ARTICLE pubs.acs.org/JPCB
Low-Temperature NMR Studies on Inosine Wobble Base Pairs Eline M. Basílio Janke,† Fanny Riechert-Krause,‡ and Klaus Weisz*,‡ † ‡
Institut f€ur Chemie, Freie Universit€at Berlin, Takustrasse 3, D-14195 Berlin, Germany Institut f€ur Biochemie, Ernst-Moritz-Arndt-Universit€at Greifswald, Felix-Hausdorff-Strasse 4, D-17487 Greifswald, Germany ABSTRACT: Base pairs formed by the inosine nucleoside (I) play an important role in many physiological processes as well as in various DNA technologies. Relative stabilities and favored base pair geometries of free inosine wobble base pairs in aprotic solvents have been determined through 1H NMR measurements at room temperature and at very low temperatures in a freonic solvent. As indicated by its significantly deshielded imino proton, the WatsonCrick-type I 3 C base pair forms a remarkably strong NHN hydrogen bond. For the thermodynamically less stable I 3 A wobble base pair, two configurations of similar population coexist at 133 K in the slow hydrogen bond exchange regime, namely a WatsonCrick(I)WatsonCrick(A) geometry and a WatsonCrick(I)Hoogsteen(A) geometry. I 3 U base pairs are stabilized by two rather weak hydrogen bonds and are significantly disfavored over inosine self-associates in a low-temperature Freon solution.
’ INTRODUCTION The hypoxanthine nucleoside inosine plays an important biological role as biosynthetic precursor to the major purine nucleosides adenosine and guanosine. In addition, inosine is frequently found in various RNA molecules mostly as a result of adenine deamination reactions. In its most prominent role, the hypoxanthine base of inosine appears at the wobble position as the 50 base of the anticodon in some tRNAs, recognizing cytosine, adenine, and uracil in the third codon position on mRNA during translation at the ribosome.1 The broad pairing specificity exhibited by the hypoxanthine base has also found its way into various DNA technological applications by employing hypoxanthine as a nondiscriminatory base analogue or universal base. Examples include inosine containing primers for the detection, isolation, and sequencing of genes2 and its use in DNA microarray hybridization.3 On the basis of the three-dimensional structure of doublehelical nucleic acids and of the ribosomal translation machinery with its steric restraints, studies have been primarily focused on the geometric requirements for the nonstandard WatsonCrick base pairs formed in regular duplexes and in particular at the first anticodon position.4,5 On the other hand, little is known about the inherent energetics of unusual base pairs, expected to be closely associated with the strength of formed hydrogen bonds between the bases. However, due to the various superimposed interactions effective in the biomolecular systems, H-bondmediated base pairing separated from additional contributions is difficult to assess and has received only little attention. Experimentally, NMR spectroscopy has been shown to be an effective tool for studying the strength of hydrogen bonds in solution.6 However, a detailed characterization of exclusively hydrogen-bond-directed base pair formation in solution requires the elimination of any contributions from additional stacking or steric interactions within a regular double helix or within the r 2011 American Chemical Society
codonanticodon minihelix at the ribosome through the use of free nucleosides. Because water will compete for hydrogen bond donor and acceptor sites in the nucleobases, base pair formation with free nucleosides in an aqueous environment will mostly be abolished and calls for the use of aprotic solvents. Finally, NMR signals at higher temperature generally correspond to an average over fast exchanging hydrogen bonded species, thus restricting a detailed characterization of hydrogen bonds for individual complexes. To circumvent problems arising from dissociation and fast exchange in weakly hydrogen bonded systems, measurements have to be performed at very low temperatures where the regime of slow hydrogen bond exchange within the NMR time scale is reached. In the liquid state, this has been achieved by employing deuterated freonic mixtures as NMR solvents. Such solvents allow high resolution NMR measurements down to 100 K and enable a detailed characterization of hydrogen bonded associates in the slow exchange regime.7,8 Also, the temperaturedependent dielectric constant of the Freon solvent of up to 40 D at very low temperatures9 may closely mimic the more hydrophobic microenvironment in many binding domains,10 allowing insight into the true strength of hydrogen bond interactions within the biomolecular complexes. In the present study we have employed NMR meaurements to determine the relative stabilities of the three nonstandard WatsonCrick base pairs I 3 C, I 3 A, and I 3 U in apolar solvents. In particular, low-temperature NMR measurements in freonic solvent mixtures were employed to gain more insight into favored complex geometries and the strength of formed hydrogen bonds for these biologically as well as technologically important base pairs.
Received: January 26, 2011 Revised: May 16, 2011 Published: June 06, 2011 8569
dx.doi.org/10.1021/jp200840j | J. Phys. Chem. B 2011, 115, 8569–8574
The Journal of Physical Chemistry B
’ EXPERIMENTAL SECTION Materials. The deuterated Freon mixture CDClF2/CDF3 was prepared as described11 and handled on a vacuum line that was also used for the sample preparation. Reagents of the highest quality available were purchased from Sigma-Aldrich, Deisenhofen, Germany. The sugar hydroxyl groups of the free nucleosides were either O-acetyl- or O-tert-butyldimethylsilyl (TBDMS) protected according to established procedures,1215 and the course of the reactions was controlled by thin layer chromatography (TLC) on silica gel plates (Merck silica gel 60 F254). If necessary, solvents were dried by standard procedures prior to use. Oligonucleotides for the UV studies were purchased from TIB MOLBIOL, Berlin, Germany. They were synthesized by the standard phosphoramidite method on a 0.2 μmol scale and finally subjected to gel filtration. UV Melting Experiments. UV experiments were performed on a Cary 100 spectrophotometer equipped with a Peltier temperature control unit (Varian, Darmstadt). The melting curves were recorded by measuring the absorption of the solution at 260 nm with 1 data point/C in 10 mm quartz cuvettes. The UV melting experiments were performed by using a protocol consisting of a heating cycle followed by cooling the sample to the initial temperature. After a waiting period of 5 min, another heating ramp was started. Heating and cooling rates of 0.5 C/min were employed. The concentration of each DNA duplex in a buffer with 100 mM NaCl and 20 mM phosphate, pH 7, was 2.5 μM. Melting temperatures were determined by the maximum of the first derivative plot of the smoothed melting curves and measured in triplicate. NMR Spectroscopy. NMR experiments were performed in CD2Cl2 or in a deuterated Freon solvent on a Bruker AMX500 spectrometer. Temperatures were adjusted by a Eurotherm Variable Temperature Unit to an accuracy of (1.0 C. Temperature calibration was performed with a sample of methanol in MeOH-d4, and the calibration curve was extrapolated for temperatures outside the range covered by the methanol sample. 1H chemical shifts in the Freon mixture were referenced relative to CHClF2 (δH = 7.13 ppm). Phase-sensitive two-dimensional nuclear Overhauser effect (2D NOE) spectra at low temperatures were recorded with a mixing time of 60 ms and a recycle delay of 1 s. A total of 1K free-induction decays (FIDs) of 2K complex data points were collected using the TPPI method and the t1 and t2 FIDs zero-filled to give a final matrix of 2K 3 2K real data points prior to Fourier transformation.
’ RESULTS AND DISCUSSION Stability of Inosine Base Pairs. To obtain heteroassociation constants between the O-protected nucleosides, we measured the concentration-dependent imino proton chemical shift of inosine at 23 C in a methylene chloride solution with the second nucleoside present in large excess. O-Silylation or O-acetylation of the free sugar OH protons in the nucleosides prevent their potential participation in hydrogen bond formation and are a prerequisite for enhancing nucleoside solubility in apolar solvents. Because inosine exhibits significant self-association in CD2Cl2 with a Ka of 200 M1 at 296 K,15 a 3040-fold excess of C, A, or U nucleosides over inosine suppresses the formation of inosine dimers and drives the equilibrium to heterocomplex formation. On the other hand, self-association of the cytidine, adenosine, and uridine nucleosides have been shown in preliminary experiments
ARTICLE
Table 1. Heteroassociation Constants Ka (M1) of Free Nucleosides X and Ya and UV Melting Temperatures Tm (C) of Duplexes with a Central X 3 Y Base Pairb A 3 U/T
I3C
c
I3A
I 3 U/T
Ka
70
1105
57
69
Tm
56.3 ( 0.8
55.3 ( 0.4
53.2 ( 0.1
48.7 ( 0.2
a
From the concentration dependence of 1H imino chemical shifts in CD2Cl2 at 293 K. b Average Tm values with standard deviations from three independent measurements. c From ref 18, measured in CDCl3 at 293 K.
to be negligible in a first approximation (e10 M1). Thus, with an intrinsic imino proton chemical shift of inosine monomers of δ = 8.97 ppm determined for an inosine solution at concentration c f 0, a good fit of the experimental data for the binary nucleoside mixtures was obtained with a simple 1:1 heteroassociation model and only two adjustable parameters, namely, the association constant for the heterodimer formation and the limiting chemical shift of the inosine imino proton in the heterodimer at c f ¥. Association constants as determined by a least-squares fit of the data are summarized in Table 1. Disregarding solvation effects upon base pair formation in the apolar solvent, the association constants are expected to largely reflect the interaction energy associated with the intermolecular hydrogen bond formation. Interestingly, with Ka > 103 M1, the I 3 C base pair formation is significantly favored over the I 3 A, I 3 U, and the standard A 3 U WatsonCrick base pair, with association constants in methylene chloride ranging from 57 to 70 M1. This high thermodynamic stability points to strong hydrogen bond interactions between inosine and cytidine in the free base pair (vide infra). In order to relate base pair stabilities for the free associates to the duplex stabilizing effect of corresponding inosine wobble pairs when incorporated into a double-helical nucleic acid, UV melting temperatures were recorded for the nonself-complementary DNA duplexes DIC, DIA and DIT as well as for the canonical duplex DAT in an aqueous buffer (Figure 1). Note, that the term wobble base pair is used here for all nonstandard WatsonCrick inosine pairs known to occur in the codonanticodon wobble position irrespective of their particular geometric arrangement. Upon the transition from duplex to single strands, a hyperchromicity is observed at 260 nm in temperature-dependent UV measurements, and its midpoint defines the corresponding melting temperature Tm. Any change in the melting temperature of the canonical duplex upon the incorporation of an inosine base pair may be used as a measure for its relative stabilizing effect on the duplex secondary structure. Table 1 summarizes the melting temperatures for the three duplexes and the reference duplex. Stabilization follows the order I 3 C > I 3 A > I 3 T, with the I 3 C base pair exhibiting the same contribution to duplex stabilization as compared to the canonical A 3 T base pair. Although significant nearest neighbor effects are anticipated for the absolute base pair stabilities, the relative stability of inosine pairs is in line with general stability trends found previously in DNA duplexes.16,17 Interestingly, among the three inosine base pairs studied, the I 3 C base pair is the most stable pair, irrespective of being dissolved in the aprotic solvent or incorporated into a nucleic acid duplex. However, strong discrepancies are apparent in the relative stability of I 3 C and A 3 U/T base pairs, the latter being at least equally stabilizing in a 8570
dx.doi.org/10.1021/jp200840j |J. Phys. Chem. B 2011, 115, 8569–8574
The Journal of Physical Chemistry B
ARTICLE
Figure 1. Oligodeoxyribonucleotide duplexes DAT, DIC, DIA, and DIT.
Figure 2. Temperature-dependent 1H NMR spectra showing the imino proton spectral region for a mixture of 20 ,30 ,50 -tri-O-(tertbutyldimethylsilyl)-inosine and 20 ,30 ,50 -tri-O-(tert-butyldimethylsilyl)adenosine in Freon.
duplex as compared to I 3 C. Clearly, direct basebase interactions mostly through hydrogen bonding largely determine base pair formation in the aprotic solvent but only partially contribute to the free energy of a duplex in an aqueous environment. Additional interactions, i.e. steric restrictions imposed by the sugar phosphate backbone on the geometry, hydration/dehydration effects, as well as stacking interactions between adjacent base pairs, superimpose to give a complex interplay of stabilizing and destabilizing forces. Being isosteric with the I 3 C base pair, the enhanced stabilizing effect of an A 3 U/T pair within a double helix suggests more favorable stacking and/or solvation effects compared to I 3 C base pairs, compensating for differences in the hydrogen-bond-mediated intrinsic base pair stability. Geometry of InosineAdenosine Complexes. To characterize preferred base pair geometries and formed hydrogen bonds between inosine and adenosine in more detail, lowtemperature NMR measurements were performed on the free nucleosides in a deuterated freonic mixture. In Figure 2, imino resonances of the 1:1 inosineadenosine mixture in Freon are plotted as a function of temperature. Apparently, the single imino signal of the hypoxanthine base shifts downfield upon cooling due to increased formation of hydrogen-bonded complexes. Below the coalescence point at about 173 K, individual imino resonances in slow exchange on the chemical shift time scale appear in the 1H NMR spectrum. These include signals of imino protons in I 3 A complexes Ha and Hb resonating at 133 K at
Figure 3. (A) Structure of I 3 I and I 3 A base pairs with NOE contacts indicated by arrows. (B) Portions of a 2D NOE spectrum for a mixture of 20 ,30 ,50 -tri-O-(tert-butyldimethylsilyl)-inosine and 20 ,30 ,50 -tri-O-(tertbutyldimethylsilyl)-adenosine in Freon showing regions with baseimino and basesugar proton crosspeaks. The spectrum was acquired at 133 K with a 60 ms mixing time.
15.7 ppm and 15.1 ppm, respectively, and a less intense, more upfield shifted resonance Hc at 13.9 ppm, which can be attributed to homodimers of inosine by comparison with previous inosine self-association studies.15 Assignment of the A 3 I imino protons Ha and Hb to a particular base pair geometry was accomplished through characteristic 1 H1H NOE contacts of the imino resonances under slow exchange conditions. Two I 3 A base pair configurations stabilized by two cyclic hydrogen bonds each are conceivable, namely a WatsonCrick(I)WatsonCrick(A) type designated I 3 A(WC), and a WatsonCrick(I)Hoogsteen(A) type designated I 3 A(H) (Figure 3A). In the former, hypoxanthine and adenine bases 8571
dx.doi.org/10.1021/jp200840j |J. Phys. Chem. B 2011, 115, 8569–8574
The Journal of Physical Chemistry B are in a head-to-head orientation with A N1 engaged as hydrogen bond acceptor in the H-bond to the inosine imino group. In contrast, A N7 at the adenosine Hoogsteen face is involved as acceptor for the corresponding hydrogen bond in the I 3 A(H) base pair. As is apparent from Figure 3A, NOE contacts in I 3 A(WC) are expected for the hydrogen-bonded inosine imino to its own H2 proton as well as to H2 and amino protons of adenosine. Furthermore, in this configuration, mutual NOE contacts are also expected for the H2 protons of inosine and adenosine because of their close spatial proximity. Portions of a 2D NOE spectrum of the inosine-adenosine mixture in Freon at 133 K are shown in Figure 3B. Several crosspeaks connect the imino resonances with H2 and H8 protons of the hypoxanthine and adenine base. The latter can be distinguished from H2 by their contacts to ribose sugar protons. Whereas corresponding NOE crosspeaks of moderate and high intensity identify H8, no significant contacts to sugar protons are found for the H2 protons. Note, that even a syn glycosidic torsion angle will not enable H2 to be in close contact with sugar protons and, in particular, with the anomeric H10 proton. On the other hand, a stronger crosspeak between H8 and H20 when compared to H8 and H10 clearly establishes an anti conformation for both nucleosides.19 As expected, all inosine imino protons exhibit strong intrabase contacts to their own H2 protons. However, unlike the Hc resonance of the inosine homodimers, Ha and Hb imino resonances show another crosspeak to an adenine H2 at 8.3 ppm and to an adenine H8 at 8.7 ppm, respectively. Also, there is a strong crosspeak between the two A H2 and I H2 resonances, which connect to the Ha imino, as well as between the A H8 and I H2 resonances, which connect to the Hb imino (not shown). Severe line-broadening at 133 K hampers the observation of additional crosspeaks to adenine amino protons. Taken together, these results unambiguously identify Ha being located in the I 3 A(WC) base pair. Correspondingly, Hb indicates the coexistence of the I 3 A(H) complex under these conditions. Noticeable shoulders at the two Ha and Hb imino resonances may be attributed to the formation of additional ternary complexes with a second adenine base occupying the free Hoogsteen or WatsonCrick face of adenosine in the I 3 A base pairs. Such I 3 A 3 A base triplets are also compatible with the presence of inosine dimers in the equimolar mixture. Being a rather weak interaction, ternary complexes remain elusive in our NOE experiments. However, they have been observed before in adenosine acetic acid complexes.14 There have been differing reports as for the geometry of inosine-adenosine base pairs in nucleic acid duplexes and in the wobble position of the ribosomal decoding center.2022 Crick originally proposed an I 3 A(WC) wobble pairing with hydrogen bond interactions between the WatsonCrick faces and with both bases in anti orientation with respect to the furanose moiety.1 A purine 3 purine base pair of this geometry has a stretched C-10 C-l0 distance of about 11 Å, and raises the question of whether it can be accommodated within a regular duplex or at the wobble position in the ribosomal complex. On the other hand, the I 3 A(H) geometry with N7 of adenosine engaged as acceptor in a hydrogen bond to NH of inosine and an associated adenosine syn conformation may be more compatible with the dimensions of canonical WatsonCrick base pairs, minimizing structural perturbations upon its incorporation.5 In fact, an I 3 A(H) base pair geometry has been found in a crystal structure of a B-DNA duplex with an inosineadenosine base pair.20 However, other crystal structure analyses on I 3 A containing nucleic acid
ARTICLE
Figure 4. (A) Structure of the I 3 C base pair with NOE contacts indicated by arrows. (B) Portion of a 2D NOE spectrum for a mixture of 20 ,30 ,50 -tri-O-(tert-butyldimethylsilyl)-inosine and 30 ,50 -di-O-acetyl20 -deoxycytidine in Freon with inosine imino contacts in the base proton spectral region. The spectrum was acquired at 133 K with a 60 ms mixing time. The corresponding 1D spectral region is shown on top.
duplexes21 as well as of an I 3 A base pair in the context of the ribosomal decoding center report on an I 3 A(WC) pairing, and suggest its easy accommodation within a regular duplex or the codonanticodon minihelix.22 According to the present studies, both geometries coexist under the low-temperature solution conditions and seem to be of about equal thermodynamic stability based on signal intensity. Consequently, the preferred geometry will be mostly structure and sequence dependent with no significant impact from the particular hydrogen bonding pattern. Geometry of InosineCytidine Complexes. For a 1:1 mixture of O-silylated inosine and O-acetylated cytidine, only one predominant complex is observed at 133 K in a freonic solvent as revealed by an inspection of the imino proton spectral region (Figure 4B top). Due to their low signal intensity and the lack of any NOE crosspeaks, we did not attempt to structurally characterize minor associates but rather focused on the major complex formed between inosine and cytidine. However, based on an imino proton chemical shift of 13.9 ppm, minor species again include inosine homodimers. As shown in Figure 4B, a 2D NOE spectrum acquired for the inosine-cytidine mixture at 133 K shows crosspeaks between the rather downfield shifted I NH signal at 16.6 ppm to its own H2 at 8.35 ppm and another contact of lower intensity to a resonance at 8.8 ppm. The latter is easily identified as a cytosine amino signal exhibiting only two other NOE contacts, namely to the second amino proton at 8.4 ppm partially overlapping I H2 and to its own H5 proton. Such an NOE pattern confirms a WatsonCrick-type I 3 C base pair as being the major species with an anti conformation for both nucleosides as derived from crosspeaks observed between I H8 and C H6 to their sugar protons (not shown). As expected, such an I 3 C base pair, which corresponds to the G 3 C WatsonCrick base pair except for the lack of an N2O2 hydrogen bond (see Figure 4A), is also exclusively formed within a nucleic acid environment.23 8572
dx.doi.org/10.1021/jp200840j |J. Phys. Chem. B 2011, 115, 8569–8574
The Journal of Physical Chemistry B
Figure 5. (A) Structure of I 3 U base pairs with NOE contacts indicated by arrows. (B) Portion of a 2D NOE spectrum for a mixture of 20 ,30 ,50 tri-O-(tert-butyldimethylsilyl)-inosine and 30 ,50 -di-O-acetyl-20 -deoxyuridine in Freon showing imino contacts of uridine and inosine nucleosides with iminoimino exchange crosspeaks indicated by circles. The spectrum was acquired at 123 K with a 60 ms mixing time. The corresponding 1D spectral region is shown on top.
Geometry of InosineUridine Complexes. In a 1:1 mixture of sugar OH-protected inosine and uridine, the two imino signals of inosine and uridine split into several imino resonances under slow exchange conditions below 153 K (Figure 5B top). Whereas the most downfield shifted Ha resonance at 13.85 ppm at 123 K can be assigned to inosine homodimers through its chemical shift and its exclusive NOE contact to the own H2 proton, the four upfield shifted signals Hd located between 12.3 and 11.7 ppm are assigned to the two symmetric and one asymmetric uridine homodimers based on previous NOE experiments on the uridine homoassociation.24 Therefore, the low-intensity Hb and Hc signals at 13.3 and 12.85 ppm are the only remaining candidates for imino protons in inosine 3 uridine base pairs (Figure 5A). In fact, as shown by the portion of the 2D NOE spectrum in Figure 5B, there is a noticeable mutual NOE contact between these two imino signals, only compatible with an inosine 3 uridine heterodimeric species. With another connectivity of Hb to inosine H2 and exchange crosspeaks observed between Hc and the imino signals of uridine homodimers, the former is identified as inosine NH and the latter as uridine NH. It has to be noted that the
ARTICLE
exchange crosspeaks still detected at 123 K for uridine imino protons and attributable to their increased dissociation rate disappear at 113 K (not shown). Apparently, I 3 U heteroassociation is restricted by the higher thermodynamic stability and preferred formation of inosine homodimers. As a result, uridine mostly self-associates, and only about 1 out of 20 uridine nucleosides pairs with residual inosine based on signal integration. The I 3 U/T base pairs within a WatsonCrick duplex or in a codonanticodon interaction are mostly found to adopt a conformation isosteric with the well characterized G 3 U base pair, i.e., with two cyclic hydrogen bonds involving NH3 and O2 of the uracil base as the hydrogen bond donor and acceptor, respectively.25,26 However, another I 3 U base pair geometry involving uracil O4 as the acceptor atom is easily conceivable provided that there are no additional steric restrictions (see Figure 5A). Unfortunately, a possible discrimination of the two I 3 U base pair configurations with NH(I) 3 3 3 O2(U) and NH(I) 3 3 3 O4(U) H-bonding is hampered by the lack of observable contacts between inosine H2 and uridine H5 and H10 protons due to their low populations and signal intensities. On the basis of the equal participation of uridine 2-carbonyl and 4-carbonyl functionalities as hydrogen bond acceptors within the free uridine homodimers, both I 3 U base pair geometries may likely form and coexist in freonic solutions without being spectrally resolved at the low temperatures. However, because the NHO4 configuration requires a syn glycosidic torsion angle for one of the nucleosides in an antiparallel strand arrangement, geometric aberrations and associated distortions renders it unlikely to be formed in standard nucleic acid structures including codonanticodon interactions.5 Hydrogen Bond Strength. NMR constitutes a powerful tool for a more detailed characterization of hydrogen bond contacts. Thus, the proton chemical shift can be used as a sensitive measure for the relative strength of H-bonding. A more deshielded proton in the DH 3 3 3 A hydrogen bridge originates from a displacement of the proton from donor atom D toward the acceptor atom A and is associated with a shortening of the hydrogen bond.27,28 The proton transfer in the case of a neutral donor and acceptor is also accompanied by a partial charge separation and ultimately results in a zwitterionic complex upon complete proton transfer. Consequently, a considerable influence of the dielectric properties of the solvent or the particular microenvironment on the hydrogen bond geometry is expected and observed.9 Because the local dielectric constant experienced by a base pair within a nucleic acid duplex is difficult to assess, it is interesting to relate the imino proton chemical shift of typically 1314 ppm observed for a WatsonCrick A 3 U base pair in a duplex to the chemical shift of 15.1 ppm determined here for the corresponding free base pair in the low temperature Freon solution. It can be assumed, that more upfield shifted Watson Crick imino protons within the duplex will be a consequence of some distortions from the ideal base pair geometry and from additional stacking interactions between base pairs. Clearly, hydrogen bonds involving adenine or cytosine amino protons are only weak, but inosine imino NH 3 3 3 N or NH 3 3 3 O hydrogen bonds exhibit strongly deshielded protons in medium to strong hydrogen bonds. In particular, a strong hydrogen bond for the I 3 C base pair is indicated by its remarkable proton downfield chemical shift of 16.6 ppm. Again, the more upfield shifted imino protons observed for I 3 C base pairs in RNA and DNA duplexes with chemical shifts of 1515.5 ppm29,30 may be attributed to base stacking and distortions enforced by 8573
dx.doi.org/10.1021/jp200840j |J. Phys. Chem. B 2011, 115, 8569–8574
The Journal of Physical Chemistry B the various interactions within the duplex. The large association constant as determined for the I 3 C base pair formation (see Table 1) can be associated with the strong NH 3 3 3 N hydrogen bond stabilizing the I 3 C pair. On the other hand, given the same WatsonCrick geometry for the I 3 C and canonical A 3 U base pair, their similar duplex stabilization must arise from more favorable (de)solvation and/or stacking interactions of the AU pair in the nucleic acid environment. Whereas a chemical shift of 13.3 and 12.8 ppm for the inosine and uridine imino proton in the free I 3 U wobble pairs only indicate NH 3 3 3 O hydrogen bonds of moderate strength, the NH 3 3 3 N1 hydrogen bond between inosine and adenosine nucleosides again exhibits a significantly deshielded imino proton resonating at 15.7 ppm. In comparison, the proton of the NH 3 3 3 N7 hydrogen bond in the I 3 A(H) geometry is upfield shifted in line with a weaker Hoogsteen-type hydrogen bond and in close agreement with previous observations on WatsonCrick and Hoogsteen A 3 U base pairs.18
’ CONCLUSIONS In the present study, base pair thermodynamic stabilities and favored base pair geometries for three inosine wobble base pairs have been determined through NMR experiments. Because suitably sugar-protected free nucleosides are employed in aprotic solvents, base pair formation is largely determined by the formation of specific hydrogen bonds between the nucleobases and additional steric and stacking interactions, and hydration effects present in nucleic acids under physiological conditions are effectively eliminated. In fact, hydrogen bond preferences are frequently overwritten by the additional interactions in duplexes or in the codonanticoden minihelix formed at the ribosome. Thus, many studies have elaborated on the geometric parameters of base pairs as important determinants of base pair formation in double-helical nucleic acids. On the contrary, detailed information on the hydrogen bond energies and the hydrogen-bondmediated base pairing is missing or mostly derived from computational studies. Yet, only a comprehensive knowledge of all possible contributions to inosine base pair formation will allow a refined and more reliable prediction and understanding of the pairing configurations found in the various biological systems.
ARTICLE
(7) Siegel, J. S.; Anet, F. A. L. J. Org. Chem. 1988, 53, 2629–2630. (8) Golubev, N. S.; Denisov, G. S. J. Mol. Struct. 1992, 270, 263–276. (9) Shenderovich, I. G.; Burtsev, A. P.; Denisov, G. S.; Golubev, N. S.; Limbach, H.-H. Magn. Reson. Chem. 2001, 39, 91–99. (10) Florian, J.; Leszczynski, J. J. Phys. Chem. A 1996, 103, 8516–8523. (11) Golubev, N. S.; Smirnov, S. N.; Gindin, V. A.; Denisov, G. S.; Benedict, H.; Limbach, H.-H. J. Am. Chem. Soc. 1994, 116, 12055–12056. (12) Ogilvie, K. K.; Thompson, E. A.; Quilliam, M. A.; Westmore, J. B. Tetrahedron Lett. 1974, 33, 2865–2868. (13) Dunger, A.; Limbach, H.-H.; Weisz, K. Chem.—Eur. J. 1998, 4, 621–628. (14) Basílio Janke, E. M.; Limbach, H.-H.; Weisz, K. J. Am. Chem. Soc. 2004, 126, 2135–2141. (15) Hupp, T.; Sturm, C.; Basílio Janke, E. M.; Perez Cabre, M.; Weisz, K.; Engels, B. J. Phys. Chem. A 2005, 109, 1703–1712. (16) Martin, F. H.; Castro, M. M.; Aboul-ela, F.; Tinoco, I., Jr Nucleic Acids Res. 1985, 13, 8927–8938. (17) Watkins, N. E., Jr.; Santa Lucia, J., Jr. Nucleic Acids Res. 2005, 33, 6258–6267. (18) Dunger, A.; Limbach, H.-H.; Weisz, K. J. Am. Chem. Soc. 2000, 122, 10109–10114. (19) W€uthrich, K. NMR of Proteins and Nucleic Acids; Wiley Interscience: New York, 1986; pp 205214. (20) Corfield, P. W. R.; Hunter, W. N.; Brown, T.; Robinson, P.; Kennard, O. Nucleic Acids Res. 1987, 19, 7935–7949. (21) Carter, R. J.; Baeyens, K. J.; SantaLucia, J.; Turner, D. H.; Holbrook, S. R. Nucleic Acids Res. 1997, 25, 4117–4122. (22) Murphy, F. V.; Ramakrishnan, V. Nat. Struct. Mol. Biol. 2004, 11, 1251–1252. (23) Xuan, J.-C.; Weber, I. T. Nucleic Acids Res. 1992, 20, 5457–5464. (24) Weisz, K.; J€ahnchen, J.; Limbach, H.-H. J. Am. Chem. Soc. 1997, 119, 6436–6437. (25) Cruse, W. B. T.; Aymani, J.; Kennard, O.; Brown, T.; Jack, A. G. C.; Leonard, G. A. Nucleic Acids Res. 1989, 17, 55–72. (26) Pan, B.; Mitra, S. N.; Sun, L.; Hart, D.; Sundaralingam, M. Nucleic Acids Res. 1998, 26, 5699–5706. (27) Smirnov, S. N.; Benedict, H.; Golubev, N. S.; Denisov, G. S.; Kreevoy, M. M.; Schowen, R. L.; Limbach, H.-H. Can. J. Chem. 1999, 77, 943–949. (28) Dingley, A. J.; Masse, J. E.; Peterson, R. D.; Barfield, M.; Feigon, J.; Grzesiek, S. J. Am. Chem. Soc. 1999, 121, 6019–6027. (29) Patel, D. J. Eur. J. Biochem. 1978, 83, 453–464. (30) Mirau, P. A.; Kearns, D. R. J. Mol. Biol. 1984, 177, 207–227.
’ AUTHOR INFORMATION Corresponding Author
*Fax: 49 (0)3834 864427. Tel: 49 (0)3834 864426. E-mail:
[email protected].
’ ACKNOWLEDGMENT We thank the Deutsche Forschungsgemeinschaft, Bonn-Bad Godesberg, for financial support (Grant Number WE 1933/10-1). ’ REFERENCES (1) Crick, F. H. C. J. Mol. Biol. 1966, 19, 548–555. (2) Palva, A.; Vidgren, G.; Paulin, L. J. Microbiol. Methods 1994, 19, 315–321. (3) Chizhikov, V.; Wagner, M.; Ivshina, A.; Hoshino, Y.; Kapikian, A. Z.; Chumakov, K. J. Clin. Microbiol. 2002, 40, 2398–2407. (4) Takai, K. J. Theor. Biol. 2006, 242, 564–580. (5) Das, G.; Lyngdoh, D. J. Mol. Struct. (THEOCHEM) 2008, 851, 319–334. (6) Hibbert, F.; Emsley, J. Adv. Phys. Org. Chem. 1990, 26, 255–279. 8574
dx.doi.org/10.1021/jp200840j |J. Phys. Chem. B 2011, 115, 8569–8574