On Facts and Artefacts: The Difficulty to Evaluate an Artificial Nuclease

Article abstract of DOI:10.1002/hlca.200390316

A number of promising synthetic catalysts for the hydrolytic degradation of RNA have been developed in recent years. Some of them show remarkable selectivity for pyrimidine nucleotides. The general problem of all these studies is to distinguish between real effects and artefacts caused by traces of contaminating natural ribonucleases. We show that methods representing the current state of the art (diethylpyrocarbonate treatment, sterilization, ultrafiltration, etc.) do not sufficiently protect against severe artefacts. However, an incorruptible assay could be found by comparing the cleavage of RNA and its mirror image. Enantiomeric RNA is completely resistant to enzymatic degradation, whereas achiral nonpeptide catalysts, by fundamental laws of symmetry, cannot distinguish between enantiomers and will induce exactly the same cleavage pattern with both substrates.

Full text of DOI:10.1002/hlca.200390316

3740
Helvetica Chimica Acta Vol. 86 (2003)
On Facts and Artefacts: The Difficulty to Evaluate an Artificial Nuclease
b
by Stefan Pitscha), Ute Schefferb), Markus Heyb), Andreas Strickb), and Michael W.Gˆbel*
)
a
¬
¬ ¬
) Institut de Chimie Moleculaire et Biologique, Ecole Polytechnique Federale de Lausanne (EPFL),
BCH-LCAN, CH-1015 Lausanne (fax: ‡41216939355; e-mail: stefan.pitsch@epfl.ch)
b) Institut f¸r Organische Chemie und Chemische Biologie, Johann Wolfgang Goethe-Universit‰t Frankfurt,
Marie-Curie-Str. 11, D-60439 Frankfurt am Main
(fax: ‡496979829464; e-mail: M.Goebel@chemie.uni-frankfurt.de)
Dedicated to Professor Duilio Arigoni on the occasion of his 75th birthday
A number of promising synthetic catalysts for the hydrolytic degradation of RNA have been developed in
recent years. Some of them show remarkable selectivity for pyrimidine nucleotides. The general problem of all
these studies is to distinguish between real effects and artefacts caused by traces of contaminating natural
ribonucleases. We show that methods representing the current state of the art (diethylpyrocarbonate treatment,
sterilization, ultrafiltration, etc.) do not sufficiently protect against severe artefacts. However, an incorruptible
assay could be found by comparing the cleavage of RNA and its mirror image. Enantiomeric RNA is completely
resistant to enzymatic degradation, whereas achiral nonpeptide catalysts, by fundamental laws of symmetry,
cannot distinguish between enantiomers and will induce exactly the same cleavage pattern with both substrates.
Introduction. Catalysts that induce the hydrolytic degradation of RNA are an
important issue of chemical research. Such artificial ribonucleases are mostly based on
metal complexes [1], however, some metal-free systems are known as well [2 4]. They
generally consist of oligo amines [4], guanidines, or structurally analogous heterocycles
[2][3]. In recent years, we have studied the catalysis of phosphoryl-transfer reactions
by guanidinium salts {3-[(guanidino)methyl]phenyl}methylguanidine (1b) and 2-[(3-
{[(4,5-dihydro-1H-imidazol-2-yl)amino]methyl}phenyl)methyl]amino-4,5-dihydro-1H-
imidazole (2b) [5][6]. With highly reactive model substrates in dipolar aprotic solvents,
large rate increases could be observed. In aqueous solution and with RNA substrates,
however, the catalytic efficiency decreased to low values. In an attempt to optimize
compounds 1b and 2b for RNA cleavage, a series of guanidine mimics have been
synthesized recently [7]. The most striking effects in this series were observed with 2-
[(3-{[(1H-benzimidazol-2-yl)amino]methyl}phenyl)methyl]amino-1H-benzimidazole
(3b), a close structural analogue of 1b and 2b. The increase of catalytic potential has
been attributed to a change in pKa from ꢀ 14 to 7. A further jump in reactivity was seen
with the HCl salt of N-(1H-benzimidazol-2-yl)-N',N'-bis[2-(1H-benzimidazol-2-ylami-
no)ethyl]ethane-1,2-diamine (4b), which is equipped with three aminobenzimidazole
units. Both compounds 3b and 4b led to well-reproducible cleavage data and did not
show a significant preference for certain nucleotides. In contrast, other candidates
investigated during recent years, for example, the HCl salt of N-(4,5-dihydro-1H-
imidazol-2-yl)-N',N'-bis[(4,5-dihydro-1H-imidazol-2-ylamino)ethyl]ethane-1,2-diamine
(5b), occasionally showed irregular concentration-effect correlations. These observa-
tions and the pronounced selectivity for pyrimidines typical for those compounds led to

Helvetica Chimica Acta Vol. 86 (2003)
3741
the suspicion that contamination with natural ribonucleases might be the true source of
RNA cleavage activity (the majority of ribonucleases seem to be specific for
pyrimidines [8]). We will present below experiments designed to distinguish between
effects of true catalyst-induced RNA cleavage and those of protein contaminants. This
has proven to be no trivial task. A definitive answer finally was derived from cleavage
experiments with enantiomeric-RNA oligonucleotides [9][10], which are completely
resistant to chiral enzyme catalysts [9]. Achiral catalysts, however, are expected to
produce identical cleavage patterns with both substrate enantiomers.
Results.
Trisguanidinium picrate 5a was prepared in one step from tris-(2-
aminoethyl)amine and the 4,5-dihydroimidazole-2-sulfonic acid (6; 26%) [11]. Similar
oligoguanidines had been reported before as anion receptors by Lehn and co-workers
[12]. Ion exchange easily led to the water-soluble hydrochloride 5b.
RNA Hydrolysis induced by 5b was studied initially with the dye labeled substrate 7.
The fluorescent cyanine dye allows detection of the RNA and of all degradation
products with high sensitivity on a DNA sequencer [7][13 15]. A short stretch of
DNA (T10) was placed between the dye and the RNA sequence to prevent the
formation of very short fragments that would complicate the electrophoretic separation
(the RNA part of oligonucleotides 7 and 8 is adopted from [16]). RNA Hydrolysis is
normally initiated by nucleophilic attack of 2'-OH at the P-atom [3]; DNA, lacking the
2'-OH group, is known to be stable under the conditions of RNA cleavage. In an
improved version of the substrate, a T4 fragment was attached to the 3'-end (8). Fig. 1
gives a survey of the oligonucleotides used in this study. A second type of RNA
substrate (13 15), constructed in a similar way, contained a stem-loop element from
HIV-1 TAR [17].
When incubated with a 10 mm solution of the salt 5b (pH 7, 378, 16 h), substrate 7
was hydrolyzed to a considerable degree. In Fig. 2, lane b, an example showing 54%
conversion of 7 is given. However, most other experiments ended with more than 95%
substrate degradation. At 1 mm 5b, typically 35% of the substrate was hydrolyzed and,

3742
Helvetica Chimica Acta Vol. 86 (2003)
Fig. 1. Top: Oligonucleotides used in this study. Bottom: Secondary-structure model of the TAR-derived
sequences 14 and 15. Enantiomeric nucleotides are symbolized by grey letters.
at 0.1 mm, 3.4%. By coincidence of peaks with a ladder of all possible hydrolytic
fragments, cleavage sites could be localized [7][13][14]. All fragmentations occurred
after pyrimidine nucleotides. Since the conformational requirements for an in-line
attack of 2'-OH are not given in regular duplex structures, cleavage normally occurs at
sites of increased flexibility such as bulges or loops. The observed pattern fits with the
secondary-structure model depicted in Fig. 11). Thus, compound 5b in contrast to the
benzimidazole 4b [7], does not denature RNA secondary structures.
1
)
An interesting, yet unexplained observation is cleavage after C(19), a nucleotide expected to participate in
the lower stem region of TAR (see also [6][7]).

Helvetica Chimica Acta Vol. 86 (2003)
3743
Fig. 2. Cleavage of substrate 7 induced by compound 5b. Control: Incubation of 7 in buffer for 16 h at 378 (lane
a). RNA Cleavage shown in lane b was performed at 10 mm catalyst concentration in 50 mm Tris-HCl, 0.01%
SDS buffer (pH 7.0). A base ladder of substrate 7 is depicted in lane c. RNA fragmentation was analyzed as
described in the Exper. Part. Cleavage sites indicated by arrows are given in bold letters: Cy5-T10-
CUAGCCGACUGCCGAUCUCGCUGACUGAC.
Well aware of the danger of RNase contamination, a wide range of precautions had
been taken. Reaction vessels, buffers, and all robust constituents were treated with
diethylpyrocarbonate (DEPC) prior to sterilization. Solutions of 5b and other catalyst
candidates were purified by ultrafiltration. This technique should remove globular
molecules larger than 3 kDa [7][18][19]. As a test, RNase A (1 ng/ml) was added to
RNA substrate 7 to produce a highly contaminated sample. Very short incubation times
were sufficient for complete degradation of 7 (Fig. 3, lane c). The same RNase A
solution was then subjected to ultrafiltration. When RNA 7 was mixed with the filtrate,
no cleavage occurred even after long incubation times (lane b). Purification of a
solution of trisguanidine 5b by this method followed by incubation with RNA 7, led to
cleavage patterns indistinguishable from those shown in Fig. 2. It seemed rather safe to
conclude, at this point, that trisguanidine 5b itself is capable of catalysis of RNA
degradation.
Fig. 3. Ultrafiltration to remove RNase contaminants. Lane a: Substrate 7 incubated for 20 h at 378. Lane c:
Substrate 7 after incubation with RNAse A (1 ng/ml enzyme, 5 min, 378). Lane b: The same solution of RNAse
A was subjected to ultrafiltration and incubated with 7 for 20 h at 378. A base ladder of 7 is shown in lane d.
In the next set of experiments, the influence of several additives was investigated
(Fig. 4). High concentrations of NaCl are expected to prevent ion-pair association of
trisguanidinium cations and RNA. As foreseen, 1m NaCl completely abolished cleavage

3744
Helvetica Chimica Acta Vol. 86 (2003)
Fig. 4. RNA Cleavage by compound 5b and RNase A in the presence of different salts (50 mm Tris-HCl/0.01%
SDS buffer, pH 7.0, 24 h, 378). Lane a: Untreated substrate 8 as a negative control. Substrate 8 was incubated
with catalyst 5b (4 mm) in the absence (lane b) or presence of 1m NaCl (lane c), 5m urea (lane d), or 4m
guanidinium hydrochloride (Gdn-HCl, lane e). The same set of experiments was repeated with 10 ng RNase A:
1m NaCl (lane g), 5m urea (lane h), 4m Gdn-HCl (lane i).
induced by 5b, whereas RNase A was not inhibited. Addition of 5m urea led to
comparable results. As a classical reagent for denaturing proteins, 4m guanidinium
chloride blocked RNase A. That catalysis by 5b was inhibited as well may be explained
by ion-pair disruption. While denaturation of a contaminating RNase cannot be ruled
out, these data still favor the interpretation that 5b is the active catalyst. This view was
challenged for the first time when RNase inhibitors were found to suppress strand
degradation (Fig. 5).
The experiment designed to unravel the conflicting data was to compare cleavage of
the oligonucleotides 9 and 10. Substrate 9 contains a single 3',5'-linked ribonucleotide
in a stretch of DNA, while, in compound 10, the ribonucleotide has a 2',5' linkage.
Furthermore, the total length of the DNA spacer of 10 is extended relative to 9,
allowing simultaneous monitoring of both substrates in a single analysis (Fig. 6,a, lane
a). As shown in lane c, both substrates are sensitive to base-induced hydrolysis. In
contrast, RNase A attacks only the natural 3',5' bond: the small amount of cleavage
seen with substrate 10 did not increase with time (Fig.6,b, lanes a and b) or with
increasing amounts of RNase A (data not shown). We explain this as cleavage of
contaminating isomeric 3',5' strands present in 10. It is known that 2'-or 3 '-silylated
ribonucleosides may isomerize to a small extent during the process of phosphoramidite
formation [20]. Unfortunately, the amount of cleavage induced by 5b was too small
even after 20 h to decide whether substrate selectivity exists or not (Fig. 6,a, lane b).

Helvetica Chimica Acta Vol. 86 (2003)
3745
Fig. 5. Influence of RNase inhibitor. Substrates 7 and 13 were incubated with trisguanidine 5b (10 mm) in
presence (lanes a and c) or absence (lanes b and d) of 1 unit/ml RNase inhibitor. Samples were run in parallel on
denaturing PAGE.
In a similar way, we constructed the substrate pair 11 and 12, the first containing six
3',5'-linked uridines, the second the corresponding 2',5'-hexa-uridine sequence
(Fig. 6,a, lane d). Upon base-induced cleavage, the 2',5'-linked substrate was distinctly
more reactive (lane f). However, in the presence of 5b, selective cleavage of the 3',5'-
linked substrate 11 occurred. 3',5'-Selective hydrolysis is also induced by RNase A
(Fig. 6,b, lanes c and d). Here, again, the digestion of the 2',5'-substrate stopped at low
levels, irrespective of enzyme concentration and reaction time (data not shown). A
positively charged anion receptor such as 5b may catalyze RNA hydrolysis by
electrostatic transition-state stabilization [2][3][7], and, while it is reasonable not to
expect high levels of 3',5' selectivity from this mechanism, it cannot be excluded in a
strictly logical sense. The experiments shown in Fig. 6, therefore, give strong evidence
for, but not proof of, contamination effects.
At this stage, two groups of assays convincingly advocated contradicting explan-
ations. To corroborate the suspicion of artefacts, an incorruptible test was needed. The
assay designed to meet this requirement makes use of enantiomeric RNA substrates.
Since 5b is an achiral molecule, it forms enantiomorphous complexes with RNA and its
mirror image. The transition states of catalysis are also expected to be enantiomor-
phous, leading to indistinguishable cleavage patterns for the enantiomeric RNA
substrates. The RNA part of substrate 15 is the exact mirror image of the RNA parts of
compounds 13 and 14. Enantiomeric nucleoside building blocks were prepared from d-
glucose according to published procedures [10] and assembled as 2'-O-tom-protected
phosphoramidites [21]. The enantio-RNA is embedded into DNA (natural config-
uration) and has a fluorescent dye attached to the 5' end via conventional amino linker/
active ester chemistry. After isolation by HPLC, 15 still contained traces of shorter
fragments that were absent in the PAGE-purified sample of 14 (Fig. 7,a, lanes a and d).
Base-induced hydrolysis of 14 and 15, as expected, led to identical cleavage patterns
(lanes c and f). In contrast, when 5b was added, RNA hydrolysis was found to be
enantioselective (lanes b and e) and, as seen with 7 (see Fig. 2), pyrimidine specific.
These findings clearly demonstrate that the active catalyst in the sample of 5b must be a
chiral molecule, e.g., a ribonuclease or another protein, but not 5b itself. Controls with

3746
Helvetica Chimica Acta Vol. 86 (2003)
Fig. 6. a) Cleavage of the 3',5'-linked substrates 9 (3',5'-rU), 11 (3',5'-rU6), and their 2',5'-linked counterparts 10
(2',5'-rU) and 12 (2',5'-rU6) by 5b (cf. Fig. 1). Different spacer lengths allow analysis of the reaction of both
substrate types in the same experiment; 5b (10 mm) was incubated for 20 h at 378 with either a mixture of 9 and
10 (lane b) or 11 and 12 (lane e) in 50 mm Tris-HCl/0.01% SDS buffer (pH 7.0). Untreated controls are shown in
lanes a (9 and 10) and d (11 and 12), base ladders are depicted in lane c (9 and 10) and f (11 and 12). b) RNase A
Treatment of substrates 9 12. Oligonucleotides were incubated with 1 ng/ml RNase A for 1 min (9 and 10: lane
a; 11 and 12: lane c) or 5 min (9 and 10: lane b; 11 and 12: lane d) at 378. Samples were run in parallel on
denaturing PAGE.
RNase A and RNase T1 are shown in Fig. 7,b. Both enzymes are specific for the natural
stereoisomer of RNA. Fig. 7,c shows RNA cleavage induced by the catalyst 4b. The
patterns of both stereoisomers are indistinguishable. It can be safely concluded,
therefore, that the active catalyst is compound 4b itself and not a protein contaminant.

Helvetica Chimica Acta Vol. 86 (2003)
3747
Fig. 7. Cy5-TAR RNA and Cy5-enantio-TAR RNA. a) The cleavage pattern produced by 5b (10 mm) with RNA
14 (lane b) is compared to that with enantio-RNA 15 (lane e). Controls contain only Cy5-labeled RNA
substrates (14: lane a; 15: lane d). Base ladders are shown in lane c (14) and in lane f (15). b) Cy5-labeled TAR
(lanes a and b) and enantio-TAR RNA ( lanes c and d) were incubated with 10 ng RNase A and 1 unit RNase T1,
respectively. c) Nonspecific base-cleavage pattern produced by 4b with TAR RNA 14 or enantio-TAR RNA 15.
Samples were run in parallel on denaturing PAGE.
Discussion. At first glance, the characterization of an artificial ribonuclease seems
not to be a challenging task; many examples have been described and major problems
have rarely been reported. However, as shown above, the risk of producing artefacts is
high. While proper laboratory techniques and ultrafiltration help to minimize protein-
related artefacts, contamination may not be strictly ruled out. Interpretation of

3748
Helvetica Chimica Acta Vol. 86 (2003)
experiments performed with low concentrations of long RNA×s incubated over many
hours is treacherous. When selectivity for pyrimidines is observed, as is common in the
literature, contamination should be suspected. Enantiomeric RNA is a powerful tool in
these cases to separate true catalysis from artefacts when achiral catalysts are used.
Studies with chiral catalysts ideally should compare the enantiomorphous systems
catalyst/RNA and ent-catalyst/ent-RNA leading to identical cleavage patterns.
Unfortunately, enantiomeric RNA is a precious material not routinely accessible to
every researcher. As an alternative, experiments with RNase inhibitors may be
considered, although variable degrees of inhibition render data interpretation less
straightforward.
A disquieting question remains to be answered. What is the agent responsible for
RNA degradation in the presence of trisguanidine 5b, and what is its source? RNA
Samples showed no trace of hydrolysis without 5b. Buffer and substrate, therefore,
seem to be free of ribonucleases. The amount of cleavage observed correlates with the
concentration of −catalyst×. Ultrafiltration, on the other hand, should remove proteins
from solutions of 5b. A possible hint comes from experiments with the benzimidazole
catalyst 4b. When RNA degradation was measured as a function of catalyst
concentration, a sudden loss of activity was found within a narrow concentration
range around 100 mm [7]. Above the borderline, formation of large aggregates of 4b
and the stable RNA model 17 could be observed by fluorescence correlation
spectroscopy (FCS) [7][22]. The diffusion time of 17 calculated from the autocorre-
lation function was 160 ms in pure buffer (Fig. 8,a). The data could be nicely fitted to a
one-component model (Fig. 8,b and c). Upon addition of 4b, this value increased
dramatically (Fig. 8,d).
Because of the heterogeneous nature of the aggregates, a fit of experimental and
calculated data was no longer possible (Fig. 8,e and f). Aggregation is accompanied by
denaturation of RNA secondary structure. At the borderline, however, the non-
selective-cleavage pattern changes to a pyrimidine-selective pattern, reflecting the
stem-bulge-loop secondary structure of 13 (Fig. 9,b). At slightly reduced concen-
trations of 4b, all cleavage effects disappear. When repeated with enantiomeric RNA
15 (not shown), the same correlation of catalyst concentration and cleavage was
observed, however, with one exception. Pyrimidine selective hydrolysis did not occur at
any stage. A possible explanation is based on the assumption that minor traces of
ribonuclease invisible in the controls might be captured and enriched to significant
local concentration by the aggregates of RNA and compound 4b. At high concen-
trations, ribonucleases might be either denatured by the aggregates or their activity is
surpassed by the efficient catalyst 4b. At low concentration, aggregates disappear, thus
abolishing colocalization of ribonucleases and their substrates. However, when the
interaction of 5b and oligonucleotide 17 (or a mixture of 16 and 17) was studied by
FCS, no aggregates were observed up to 10 mm of 5b (not shown). A second tentative
explanation is based on the assumption that covering a strand of RNA with
guanidinium salts such as 4b and 5b promotes in some way the complexation with
ribonucleases. What both working hypotheses have in common is the idea of induced
colocalization of ribonucleases and nucleic acids. This phenomenon should be
detectable by FCS. The objectives of our present work are to investigate the
mechanism of benzimidazole catalyzed RNA hydrolysis, to synthesize conjugates of

Helvetica Chimica Acta Vol. 86 (2003)
3749
Fig. 8. Fluorescence-correlation spectroscopy (FCS). Testing for aggregation of oligonucleotides by 4b and 5b.
Diffusion of Cy5-labeled T20U molecules (17) through the confocal volume element was measured as
fluorescence-intensity fluctuations. Count rate vs. time graphs are shown in a) (buffer control) and d) (250 mm
*
4b), and the corresponding autocorrelation curves (– observed; calculated) are shown in b) (buffer control)
and e) (250 mm 4b). The difference between observed and calculated data points is presented as residuals in c)
and f). Upon addition of 5b up to 10 mm (not shown), the count rate vs. time graph shown in a) did not change
significantly. The constant diffusion times rule out aggregation phenomena.
benzimidazoles and specific RNA ligands, and to decipher the exact cause of the
artefacts described above.
Experimental Part
General. All reagents were of the highest grade commercially available. Chemicals for polyacrylamide gel
electrophoresis (PAGE) were purchased from Roth (Karlsruhe, Germany). NAP-10 Columns, blue dextran
2000, and the N-hydroxysuccinimide active ester of Cy5 (PA15101) were purchased from Amersham Biosciences
(Uppsala, Sweden). Amino linker was obtained from Glen Research (Sterling, VA, USA), recombinant
¾
ribonuclease inhibitor RNasin (N2511) was from Promega, RNase A from Sigma-Aldrich, and RNase T1 from
MBI Fermentas. TLC: glass plates precoated with silica gel F 254/366 (0.25 mm, Macherey-Nagel). M.p.: Kofler
hot-plate microscope, uncorrected. FT-IR: Perkin-Elmer 1600; n in cmÀ1. 1H-NMR: Bruker AM-250 or AMX-
400; chemical shifts (d) in ppm relative to Me4Si (0.00 ppm) or (D5)DMSO (2.50 ppm) as internal standards, J in
Hz. ESI-MS: Fisons VG Platform II. Elemental analysis: Heraeus HCN-Rapid.

3750
Helvetica Chimica Acta Vol. 86 (2003)
Fig. 9. Concentration-dependent cleavage pattern induced by 4b. a) Substrate 7. All reactions shown were
performed at 120 mm (lane a) and 100 mm (lane b) catalyst concentration in 50 mm Tris-HCl, 0.01% SDS buffer
(pH 6.0) under standard RNA-cleavage-assay conditions. b) Analogous experiments with TAR-RNA 13. RNA
fragmentation was analyzed as described in Exper. Part. Cleavage sites indicated by arrows are given in bold
letters. 7: Cy5-T10-CUAGCCGACUGCCGAUCUCGCUGACUGAC; 13: Cy5-T20-GGCCAGAUCUGAGC-
CUGGGAGCUCUCUGGCC.
Tris{2-[(4,5-dihydro-1H-imidazol-2-yl)amino]ethyl}amine, Salt with 2,4,6-trinitrophenol (5a). Tris(2-amino-
ethyl)amine (0.150 ml, 1.00 mmol) was added at r.t. to a soln. of 4,5-dihydro-1H-imidazole-2-sulfonic acid (6)
[11] (496 mg, 3.30 mmol) in 5 ml MeOH and 3 ml Et3N. The mixture was stirred overnight and then diluted with
10 ml H2O and 4 ml aq. NaOH soln. Extraction with CH2Cl2 removed Et3N. The aq. phase was concentrated in
vacuo to 50% and acidified with 5 ml 6m HCl (with release of SO2). Complete evaporation and drying in vacuo
yielded a mixture of product and NaCl. Product was extracted from the solid with 3 portions of boiling EtOH.
The filtered EtOH solns. were then evaporated. After dissolving the residue in a small volume of MeOH, a sat.
hot soln. of 2,4,6-trinitrophenol (1.18 g, 40% H2O content, 3.10 mmol) in MeOH was added. The yellow
precipitate was isolated, suspended in acetone, isolated again, and dissolved in MeOH. Insoluble, sticky material
was discarded. Three recrystallizations from MeOH/H2O finally yielded 272 mg of pure 5a (0.260 mmol, 26%).
M.p. 2478. IR (KBr): 3288m, 3086w, 2958w, 2901w, 1674s, 1637s, 1602s, 1566s, 1430m, 1364m, 1333s, 1267s,
1162m, 1078m, 909w, 787w, 711m. 1H-NMR ((D6)DMSO, 250 MHz): 2.63 (t, J ˆ 6.5, 3 CH2); 3.61 (dt, J ˆ 6.4/6.1,
t after addition of D2O, 3 CH2); 3.60 (s, 6 imid. CH2); 7.4 8.5 (br. s, exchangeable with D2O, NH); 7.90 (t, J ˆ 5.3,
exchangeable with D2O, 3 NH); 8.58 (s, 6 arom. H, 2,4,6-trinitrophenol). Anal. calc. for C33H39N19O21: C 38.19,
H 3.79, N 25.64; found: C 37.99, H 3.91, N 25.48.
Tris{2-[(4,5-dihydro-1H-imidazol-2-yl)amino]ethyl}amine, Hydrochloride (5b). A small column was filled
with anion-exchange resin (Dowex 1 Â 8, ClÀ, 200 400 mesh) and rinsed with MeOH. Filtration of 100 mg
2,4,6-trinitrophenol salt 5a (0.096 mmol), dissolved in a small volume of MeOH, through this column afforded a
colorless soln. of hydrochloride 5b (43 mg after drying in vacuo; 96%). Colorless glass. 1H-NMR ((D6)DMSO,
400 MHz): 2.62 (t, J ˆ 6.0, 3 CH2); 3.30 (q, J ˆ 5.9, t after addition of D2O, 3 CH2); 3.60 (s, 6 imid. CH2); 7.7 9.0
(br. s, exchangeable with D2O, 6 NH); 8.26 (t, J ˆ 5.5, exchangeable with D2O, 3 NH). Anal. calc. for
C15H33Cl3N10: C 39.18, H 7.23, N 30.46; found: C 38.15, H 7.38, N 27.39. Note: correct analyses could not be
obtained due to the hygroscopic nature of 5b.
5'-O-[Bis(4-methoxyphenyl)(phenyl)methyl]-2'-O-[P-(2-cyanethoxy)-P-(N,N-diisopropylamino)phosphi-
no]-3'-O-[dimethyl(tert-butyl)silyl]uridine. A mixture of 5'-O-[bis(4-methoxyphenyl)(phenyl)methyl]-3'-O-
[dimethyl(tert-butyl)silyl]uridine [23] (3.30 g, 5.00 mmol) and 4-dimethylaminopyridine (60 mg, 0.50 mmol)
was coevaporated with dry THF (5 ml). The solid was dissolved in dry THF (20 ml) under Ar and freshly
distilled H¸nig×s base (3.40 ml, 16.0 mmol) was added. After the dropwise addition of 2-cyanoethoxydiisopro-

Helvetica Chimica Acta Vol. 86 (2003)
3751
pylaminophosphine (1.41 g, 6.00 mmol), the soln. was stirred for 4 h at r.t. 150 ml 5% aq. NaHCO3 soln. was
added and the mixture was extracted three times with 100 ml AcOEt. The combined org. layers were dried
(MgSO4), the solvent was evaporated and the residue was purified by flash chromatography (FC) (CH2Cl2/
hexane 1 :1 ‡ 5% NEt3). After drying in vacuo, 3.77 g (4.30 mmol, 86%) of the title compound (as a mixture of
two diastereoisomers) was obtained as a colorless foam. TLC: AcOEt/hexanes 1:1 ‡ 3% NEt3, Rf ˆ 0.43, 0.50.
IR (KBr): 3412w, 3199m, 3061w, 2963s, 2931s, 2857w, 1696s, 1608m, 1582w, 1560w, 1509s, 1460s, 1202w, 1178m,
1153w, 1111w, 1069w, 1036m, 980m, 899w, 864w, 834s, 779m, 726w, 703w, 671w, 640w, 583m. 1H-NMR
((D6)DMSO, 400 MHz, 2 diastereoisomers): À0.04, À0.02 (2s, SiMe2); 0.02, 0.07 (2s, SiMe2); 0.76, 0.77 (2s, 2 t-
Bu); 1.08, 1.12 (2d, J ˆ 6.7, 4 CH(Me)2); 2.71 (2t, J ˆ 5.8, 2, CH2CN); 3.19, 3.41 (2m, 4 CH(Me)2); 3.59 (2m, 2
H2C(5')); 3.74 (2s, 4 MeO); 3.95 (2m, 2 OCH2); 4.23 (2m, 2 HÀC(4')); 4.31 (2m, 2 HÀC(3')); 4.44 (2m, 2
HÀC(2')); 5.38, 5.44 (2d, J ˆ 8.0, 2 HÀC(5)); 5.87, 5.94 (2d, J ˆ 4.0, 2 HÀC(1')); 6.90 (2d, J ˆ 8.5, 8 arom. H);
7.20 7.38 (2m, 18 arom. H); 7.78, 7.84 (2d, J ˆ 8.0 Hz, 2 HÀC(6)); 11.37 (s, 2 HÀN(3)). 31P-NMR ((D6)DMSO,
400 MHz): 150.6, 150.5 ratio 1 :1. ESI-MS: 861.8 (calc. 861.1). Anal. calc. for C46H65N4O9PSi: C 62.99, H 7.47,
N 6.39; found C 62.90, H 7.17, N 6.51.
Assembly of Oligonucleotides. Oligonucleotides 9 12, 16, and 17 were assembled on an Applied Biosystems
381A DNA synthesizer by standard phosphoramidite chemistry with 2'-silylated monomers, and with 3'-silylated
reagent (see above) for 10 and 12. Compounds 7, 8, 13, and 14 were purchased from Biospring (Frankfurt,
Germany). PAGE was used to purify all oligonucleotides except 15. Compound 15, containing the enantio-RNA
sequence was synthesized from 2'-O-tom-protected monomers [21] with an C6-amino linker attached to the 5'
end. Coupling conditions for the active ester of Cy5: aq. borate buffer pH 8.5 (H2O/DMF 1:1) 258, 1 h. The final
conjugate was purified by anion-exchange HPLC.
Inactivation of RNases [14]. RNA Cleavage was performed in an RNase-free environment. Glassware was
baked for 6 h at 1808. During all experimental steps, gloves were worn and, when handling the reagents, sterile
techniques were used. Plasticware and tubes were treated with DEPC to ensure that they were RNase-free.
Solns. were prepared by mixing molecular-biology-grade powdered reagents in DEPC-treated ultrapure H2O.
Ultrafiltration. To remove possible RNase contamination from catalysts, an ultrafiltration step (Microcon
YM3, Millipore) was carried out as discribed by the manufacturer. After ultrafiltration, the filtrate was used for
RNA-cleavage experiments. The retentate was discarded.
Purification of Cy5-Labeled Oligonucleotides. The Cy5-labeled oligonucleotides were purified by
denaturing PAGE (16% monomer, 7m urea) to remove prematurely terminated products. After PAGE, the
bands of interest were excised; the gel fragments transferred to a nuclease-free tube and submerged with elution
buffer (500 mm NH4OAc, 0.1% SDS, 2 mm EDTA) and incubated under vigorous shaking overnight at r.t. To
remove the gel fragments, −Quantum Prep Freeze×N Squeeze× spin columns (BioRad, Munich, Germany) were
used. Desalting of the Cy5-labeled oligonucleotides was carried out in two steps: after YM3 ultrafiltration, the
retentate was diluted to 1 ml with DEPC-treated H2O and loaded onto a NAP-10 column. The pooled fractions
were lyophilized to dryness, and the pellet was dissolved in DEPC-treated H2O to give a conc. of ca. 0.5 mg/ml.
PAGE. Prior to electrophoresis, one volume of loading buffer (5 mg/ml blue dextran in formamide) was
added to each sample, and 10 ml aliquots were loaded on the gel. The oligonucleotide fragments were separated
by denaturing PAGE (16% monomer, 7m urea) on a DNA-sequencing device (ALFexpress, Amersham
Biosciences). Running conditions were 1500 Vand 60 mA maximum at constant 30 W, 558, 2 s sampling interval,
and 350 min running time. Electropherograms were analyzed with the AlleleLinks 1.01 software package
(Amersham Biosciences, Uppsala, Sweden). The peak areas under the curves were added, and the percentage of
degraded RNA was calculated. Multiple cleavage reactions were disregarded in this system. All data were
averaged over a minimum of two experiments.
RNA-Cleavage Assay. The 10 ml assays contained the indicated catalyst concentration (0.1 10 mm) and
120 140 nm of the appropriate Cy5-labeled RNA (7 15). A 50 mm Tris-HCl buffer containing 0.01% SDS was
used: cleavage reactions with guanidinium compounds were carried out at pH 7.0, and with benzimidazole
compounds at pH 6.0. The SDS prevents nonspecific binding of the RNA to the tube walls. Incubation was
¾
performed at 378 for 16 20 h. The final conc. of ribonuclease inhibitor (RNasin ), if applied, was 1 unit per ml.
RNase T1 Treatment: To 0.5 pmol Cy5-labeled RNA in RNase T1 incubation buffer (10 mm Tris-HCl,
0.02 mm EDTA, 0.3m NaCl; pH 8.0), 1 unit RNase T1 was added. After incubation for 12 min at 378, the
reaction was stopped by adding 10 ml of ALF loading buffer; 10 ml aliquots were loaded on the gel (denaturing
16% polyacrylamide (PAA) gel, electrophoresis conditions as described above).
RNase A Treatment. In a final volume of 10 ml, 0.5 pmol Cy5-labeled RNA was mixed with 10 ng RNase A
in RNase A incubation buffer (6 mm Tris-HCl, 14% DMSO, pH 7.6). After incubation for 5 min at 378, the
reaction was stopped by adding 10 ml of ALF loading buffer, and 10 ml aliquots were loaded on the gel.

3752
Helvetica Chimica Acta Vol. 86 (2003)
FCS Instrument. FCS Measurements were carried out with a ConfoCor 2 (Carl Zeiss, Jena, Germany)
equipped with an Axiovert 200 M microscope containing a laser-adapted Zeiss C-Apochromat 40 Â /1.2 W corr
H2O-immersion objective. Cover slips (24 Â 60 mm, Roth, Karlsruhe, Germany) served as sample carriers, and,
for calibration of the instrument, free Cy5 dye was used. Fluctuation measurements, which are calculated in real
time to give an autocorrelation curve, and further analysis, like determination of the average number and the
diffusion time of the fluorescent particles in the confocal volume, were performed with the Fluorescence
Correlation Microscope ConfoCor 2 Software version 3.2 SP1.
Sample Measurements. FCS Measurements were performed under conditions comparable to those used for
RNA-cleavage experiments. The RNA substrates were replaced by a mixture of a Cy5-labeled T20U probe 17
and an unlabeled DNA oligonucleotide 16. A final volume of 30 ml contained 25 nm Cy5-labeled T20U, 175 nm
unlabeled DNA oligonucleotide, 500 62.5 mm 4 or 10 1.25 mm 5b in 50 mm Tris-HCl, 0.01% SDS buffer
(pH 6.0). The incubation step (20 h, 378) was omitted, and all FCS measurements were performed at 248. A He/
Ne laser at 633 nm was used as the excitation source. Each sample (30 ml droplet) was measured either five times
for 45 s or ten times for 10 s.
REFERENCES
[1] B. Trawick, A. T. Daniher, J. K. Bashkin, Chem. Rev. 1998, 98, 939.
[2] D. M. Perreault, L. A. Cabell, E. V. Anslyn, Bioorg. Med. Chem. 1997, 5, 1209.
[3] D. M. Perreault, E. V. Anslyn, Angew. Chem. 1997, 109, 471; Angew. Chem., Int. Ed. 1997, 36, 432.
[4] J. C. Verheijen, B. A. L. M. Deiman, E. Yeheskiely, G. A. van der Marel, J. H. van Boom, Angew. Chem.
2000, 112, 377; Angew. Chem., Int. Ed. 2000, 39, 369.
[5] R. Gross, G. D¸rner, M. W. Gˆbel, Liebigs Ann. Chem. 1994, 49.
[6] K. Kurz, M. W. Gˆbel, Helv. Chim. Acta 1996, 79, 1967.
[7] U. Scheffer, A. Strick, E. Kalden, M. W. Gˆbel, in preparation.
[8] D. Brown, B. L. Pasloske, Meth. Enzymol. 2001, 341, 648.
[9] S. Klussmann, A. Nolte, R. Bald, V. A. Erdmann, J. P. F¸rste, Nature Biotechnol. 1996, 14, 1112; A. Nolte, S.
Klussmann, R. Bald, V. A. Erdmann, J. P. F¸rste, Nature Biotechnol. 1996, 14, 1116.
[10] S. Pitsch, Helv. Chim. Acta 1997, 80, 2286.
[11] M.-S. Muche, M. W. Gˆbel, Angew. Chem. 1996, 108, 2263; Angew. Chem., Int. Ed. 1996, 35, 2126.
[12] B. Dietrich, D. L. Fyles, T. M. Fyles, J.-M. Lehn, Helv. Chim. Acta 1979, 62, 2763.
[13] A. Scarso, U. Scheffer, M. W. Gˆbel, Q. B. Broxterman, B. Kaptein, F. Formaggio, C. Toniolo, P. Scrimin,
Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5144.
[14] U. Scheffer, M. W. Gˆbel, in −DNA Synthesis: Methods and Protocols×, −Methods in Molecular Biology×
Series, Ed. P. Herdewijn, Humana Press, Totowa, in press.
[15] C. Schmidt, R. Welz, S. M¸ller, Nucleic Acids Res. 2000, 28, 886.
[16] J. Hall, D. H¸sken, U. Pieles, H. E. Moser, R. H‰ner, Chem. Biol. 1994, 1, 185.
[17] J. Karn, J. Mol. Biol. 1999, 293, 235.
¬
[18] M. A. Podyminogin, V. V. Vlassow, R. Giege, Nucleic Acids Res. 1993, 21, 5950.
¬
[19] M. A. Zenkova, N. G. Beloglazova, V. N. Sil×nikov, V. V. Vlassov, R. Giege, Meth. Enzymol. 2001, 341, 468.
[20] N. Q. Nguyen-Trung, O. Botta, S. Terenzi, P. Strazewski, J. Org. Chem. 2003, 68, 2038.
[21] S. Pitsch, P. A. Weiss, L. Jenny, A. Stutz, X. Wu, Helv. Chim. Acta 2001, 84, 3773.
[22] J. D. M¸ller, Y. Chen, E. Gratton, Meth. Enzymol. 2003, 361, 69.
[23] K. K. Ogilvie, S. L. Beaucage, A. L. Schifman, N. Y. Theriault, K. L. Sadana, Can. J. Chem. 1978, 56, 2768.
Received August 25, 2003