclease inhibitor, 3 mM a-
temperature. After 1 h, the precipitate was ®ltered o€
and the ®ltrate was concentrated in vacuo to yield a
white solid. Recrystallization from 2-propanol a€orded
3
2
[
P]-GTP (10 mCi/mL) and 250 U of T7 RNA polymerase.
6
(0.100 g, 78%). The NHS-activated benzoic acid 6
T7 transcriptions with the initiator nucleotide 4a were
carried out to obtain 4b under the same conditions
(0.100g, 0.456 mmol) was dissolved in triethylamine
(0.762 mL, 5.47 mmol), and DMSO (3 mL). To the
solution, S-(-2-thiopyridyl)-l-cysteine 6 (0.147 g, 0.547
mmol) was added and the reaction mixture was stirred
at room temperature. After 4 h, the reaction mixture
was washed with 1 M hydrochloric acid, and the aqu-
eous layer was extracted with CH Cl (3Â5 mL). The
without DTT, but with various concentrations of Mg2
+
,
DNA-template, GTP and 4a as mentioned above. The
ꢀ
reaction mixtures were incubated at 37 C overnight.
After transcription the RNA was precipitated by adding
ammonium acetate/ethanol, re-suspended in 9 M urea,
2
2
50 mM EDTA pH 8.0 and loaded on a preparative 8%
combined organic layers were washed with H O (5Â5
2
(94 nt transcript), 5% (213/158 nt transcripts) denaturating
mL) to remove DMSO, dried over MgSO , ®ltered, and
4
polyacrylamide gel. Products were separated on a 24 cm
vertical gel apparatus, the band was located by UV-
shadowing, cut out, eluted for 1.5 h in 0.3 M NaOAc
concentrated in vacuo. Recrystallization from 2-propanol
and ether a€orded S-(2-thiopyridyl)-l-N-Bz-cysteine 8
(0.65 g, 43%) as a white solid. H NMR (500 M Hz,
1
ꢀ
at 65 C, precipitated by adding NaOAc/ethanol and re-
CD OD) d=8.34 (app d, J=5.0, 1H), 7.83 (d, J=7.1
3
dissolved in distilled water. The 100 mL reaction mixture
containing 4a yielded under the optimized conditions of
Hz, 2H), 7.78 ( d, J=8.1 Hz, 1H), 7.68 (t, J=8.5 Hz,
1H), 7.56 (t, J=7.5 Hz, 1H), 7.47 (t, J=7.5 Hz, 2H),
7.16 (dd, J=10.0, 4.5 Hz, 1H), 4.88 (dd, J=9.5, 4.2 Hz,
1H), 3.48 (dd, J=14.0, 4.2 Hz, 1H), 3.33 (dd, J=14.0,
2
.4 mM of DNA template, 12 mM of MgCl and a ratio
2
between 4a and GTP of 3 and 8 mM an average of 5.80
mg of modi®ed transcripts, which corresponds to 0.38
mol of modi®ed RNA per mol of DNA template.
9.5 Hz, 1H). 13C NMR (125 M Hz, CD OD) d=173.5,
3
170.4, 160.8, 150.6, 139.2, 135.3, 133.1, 129.7, 128.7,
1
N O S (M+H) , calcd 335.0524, found 335.0525.
22.7, 121.7, 53.9, 41.9. HRMS (FAB) m/e for C H
15 15
+
2
3 2
HPLC-stability studies
0
The stability of 1a, 3a and 4a was determined by HPLC
on a C18 reversed-phase column using gradient elution
with 0±40 min H O, 0.1% TFA, 40±50 min CH CN.
BzCys-GMPS (1a). A solution of guanosine-5 -phos-
phorothioate 9 (GMPS, 5.0 mg, 0.012 mmol) and S-(2-
thiopyridyl)-l-benzoylcysteine 8 (25.0 mg, 0.072 mmol)
in 50 mM TRIS-HCl (1 mL, pH=7.4) was incubated at
room temperature. After 1 h, the reaction mixture was
injected into an HPLC apparatus equipped with a UV
detector (C18 reversed phase column, 17:83, methanol:
2
3
ꢀ
The samples were incubated at 25 C and every 3 h a
10 mL aliquot was automatically taken and immediately
injected.
0.1% tri¯uoroacetic acid in H O, 25 mL/min). The peak
2
with the retention time of 13.4 min was isolated and
Cleavage studies on RNA-matrices
Coupling to the matrices thiopropyl-sepharose 6B
Pharmacia)/streptavidin-agarose (Pierce) was achieved
concentrated in vacuo to a€ord 8 mg (62%) of 1a as a
1
(
by dissolving GMPS-primed
white solid. H NMR (500 M Hz, CD CN:D O, 10:90)
3
2
32
P-radiolabeled RNA
d=8.86 (d, J=5.3 Hz, 1H), 7.82 (d, J=8.0 Hz, 2H),
7.66 (t, J=7.4, 1H), 7.55 (t, J=7.7 Hz, 2H), 5.96 (d,
J=5.2 Hz, 1H), 4.91 (dd, J=8.5, 5.0 Hz, 1H), 4.78±4.76
(m, 1H), 4.52 (t, J=4.6 Hz, 1H), 4.40 (app t, J=3.2 Hz,
1H), 4.30±4.28 (m, 1H), 4.25±4.22 (m, 1H), 3.49 (dd,
J=14.0, 5.0 Hz, 1H), 3.32 ( dd, J=14.2, 8.5 Hz, 1H).
respectively 2b, 4b, 4b* in each coupling bu€er (for
thiopropyl-sepharose 6B: 25 mM HEPES, 1 mM EDTA,
pH 7.4, for streptavidin-agarose: 25 mM Na HPO /
2
4
NaH PO 150 mM NaCl, pH 6.9) and then incubated
4
2
for 30 min with the double volume of a 50% slurry of

1
328
G. Sengle et al. / Bioorg. Med. Chem. 8 (2000) 1317±1329
LRMS (FAB) m/e for C H N O PS (M+H)+,
calcd 603, found 603.
incubated in 1.0 mL 80% acetic acid for 45 min at
ꢀ
20
24
6
10
2
23 C. Excess acetic acid was evaporated in vacuo, the
product suspended in 1.0 mL water and then dried in
vacuo for 20 h. Final puri®cation was achieved by dis-
solving the colourless powder in 2.0 mL bu€er A (100
mM triethylammonium acetate, pH 7.0) and pre-
parative reversed phase HPLC under the following
conditions: 0±15 min: 85% bu€er A (100 mM triethy-
lammonium acetate, pH 7.0), 15% bu€er B (100 mM
Biotin-Cys-GMPS-primed RNA (2b). A 10 mM solution
of thiopyridine-activated biotinylated cysteine 10 in
1
3
0% DMF/50 mM TRIS (pH 7.4) was incubated with
3 mM GMPS primed P labeled 94mer RNA (TRIS
32
bu€er: 50 mM TRIS, 5 mM MgCl , 250 mM NaCl).
2
The amount of disul®de formation was determined as
quantitative by measuring the UV-absorbance of the
leaving-group thiopyridine at 343 nm after 30 min. The
reacted mixture RNA was precipitated, redesolved in
streptavidin-binding bu€er and added to column con-
taining 250 mL streptavidin-agarose. After an additional
triethylammonium acetate pH 7.0 in 80% CH CN).
3
Combined fractions were lyophilized in vacuo to yield
+
24.2 mg of a colourless powder (24%). MS (ESI ): m/e
calcd for C H O N P: 463.3, found: 463.3 MH;
16
26
8
6
HPLC (analytical): 10.69 min (7.98 min for triiso-
butyrylguanosine); UV (H O): lmax=252.2 nm, 272 nm
(shoulder).
30 min, the unbound RNA is washed o€ the column
2
with TRIS bu€er. Approximately 70% of the RNA
remained bound to streptavidine-agarose could be
detected by scintillation counting. This result implies,
that about 70% of the 94mer RNA was primed with
GMPS during in vitro transcription.
Initiator nucleotide (4a). 24.2 mg (52.4 mmol) amino-
hexyl-guanosine-5 -monophosphate 17 and 194 mg (340
0
mmol) sulfosuccinimidyl-(biotinamido)-ethyl-1,3-dithio-
propionate 18 were dissolved in 35 mL 0.02 M KH PO
2
4
ꢀ
Biotin-Cys-S-Py (9). A solution of d-biotin-N-hydroxy-
succinimide ester (0.165 g, 0.484 mmol) in DMF (2 mL)
was added to a solution containing S-(2-thiopyridyl)-l-
cysteine 6 (0.100g, 0.376 mmol), triethylamine (1.04 mL,
pH 8.1 and stirred at 23 C overnight. The solution was
concentrated to 7.7 mL in vacuo and puri®ed by pre-
parative reversed-phase HPLC as described for amino-
hexyl-guanosine-5 -monophosphate. Combined fractions
were lyophylized in vacuo to yield 26.4 mg of a white
0
7
.50 mmol), H O (1 mL), and CH CN (2 mL). The
2
3
resulting mixture was stirred for 4 h and then con-
centrated by rotary evaporation to obtain a pale white
solid. Recrystallization from methanol and ether a€or-
ded 0.138g (80%) of thiopyridyl-biotin 10 as a white
solid (31 mmol, 50%). HABA-Test (SIGMA): 27.3 mg;
1
H NMR (400 M Hz, D O): d=8.07 (s, 1H), 5.88 (d,
2
J=6 Hz, 1H), 4.82 (m, 1H), 4.55 (m 1H), 4.48 (t, J=4
Hz, 1H), 4.36 (m, 1H), 4.29 (m, 1H), 4.04 (s, 2H), 3.69
(t, J=6.5 Hz, 2H), 3.47 (t, J=6.5 Hz, 2H), 3.36 (m,
1H), 3.07 (t, J=7 Hz, 2H), 2.91 (m, 3H), 2.82 (m, 2H),
2.74 (m, 1H), 2.60 (t, J=6.5 Hz, 2H), 2.29 (t, J=7.75
Hz, 2H), 1.70±1.30 (m, 10H), 1.12 (m, 4H); ); 13C NMR
1
solid. H NMR (500 M Hz, D O:CD OD, 10:90)
3
2
d=8.44±8.43 (m, 1H); 7.87±7.82 (m, 2H); 7.30±7.27 (m,
1H); 4.71 (dd, J=9.1, 4.2 Hz, 1H); 4.54±4.51 (m, 1H);
4.35±4.32 (m, 1H); 3.34 (dd, J=13.0, 4.3 Hz, 1H); 3.25±
3.21 (m, 1H); 3.18 (dd, J=14.0, 9.0 Hz, 1H); 2.94 (dd,
(500 M Hz, D O): d=177.2, 174.2, 165.6, 159.2, 154.3,
2
J=12.8, 5.0 Hz, 1H); 2.73±2.70 (m, 1H); 2.33±2.24 (m,
1
152.3, 137.8, 116.6, 87.2, 84.1 (d, JCP=9 Hz), 73.8, 70.8,
66.6 (d, JCP=5.5 Hz), 65.2 (d, JCP=5 Hz), 62.4, 60.6,
55.8, 40.1, 39.6, 38.2, 37.2, 35.8, 35.4, 33.8, 30.0 (d,
JCP=7 Hz), 28.5, 28.2, 28.0, 26.1, 25.5, 24.9; 31P NMR
3
2
H); 1.76±1.59 (m, 4H); 1.50±1.46 (m, 2H). C NMR
(
100 M Hz, D O:CD OD, 10:90) d=176.6, 173.9, 166.3,
2
3
1
4
60.5, 150.6, 139.6, 123.0, 121.8, 63.4, 61.7, 57.0, 53.1,
1.3, 41.2, 36.5, 29.6. 29.4, 26.8. HRMS (FAB) m/e
(121.5 M Hz, D O): d=� 0.1 (s, 1P); MS (MALDI): m/z
2
+
for C H N O S (M+H) , calcd 457.1038, found
1
calcd for C H O S N P: 853.6, found: 853 MH;
9
8
25
4
4
3
31 50 11
3
4
57.1049.
HPLC (preparative): 17.21 min (10.51 min for amino-
hexyl-guanosine-5 -monophosphate); UV (H O): lmax=
0
252.2 nm, 270 nm (shoulder).
2
0
0.44 mmol) 2 -O, 3 -O, N -triisobutyrylguanosine 15
Aminohexyl-guanosine-5 -monophosphate (17). 220 mg
0 0
2
(
were co-evaporated three times from pyridine, dried in
vacuo over P O , and dissolved in 3.0 mL dry CH CN
Alkylsul®de-primed RNA (4b*). 4a-initiated 94mer
RNA was incubated with the same volume of a solution
containing 200 mM DTT, pH 7.4 overnight at 23 C,
afterwards precipitated by addition of NaOAc/ethanol,
washed several times with 70% ethanol and redissolved
in the required bu€er.
4
10
3
ꢀ
under argon atmosphere. To this, a solution of 150 mg
0.22 mmol) N-monomethoxytrityl-aminohexyl-cya-
noethyl-N,N-diisopropyl phosphoramidite in 3.0 mL
(
dry CH CN, and, subsequently, 1.9 mL of a 0.47 M
3
solution of tetrazol in pyridine were added under argon
atmosphere. After 20 min, 3.0 mL of 0.1 M I , dissolved
in THF:pyridine:H O (2:2:1 v/v) was added. Excess I2
2
L-Citrullyl-cysteine-tri¯uoroacetate (13). To a solution
a
2
of 1.98 g (5.00 mmol) of the activated ester N -Boc-
was reduced by adding 10.0 mL of a saturated Na S O
3
citrulline-p-nitrophenylester 11 in 10 mL DMF a sus-
pension of 1.82 g (5.00 mmol) S-Tritylcysteine 12 in
20 mL of triethylamin:dimethylformamid (1:1 v/v) was
added dropwise. The solution was stirred at ambient
temperature for 16 h, concentrated in vacuo, and the
product was puri®ed by ¯ash chromatography CHCl3:
2
2
solution, the product extracted with 50.0 mL CH Cl .
2
2
The organic phase was washed with 40 mL of saturated
NaHCO , dried with Na SO , and concentrated in
3
2
4
vacuo. Puri®cation was achieved by ¯ash chromato-
graphy (silica gel 0.063±0.200 mm, CH Cl :CH OH
2
2
3
(95:5 v/v)). The product was dissolved in 0.3 mL ethanol
and incubated overnight in 1.0 mL 33% NH at 55 C.
CH OH (1:9 v/v). The product was dissolved in a
3
mixture of 5 mL of tri¯uoroacidic acid and 5 mL of
ethylmercaptane. After 2 h of stirring, the product was
ꢀ
After removal of excess NH in vacuo the product was
3
3

G. Sengle et al. / Bioorg. Med. Chem. 8 (2000) 1317±1329
1329
precipitated by adding 50 mL of diethylether, subse-
quently ®ltered and washed with diethylether. Recrys-
tallization from methanol/diethylether a€orded 0.73g
form of awards: the Camille and Henry Dreyfus Foun-
dation (Teacher-Scholar Award), the Alfred P. Sloan
Foundation (Alfred P. Sloan Research Fellowship),
and the American Chemical Society (Arthur C. Cope
Scholar Award).
1
(
61%) of a white solid. H NMR (300 M Hz, D O):
2
3
3
d=4.48 (dd, J=5 Hz, J=5 Hz,1H, CH CH CH),
4
(t, J=7 Hz, 2H, NDCH CH ), 3.03±2.91 (m, 2H,
2 2
NDCHCH ), 1.98±1.90 (m, 2H, CH CH CH), 1.66±
2
2
3
3
.11 (dd, J=7 Hz, J=7 Hz, 1H, NDCHCH ), 3.14
3
2
References
2
2
2
1
.55 (m, 2H, CH CH CH ).
2
2
2
1. Dewey, T. M.; Zyzniewski, M. C.; Eaton, B. E. Nucleosides
Nucleotides 1996, 15, 1611.
L-Citrullyl-S-(2-thiopyridyl)-cysteine tri¯uoroacetate (14).
a To 0.78 g (2.00 mmol) of the dipeptide 13 in 15 mL
methanol:water (1:1 v/v) were added 2.2 g (10.0 mmol)
2,2 -dipyridyl disul®de. The solution was stirred at
ambient temperature for 16 h, concentrated in vacuo to
2
1
3
4
1
5
. Aurup, H.; Williams, D. M.; Eckstein, F. Biochemistry
992, 31, 9636.
. Burgin, A. B.; Pace, N. R. EMBO J. 1990, 9, 4111.
. Burgstaller, P.; Famulok, M. Angew. Chem., Int. Ed. Engl.
995, 34, 1189.
0
2
±3 mL and then extracted water (5Â10 mL). The aqu-
. Breaker, R. R. Chem. Rev. 1997, 97, 371.
eous phases were combined, washed with three 30 mL
portions of CH Cl and then lyophylized. The impu-
6. Pan, T. Curr. Opin. Chem. Biol. 1997, 1, 17.
7. Jaschke, A. In The Many Faces of RNA; Eggleston, D. S.;
2
2
È
rities in the crude product were removed by ¯ash chro-
matography CHCl :CH OH (1:9 v/v) and the puri®ed
Prescott, C. D.; Pearson, N. D., Eds.; Academic Press: Lon-
don, 1998; pp 179±190.
3
3
8
9
1
1
È
. Hausch, F.; Jaschke, A. Bioconjug. Chem. 1997, 8, 885.
. Zhang, B.; Cech, T. R. Nature 1997, 390, 96.
0. Zhang, B.; Cech, T. R. Chem. Biol. 1998, 5, 539.
1. Sampson, J. R.; Uhlenbeck, O. C. Proc. Natl. Acad. Sci.
product was eluted,with water. After lyophylization of
the combined aqueous phases 0.79 g (79.2%) of the
1
product were obtained as a pale yellow powder. H
3
NMR (400 M Hz, D O): d=8.67 (d, J=7 Hz, 1H,
2
USA 1988, 85, 1033.
2. Connolly, B. A.; Romaniuk, P. J.; Eckstein, F. Biochem-
istry 1982, 21, 1983.
13. Seelig, B.; Jaschke, A. Terahedron Lett. 1997, 38, 7729.
3
3
heteroaryl-H), 8.42 (dd, J=9 Hz, J=8 Hz, 1H, het-
3
1
eroaryl-H), 8.25 (d, J=9 Hz, 1H, heteroaryl-H), 7.83
3
3
(
dd, J=7 Hz, J=8 Hz, 1H, heteroaryl-H), 4.11 (dd,
2
È
3
3
J=8 Hz, J=5 Hz, 1H, NDCHCH ), 3.50 (dd, J=15
2
14. Pitulle, C.; Kleineidam, R. G.; Sproat, B.; Krupp, G. Gene
1992, 112, 101.
15. Eisele, K.; Hoppe Seylers, Z. Physiol. Chem. 1975, 356,
3
2
3
Hz, J=5 Hz, 1H, CH S), 3.33 (dd, J=15 Hz, J=8
2
Hz, 1H, CH S), 3.14 (t, J=9 Hz, 2H, NDCH ), 1.99±
2
3
2
8
1
45.
6. Hnatowich, D. N.; Virzi, F.; Ruschkowski, M. J. Nucl.
1.93 (m, 2H, CH CH CH), 1.66±1.55 (m, 2H, CH CH
CH ); MS (FD): m/z calcd for C H O S N [M+
2 2 2 2
2
14 22
4
2
5
Med. 1987, 28, 1294.
-
CF COO-]=388, found 388.
3
1
1
1
2
7. Green, N. M. Meth. Enzymol. 1970, 18A, 418.
8. Igloi, G. L. Biochemistry 1988, 27, 3842.
9. Wecker, M.; Smith, D.; Gold, L. RNA 1996, 2, 982.
0. Vaish, N. K.; Heaton, P. A.; Fedorova, O.; Eckstein, F.
Acknowledgements
Proc. Natl. Acad. Sci. USA 1998, 95, 2158.
21. Axelrod, V. D.; Kramer, F. D. Biochemistry 1985, 24,
5716.
This study was supported by grants of the Deutsche
Forschungsgemeinschaft and the Fonds der Che-
mischen Industrie to M.F. and by a NATO travel grant
to M. F. and J. S. N. We thank Nicole Rubner for
operating HPLC-stability studies, Carol Zehetmeier for
2
2
2
2
2. Seelig, B.; Ja
3. Unrau, P. J.; Bartel, D. P. Nature 1998, 395, 260.
4. Seelig, B.; Jaschke, A. Chem. Biol. 1999, 6, 167.
5. Geiger, A.; Burgstaller, P.; von der Eltz, H.; Roeder, A.;
È
schke, A. Bioconjug. Chem. 1999, 10, 371.
È
È
excellent technical assistance, Michael Blind, Gunter
Famulok, M. Nucleic Acids Res. 1996, 24, 1029.
Mayer, Felix Hausch, Barbara Gatto, Nikolai Ra‚er
and Oliver Thum for helpful discussions and Wolfgang
Steglich and Ernst-L. Winnacker for support. A. J. and
B. S. express their gratitude to the DFG for support. J.
S. N. thanks the following agencies for support in the
26. Bartel, D. P.; Szostak, J. W. Science 1993, 261, 1411.
27. Milligan, J. F.; Groebe, D. R.; Witherell, G. W.; Uhlenbeck,
O. C. Nucleic Acids Res. 1987, 15, 8783.
28. Milligan, J. F.; Uhlenbeck, O. C. Meth. Enzymol. 1989,
180, 51.