Expanding functionality of RNA: Synthesis and properties of RNA containing imidazole modified tandem G-U wobble base pairs

Article abstract of DOI:10.1039/b510846b

Imidazole modification at C-5 of uridine that is part of tandem G-U wobble base pairs causes slight reduction of thermal stability (ΔΔG ⁰³¹⁰ < 0.4 kcal mol^-1) and relatively small change in hydration of short RNA helices. The Royal S

Full text of DOI:10.1039/b510846b

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Expanding functionality of RNA: synthesis and properties of RNA
containing imidazole modified tandem G–U wobble base pairs{
Eriks Rozners,* Romualdas Smicius and Chika Uchiyama
Received (in Cambridge, UK) 29th July 2005, Accepted 27th September 2005
First published as an Advance Article on the web 20th October 2005
DOI: 10.1039/b510846b
distort the regular helices, exposing heterocyclic bases. Such motifs
have certain functional groups, which are usually involved in the
Watson–Crick base pairing, available in the expanded grooves.
Therefore, we propose that non-canonical base pairs are ideal for
engineering of functional groups for catalysis and molecular
recognition. Herein, we report the synthesis and properties of
RNA containing imidazole attached directly to 5-position of
uracyl heterocycles of tandem G–U wobble base pairs (Fig. 1).
Although a variety of functional groups have been conjugated with
heterocycles of Watson–Crick base pairs,2–5 chemical modifica-
tions of non-canonical base pairs are virtually unexplored.
The modified uridine was prepared using a palladium catalyzed
coupling of 5-iodouridine 1 and 4-tributylstannyl imidazole 2
(Scheme 1).5,6 Both (Ph3P)2PdCl2 and (Ph3P)4Pd were good
catalysts with the latter giving slightly higher yields and cleaner
coupling. In contrast to similar literature precedents,5a efficient
coupling was achieved without silver(I) oxide additive. For the
protection of imidazole we chose the trityl group. In 3%
dichloroacetic acid, which is used to remove the 59-O-methoxytrityl
group during the RNA solid-phase synthesis, only traces of
deprotected 3 could be observed after 30 min (TLC). On the other
hand, in 2% trifluoroacetic acid the half-life of the N-trityl
protection in 3 was less than 15 min (TLC). Thus, we expected that
the N-trityl group would be stable during the solid phase synthesis
and could be removed using a stronger acid at the end of the
synthesis.
Imidazole modification at C-5 of uridine that is part of tandem
G–U wobble base pairs causes slight reduction of thermal
stability (DDG0310 , 0.4 kcal mol21) and relatively small change
in hydration of short RNA helices.
The ability of RNA to catalyze chemical reactions is truly
remarkable. However, compared to proteins RNA is an inherently
much less efficient catalyst.1 Whereas proteins are built with
twenty amino acids having a variety of functional groups, RNA is
built with only four nucleosides with limited functionality. Most
notably, RNA lacks positively charged functional groups, which
are important for catalysis in proteins. Several laboratories have
designed modified nucleoside triphosphates (mostly 5-substituted
uridines) to incorporate amino2,3 and imidazole3,4 groups in RNA
and DNA via polymerase-catalyzed reactions and artificial in vitro
selection techniques. In several instances, the activity of the selected
nucleic acid catalysts for RNA cleavage4a and amide synthesis4b
reactions critically depended on the presence of the imidazole-
modified uridines. The limitation of such an approach is that the
polymerase replaces every uridine with the modified nucleoside,
which may have an undesirable effect on folding of a large nucleic
acid.
Due to rapid advances in understanding of nucleic acid structure
and folding, an alternative approach of rational engineering of
nucleic acid functionality at selected sites becomes a conceivable
and attractive goal. However, the nucleoside heterocyclic bases are
buried deep inside the canonical double helix. To reach the surface
of the helix, the functional groups attached to the heterocycles of
Watson–Crick bases pairs require relatively long linkers2–4 that
may not be ideal for rational structure engineering. In contrast,
non-canonical base pairs (e.g., G–U wobble in Fig. 1) tend to
The 59-OH of the 5-imidazolyl uridine 3 was protected with the
4-monomethoxytrityl group (MMT). Selective benzoylation of the
29-OH in 4 was immediately followed by phosphonylation of
the 39-OH in the same reaction mixture following our previously
published procedures.7,8 Isomerically pure 29-O-(2-chloroben-
zoyl)-39-O-(H-phosphonate) 5 was obtained after silica gel column
chromatography.
Fig. 1 Compared to an A–U base pair, 5-imidazole of a G–U wobble
pair is shifted out in the major groove.
Department of Chemistry and Chemical Biology, Northeastern
University, Boston, Massachusetts, 02115, USA.
Scheme 1 Synthesis of H-phosphonate 5. Reagents and conditions: (a)
(Ph3P)4Pd, 2, dioxane, 80 uC, 46 h, 74%; (b) 4-methoxytrityl chloride,
pyridine, rt, overnight, 72%; (c) 2-chlorobenzoyl chloride, pyridine/CH2Cl2
(1:9), 278 uC, 45 min; (d) imidazole, PCl3, NEt3, CH2Cl2, 278 uC, 1 h,
63% (two steps).
E-mail: e.rozners@neu.edu; Fax: 1-617-373-8795; Tel: 1-617-373-5826
{ Electronic supplementary information (ESI) available: Experimental
procedures, detailed results of CD, thermal melting, and osmotic stressing
1
experiments, and H and 13C NMR data. See DOI: 10.1039/b510846b
5778 | Chem. Commun., 2005, 5778–5780
This journal is ß The Royal Society of Chemistry 2005

View Article Online
To test our hypothesis that the imidazole heterocycle can be
engineered in the major groove of the non-canonical tandem G–U
wobble base pairs (Fig. 1) we prepared the self complementary
sequences 6i–9i (Table 1) using standard H-phosphonate oligori-
bonucleotide procedures7–9 and the imidazole modified monomer
5. The average coupling yields (typically .90% per step) were
similar for 5 and the unmodified H-phosphonates. Deprotection of
the imidazole was achieved by a prolonged treatment with 3%
trifluoroacetic acid in the presence of triethylsilane. The overall
yields, HPLC traces, and purity of the modified oligoribonucleo-
tides were similar to native RNA prepared in our laboratory; the
chemical composition of imidazole modified RNA was confirmed
by MALDI-TOF mass spectrometry (for experimental details,
see ESI{).
We chose 6–9 as the model compounds because Turner and co-
workers have previously shown that these sequences display well-
organized structure by NMR and a well-defined two-state melting
behavior by UV spectroscopy.10 In general, the 59-UG motifs are
thermally more stable than the 59-GU motifs.10 The NMR
structure of 7 shows that the 59-CGUG motif has only one
hydrogen bond per G?U base pair.10a Consistent with previous
studies,10 the UV thermal melting of 6–9 and 6i–9i gave good
quality two-state melting curves. The results (Table 1) revealed that
the imidazole modifications in 6i, 7i, and 9i caused only a slight
destabilization of 0.5–0.8 kcal mol21 (Dtm 5 20.9 to 21.7 uC, per
modification). An exception was 8i where the imidazole modifica-
tions resulted in a 2.2 kcal mol21 loss of thermal stability
(Dtm 5 24.6 uC, per modification). These results were somewhat
unexpected, because a structurally similar imidazole modification
in DNA was reported to increase the stability of RNA–DNA
heteroduplexes (Dtm 5 +0.7 uC, per modification).5a We have
previously observed that chemical modifications cause remarkably
different responses in RNA vs. DNA.8 To gain more insight into
the effect of the imidazole modification on the tandem G–U
motifs, we probed hydration of 6–9 using osmotic stress.
where 2DH is the enthalpy determined from the width at the half-
height of differentiated melting curves,13 R is the universal gas
constant (1.986 cal mol21 K), and d(Tm21)/d(ln aW) is the slope of
the plot of reciprocal melting temperature vs. the logarithm of
water activity. Water activity is modulated by addition of low
molecular weight co-solutes as osmotic stressors: ethylene glycol,
glycerol, and acetamide (for details, see ESI{).
Spink and Chaires used osmotic stress to probe the hydration of
long DNA molecules and found that about four water molecules
per base pair were released upon melting of long DNA duplexes
(E. coli DNA and poly(dA)–poly(dT)).12 Recently, we demon-
strated that osmotic stress can also detect relative changes of
hydration caused by subtle structural differences in short DNA,
RNA and 29-OMe oligonucleotides.14 Although the absolute
numbers of water molecules obtained using different osmotic
stressors is still under debate,15 osmotic stress generally gives
reliable relative estimates of the trends in hydration for a series of
similar compounds.
Consistent with the previous studies,12,14 we observed a good
correlation of DnW for ethylene glycol and glycerol, whereas
acetamide gave somewhat higher DnW numbers (Table 1). The
differences in DnW numbers obtained with different osmotic
stressors may be in part because of specific solute–oligonucleotide
interactions.15 However, comparisons within each series of osmotic
stressors were consistent with the general trends discussed bellow.
The errors for DnW were estimated using standard deviations of
2DH0 and alternative graph fitting as described in the ESI.{
Although these conservative error estimates are somewhat large,
the consistency of the data among several osmotic stressors
suggests that the differences are meaningful and allows for the
following qualitative discussion of the trends.
The non-covalent interactions that determine the conformation
and stability of RNA structure are not completely understood.
NMR structures obtained by Turner and co-workers10 show that
7–9 form similar helices, where the only significant departure from
the A-type geometry is a slight overtwisting 59 of the G and
displacement of the G–U wobble pairs toward the major groove
(Fig. 1). The different sequence dependent thermal stabilities
appear to be due to subtle differences in electrostatic stacking
interactions and hydrogen bonding. On the other hand, the
importance of hydration for the conformation and stability of the
G–U wobble pairs is little studied. Crystal structures of
oligoribonucleotides show that the G–U wobble pairs are
extensively hydrated in both grooves.16 Analysis of the hydration
pattern in the major groove,16 suggests that the 5-imidazole
The osmotic stress is a thermodynamic method to evaluate the
trends in hydration associated with biologically relevant equilibria
from the dependence of the equilibrium constant on osmotic
pressure (water activity).11 For nucleic acids,12 osmotic stress
monitors the depression of the melting temperature upon
decreasing of the water activity and calculates the number of
water molecules (Table 1, DnW) associated with the double helix
and released from the single strands upon melting:
DnW 5 (2DH/R)[d(Tm21)/d(ln aW)]
(1)
Table 1 Thermal melting and hydration of oligoribonucleotides having tandem G–U base pairs
a
310
kcal mol21
2DG0
/
2DH0a
/
2DS0a
/
Ethylene
glycol DnW
Glycerol
DnW
Acetamide
DnW
Sequence
tm a/uC
kcal mol21
kcal mol21 K21
b
b
b
6
6i
7
7i
8
8i
9
GGCUGGCC
GGCUImGGCC
GGCGUGCC
GGCGUImGCC
GAGUGCUC
GAGUImGCUC
GAGGUCUC
GAGGUImCUC
51.8 ¡ 0.1
49.6 ¡ 0.2
42.9 ¡ 0.2
41.1 ¡ 0.3
39.3 ¡ 0.3
30.1 ¡ 0.2
32.6 ¡ 0.3
29.2 ¡ 0.2
12.1 ¡ 0.1
11.3 ¡ 0.2
9.5 ¡ 0.2
9.0 ¡ 0.2
8.6 ¡ 0.2
6.4 ¡ 0.3
6.8 ¡ 0.1
6.0 ¡ 0.1
84.5 ¡ 2.3
80.5 ¡ 3.0
75.1 ¡ 2.7
67.2 ¡ 4.7
75.2 ¡ 3.4
69.3 ¡ 4.6
78.7 ¡ 2.8
74.3 ¡ 2.6
245 ¡ 5
236 ¡ 10
220 ¡ 9
197 ¡ 15
233 ¡ 14
238 ¡ 17
267 ¡ 13
235 ¡ 10
30.4 ¡ 2.2
27.2 ¡ 2.2
26.7 ¡ 3.7
29.4 ¡ 4.7
27.7 ¡ 4.1
20.1 ¡ 3.4
22.4 ¡ 4.6
27.2 ¡ 4.9
27.2 ¡ 4.0
21.4 ¡ 3.7
31.1 ¡ 4.8
31.6 ¡ 5.8
24.6 ¡ 6.1
23.7 ¡ 4.1
20.3 ¡ 5.0
21.6 ¡ 6.3
56.7 ¡ 2.7
43.0 ¡ 3.2
58.2 ¡ 3.4
56.3 ¡ 5.5
51.5 ¡ 5.0
33.1 ¡ 4.4
52.1 ¡ 4.8
70.4 ¡ 4.4
9i
a
b
Oligonucleotide (2 mM) in 10 mM sodium cacodylate, 0.1 mM EDTA, and 300 mM NaCl. Results ¡ standard deviations. Results ¡ error
estimates.
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 5778–5780 | 5779

View Article Online
modification may sterically exclude at least one water molecule out
of the five ‘‘in-plane’’ hydration sites (see discussion and Fig. S2
in ESI{).
We thank Northeastern University, The Petroleum Research
Fund, administered by the American Chemical Society (37599-
AC1) and NSF-NATO (postdoctoral fellowship DGE-0209488 to
R. S.) for support of this research. We thank Heather Brodkin,
James Glick, Dr Norman Chiu and Dr Paul Vouros for MALDI-
TOF MS analysis and Dr James Manning for help with CD
spectroscopy.
In general, the changes in hydration caused by the imidazole
modification were relatively small. For the relatively more stable
59-UG motifs 6 and 8, the imidazole modification resulted in a loss
of several water molecules per G–U base pair. For the 59-GU
motifs, imidazole modification caused either small changes (7 to 7i)
or increased the hydration (9 to 9i). It is conceivable that more
extensive hydration may contribute to higher thermal stability and
conformational rigidity of 59-UG compared to 59-GU motifs.
Interestingly, the most significant loss in thermal stability (8 to 8i)
was also accompanied by the overall largest decrease in DnW.
However, the relationship between tm and DnW was not entirely
consistent across the Table 1. More structural data will have to be
obtained on hydration of G–U pairs in different sequence contexts
to probe the role of hydration in conformation and thermal
stability of these motifs.
Notes and references
1 G. J. Narlikar and D. Herschlag, Annu. Rev. Biochem., 1997, 66, 19.
2 (a) A. Roychowdhury, H. Illangkoon, C. L. Hendrickson and
S. A. Benner, Org. Lett., 2004, 6, 489; (b) T. R. Battersby, D. N. Ang,
P. Burgstaller, S. C. Jurczyk, M. T. Bowser, D. D. Buchanan,
R. T. Kennedy and S. A. Benner, J. Am. Chem. Soc., 1999, 121, 9781;
(c) N. K. Vaish, R. Larralde, A. W. Fraley, J. W. Szostak and
L. W. McLaughlin, Biochemistry, 2003, 42, 8842; (d) N. K. Vaish,
A. W. Fraley, J. W. Szostak and L. W. McLaughlin, Nucleic Acids Res.,
2000, 28, 3316.
3 (a) S. E. Lee, A. Sidorov, H. Gourlain, N. Mignet, S. J. Thorpe,
J. A. Brazier, M. J. Dickman, D. P. Hornby, J. A. Grasby and
D. M. Williams, Nucleic Acids Res., 2001, 29, 1565; (b) T. Gourlain,
A. Sidorov, N. Mignet, S. J. Thorpe, S. E. Lee, J. A. Grasby and
D. M. Williams, Nucleic Acids Res., 2001, 29, 1898; (c) L. Lermer,
Y. Roupioz, R. Ting and D. M. Perrin, J. Am. Chem. Soc., 2002, 124,
9960; (d) D. M. Perrin, T. Garestier and C. Helene, J. Am. Chem. Soc.,
2001, 123, 1556; (e) J. D. Vaught, T. Dewey and B. E. Eaton, J. Am.
Chem. Soc., 2004, 126, 11231; (f) K. Sakthivel and C. F. Barbas III,
Angew. Chem., Int. Ed., 1998, 37, 2872.
4 (a) S. W. Santoro, G. F. Joyce, K. Sakthivel, S. Gramatikova and
C. F. Barbas III, J. Am. Chem. Soc., 2000, 122, 2433; (b) T. W. Wiegand,
R. C. Janssen and B. E. Eaton, Chem. Biol., 1997, 4, 675.
5 (a) A. J. Gutierrez, T. J. Terhorst, M. D. Matteucci and B. C. Froehler,
J. Am. Chem. Soc., 1994, 116, 5540; (b) for a recent review on palladium
assisted synthesis of nucleosides, see: L. A. Agrofoglio, I. Gillaizeau and
Y. Saito, Chem. Rev., 2003, 103, 1875.
6 M. C. Jetter and A. B. Reitz, Synthesis, 1998, 829.
7 E. Rozners, R. Renhofa, M. Petrova, Y. Popelis, V. Kumpins and
E. Bizdena, Nucleosides Nucleotides, 1992, 11, 1579.
8 (a) E. Rozners, D. Katkevica, E. Bizdena and R. Stro¨mberg, J. Am.
Chem. Soc., 2003, 125, 12125; (b) E. Rozners and R. Stro¨mberg, J. Org.
Chem., 1997, 62, 1846.
9 (a) E. Westman, S. Sigurdsson, J. Stawinski and R. Stro¨mberg, Nucleic
Acids Symp. Ser., 1994, 31, 25; (b) J. Stawinski and R. Stro¨mberg, in
Oligonucleotide Synthesis, Methods and Applications, Methods in
Molecular Biology, ed. P. Herdewijn, Humana Press, Totowa, NJ,
2005, vol. 288, ch. 6, pp. 81–100.
10 (a) X. Chen, J. A. McDowell, R. Kierzek, T. R. Krugh and
D. H. Turner, Biochemistry, 2000, 39, 8970; (b) J. A. McDowell and
D. H. Turner, Biochemistry, 1996, 35, 14077; (c) L. He, R. Kierzek,
J. SantaLucia, Jr., A. E. Walter and D. H. Turner, Biochemistry, 1991,
30, 11124.
11 (a) V. A. Parsegian, R. P. Rand and D. C. Rau, Methods Enzymol.,
1995, 259, 43; (b) C. R. Robinson and S. G. Sligar, Methods Enzymol.,
1995, 259, 395.
To gain more insight into the effect of imidazole modification
on tandem G–U wobble base pairs we compared the CD spectra
of 6–8 with 6i–8i (Fig. S3 in ESI{). Most notably, imidazole
modifications that resulted in loss of hydration in 6 and 8 (Table 1)
also caused significant changes in the CD spectra of these
sequences. For 7 and 7i, where imidazole caused the smallest
changes in tm, DG and DnW (Table 1), the CD spectra were also
similar. Thus, our results suggest that the imidazole modification
may fit best in the structurally more flexible G–U motifs, such as
59-CGUG (7), which has only one hydrogen bond per G?U base
pair according to the NMR structure.10a However, more studies
are certainly needed to confirm the generality of this observation.
In summary, our results show that 5-imidazole modification of a
G–U wobble pair is well accepted in most RNA sequence contexts,
except 8. The loss of thermal stability in 6i, 7i and 9i is relatively
small (per modification DDG0310 , 0.4 kcal mol21). The imidazole
modification also causes a small but significant rearrangement of
the hydration of RNA in 6i, 7i and 9i. Sequence 8i represents an
exception where imidazole causes a more significant loss of
thermal stability and hydration. These results suggest that
hydration effects are important to consider when designing
chemically modified nucleic acids. It is possible that the loss of
thermal stability frequently observed for non-polar modifications17
is actually due to poor hydration rather than incompatible
conformation or steric hindrance of the backbone. It is conceivable
that the imidazole modification (which is likely to be partially
protonated due to pKa in vicinity of 7) uses its hydrogen bond
donor/acceptor sites to rearrange the major groove water structure
of G–U base pairs without disrupting the overall hydration of the
duplex. In conclusion, G–U wobble pairs (and perhaps other non-
canonical base pairs as well) can be used to rationally engineer
imidazole incorporation at selected sites in RNA. This may open
new avenues for design of more active nucleic acid catalysts
(ribozymes) and receptors (aptamers) for biomedical and industrial
applications.
12 C. H. Spink and J. B. Chaires, Biochemistry, 1999, 38, 496.
13 K. J. Breslauer, Methods Enzymol., 1995, 259, 221.
14 E. Rozners and J. Moulder, Nucleic Acids Res., 2004, 32, 248.
15 E. S. Courtenay, M. W. Capp, C. F. Anderson and M. T. Record, Jr.,
Biochemistry, 2000, 39, 4455.
16 (a) P. Auffinger and E. Westhof, J. Biomol. Struct. Dyn., 1998, 16, 693;
(b) R. Biswas, M. C. Wahl, C. Ban and M. Sundaralingam, J. Mol.
Biol., 1997, 267, 1149; (c) S. R. Holbrook, C. Cheong, I. Tinoco, Jr. and
S. H. Kim, Nature, 1991, 353, 579.
17 S. M. Freier and K.-H. Altmann, Nucleic Acids Res., 1997, 25, 4429.
5780 | Chem. Commun., 2005, 5778–5780
This journal is ß The Royal Society of Chemistry 2005