Solution structure of 32-modified anticodon stem–loop of Escherichia c
http://www.100md.com
《核酸研究医学期刊》
Department of Biochemistry and Cell Biology, Rice University Houston, TX 77251-1892, USA
*To whom correspondence should be addressed. Tel: +1 713 348 4912; Fax +1 713 348 5154; Email: edn@bioc.rice.edu
ABSTRACT
Nucleoside base modifications can alter the structures and dynamics of RNA molecules and are important in tRNAs for maintaining translational fidelity and efficiency. The unmodified anticodon stem–loop from Escherichia coli tRNAPhe forms a trinucleotide loop in solution, but Mg2+ and dimethylallyl modification of A37 N6 destabilize the loop-proximal base pairs and increase the mobility of the loop nucleotides. The anticodon arm has three additional modifications, 32, 39, and A37 C2-thiomethyl. We have used NMR spectroscopy to investigate the structural and dynamical effects of 32 on the anticodon stem-loop from E.coli tRNAPhe. The 32 modification does not significantly alter the structure of the anticodon stem–loop relative to the unmodified parent molecule. The stem of the RNA molecule includes base pairs 32-A38 and U33–A37 and the base of 32 stacks between U33 and A31. The glycosidic bond of 32 is in the anti configuration and is paired with A38 in a Watson–Crick geometry, unlike residue 32 in most crystal structures of tRNA. The 32 modification increases the melting temperature of the stem by 3.5°C, although the 32 and U33 imino resonances are exchange broadened. The results suggest that 32 functions to preserve the stem integrity in the presence of additional loop modifications or after reorganization of the loop into a translationally functional conformation.
INTRODUCTION
Posttranscriptional modification of RNA molecules occurs in all cells (1–3). The modifications, which are primarily localized to the nucleotide bases, can alter the chemical, structural, and thermodynamic properties of RNAs and thereby contribute to RNA function (4). Many nucleotide base modifications have been chemically well characterized but the impact of these modifications on RNA structure and stability has been less thoroughly examined. Pseudouridine is the most common nucleotide base modification and is found in the tRNA and rRNA of all cells (5,6). This modification is formed by an isomerization of uridine resulting in a C5-C1' base–ribose glycosidic bond and a second imino (NH) functionality. Pseudouridine preferentially adopts the syn conformation about the glycosidic bond as the free nucleotide (7,8), but the anti configuration has been most frequently observed for pseudouridine within oligonucleotides and double helices (9). Consequently pseudouridine tends to base pair with adenosine through the N3 imino and the C2 carbonyl groups (Figure 1) (10–13) and imparts thermodynamic stability to helices when located in the interior of an RNA duplex or in a single strand region adjacent to a duplex (8).
Figure 1 Sequences of (A) the 32-modified and (B) fully modified RNA hairpins corresponding to the anticodon arm of E.coli tRNAPhe. Nucleotide numbering corresponds to the full-length tRNAPhe molecule. designates pseudouridine and ms2i6A designates (2-thiomethyl, N6-dimethylallyl)-adenine. 32-A38 base arrangements for (C) the Watson–Crick base pair and (D) the bifurcated hydrogen bond interaction.
In addition to the nearly universally conserved 55, pseudouridine occurs frequently at several other positions in tRNA including residues 13, 32, 39, and 40 (14), although the frequency of occurrence at these positions varies among species. 55, located in the T-loop, serves only as an acceptor in its hydrogen bonding with the imino and amino groups of G18. The 55-G18 pair stacks between the flanking base pairs and contributes to maintenance of the tertiary fold of tRNA (15,16). In yeast tRNAPhe and other tRNAs or tRNA complexes, 39, at the bottom of the anticodon stem, generally aligns with A31 in a standard A–U Watson–Crick hydrogen bond geometry (15–20), although the hydrogen bonds tend to be somewhat long (2.1–2.3 ?). 32 also is located in the anticodon arm of six tRNA species in Escherichia coli and adopts a syn configuration about its glycosidic bond in tRNACys (17). In yeast tRNAAsp, 32 forms a bifurcated base pair with C38 involving 32 O4 and the exocyclic amino hydrogens of C38 (18,21).
Residue 32 in the anticodon arm is important for the translation function of tRNA. A central element of the extended anticodon hypothesis is that nucleotides at the stem–loop junction in the anticodon arm contribute to the translational efficiency of tRNA (22,23). The functional importance of residues 32 and 38 was first demonstrated using amber suppressor tRNAs Su2 and Su7 that incorporate glutamine and tryptophan, respectively, at the stop codon 5'-UAG-3'. These suppressor tRNAs share the same anticodon nucleotide sequence, but Su2 is a weak suppressor whereas Su7 is a strong suppressor. The different suppression efficiencies of these tRNAs originate mainly from the identities of nucleotides 32 and 38 (23). In addition to the suppressor tRNA studies of Yarus and coworkers (23), the functional significance of the 32–38 bp is further illustrated by nucleotide substitution studies of glycyl tRNAs of E.coli and Mycoplasma mycoides (24,25). The genome of M.mycoides encodes one tRNAGly (anticodon 5'-UCC-3') that is used to decode all four glycine codons and has a C32–A38 mismatch in the anticodon arm. In E.coli, tRNAGly,2 has the anticodon 5'-UCC-3', but has the base pair U32–A38. When mutated to a U32–A38 base pair, the ability of the M.mycoides tRNAGly to decode non-cognate codons dramatically diminishes (25). Similarly, mutation of the U32–A38 base pair of E.coli tRNAGly,2 to C32–A38 leads to decreased fidelity and –1 frameshifting at 5'-GGG-3' codons (26). These studies demonstrate the ability of the 32–38 interaction to modulate the wobble properties of the anticodon. Thus, the types of interactions formed by residues 32 and 38 may alter the conformation and dynamics of the anticodon loop and/or the interactions within the codon-anticodon complex at the ribosome.
We have used heteronuclear NMR spectroscopy to examine the solution structure of the 32-modified form of the anticodon arm of E.coli tRNAPhe. A 17 nt RNA molecule that forms a stem–loop secondary structure in solution and corresponds to the anticodon arm of tRNAPhe was used (Figure 1). Our results demonstrate that the loop of the modified RNA molecule is composed of three nucleotides and lacks the characteristic U-turn motif. The anticodon arm is extended by two base pairs as in the unmodified parent molecule and 32 increases the stability of the stem. The 32 base forms a Watson–Crick type base pair with A38 and not the bifurcated hydrogen bond configuration found in most crystal structures of tRNAs.
MATERIALS AND METHODS
All enzymes were purchased from Sigma Chemical, except for T7 RNA polymerase and RluA enzymes, which were prepared as described (27,28). Deoxyribonuclease I type II, pyruvate kinase, adenylate kinase, and nucleotide monophosphate kinase were obtained as powders, dissolved in 15% glycerol, 1 mM dithiothreitol, and 10 mM Tris–HCl, pH 7.4, and stored at –20°C. Guanylate kinase and nuclease P1 were obtained as solutions and stored at –20°C. Unlabeled 5' nucleoside triphosphates (5'-NTPs) were purchased from Sigma, phosphoenolpyruvate (potassium salt) was purchased from Bachem, and 99% {15N}-ammonium sulfate and 99% {13C}- glucose were purchased from Isotec.
Preparation of RNA samples
The RNA sequence for E.coli ACSLPhe shown in Figure 1 was synthesized in vitro using T7 RNA polymerase and a synthetic DNA template. The nucleotide sequence of the stem corresponds to residues G27–C43 of full-length E.coli tRNAPhe. Isotopically labeled RNA molecules were prepared from 10 ml transcription reactions using 3 mM uniformly 15N-enriched and 13C-enriched 5'-NTPs as described (29). The RNA molecules were purified by passage through 20% (w/v) preparative PAGE, electroeluted (Schleicher and Schuell), and precipitated with ethanol. The purified oligonucleotides were dissolved in 1.0 M NaCl, 20 mM potassium phosphate, pH 6.8, and 2.0 mM EDTA, and dialyzed extensively against 10 mM NaCl, 10 mM potassium phosphate, pH 6.8, and 0.05 mM EDTA, using a Centricon-3 concentrator (Amicon Inc.). The samples were diluted with buffer to a volume of 0.2 ml and lyophilized to a powder. For experiments involving the non-exchangeable protons, the samples were exchanged twice with 99.9% D2O and then resuspended in 0.2 ml of 99.96% D2O. For experiments involving detection of the exchangeable protons, the samples were resuspended in 0.2 ml of 90% H2O/10% D2O. The samples contained 80 and 100 A260 OD units of 15N-labeled and 13C-labeled, respectively, RNA oligonucleotides (2.5–3.4 mM).
Preparation of the 32-ACSLPhe
Pseudouridine was introduced at position U32 of purified ACSLPhe using the pseudouridine synthase RluA (RluA). Histidine-tagged RluA was expressed in E.coli and purified using Ni2+ affinity resin as described (28). The pseudouridylation reaction was carried out using a mole ratio of RluA:RNA of 1:48. The reaction conditions were 50 mM HEPES pH 7.5, 100 mM NH4Cl, 0.03 mg/ml RluA, 0.06 mM ACSLPhe. The reactions were allowed to proceed overnight at 37°C. The RNA was purified from the RluA enzyme by heating the reaction to 90°C for 2 min followed by centrifugation to remove the precipitated protein. The supernatant was dialyzed with NMR buffer. Completion was determined by the disappearance of the C5–H5 resonance of U32 from 2D 13C-1H HMQC spectra (Supplementary Figure S1). The reactions could not be monitored using denaturing PAGE because the modified ACSLPhe migrates at the same rate as the unmodified ACSLPhe.
NMR spectroscopy
All NMR spectra were acquired on a Bruker AMX-500 spectrometer equipped with a 1H-{X} broadband probe, except for the 31P-decoupled 13C-1H constant time HSQC experiment, which was collected with a 1H-{13C, 31P} triple resonance probe. Broadband decoupling of the carbon and nitrogen resonances was achieved using GARP with B2 = 3125 Hz for carbon and B2 = 1570 Hz for nitrogen. H2O spectra were collected at 12°C with solvent suppression using either spin lock pulses or binomial read pulses with maximum excitation at 12.5 p.p.m. D2O spectra were collected at 25°C with presaturation or spin lock pulses to suppress the residual HDO peak. Quadrature detection was achieved using the States-TPPI method, and acquisition was delayed by a half-dwell in all indirectly detected dimensions. Typically, the data points were extended by 25% using linear prediction for the indirectly detected dimensions and the data were apodized using 1 Hz line broadening and 65° shifted sinebell functions. 1H spectra were referenced relative to DSS (0.00 p.p.m.). References for the 13C and 15N spectra were calculated using the spectrometer frequencies as reported (30). The 31P spectra were referenced to an external standard of TMP which was set at 0.00 p.p.m. All spectra were processed and analyzed with Felix 98.0 (Accelerys, Inc.).
2D 13C-1H HMQC and HSQC spectra were collected to identify 13C-1H chemical shift correlations. 2D HCCH-COSY and 3D HCCH-TOCSY (24 ms DIPSI-3 spin lock) experiments optimized for polarization transfer through the ribose carbons and a 2D 13C-1H HCCH-TOCSY (52 ms DIPSI-3 spin lock) optimized for polarization transfer through the adenine bases were collected in D2O to identify ribose spin systems and H8-H2 correlations, respectively (31,32). To identify intra-residue base-sugar correlations, a 2D 15N-1H HSQC experiment was acquired in D2O and optimized for two- and three-bond correlations. A J(N, N)-HNN COSY experiment was acquired in D2O to confirm the presence of A?U base pairs (33) and 31P assignments were obtained using a 31P/1H hetero-TOCSY-NOESY spectrum (34).
Distance constraints for the non-exchangeable resonances of 32-modified ACSLPhe were derived at 25°C from 2D 1H-1H NOESY spectra (80, 120, 180, 360, and 480 ms mixing times), 13C-edited 3D NOESY-HMQC spectra (180 and 360 ms mixing times), and 13C-edited 3D NOESY-ctHSQC spectra (80, 180, and 360 ms mixing times) optimized for the ribose resonances in 2 and 3. For the exchangeable resonances, 2D 15N-1H HSQC spectra were collected to identify 15N-1H chemical shift correlations. 2D 1H-1H NOESY experiments optimized for imino (NH) proton resonances were acquired at 60 and 360 ms mixing time in 90% H2O to obtain distance restraints involving the exchangeable protons.
Backbone torsion angle constraints were derived from 1H-1H and 31P-1H coupling constants obtained from the following experiments. A 31P-decoupled DQF-COSY experiment and a 2D 31P-1H HetCor experiment were acquired in D2O with unlabeled RNA samples.
Interproton distance constraints
Semi-quantitative distance constraints between non-exchangeable protons were estimated from cross peak intensities in 2D NOESY and 3D 13C-edited NOESY spectra. Using the covalently fixed pyrimidine H5-H6 distance (2.4 ?) and the conformationally restricted sugar H1'-H2' distance (2.8–3.0 ?) as references, peak intensities were classified as strong, medium, weak, or very weak and their corresponding proton pairs given upper bound distance constraints of 3.0, 4.0, 5.0, or 6.0 ?, respectively. Cross peaks observed only at mixing times >180 ms were classified as extremely weak and given 7.0 ? upper bound distance constraints to account for the possibility of spin diffusion. All distance constraints were given lower bounds of 1.8 ?. Distance constraints involving exchangeable protons were estimated from 360 ms mixing time NOESY spectra and were classified as either weak, very weak, or extremely weak, except for the intra-base pair distances A?U H2–NH and G?C NH–NH2, which were classified as strong constraints. Only intra-residue sugar-to-sugar constraints involving H5' and H5'' resonances were included in the calculations.
An initial set of structures was calculated using a shortened version of the simulated annealing protocol (described below). A list of all proton pairs in the calculated structures closer than 4.0 ? (representing expected NOEs) was compared to the list of constraints. The NOESY spectra were then re-examined for predicted NOEs absent from the constraint list. In some cases, this allowed the unambiguous assignment of previously unidentified NOEs, but, in other cases, the predicted NOEs were unobservable due to spectral overlap or the broadening of resonances by exchange with solvent. After the final calculations, virtually all predicted NOEs not in the list could be accounted for by spectral overlap or exchange broadening.
Hydrogen bonding constraints
Watson–Crick base pairs were identified using two criteria: the observation of a significantly downfield shifted NH or NH2 proton resonance and the observation of strong G?C NH–NH2 or A?U H2–NH NOEs. The A37?U33 base pair was identified by observation of a cross peak between A37H2 and U33N3 in the J(N, N)-HNN COSY spectrum. Hydrogen bonds were introduced as distance restraints of 2.9 ± 0.3 ? between donor and acceptor heavy atoms and 2.0 ± 0.2 ? between acceptor and hydrogen atoms. Constraints identified in this way were included in the calculations for base pairs G27?C43, G28?C42, G29?C41, G30?C40, A31?U39, and 32?A38. The U33?A37 base pair constraint was set to 2.9 ± 1.2 ? and 2.0 ± 1.2 ? between donor and acceptor heavy atoms and acceptor and hydrogen atoms, respectively, to permit conformational freedom of loop residues.
Dihedral angle constraints
Constraints on the ribose ring and backbone dihedral angles were derived from semi-quantitative measurements of 3JH-H and 3JH-P couplings (35,36). Sugar pucker conformations were determined from 3JH1'-H2' couplings in 31P-decoupled 2D DQF-COSY spectra. Residues with H1'-H2' couplings >7 Hz were constrained to the C2'-endo conformation through two of the torsion angles in the ribose sugar ring (37). Independent confirmation of sugar pucker conformation was provided by the observation of weak (<5 Hz) 3JH3'-H4' couplings, C3' resonances shifted downfield to 76–80 p.p.m. from the main cluster at 70–72 p.p.m., and C4' resonances shifted downfield to 85–86 p.p.m. from the main cluster at 82–84 p.p.m. Residues with weak (<5 Hz) 3JH1'-H2' couplings were constrained to the C3'-endo conformation. Residues with intermediate 3JH1'-H2' couplings were left unconstrained to reflect the possibility of conformational averaging.
Dihedral angle constraints for the torsion angle were derived from 3JH4'-H5' and 3JH4'-H5'' couplings in the DQF-COSY spectrum and intra-residue H4'-H5' and H4'-H5'' cross peak intensities in the 80 ms mixing time 3D NOESY-ctHSQC spectrum. For residues in which H4'-H5' and H4'-H5'' peaks in the DQF-COSY spectra were clearly absent, representing couplings <5 Hz, was constrained to the gauche+ conformation (60 ± 20) (35,36). For residues with clear 3JH4'-H5' or 3JH4'-H5'' couplings >5 Hz and unequal H4'-H5' and H4'-H5'' NOE intensities, was constrained to include both the trans and gauche– conformations (–120 ± 120), reflecting the lack of stereospecific assignments for the H5' and H5'' resonances. For residues with only weak or unobservable 3JH4'-H5' or 3JH4'-H5'' couplings and unequal H4'-H5' and H4'-H5'' NOE intensities, was left unconstrained to reflect the possibility of conformational averaging.
Dihedral angle restraints for the ? torsion angles were derived from 3JP-H5' and 3JP-H5'' couplings measured in 2D 31P-1H HetCor spectra. ? was constrained to the trans conformation (180 ± 40) for residues in which P-H5' and P-H5'' peaks in the HetCor spectra were clearly absent, representing couplings <5 Hz (35,36). For residues in which P-H5' and P-H5'' peaks could be observed, ? was left unconstrained to reflect the lack of stereospecific assignments and the possibility of conformational averaging. All P-H3' couplings that could be clearly identified were >5 Hz which allows for both trans and gauche- conformations for the torsional angle. A refinement that constrained angles in the stem (G27-C43 to A31-U39) to –125 ± 80 had small improvements (relative to not constraining ) on the quality of the stem regions of the structures and were therefore included for the structures reported here.
Dihedral angle restraints for and were derived from the observation that a trans conformation of either dihedral angle is generally associated with a large downfield shift of the bridging 31P resonance (38). Because no such shift is observed for any of the 31P resonances in the RNA molecules, and were loosely constrained to exclude the trans conformation (0 ± 120) for all residues except those in the loop regions (nucleotides U33 to A37), which were left unconstrained. No dihedral angle constraints were used for the glycosidic angle . A total of 48 restraints (11 , 8 ?, 9 , 9, and 11 ) were used constrain the phosphate backbone dihedral angles in the calculations.
Structure calculations
All calculations were carried out on Silicon Graphics O2 work stations using X-PLOR 3.851 (39). The dihedral angles of a linear starting structure (generated using Insight II, Molecular Simulations, Inc.) were randomized to generate 75 structures with randomized coordinates which were used in a simulated annealing/restrained molecular dynamics (rMD) routine (36,37). The calculation protocol was divided into three stages: global fold, refinement, and final minimization. The global fold step consisted of 1000 cycles of unconstrained energy minimization, 10 ps of rMD at 1000 K using only hydrogen bond and NOE constraints, 9 ps of rMD at 1000 K during which repulsive van der Waals forces were introduced, 14 ps of rMD while cooling to 300 K, and 1000 cycles of constrained minimization. The structures were then refined with 500 cycles of constrained minimization, 5 ps of rMD at 1200 K during which the , ?, , , , and sugar ring dihedral constraints were slowly introduced followed by 5 ps of rMD while cooling to 300 K, and 1000 cycles of constrained minimization. The final minimization step consisted of 1000 cycles of conjugate gradient energy minimization using all constraints and repulsive van der Waals potentials. To determine the consistency of the NMR data with the tri-loop conformation of the unmodified anticodon stem–loop, an additional set of calculations was performed using constraints involving overlapped resonances that were derived from spectra of the unmodified RNA molecule. These calculations were performed beginning with the coordinates of converged structures from the global-fold rMD simulation. Structures were viewed using Insight II (Accelerys, Inc.). The structure coordinates have been deposited in the rcsb with accession number 2AWQ.
Thermal stability
UV melting studies were performed using 2.2 μM RNA hairpin dissolved in NMR buffer (10 mM NaCl, 10 mM potassium phosphate, pH 6.8, and 0.05 mM EDTA). The samples were heated to 90°C for two minutes and snap cooled on ice before each melt experiment. A260 absorbance spectra from 20–95°C and from 95–20°C were acquired (1.0°C per minute) on a Pharmacia Ultrospec 2000 UV-Visible spectrophotometer equipped with a peltier melting heating apparatus. The melting curves were acquired in triplicate but could not be fit to a two-state model.
RESULTS
Effect of 32 modification on RNA stability
The thermal stability of the ACSLPhe and 32 ACSLPhe RNA hairpins (Figure 1) was investigated using UV melting experiments to determine overall molecular stability (Tm). The UV thermal denaturation curves indicate that the hairpins melt in two stages (Figure 2). The lower temperature (<50°C) transitions presumably correspond to the destacking of the loop nucleotides. The pseudouridine modification shifts the transition midpoint for the hairpin 3.5°C ± 0.5°C higher. This indicates that 32 increases the stability of the stem. The anneal spectra from 95–20°C also were acquired and showed 1.2°C ± 0.5°C of hysteresis at the Tm. Slow cooling of the RNA can lead to duplex formation and to hysteresis. The increased thermal stability of the modified RNA hairpin is consistent with other studies that predict pseudouridine enhances stability when located in a stem or at a loop–stem junction (8,9). Increased stability of the upper stem and slightly lower stability of the loop is supported by the NH spectrum (Figure 3). The guanine NH protons sharpen with the introduction of 32 and the U33 resonance is broadened.
Figure 2 Overlay of the UV melting curves of the unmodified (ACSL) and 32–modified RNA (32-ACSL) hairpins. Each of the hairpins appears to exhibit two melting transitions—the broad lower temperature transitions presumably corresponding to destacking of the loop nucleotide bases. The apparent melting temperatures of the unmodified and 32–modified RNA hairpins are estimated to be 77.0°C and 73.5°C, respectively.
Figure 3 Imino proton spectra of the (A) unmodified and (B) 32-modified ACSLPhe. The imino resonances of 32 and U33 have chemical shifts consistent with Watson–Crick base pairs but are exchange broadened in 32-ACSLPhe. This suggests that the pseudouridine has a small destabilizing effect on the loop leading to increased solvent exposure of the imino groups.
Resonance assignments of 32-ACSLPhe
The non-exchangeable 1H and 13C resonances of 32-ACSLPhe were assigned using standard heteronuclear methods (31,40). Most of the base and ribose 1H-13C correlations are resolved, and none of the resonances have spectral characteristics indicative of intermediate exchange. All 17 ribose spin systems, except for the incompletely labeled 5'-terminal nucleotide, were identified using 2D HCCH-COSY and 3D HCCH-TOCSY experiments. The five adenine intra-base H8-H2 correlations also were identified using the HCCH-TOCSY experiment. Intra-residue base-to-sugar correlations were identified using 2D 15N-1H HSQC experiments optimized to yield the multiple bond correlations H5/6-N1, H8-N9, and H1'-N1/N9 (41). All purine correlations and four of six pyrimidine correlations were identified in these spectra.
Sequence specific assignments were determined using two-dimensional (2D) NOESY (Figure 4) and 3D 13C-edited NOESY experiments (31). The sequential base-1' NOE connectivities are continuous through all 17 nt in the 180 ms NOESY spectrum, but the connectivity between nucleotides U33 and G34 is weak. With the exception of the A35 H8 to A36 H8 step, the base-2' and base–base inter-residue connectivities also are continuous in the 180 ms spectrum. Additionally, the base-2' inter-residue connectivities from U32 to A37 are weak, consistent with the non-C3'-endo ribose ring conformations of these residues (described below).
Figure 4 H6/8-to-H1' region of the 2D 400 ms NOESY spectrum. The base-1' proton sequential walk is traced with intra-residue peaks labeled. The 32 H1' resonance has a chemical shift of 4.56 p.p.m. and is data not shown. The arrows (a) points to the inter-residue sequential NOE between U33 H1' and G34 H8 and (b) non-sequential NOE between U33 H1' and A35 H8. The presence of the sequential NOE is not compatible with a U-turn motif for the loop whereas the non-sequential NOE can be produced by non-U-turn loop conformations.
The exchangeable NH and NH2 resonances were assigned using 2D NOESY experiments. The NH proton spectrum is shown in Figure 3. The three strong G NH resonances corresponding to G?C base pairs and one U NH resonance corresponding to an A?U base pair are connected by NOE cross peaks between NH proton resonances of adjacent base pairs. These connectivities are continuous in the helix from G28 to U39. The weak NH resonance of the terminal G?C base pair does not yield cross peaks in the NOESY spectrum. The U33 NH resonance is weak (Figure 3) and could only be assigned through a U33 N3–A37 H2 cross peak in the J(N, N)-HNN COSY spectrum (Figure 5) (33). A weak G NH resonance, which has a chemical shift that corresponds to a non-base paired guanine, was assigned to G34. The cytidine NH2 resonances were assigned using the strong intra-base pair C NH2 to G NH NOE cross peaks. The NH2 resonances of A37 and of all guanine nucleotides except G34 were not observed. The A31 NH2 resonances were assigned based on their NOE cross peaks with U39 H3. The NH2 15N resonances of A35, A36 and A38 were assigned based on intrabase H2 to N6 correlations in the J(N, N)-HNN COSY spectrum. The NH2 proton resonances of A35 and A36 are broad and have chemical shifts indicative of solvent-exposed protons.
Figure 5 (A) HNN-COSY and (B) multiple-bond 15N-1H HSQC spectra showing intra-residue U H5 to N3 and H6 to N1 correlations. The HNN-COSY shows cross-strand H2-N3 crosspeaks for residues 31–39, 38–32, and 37–33 produced by the Watson–Crick base pair configurations. The syn configuration about the 32 glycosidic bond and participation of 32 N1H in the 32-A38 base pair would produce a 32 N1-A38 H2 crosspeak (dashed circle). The U33–A37 crosspeak is weak and is indicative of a weak hydrogen bond.
Resonance assignments for 32 were accomplished using conventional heteronuclear methods. The 32 C1' resonance is shifted upfield into the region of the C4' resonances at 82.78 p.p.m. This position is consistent with the carbon–carbon glycosidic bond that replaces the more electronegative nitrogen–carbon bond of uridine. The 32 H6 has a sequential NOE cross peak with A31 H1' and the 32 H1' has a cross peak with the U33 H6. The 32 H1' assignment was confirmed through direct correlation with the 32 H6 using an HCCH-RELAY experiment. The 32 H1 and N1 imino resonances give rise to a cross peak in the 2D 15N-1H HMQC spectrum at 10.31 (1H) and 128.59 (15N) p.p.m. The identity of the N1 imino resonance is confirmed by the two-bond correlation with H6 in the multiple-bond HSQC spectrum (Figure 5). An intense intrabase NOE crosspeak between the N1H and H6 of 32 also supports assignment of the 32 H1. The 32 N3H resonance was assigned based on its NOE cross peak with the A38 H2 since the 32 H3 resonance is weak and does not give rise to a cross peak with the U39 NH resonance.
All of the internucleotide phosphate 31P resonances are clustered between –3.35 and –4.63 p.p.m. and were assigned using a 31P-1H hetero-TOCSY-NOESY experiment. The sequential P–H6/8 and P–H1' correlations are continuous throughout the molecule. The P–H3' correlations and several P–H4' and P–H5'/H5'' correlations are present in 31P–1H HetCor spectra and provide independent confirmation of the 31P assignments. The 5' terminal phosphate resonates at –1.30 p.p.m. and was assigned through cross peaks to the ribose protons of G27.
Structure calculations
The structure of 32-ACSLPhe was calculated using a rMD routine starting from 75 sets of coordinates with randomized backbone dihedral angles. The calculations used a total of 238 NOE derived distance constraints, 36 bp constraints, and 74 dihedral angle constraints (Table 1) resulting in 10 converged structures (Figure 6). Structures were classified as converged if they had low energy, few constraint violations, and predicted only NOEs that could be experimentally verified or explained. The converged structures have an average of 5 distance constraint violations between 0.1 and 0.3 ?, most of these involving the loop region. All converged structures have no constraints violated by more than 0.3 ?. The average root mean square deviations (RMSDs) of the heavy atoms between the individual structures and the minimized mean structure is 0.67 ? for the loop region (residues 33–37) and 0.92 ? for the stem region (residues 27–32 and 38–43) (Figure 6).
Table 1 Summary of experimental constraints and structure calculation statistics for 32-ACSLPhe
Figure 6 Stereoview of the superposition of (A) the stems and (B) the loops of the 10 converged 32-ACSLPhe structures. Convergence criteria are given in the text. The views are into the major groove. Only sugar and base heavy atoms are shown and the average r.m.s. deviation for the heavy atoms between the ten structures and the average structure is 1.14 ?. The loop and stem regions are locally well defined, but the propeller twist of the A31–U39 base pair is variable among the structures and slightly increases the r.m.s.d. of the full hairpin.
Structure of the loop region of 32-ACSLPhe
The 32-ACSLPhe loop is made up of the anticodon nucleotides G34–A36 and is closed by the U33?A37 base pair (Figure 7). The base of G34 stacks against the base of U33, consistent with the observed H6-H8 and H5-H8 NOEs. The bases of G34 and A35 are approximately coplanar and are parallel. This configuration satisfies the observed G34 H8–A35 H8 NOE cross-peak. The A36 base stacks beneath A37 base but is displaced toward the minor groove edge of A37 (Figure 7). Several NOEs involving the A36 H2 resonance help position this base and include interactions with the H2' of U33, the H8 of A35, and the G34 and/or A35 H1' (these 1' resonances are degenerate). These cross-peaks indicate that the Watson–Crick base pair functional groups are oriented to the interior of the molecule and point toward the residues flanking the 5' side of the nucleotide. The conserved NOE cross-peaks A36 H2 to A37 H2 and H1' suggest that A36 is partially stacked with A37. Thus, the observed NOEs support the compact tri-loop conformation.
Figure 7 Stereoview of minimized average structure of 32-ACSLPhe (residues G30 to C40). Hydrogen bonding NH and NH2 protons of base pairs 32-A38 and A31–U39 are colored purple and the exocyclic amino nitrogens (A31 and A38) are green. The pro-R(p) phosphoryl oxygens of A31 and 32 (yellow) are predicted to form water-mediated hydrogen bonds with 32 N1H and slow exchange of the N1 proton. The bridging water molecule is shown in pink. Explicit waters were not used for the structure calculations, but upper distance bounds derived from crystal structures were applied between 32 N1 and the phosphoryl oxygens.
The loop nucleotides also have unusual sugar–phosphate backbone conformations The large H1'–H2' couplings of residues G34–A36 indicate that the ribose sugar rings have the C2'-endo conformation. The strong intra-residue and weak sequential H2' to H8 NOE cross-peaks as well as the substantial downfield shift of the C3' and C4' resonances involving residues G34–A36 also support the C2'-endo conformation. Residues U33 and A37 at the junction of the loop and stem have a mixture of C2'- and C3'-endo conformations as evidenced by their intermediate H1'–H2' couplings and the modest downfield chemical shifts of their C3' and C4' resonances. The torsion angles of A35 and A36 have the -gauche conformation not typical of A-form geometry.
A superposition of the loop regions from the ten converged structures is shown in Figure 6B and the minimized average structure is shown in Figure 7. The helical base stack is continuous along the 3' side of the loop, with G34 stacking against the U33 base. On the 3' side of the loop, the A37 base straddles the bases of A35 and A36. The A36 N6 is 3 ? from the U33 O2' in half of the converged structures, suggesting the possibility of a cross strand base–sugar hydrogen bond. However, this interaction could not be confirmed since the A36 NH2 proton resonances could not be assigned and no 2'-OH proton resonances were identified.
The loop of the 32-ACSLPhe does not contain the classical ‘U-turn’ motif. Although weak, the inter-residue U33 H1'–G34 H8 and U33 H6–G34 H8 NOE cross-peaks (Figure 4) are consistent with the U33 and G34 positions shown in Figure 7 but not with their positions in the U-turn (Figures 8B and C). A U33 H1' to A35 H8 NOE is observed, but only at long mixing time and is even less intense than the sequential U33 H1'–G34 NOE (Figure 4). The ribose puckers of the anticodon nucleotides of the U-turn also tend toward the C3'-endo conformation rather than the C2'-endo conformation observed in this study. Another feature characteristic of the U-turn motif is the trans conformation of backbone torsion angle between U33 and G34. The 31P resonance corresponding to this phosphate is located in the main cluster of 31P peaks (Supplementary Table S1). The chemical shift of this resonance is not consistent with the 31P chemical shift predicted for a phosphate having the trans conformation of the torsional angle (38).
Figure 8 Comparison of residues 31–39 of (A) the solution structure of E.coli tRNAPhe (42) and the crystal structures of (B) fully modified yeast tRNAPhe (15,16) and (C) fully modified tRNAAsp (18,21). 32 and A38 in (A) form a Watson–Crick base pair. In (B) and (C), nucleotides C32-A38 and 32-C38, respectively, form the bifurcated hydrogen bond. The anticodon loop in (A) adopts the tri-loop conformation whereas the anticodon loops in (B) and (C) adopt the U-turn motif. In E.coli tRNACys, 32 is in the syn configuration about the glycosidic bond and the loop forms a U-turn motif (17).
Structure of the stem of 32-ACSLPhe
The geometry of the hairpin stem, base pairs G27?C43 to U33?A37, is primarily A-form (Figure 6A). The sequential base-1', 2' and several base–base NOEs are continuous at 180 ms mixing time. The A31?U39 and 32?A38 base pairs are sufficiently stable to give rise to a cross-strand NH–H2 NOEs, but the NH proton of the U33?A37 base pair is not. The 32 and U33 NH proton resonances are weak, but their participation in Watson–Crick pairing schemes was confirmed using a J(N, N)-HNN-COSY spectrum (Figure 5). Nucleotides A31, A37, and A38 also produce cross-strand H2-H1' NOEs commonly observed for A?U base pairs within helices. The H2 to N3 through bond correlations clearly confirm that U39 is base paired with A31 and 32 is base paired with A38 through its N3 imino group (Figure 7).
The torsion angles of the sugar-phosphate backbone are within the limits of A-form geometry. The small (<5 Hz) H1'–H2' couplings and the 13C chemical shifts of the 3' and 4' resonances of nucleotides G28–32 and A38–C42 are indicative of the typical C3'–endo sugar pucker. None of the or torsional angles within the converged structures deviated from gauche–.
The most remarkable feature of the stem is the minimal structural perturbation caused by 32. Although the combination of A31 H8-32 H6 and A31 H2-32 H6 NOEs is unusual, the relative positions of the adjacent A31 and 32 bases that give rise to the interactions are accommodated within the A-form helix.
DISCUSSION
We are using the anticodon stem–loop of E.coli tRNAPhe as a model to probe the cumulative physical effects of sequential modifications in RNA molecules. Pseudouridine is the most common base modification found in RNA and the anticodon stem-loop of E.coli tRNAPhe has two of these nucleotides, 32 and 39. In this study, we have examined the thermodynamic and structural effects of 32 on the otherwise unmodified tRNAPhe anticodon stem–loop.
The structures of ACSLPhe and 32-ACSLPhe are similar
The unmodified anticodon stem–loop of ACSLPhe forms a highly ordered tri-loop conformation (42) and the 32-ACSLPhe adopts a similar structure. The stems of the two hairpins are continuous through base pair U33–A37, but this loop-closing base pair is weaker in the 32-modified molecule indicating a slightly less compact tri-loop. Also, fewer constraints involving G34 and A35 in 32-ACSLPhe lead to greater variability in the positions of these residues compared to the unmodified molecule, including excursions of A35 to the minor groove side of the loop. However, the lower number of experimental constraints in the loop is due to spectral overlap of G34 and A35 H1' resonances and probably does not reflect conformational heterogeneity inherent in the loop. Nonetheless, the small chemical shift differences between unmodified and 32-modified ACSLPhe molecules indicate that the loop structures are not identical. The energy difference between the converged conformations of 32-ACSLPhe is small (<15 kcal/mole), and the RMSD between the 32-ACSLPhe and ACSLPhe minimized average structures is 1.13 ?.
The pseudouridylation of U32 increases the overall stability of the RNA hairpin. This is consistent with studies that have examined the thermodynamic effects of pseudouridine incorporation (9,13,43,44). Pseudouridine within a helix can increase the melting temperature of the helix 3–5°C depending upon the neighboring base pairs and improves local base stacking (9). A contributing factor to the stabilization effect of is its ability to form a water-mediated hydrogen bond with the phosphate backbone (9). This hydrogen bond interaction has been inferred from difference maps of X-ray crystallographic studies of unmodified and fully modified tRNAGln molecules (19) and is present in molecular dynamics simulations of yeast tRNAAsp (45). The strong exchange-protected 32 N1H resonance (Figure 3) supports this interaction within the ACSLPhe and is modeled in the average structure (Figure 7). Pseudouridine also confers a smaller degree of helix stabilization when present in single strand regions adjacent to a helix (46) or is on the 3' end of the loop at loop-helix junctions such as 39 (46). 32 is adjacent to the helix-loop junction (U33–A37), but is at the 5' end of the loop. Thus, pseudouridine has a comparable stabilizing effect when positioned on either the 3' or 5' edge of a helix.
Pseudouridine tends to form a Watson–Crick base pair with adenine in helical contexts and the 32 is no exception (43). The HNN-COSY spectrum confirms the hydrogen bond configurations of the 32-A38 and U33–A37 base pairs (Figure 5). However, the H3 resonances of U33 and 32 are weaker than the corresponding H3 resonances of unmodified ACSLPhe (Figure 3). This weakening may reflect altered alignment of the hydrogen bonds of these base pairs as the base of 32 is stabilized toward the major groove by the water-mediated N1H-phosphate hydrogen bond (Figure 7). A 32 H6-A31 H2 NOE is consistent with the major groove displacement of the 32 base. This NOE is not observed in the unmodified molecule (42) and is not typical of 5'-AU-3' base stacking in helices with regular A-form geometry.
The 32–38 bp in tRNA
The nucleotide composition of residues 32 and 38 is not an equal distribution of the pyrimidine and purine bases. tRNA sequence data show that position 38 is often adenine (67%), but a significant fraction (30%) of position 38 nt are pyrimidines (14). Residue 32 is almost exclusively occupied by a pyrimidine base (>97%) and is modified in approximately a third of sequenced tRNA molecules on either the base or ribose group. The most common base modification at position 32 is pseudouridine and although this base does not alter the overall structure of the ACSLPhe loop compared to uridine, it partially destabilizes the tri-loop conformation and increases the stem melting temperature by about 3.5°C. Other modifications at position 32 are frequently those that favor the C3' endo conformation of the ribose, such as pyrimidine 2'-O-methyl and the pyrimidine C2-thiol (47), which may help to maintain the helical integrity of the stem of the anticodon arm (48). Whether these modifications can alter or stabilize anticodon loop conformations or function as pseudouridine primarily to stabilize the stem has not been investigated.
The geometry of the 32 base in solution is different from that observed for position 32 residues in crystal structures of tRNAs. The crystal structures of several tRNAs reveal a highly conserved structural motif, the bifurcated hydrogen bond, between residues 32 and 38 (49) (Figure 8). The bifurcated hydrogen bond frequently involves proton donation by an exocyclic NH2 group and acceptance by the O2 of C32 or U32. This hydrogen bond configuration imposes an underwinding of the helix geometry at the 32–38 base step. For E.coli tRNAPhe, the configuration would correspond to hydrogen bonding between A38 N6H2 and 32 O4 (adjacent to the C5-C1' glycosidic bond). This base arrangement of 32 and A38 is isosteric with other 32–38 pairs, such as C32–A38 and U32–C38, that adopt the bifurcated hydrogen bond (49) (Figure 8). Auffinger and Westhof (49) have proposed the geometry of nt 32 and 38 resulting from the bifurcated hydrogen bond creates a transition between the stem and loop in the anticodon arm and facilitates formation of the U-turn. Thus, the 32–38 interaction may allow an open conformation of the loop without distortion of the canonical A-form stem. The results of this study indicate that if the bifurcated hydrogen exists in solution and functions in U-turn stabilization, the 32 modification alone is not sufficient to organize this motif.
The crystal structure of fully modified Cys-tRNACys in complex with EF-Tu and GDPNP presents an alternative conformation for the 32-A38 base pair. In this structure, the 32 base adopts the syn configuration about the glycosidic bond and does not form any interactions with A38 (17). This configuration of the base can accommodate the 32 water-mediated hydrogen bond to the phosphate backbone and thus should retain the stabilizing effect conferred by pseudouridine in the anti configuration. Notably though, the anticodon loop in this complex adopts the characteristic U-turn motif with stacking of the anticodon bases G, C, and A. The EF-Tu protein-binds at the acceptor stem and has no contacts in the region of the anticodon arm, but the G and C nucleotides of the anticodon form intermolecular base pairs within the unit cell of the crystal. The structure of this molecule is of particular interest since the anticodon arm of tRNACys contains the same nucleotide base modifications found in fully modified E.coli tRNAPhe, specifically 32, 39, and ms2i6A37. These additional modifications could affect the conformation of the 32 base, however, neither 39 nor i6A37 alone are sufficient to induce the syn conformation of 32 in solution (J.Cabello and E.P.Nikonowicz, unpublished data).
Pseudouridylation within the anticodon arm of tRNA is important to the overall fitness of an organism and loss of this modification can lead to a spectrum of growth defects in bacteria and yeast (50–54). Disruption of the truA gene which catalyzes formation of 38, 39, and 40 decreases polypeptide chain elongation rates and reduces cell growth rate 30% (1,55). A similar effect is observed for the deg1 gene of yeast that catalyzes 38 and 39 formation (54,56). E.coli that contain a catalytically inactive truB gene which catalyzes formation of 55 exhibit no growth defects (52), even though 55 is universally conserved (14). Pseudouridylation at position 32 is intermediate in its physiological effects. Cells that lack RluA activity have near normal growth rates, but are strongly selected against when grown in competition with wild type cells (51). However, given that RluA modifies only six tRNA isoaccpetors, unlike TruA and TruB that modify their target sites in all tRNA molecules in E.coli, the physiological effects of its loss are remarkable and suggest a critical role of 32 in proper tRNA function. The specific translational defects associated with loss of 32 have not been determined, but studies of position 32 mutations in other tRNAs suggest translational fidelity and wobble base recognition are likely to be impaired (23–25,57). The studies presented here indicate that a primary role of 32 is the stabilization of the stem of the anticodon arm. Thus, 32 may serve as a buffer to preserve the stability and the architecture of the stem within the context of other loop-destabilizing nucleotide modifications (42) or after formation of the U-turn.
SUPPLEMENTARY DATA
Supplementary Data are available at NAR Online.
ACKNOWLEDGEMENTS
We thank M. Michnicka for preparation of the T7 RNA polymerase and isotopically labeled 5'-rNTPs. This work was supported by NSF grant MCB-0078501 and by a grant from the Robert A. Welch Foundation (C1277) to E.P.N. Funding to pay the Open Access publication charges for this article was provided by the Robert A. Welch Foundation.
REFERENCES
Bj?rk, G.R. (1995) Biosynthesis and function of modified nucleosides In S?ll, D. and RajBhandary, U.L. (Eds.). tRNA Structure, Biosynthesis, and Function, Washington, D.C. American Society for Microbiology Press pp. 165–206 .
Garcia, G.A. and Goodenough-Lashua, D.M. (1998) Mechanisms of modifying enzymes In Grosjean, H. and Benne, R. (Eds.). Modification and Editing of RNA: The Alteration of RNA Structure and Function, Washington, D.C. ASM Press pp. 135–168 .
Winkler, M.E. (1998) Genetics and regulation of tRNA and rRNA modification In Grosjean, H. and Benne, R. (Eds.). Modification and Editing of RNA: The Alteration of RNA Structure and Function, Washington, D.C. ASM Press pp. 441–469 .
Agris, P.F. (1996) The importance of being modified: roles of modified nucleosides and Mg2+ in RNA structure and function Prog. Nucleic Acid Res. Mol. Biol, . 53, 79–129 .
Charette, M. and Gray, M.W. (2000) Pseudouridine in RNA: what, where, how, and why IUBMB Life, 49, 341–351 .
Ofengand, J. (2002) Ribosomal RNA pseudouridines and pseudouridine synthases FEBS Lett, . 514, 17–25 .
Neumann, J.M., Bernassau, J.M., Gueron, M., Tran-Dinh, S. (1980) Comparative conformations of uridine and pseudouridine and their derivatives Eur. J. Biochem, . 108, 457–463 .
Davis, D.R. (1998) Biophysical and conformational properties of modified nucleosides in RNA (NMR Studies) In Grosjean, H. and Benne, R. (Eds.). Modification and Editing of RNA, Washington, D.C. ASM Press pp. 85–101 .
Davis, D.R. (1995) Stabilization of RNA stacking by pseudouridine Nucleic Acids Res, . 23, 5020–5026 .
Roy, S., Papastavros, M.Z., Sanchez, V., Redfield, A.G. (1984) Nitrogen-15-labeled yeast tRNAPhe: double and two-dimensional heteronuclear NMR of guanosine and uracil ring NH groups Biochemistry, 23, 4395–4400 .
Griffey, R.H., Davis, D., Yamaizumi, Z., Nishimura, S., Bax, A., Hawkins, B., Poulter, C.D. (1985) 15N-labeled Escherichia coli tRNAfMet, tRNAGlu, tRNATyr, and tRNAPhe. Double resonance and two-dimensional NMR of N1-labeled pseudouridine J. Biol. Chem, . 260, 9734–9741 .
Hruska, F.E., Grey, A.A., Smith, I.C. (1970) A nuclear magnetic resonance study of the molecular conformation of ?-pseudouridine in aqueous solution J. Am. Chem. Soc, . 92, 4088–4094 .
Yarian, C.S., Basti, M.M., Cain, R.J., Ansari, G., Guenther, R.H., Sochacka, E., Czerwinska, G., Malkiewicz, A., Agris, P.F. (1999) Structural and functional roles of the N1- and N3-protons of at tRNA's position 39 Nucleic Acids Res, . 27, 3543–3549 .
Sprinzl, M., Horn, C., Brown, M., Loudovitch, A., Steinberg, S. (1998) Compilation of tRNA sequences and sequences of tRNA genes Nucleic Acids Res, . 26, 148–153 .
Robertus, J.D., Ladner, J.E., Finch, J.R., Rhodes, D., Brown, R.S., Clark, B.F., Klug, A. (1974) Structure of yeast phenylalanine tRNA at 3 ? resolution Nature, 250, 546–551 .
Kim, S.-H., Suddath, F.L., Quigley, G.J., McPherson, A., Sussman, J.L., Wang, A.H.J., Seeman, N.C., Rich, A. (1974) Three-dimensional tertiary structure of yeast phenylalanine transfer RNA Science, 185, 435–440 .
Nissoen, P., Thirup, S., Kjeldgaard, M., Nyborg, J. (1999) The crystal structure of Cys-tRNA-EF-Tu-GDPNP reveals general and specific features in the ternary complex and in tRNA Structure, 7, 143–156 .
Moras, D., Comarmond, M.B., Fischer, J., Weiss, R., Thierry, J.C., Ebel, J.P., Giege, R. (1980) Crystal structure of yeast tRNAAsp Nature, 288, 669–674 .
Arnez, J.G. and Steitz, T.A. (1994) Crystal structure of unmodified tRNAGln complexed with glutaminyl-tRNA synthetase and ATP suggests a possible role for pseudo-uridines in stabilization of RNA structure Biochemistry, 33, 7560–7567 .
Shi, H. and Moore, P.B. (2000) The crystal structure of yeast phenylalanine tRNA at 1.93 ? resolution: a classic structure revisited RNA, 6, 1091–1105 .
Westhof, E., Dumas, P., Moras, D. (1985) Crystallographic refinement of yeast aspartic acid transfer RNA J. Mol. Biol, . 184, 119–145 .
Yarus, M. (1982) Translational efficiency of transfer RNAs: uses of an extended anticodon Science, 218, 646–652 .
Yarus, M., Cline, S., Raftery, L., Wier, P., Bradley, D. (1986) The translational efficiency of tRNA is a property of the anticodon arm J. Biol. Chem, . 261, 10496–10505 .
Lustig, F., Boren, T., Claesson, C., Simonsson, C., Barciszewska, M., Lagerkvist, U. (1993) The nucleotide in position 32 of the tRNA anticodon loop determines ability of anticodon UCC to discriminate among glycine codons Proc. Natl Acad. Sci. USA, 90, 3343–3347 .
Claesson, C., Lustig, F., Boren, T., Simonsson, C., Barciszewska, M., Lagerkvist, U. (1995) Glycine codon discrimination and the nucleotide in position 32 of the anticodon loop J. Mol. Biol, . 247, 191–196 .
O'Connor, M. (1998) tRNA imbalance promotes -1 frameshifting via near-cognate decoding J. Mol. Biol, . 279, 727–736 .
Davanloo, P., Rosenberg, A.H., Dunn, J.J., Sturdier, F.W. (1984) Cloning and expression of the gene for bacteriophage T7 RNA polymerase Proc. Natl Acad. Sci. USA, 81, 2035–2039 .
Wrzesinski, J., Nurse, K., Bakin, A., Lane, B.G., Ofengand, J. (1995) A dual-specificity pseudouridine synthase: an Escherichia coli synthase purified and cloned on the basis of its specificity for 746 in 23S RNA is also specific for 32 in tRNAPhe RNA, 4, 437–448 .
Nikonowicz, E.P., Sirr, A., Legault, P., Jucker, F.M., Baer, L.M., Pardi, A. (1992) Preparation of 13C and 15N Labelled RNAs for heteronuclear multi-dimensional NMR studies Nucleic Acids Res, . 20, 4507–4513 .
Markley, J.L., Bax, A., Arata, Y., Hilbers, C.W., Kaptein, R., Sykes, B.D., Wright, P.E., Wuthrich, K. (1998) Recommendations for the presentation of NMR structures of proteins and nucleic acids. IUPAC-IUBMB-IUPAB Inter-Union Task Group on the Standardization of Data Bases of Protein and Nucleic Acid Structures Determined by NMR Spectroscopy J. Biomol. NMR, 12, 1–23 .
Nikonowicz, E.P. and Pardi, A. (1993) An efficient procedure for assignment of the proton, carbon, and nitrogen resonances in 13C/15N labeled nucleic acids J. Mol. Biol, . 232, 1141–1156 .
Legault, P., Farmer, B.T., II, Mueller, L., Pardi, A. (1994) Through-bond correlation of adenine protons in a 13C-labeled ribozyme J. Am. Chem. Soc, . 116, 2203–2204 .
Hennig, M. and Williamson, J.R. (2000) Detection of N-H???N hydrogen bonding in RNA via scalar couplings in the absence of observable imino proton resonances Nucleic Acids Res, . 28, 1585–1593 .
Kellogg, G.W. and Schweitzer, B.I. (1993) Two- and three-dimensional 31P-driven NMR procedures for complete assignment of backbone resonances in oligodeoxyribonucleotides J. Biomol. NMR, 3, 577–595 .
Altona, C. (1982) Conformational analysis of nucleic acids. Determination of backbone geometry of single-helical RNA and DNA in aqueous solution Recl. Trav. Chim. Pays-Bas, 101, 413–433 .
Smith, J.S. and Nikonowicz, E.P. (1998) NMR structure and dynamics of an RNA motif common to the spliceosome branch-point helix and the RNA-binding site for phage GA coat protein Biochemistry, 37, 13486–13498 .
Varani, G., Aboul-ela, F., Allain, F.H.-T. (1996) NMR investigation of RNA structure Prog. NMR Spectroscopy, 29, 51–127 .
Gorenstein, D.G. (1981) Nucleotide conformational analysis by 31P nuclear magnetic resonance spectroscopy Annu Rev Biophys Bioeng, . 10, 355–386 .
Brünger, A.T. X-PLOR Version 3.1 Manual, (1992) NH Yale University .
Pardi, A. (1995) Multidimensional heteronuclear NMR experiments for structure determination of isotopically labeled RNA Methods Enzymol, . 261, 350–380 .
Sklenar, V., Peterson, R.D., Rejante, M.R., Feigon, J. (1994) Correlation of nucleotide base and sugar protons in a 15N-labeled HIV-1 RNA oligonucleotide by 1H-15N HSQC experiments J. Biomol. NMR, 4, 117–122 .
Cabello-Villegas, J., Winkler, M.E., Nikonowicz, E.P. (2002) Solution conformations of unmodified and A37N6-dimethylallyl modified anticodon stem–loops of Escherichia coli tRNAPhe J. Mol. Biol, . 319, 1015–1034 .
Hall, K.B. and McLaughlin, L.W. (1992) Properties of pseudouridine N1 imino protons located in the major groove of an A-form RNA duplex Nucleic Acids Res, . 20, 1883–1889 .
Meroueh, M., Grohar, P.J., Qiu, J., SantaLucia, J., Jr, Scaringe, S.A., Chow, C.S. (2000) Unique structural and stabilizing roles for the individual pseudouridine residues in the 1920 region of Escherichia coli 23S rRNA Nucleic Acids Res, . 28, 2075–2083 .
Auffinger, P. and Westhof, E. (1997) RNA hydration: three nanoseconds of multiple molecular dynamics simulations of the solvated tRNAAsp anticodon hairpin J. Mol. Biol, . 269, 326–341 .
Davis, D.R., Veltri, C.A., Nielsen, L. (1998) An RNA model system for investigation of pseudouridine stabilization of the codon-anticodon interaction in tRNALys, tRNAHis and tRNATyr J. Biomol. Struct. Dyn, . 15, 1121–1132 .
Yokoyama, S. and Nishimura, S. (1995) Modified nucleosides and codon recognition In S?ll, D. and RajBhandary, U.L. (Eds.). tRNA Structure, Biosynthesis, and Function, Washington, D.C. American Society for Microbiology Press pp. 207–224 .
Kawai, G., Yamamoto, Y., Kamimura, T., Masegi, T., Sekine, M., Hata, T., Iimori, T., Watanabe, T., Miyazawa, T., Yokoyama, S. (1992) Conformational rigidity of specific pyrimidine residues in tRNA arises from posttranscriptional modifications that enhance steric interaction between the base and the 2'-hydroxyl group Biochemistry, 31, 1040–1046 .
Auffinger, P. and Westhof, E. (1999) Singly and bifurcated hydrogen-bonded base-pairs in tRNA anticodon hairpins and ribozymes J. Mol. Biol, . 292, 467–483 .
Tsui, H.C., Arps, P.J., Connolly, D.M., Winkler, M.E. (1991) Absence of hisT-mediated tRNA pseudouridylation results in a uracil requirement that interferes with Escherichia coli K-12 cell division J Bacteriol, 173, 7395–7400 .
Raychaudhuri, S., Niu, L., Conrad, J., Lane, B.G., Ofengand, J. (1999) Functional effect of deletion and mutation of the Escherichia coli ribosomal RNA and tRNA pseudouridine synthase RluA J. Biol. Chem, . 274, 18880–18886 .
Gutgsell, N., Englund, N., Niu, L., Kaya, Y., Lane, B.G., Ofengand, J. (2000) Deletion of the Escherichia coli pseudouridine synthase gene truB blocks formation of pseudouridine 55 in tRNA in vivo, does not affect exponential growth, but confers a strong selective disadvantage in competition with wild-type cells RNA, 6, 1870–1881 .
Del Campo, M., Kaya, Y., Ofengand, J. (2001) Identification and site of action of the remaining four putative pseudouridine synthases in Escherichia coli RNA, 7, 1603–1615 .
Lecointe, F., Simos, G., Sauer, A., Hurt, E.C., Motorin, Y., Grosjean, H. (1998) Characterization of yeast protein Deg1 as pseudouridine synthase (Pus3) catalyzing the formation of 38 and 39 in tRNA anticodon loop J. Biol. Chem, . 273, 1316–1323 .
Singer, C.E., Smith, G.R., Cortese, R., Ames, B.N. (1972) Mutant tRNAHis ineffective in repression and lacking two pseudouridine modifications Nature New Biol, . 238, 72–74 .
Carbone, M.L., Solinas, M., Sora, S., Panzeri, L. (1991) A gene tightly linked to CEN6 is important for growth of Saccharomyces cerevisiae Curr. Genet, . 19, 1–8 .
Yarus, M., Cline, S.W., Wier, P., Breeden, L., Thompson, R.C. (1986) Actions of the anticodon arm in translation on the phenotypes of RNA mutants J. Mol. Biol, . 192, 235–255 .(Javier Cabello-Villegas and Edward P. Ni)
*To whom correspondence should be addressed. Tel: +1 713 348 4912; Fax +1 713 348 5154; Email: edn@bioc.rice.edu
ABSTRACT
Nucleoside base modifications can alter the structures and dynamics of RNA molecules and are important in tRNAs for maintaining translational fidelity and efficiency. The unmodified anticodon stem–loop from Escherichia coli tRNAPhe forms a trinucleotide loop in solution, but Mg2+ and dimethylallyl modification of A37 N6 destabilize the loop-proximal base pairs and increase the mobility of the loop nucleotides. The anticodon arm has three additional modifications, 32, 39, and A37 C2-thiomethyl. We have used NMR spectroscopy to investigate the structural and dynamical effects of 32 on the anticodon stem-loop from E.coli tRNAPhe. The 32 modification does not significantly alter the structure of the anticodon stem–loop relative to the unmodified parent molecule. The stem of the RNA molecule includes base pairs 32-A38 and U33–A37 and the base of 32 stacks between U33 and A31. The glycosidic bond of 32 is in the anti configuration and is paired with A38 in a Watson–Crick geometry, unlike residue 32 in most crystal structures of tRNA. The 32 modification increases the melting temperature of the stem by 3.5°C, although the 32 and U33 imino resonances are exchange broadened. The results suggest that 32 functions to preserve the stem integrity in the presence of additional loop modifications or after reorganization of the loop into a translationally functional conformation.
INTRODUCTION
Posttranscriptional modification of RNA molecules occurs in all cells (1–3). The modifications, which are primarily localized to the nucleotide bases, can alter the chemical, structural, and thermodynamic properties of RNAs and thereby contribute to RNA function (4). Many nucleotide base modifications have been chemically well characterized but the impact of these modifications on RNA structure and stability has been less thoroughly examined. Pseudouridine is the most common nucleotide base modification and is found in the tRNA and rRNA of all cells (5,6). This modification is formed by an isomerization of uridine resulting in a C5-C1' base–ribose glycosidic bond and a second imino (NH) functionality. Pseudouridine preferentially adopts the syn conformation about the glycosidic bond as the free nucleotide (7,8), but the anti configuration has been most frequently observed for pseudouridine within oligonucleotides and double helices (9). Consequently pseudouridine tends to base pair with adenosine through the N3 imino and the C2 carbonyl groups (Figure 1) (10–13) and imparts thermodynamic stability to helices when located in the interior of an RNA duplex or in a single strand region adjacent to a duplex (8).
Figure 1 Sequences of (A) the 32-modified and (B) fully modified RNA hairpins corresponding to the anticodon arm of E.coli tRNAPhe. Nucleotide numbering corresponds to the full-length tRNAPhe molecule. designates pseudouridine and ms2i6A designates (2-thiomethyl, N6-dimethylallyl)-adenine. 32-A38 base arrangements for (C) the Watson–Crick base pair and (D) the bifurcated hydrogen bond interaction.
In addition to the nearly universally conserved 55, pseudouridine occurs frequently at several other positions in tRNA including residues 13, 32, 39, and 40 (14), although the frequency of occurrence at these positions varies among species. 55, located in the T-loop, serves only as an acceptor in its hydrogen bonding with the imino and amino groups of G18. The 55-G18 pair stacks between the flanking base pairs and contributes to maintenance of the tertiary fold of tRNA (15,16). In yeast tRNAPhe and other tRNAs or tRNA complexes, 39, at the bottom of the anticodon stem, generally aligns with A31 in a standard A–U Watson–Crick hydrogen bond geometry (15–20), although the hydrogen bonds tend to be somewhat long (2.1–2.3 ?). 32 also is located in the anticodon arm of six tRNA species in Escherichia coli and adopts a syn configuration about its glycosidic bond in tRNACys (17). In yeast tRNAAsp, 32 forms a bifurcated base pair with C38 involving 32 O4 and the exocyclic amino hydrogens of C38 (18,21).
Residue 32 in the anticodon arm is important for the translation function of tRNA. A central element of the extended anticodon hypothesis is that nucleotides at the stem–loop junction in the anticodon arm contribute to the translational efficiency of tRNA (22,23). The functional importance of residues 32 and 38 was first demonstrated using amber suppressor tRNAs Su2 and Su7 that incorporate glutamine and tryptophan, respectively, at the stop codon 5'-UAG-3'. These suppressor tRNAs share the same anticodon nucleotide sequence, but Su2 is a weak suppressor whereas Su7 is a strong suppressor. The different suppression efficiencies of these tRNAs originate mainly from the identities of nucleotides 32 and 38 (23). In addition to the suppressor tRNA studies of Yarus and coworkers (23), the functional significance of the 32–38 bp is further illustrated by nucleotide substitution studies of glycyl tRNAs of E.coli and Mycoplasma mycoides (24,25). The genome of M.mycoides encodes one tRNAGly (anticodon 5'-UCC-3') that is used to decode all four glycine codons and has a C32–A38 mismatch in the anticodon arm. In E.coli, tRNAGly,2 has the anticodon 5'-UCC-3', but has the base pair U32–A38. When mutated to a U32–A38 base pair, the ability of the M.mycoides tRNAGly to decode non-cognate codons dramatically diminishes (25). Similarly, mutation of the U32–A38 base pair of E.coli tRNAGly,2 to C32–A38 leads to decreased fidelity and –1 frameshifting at 5'-GGG-3' codons (26). These studies demonstrate the ability of the 32–38 interaction to modulate the wobble properties of the anticodon. Thus, the types of interactions formed by residues 32 and 38 may alter the conformation and dynamics of the anticodon loop and/or the interactions within the codon-anticodon complex at the ribosome.
We have used heteronuclear NMR spectroscopy to examine the solution structure of the 32-modified form of the anticodon arm of E.coli tRNAPhe. A 17 nt RNA molecule that forms a stem–loop secondary structure in solution and corresponds to the anticodon arm of tRNAPhe was used (Figure 1). Our results demonstrate that the loop of the modified RNA molecule is composed of three nucleotides and lacks the characteristic U-turn motif. The anticodon arm is extended by two base pairs as in the unmodified parent molecule and 32 increases the stability of the stem. The 32 base forms a Watson–Crick type base pair with A38 and not the bifurcated hydrogen bond configuration found in most crystal structures of tRNAs.
MATERIALS AND METHODS
All enzymes were purchased from Sigma Chemical, except for T7 RNA polymerase and RluA enzymes, which were prepared as described (27,28). Deoxyribonuclease I type II, pyruvate kinase, adenylate kinase, and nucleotide monophosphate kinase were obtained as powders, dissolved in 15% glycerol, 1 mM dithiothreitol, and 10 mM Tris–HCl, pH 7.4, and stored at –20°C. Guanylate kinase and nuclease P1 were obtained as solutions and stored at –20°C. Unlabeled 5' nucleoside triphosphates (5'-NTPs) were purchased from Sigma, phosphoenolpyruvate (potassium salt) was purchased from Bachem, and 99% {15N}-ammonium sulfate and 99% {13C}- glucose were purchased from Isotec.
Preparation of RNA samples
The RNA sequence for E.coli ACSLPhe shown in Figure 1 was synthesized in vitro using T7 RNA polymerase and a synthetic DNA template. The nucleotide sequence of the stem corresponds to residues G27–C43 of full-length E.coli tRNAPhe. Isotopically labeled RNA molecules were prepared from 10 ml transcription reactions using 3 mM uniformly 15N-enriched and 13C-enriched 5'-NTPs as described (29). The RNA molecules were purified by passage through 20% (w/v) preparative PAGE, electroeluted (Schleicher and Schuell), and precipitated with ethanol. The purified oligonucleotides were dissolved in 1.0 M NaCl, 20 mM potassium phosphate, pH 6.8, and 2.0 mM EDTA, and dialyzed extensively against 10 mM NaCl, 10 mM potassium phosphate, pH 6.8, and 0.05 mM EDTA, using a Centricon-3 concentrator (Amicon Inc.). The samples were diluted with buffer to a volume of 0.2 ml and lyophilized to a powder. For experiments involving the non-exchangeable protons, the samples were exchanged twice with 99.9% D2O and then resuspended in 0.2 ml of 99.96% D2O. For experiments involving detection of the exchangeable protons, the samples were resuspended in 0.2 ml of 90% H2O/10% D2O. The samples contained 80 and 100 A260 OD units of 15N-labeled and 13C-labeled, respectively, RNA oligonucleotides (2.5–3.4 mM).
Preparation of the 32-ACSLPhe
Pseudouridine was introduced at position U32 of purified ACSLPhe using the pseudouridine synthase RluA (RluA). Histidine-tagged RluA was expressed in E.coli and purified using Ni2+ affinity resin as described (28). The pseudouridylation reaction was carried out using a mole ratio of RluA:RNA of 1:48. The reaction conditions were 50 mM HEPES pH 7.5, 100 mM NH4Cl, 0.03 mg/ml RluA, 0.06 mM ACSLPhe. The reactions were allowed to proceed overnight at 37°C. The RNA was purified from the RluA enzyme by heating the reaction to 90°C for 2 min followed by centrifugation to remove the precipitated protein. The supernatant was dialyzed with NMR buffer. Completion was determined by the disappearance of the C5–H5 resonance of U32 from 2D 13C-1H HMQC spectra (Supplementary Figure S1). The reactions could not be monitored using denaturing PAGE because the modified ACSLPhe migrates at the same rate as the unmodified ACSLPhe.
NMR spectroscopy
All NMR spectra were acquired on a Bruker AMX-500 spectrometer equipped with a 1H-{X} broadband probe, except for the 31P-decoupled 13C-1H constant time HSQC experiment, which was collected with a 1H-{13C, 31P} triple resonance probe. Broadband decoupling of the carbon and nitrogen resonances was achieved using GARP with B2 = 3125 Hz for carbon and B2 = 1570 Hz for nitrogen. H2O spectra were collected at 12°C with solvent suppression using either spin lock pulses or binomial read pulses with maximum excitation at 12.5 p.p.m. D2O spectra were collected at 25°C with presaturation or spin lock pulses to suppress the residual HDO peak. Quadrature detection was achieved using the States-TPPI method, and acquisition was delayed by a half-dwell in all indirectly detected dimensions. Typically, the data points were extended by 25% using linear prediction for the indirectly detected dimensions and the data were apodized using 1 Hz line broadening and 65° shifted sinebell functions. 1H spectra were referenced relative to DSS (0.00 p.p.m.). References for the 13C and 15N spectra were calculated using the spectrometer frequencies as reported (30). The 31P spectra were referenced to an external standard of TMP which was set at 0.00 p.p.m. All spectra were processed and analyzed with Felix 98.0 (Accelerys, Inc.).
2D 13C-1H HMQC and HSQC spectra were collected to identify 13C-1H chemical shift correlations. 2D HCCH-COSY and 3D HCCH-TOCSY (24 ms DIPSI-3 spin lock) experiments optimized for polarization transfer through the ribose carbons and a 2D 13C-1H HCCH-TOCSY (52 ms DIPSI-3 spin lock) optimized for polarization transfer through the adenine bases were collected in D2O to identify ribose spin systems and H8-H2 correlations, respectively (31,32). To identify intra-residue base-sugar correlations, a 2D 15N-1H HSQC experiment was acquired in D2O and optimized for two- and three-bond correlations. A J(N, N)-HNN COSY experiment was acquired in D2O to confirm the presence of A?U base pairs (33) and 31P assignments were obtained using a 31P/1H hetero-TOCSY-NOESY spectrum (34).
Distance constraints for the non-exchangeable resonances of 32-modified ACSLPhe were derived at 25°C from 2D 1H-1H NOESY spectra (80, 120, 180, 360, and 480 ms mixing times), 13C-edited 3D NOESY-HMQC spectra (180 and 360 ms mixing times), and 13C-edited 3D NOESY-ctHSQC spectra (80, 180, and 360 ms mixing times) optimized for the ribose resonances in 2 and 3. For the exchangeable resonances, 2D 15N-1H HSQC spectra were collected to identify 15N-1H chemical shift correlations. 2D 1H-1H NOESY experiments optimized for imino (NH) proton resonances were acquired at 60 and 360 ms mixing time in 90% H2O to obtain distance restraints involving the exchangeable protons.
Backbone torsion angle constraints were derived from 1H-1H and 31P-1H coupling constants obtained from the following experiments. A 31P-decoupled DQF-COSY experiment and a 2D 31P-1H HetCor experiment were acquired in D2O with unlabeled RNA samples.
Interproton distance constraints
Semi-quantitative distance constraints between non-exchangeable protons were estimated from cross peak intensities in 2D NOESY and 3D 13C-edited NOESY spectra. Using the covalently fixed pyrimidine H5-H6 distance (2.4 ?) and the conformationally restricted sugar H1'-H2' distance (2.8–3.0 ?) as references, peak intensities were classified as strong, medium, weak, or very weak and their corresponding proton pairs given upper bound distance constraints of 3.0, 4.0, 5.0, or 6.0 ?, respectively. Cross peaks observed only at mixing times >180 ms were classified as extremely weak and given 7.0 ? upper bound distance constraints to account for the possibility of spin diffusion. All distance constraints were given lower bounds of 1.8 ?. Distance constraints involving exchangeable protons were estimated from 360 ms mixing time NOESY spectra and were classified as either weak, very weak, or extremely weak, except for the intra-base pair distances A?U H2–NH and G?C NH–NH2, which were classified as strong constraints. Only intra-residue sugar-to-sugar constraints involving H5' and H5'' resonances were included in the calculations.
An initial set of structures was calculated using a shortened version of the simulated annealing protocol (described below). A list of all proton pairs in the calculated structures closer than 4.0 ? (representing expected NOEs) was compared to the list of constraints. The NOESY spectra were then re-examined for predicted NOEs absent from the constraint list. In some cases, this allowed the unambiguous assignment of previously unidentified NOEs, but, in other cases, the predicted NOEs were unobservable due to spectral overlap or the broadening of resonances by exchange with solvent. After the final calculations, virtually all predicted NOEs not in the list could be accounted for by spectral overlap or exchange broadening.
Hydrogen bonding constraints
Watson–Crick base pairs were identified using two criteria: the observation of a significantly downfield shifted NH or NH2 proton resonance and the observation of strong G?C NH–NH2 or A?U H2–NH NOEs. The A37?U33 base pair was identified by observation of a cross peak between A37H2 and U33N3 in the J(N, N)-HNN COSY spectrum. Hydrogen bonds were introduced as distance restraints of 2.9 ± 0.3 ? between donor and acceptor heavy atoms and 2.0 ± 0.2 ? between acceptor and hydrogen atoms. Constraints identified in this way were included in the calculations for base pairs G27?C43, G28?C42, G29?C41, G30?C40, A31?U39, and 32?A38. The U33?A37 base pair constraint was set to 2.9 ± 1.2 ? and 2.0 ± 1.2 ? between donor and acceptor heavy atoms and acceptor and hydrogen atoms, respectively, to permit conformational freedom of loop residues.
Dihedral angle constraints
Constraints on the ribose ring and backbone dihedral angles were derived from semi-quantitative measurements of 3JH-H and 3JH-P couplings (35,36). Sugar pucker conformations were determined from 3JH1'-H2' couplings in 31P-decoupled 2D DQF-COSY spectra. Residues with H1'-H2' couplings >7 Hz were constrained to the C2'-endo conformation through two of the torsion angles in the ribose sugar ring (37). Independent confirmation of sugar pucker conformation was provided by the observation of weak (<5 Hz) 3JH3'-H4' couplings, C3' resonances shifted downfield to 76–80 p.p.m. from the main cluster at 70–72 p.p.m., and C4' resonances shifted downfield to 85–86 p.p.m. from the main cluster at 82–84 p.p.m. Residues with weak (<5 Hz) 3JH1'-H2' couplings were constrained to the C3'-endo conformation. Residues with intermediate 3JH1'-H2' couplings were left unconstrained to reflect the possibility of conformational averaging.
Dihedral angle constraints for the torsion angle were derived from 3JH4'-H5' and 3JH4'-H5'' couplings in the DQF-COSY spectrum and intra-residue H4'-H5' and H4'-H5'' cross peak intensities in the 80 ms mixing time 3D NOESY-ctHSQC spectrum. For residues in which H4'-H5' and H4'-H5'' peaks in the DQF-COSY spectra were clearly absent, representing couplings <5 Hz, was constrained to the gauche+ conformation (60 ± 20) (35,36). For residues with clear 3JH4'-H5' or 3JH4'-H5'' couplings >5 Hz and unequal H4'-H5' and H4'-H5'' NOE intensities, was constrained to include both the trans and gauche– conformations (–120 ± 120), reflecting the lack of stereospecific assignments for the H5' and H5'' resonances. For residues with only weak or unobservable 3JH4'-H5' or 3JH4'-H5'' couplings and unequal H4'-H5' and H4'-H5'' NOE intensities, was left unconstrained to reflect the possibility of conformational averaging.
Dihedral angle restraints for the ? torsion angles were derived from 3JP-H5' and 3JP-H5'' couplings measured in 2D 31P-1H HetCor spectra. ? was constrained to the trans conformation (180 ± 40) for residues in which P-H5' and P-H5'' peaks in the HetCor spectra were clearly absent, representing couplings <5 Hz (35,36). For residues in which P-H5' and P-H5'' peaks could be observed, ? was left unconstrained to reflect the lack of stereospecific assignments and the possibility of conformational averaging. All P-H3' couplings that could be clearly identified were >5 Hz which allows for both trans and gauche- conformations for the torsional angle. A refinement that constrained angles in the stem (G27-C43 to A31-U39) to –125 ± 80 had small improvements (relative to not constraining ) on the quality of the stem regions of the structures and were therefore included for the structures reported here.
Dihedral angle restraints for and were derived from the observation that a trans conformation of either dihedral angle is generally associated with a large downfield shift of the bridging 31P resonance (38). Because no such shift is observed for any of the 31P resonances in the RNA molecules, and were loosely constrained to exclude the trans conformation (0 ± 120) for all residues except those in the loop regions (nucleotides U33 to A37), which were left unconstrained. No dihedral angle constraints were used for the glycosidic angle . A total of 48 restraints (11 , 8 ?, 9 , 9, and 11 ) were used constrain the phosphate backbone dihedral angles in the calculations.
Structure calculations
All calculations were carried out on Silicon Graphics O2 work stations using X-PLOR 3.851 (39). The dihedral angles of a linear starting structure (generated using Insight II, Molecular Simulations, Inc.) were randomized to generate 75 structures with randomized coordinates which were used in a simulated annealing/restrained molecular dynamics (rMD) routine (36,37). The calculation protocol was divided into three stages: global fold, refinement, and final minimization. The global fold step consisted of 1000 cycles of unconstrained energy minimization, 10 ps of rMD at 1000 K using only hydrogen bond and NOE constraints, 9 ps of rMD at 1000 K during which repulsive van der Waals forces were introduced, 14 ps of rMD while cooling to 300 K, and 1000 cycles of constrained minimization. The structures were then refined with 500 cycles of constrained minimization, 5 ps of rMD at 1200 K during which the , ?, , , , and sugar ring dihedral constraints were slowly introduced followed by 5 ps of rMD while cooling to 300 K, and 1000 cycles of constrained minimization. The final minimization step consisted of 1000 cycles of conjugate gradient energy minimization using all constraints and repulsive van der Waals potentials. To determine the consistency of the NMR data with the tri-loop conformation of the unmodified anticodon stem–loop, an additional set of calculations was performed using constraints involving overlapped resonances that were derived from spectra of the unmodified RNA molecule. These calculations were performed beginning with the coordinates of converged structures from the global-fold rMD simulation. Structures were viewed using Insight II (Accelerys, Inc.). The structure coordinates have been deposited in the rcsb with accession number 2AWQ.
Thermal stability
UV melting studies were performed using 2.2 μM RNA hairpin dissolved in NMR buffer (10 mM NaCl, 10 mM potassium phosphate, pH 6.8, and 0.05 mM EDTA). The samples were heated to 90°C for two minutes and snap cooled on ice before each melt experiment. A260 absorbance spectra from 20–95°C and from 95–20°C were acquired (1.0°C per minute) on a Pharmacia Ultrospec 2000 UV-Visible spectrophotometer equipped with a peltier melting heating apparatus. The melting curves were acquired in triplicate but could not be fit to a two-state model.
RESULTS
Effect of 32 modification on RNA stability
The thermal stability of the ACSLPhe and 32 ACSLPhe RNA hairpins (Figure 1) was investigated using UV melting experiments to determine overall molecular stability (Tm). The UV thermal denaturation curves indicate that the hairpins melt in two stages (Figure 2). The lower temperature (<50°C) transitions presumably correspond to the destacking of the loop nucleotides. The pseudouridine modification shifts the transition midpoint for the hairpin 3.5°C ± 0.5°C higher. This indicates that 32 increases the stability of the stem. The anneal spectra from 95–20°C also were acquired and showed 1.2°C ± 0.5°C of hysteresis at the Tm. Slow cooling of the RNA can lead to duplex formation and to hysteresis. The increased thermal stability of the modified RNA hairpin is consistent with other studies that predict pseudouridine enhances stability when located in a stem or at a loop–stem junction (8,9). Increased stability of the upper stem and slightly lower stability of the loop is supported by the NH spectrum (Figure 3). The guanine NH protons sharpen with the introduction of 32 and the U33 resonance is broadened.
Figure 2 Overlay of the UV melting curves of the unmodified (ACSL) and 32–modified RNA (32-ACSL) hairpins. Each of the hairpins appears to exhibit two melting transitions—the broad lower temperature transitions presumably corresponding to destacking of the loop nucleotide bases. The apparent melting temperatures of the unmodified and 32–modified RNA hairpins are estimated to be 77.0°C and 73.5°C, respectively.
Figure 3 Imino proton spectra of the (A) unmodified and (B) 32-modified ACSLPhe. The imino resonances of 32 and U33 have chemical shifts consistent with Watson–Crick base pairs but are exchange broadened in 32-ACSLPhe. This suggests that the pseudouridine has a small destabilizing effect on the loop leading to increased solvent exposure of the imino groups.
Resonance assignments of 32-ACSLPhe
The non-exchangeable 1H and 13C resonances of 32-ACSLPhe were assigned using standard heteronuclear methods (31,40). Most of the base and ribose 1H-13C correlations are resolved, and none of the resonances have spectral characteristics indicative of intermediate exchange. All 17 ribose spin systems, except for the incompletely labeled 5'-terminal nucleotide, were identified using 2D HCCH-COSY and 3D HCCH-TOCSY experiments. The five adenine intra-base H8-H2 correlations also were identified using the HCCH-TOCSY experiment. Intra-residue base-to-sugar correlations were identified using 2D 15N-1H HSQC experiments optimized to yield the multiple bond correlations H5/6-N1, H8-N9, and H1'-N1/N9 (41). All purine correlations and four of six pyrimidine correlations were identified in these spectra.
Sequence specific assignments were determined using two-dimensional (2D) NOESY (Figure 4) and 3D 13C-edited NOESY experiments (31). The sequential base-1' NOE connectivities are continuous through all 17 nt in the 180 ms NOESY spectrum, but the connectivity between nucleotides U33 and G34 is weak. With the exception of the A35 H8 to A36 H8 step, the base-2' and base–base inter-residue connectivities also are continuous in the 180 ms spectrum. Additionally, the base-2' inter-residue connectivities from U32 to A37 are weak, consistent with the non-C3'-endo ribose ring conformations of these residues (described below).
Figure 4 H6/8-to-H1' region of the 2D 400 ms NOESY spectrum. The base-1' proton sequential walk is traced with intra-residue peaks labeled. The 32 H1' resonance has a chemical shift of 4.56 p.p.m. and is data not shown. The arrows (a) points to the inter-residue sequential NOE between U33 H1' and G34 H8 and (b) non-sequential NOE between U33 H1' and A35 H8. The presence of the sequential NOE is not compatible with a U-turn motif for the loop whereas the non-sequential NOE can be produced by non-U-turn loop conformations.
The exchangeable NH and NH2 resonances were assigned using 2D NOESY experiments. The NH proton spectrum is shown in Figure 3. The three strong G NH resonances corresponding to G?C base pairs and one U NH resonance corresponding to an A?U base pair are connected by NOE cross peaks between NH proton resonances of adjacent base pairs. These connectivities are continuous in the helix from G28 to U39. The weak NH resonance of the terminal G?C base pair does not yield cross peaks in the NOESY spectrum. The U33 NH resonance is weak (Figure 3) and could only be assigned through a U33 N3–A37 H2 cross peak in the J(N, N)-HNN COSY spectrum (Figure 5) (33). A weak G NH resonance, which has a chemical shift that corresponds to a non-base paired guanine, was assigned to G34. The cytidine NH2 resonances were assigned using the strong intra-base pair C NH2 to G NH NOE cross peaks. The NH2 resonances of A37 and of all guanine nucleotides except G34 were not observed. The A31 NH2 resonances were assigned based on their NOE cross peaks with U39 H3. The NH2 15N resonances of A35, A36 and A38 were assigned based on intrabase H2 to N6 correlations in the J(N, N)-HNN COSY spectrum. The NH2 proton resonances of A35 and A36 are broad and have chemical shifts indicative of solvent-exposed protons.
Figure 5 (A) HNN-COSY and (B) multiple-bond 15N-1H HSQC spectra showing intra-residue U H5 to N3 and H6 to N1 correlations. The HNN-COSY shows cross-strand H2-N3 crosspeaks for residues 31–39, 38–32, and 37–33 produced by the Watson–Crick base pair configurations. The syn configuration about the 32 glycosidic bond and participation of 32 N1H in the 32-A38 base pair would produce a 32 N1-A38 H2 crosspeak (dashed circle). The U33–A37 crosspeak is weak and is indicative of a weak hydrogen bond.
Resonance assignments for 32 were accomplished using conventional heteronuclear methods. The 32 C1' resonance is shifted upfield into the region of the C4' resonances at 82.78 p.p.m. This position is consistent with the carbon–carbon glycosidic bond that replaces the more electronegative nitrogen–carbon bond of uridine. The 32 H6 has a sequential NOE cross peak with A31 H1' and the 32 H1' has a cross peak with the U33 H6. The 32 H1' assignment was confirmed through direct correlation with the 32 H6 using an HCCH-RELAY experiment. The 32 H1 and N1 imino resonances give rise to a cross peak in the 2D 15N-1H HMQC spectrum at 10.31 (1H) and 128.59 (15N) p.p.m. The identity of the N1 imino resonance is confirmed by the two-bond correlation with H6 in the multiple-bond HSQC spectrum (Figure 5). An intense intrabase NOE crosspeak between the N1H and H6 of 32 also supports assignment of the 32 H1. The 32 N3H resonance was assigned based on its NOE cross peak with the A38 H2 since the 32 H3 resonance is weak and does not give rise to a cross peak with the U39 NH resonance.
All of the internucleotide phosphate 31P resonances are clustered between –3.35 and –4.63 p.p.m. and were assigned using a 31P-1H hetero-TOCSY-NOESY experiment. The sequential P–H6/8 and P–H1' correlations are continuous throughout the molecule. The P–H3' correlations and several P–H4' and P–H5'/H5'' correlations are present in 31P–1H HetCor spectra and provide independent confirmation of the 31P assignments. The 5' terminal phosphate resonates at –1.30 p.p.m. and was assigned through cross peaks to the ribose protons of G27.
Structure calculations
The structure of 32-ACSLPhe was calculated using a rMD routine starting from 75 sets of coordinates with randomized backbone dihedral angles. The calculations used a total of 238 NOE derived distance constraints, 36 bp constraints, and 74 dihedral angle constraints (Table 1) resulting in 10 converged structures (Figure 6). Structures were classified as converged if they had low energy, few constraint violations, and predicted only NOEs that could be experimentally verified or explained. The converged structures have an average of 5 distance constraint violations between 0.1 and 0.3 ?, most of these involving the loop region. All converged structures have no constraints violated by more than 0.3 ?. The average root mean square deviations (RMSDs) of the heavy atoms between the individual structures and the minimized mean structure is 0.67 ? for the loop region (residues 33–37) and 0.92 ? for the stem region (residues 27–32 and 38–43) (Figure 6).
Table 1 Summary of experimental constraints and structure calculation statistics for 32-ACSLPhe
Figure 6 Stereoview of the superposition of (A) the stems and (B) the loops of the 10 converged 32-ACSLPhe structures. Convergence criteria are given in the text. The views are into the major groove. Only sugar and base heavy atoms are shown and the average r.m.s. deviation for the heavy atoms between the ten structures and the average structure is 1.14 ?. The loop and stem regions are locally well defined, but the propeller twist of the A31–U39 base pair is variable among the structures and slightly increases the r.m.s.d. of the full hairpin.
Structure of the loop region of 32-ACSLPhe
The 32-ACSLPhe loop is made up of the anticodon nucleotides G34–A36 and is closed by the U33?A37 base pair (Figure 7). The base of G34 stacks against the base of U33, consistent with the observed H6-H8 and H5-H8 NOEs. The bases of G34 and A35 are approximately coplanar and are parallel. This configuration satisfies the observed G34 H8–A35 H8 NOE cross-peak. The A36 base stacks beneath A37 base but is displaced toward the minor groove edge of A37 (Figure 7). Several NOEs involving the A36 H2 resonance help position this base and include interactions with the H2' of U33, the H8 of A35, and the G34 and/or A35 H1' (these 1' resonances are degenerate). These cross-peaks indicate that the Watson–Crick base pair functional groups are oriented to the interior of the molecule and point toward the residues flanking the 5' side of the nucleotide. The conserved NOE cross-peaks A36 H2 to A37 H2 and H1' suggest that A36 is partially stacked with A37. Thus, the observed NOEs support the compact tri-loop conformation.
Figure 7 Stereoview of minimized average structure of 32-ACSLPhe (residues G30 to C40). Hydrogen bonding NH and NH2 protons of base pairs 32-A38 and A31–U39 are colored purple and the exocyclic amino nitrogens (A31 and A38) are green. The pro-R(p) phosphoryl oxygens of A31 and 32 (yellow) are predicted to form water-mediated hydrogen bonds with 32 N1H and slow exchange of the N1 proton. The bridging water molecule is shown in pink. Explicit waters were not used for the structure calculations, but upper distance bounds derived from crystal structures were applied between 32 N1 and the phosphoryl oxygens.
The loop nucleotides also have unusual sugar–phosphate backbone conformations The large H1'–H2' couplings of residues G34–A36 indicate that the ribose sugar rings have the C2'-endo conformation. The strong intra-residue and weak sequential H2' to H8 NOE cross-peaks as well as the substantial downfield shift of the C3' and C4' resonances involving residues G34–A36 also support the C2'-endo conformation. Residues U33 and A37 at the junction of the loop and stem have a mixture of C2'- and C3'-endo conformations as evidenced by their intermediate H1'–H2' couplings and the modest downfield chemical shifts of their C3' and C4' resonances. The torsion angles of A35 and A36 have the -gauche conformation not typical of A-form geometry.
A superposition of the loop regions from the ten converged structures is shown in Figure 6B and the minimized average structure is shown in Figure 7. The helical base stack is continuous along the 3' side of the loop, with G34 stacking against the U33 base. On the 3' side of the loop, the A37 base straddles the bases of A35 and A36. The A36 N6 is 3 ? from the U33 O2' in half of the converged structures, suggesting the possibility of a cross strand base–sugar hydrogen bond. However, this interaction could not be confirmed since the A36 NH2 proton resonances could not be assigned and no 2'-OH proton resonances were identified.
The loop of the 32-ACSLPhe does not contain the classical ‘U-turn’ motif. Although weak, the inter-residue U33 H1'–G34 H8 and U33 H6–G34 H8 NOE cross-peaks (Figure 4) are consistent with the U33 and G34 positions shown in Figure 7 but not with their positions in the U-turn (Figures 8B and C). A U33 H1' to A35 H8 NOE is observed, but only at long mixing time and is even less intense than the sequential U33 H1'–G34 NOE (Figure 4). The ribose puckers of the anticodon nucleotides of the U-turn also tend toward the C3'-endo conformation rather than the C2'-endo conformation observed in this study. Another feature characteristic of the U-turn motif is the trans conformation of backbone torsion angle between U33 and G34. The 31P resonance corresponding to this phosphate is located in the main cluster of 31P peaks (Supplementary Table S1). The chemical shift of this resonance is not consistent with the 31P chemical shift predicted for a phosphate having the trans conformation of the torsional angle (38).
Figure 8 Comparison of residues 31–39 of (A) the solution structure of E.coli tRNAPhe (42) and the crystal structures of (B) fully modified yeast tRNAPhe (15,16) and (C) fully modified tRNAAsp (18,21). 32 and A38 in (A) form a Watson–Crick base pair. In (B) and (C), nucleotides C32-A38 and 32-C38, respectively, form the bifurcated hydrogen bond. The anticodon loop in (A) adopts the tri-loop conformation whereas the anticodon loops in (B) and (C) adopt the U-turn motif. In E.coli tRNACys, 32 is in the syn configuration about the glycosidic bond and the loop forms a U-turn motif (17).
Structure of the stem of 32-ACSLPhe
The geometry of the hairpin stem, base pairs G27?C43 to U33?A37, is primarily A-form (Figure 6A). The sequential base-1', 2' and several base–base NOEs are continuous at 180 ms mixing time. The A31?U39 and 32?A38 base pairs are sufficiently stable to give rise to a cross-strand NH–H2 NOEs, but the NH proton of the U33?A37 base pair is not. The 32 and U33 NH proton resonances are weak, but their participation in Watson–Crick pairing schemes was confirmed using a J(N, N)-HNN-COSY spectrum (Figure 5). Nucleotides A31, A37, and A38 also produce cross-strand H2-H1' NOEs commonly observed for A?U base pairs within helices. The H2 to N3 through bond correlations clearly confirm that U39 is base paired with A31 and 32 is base paired with A38 through its N3 imino group (Figure 7).
The torsion angles of the sugar-phosphate backbone are within the limits of A-form geometry. The small (<5 Hz) H1'–H2' couplings and the 13C chemical shifts of the 3' and 4' resonances of nucleotides G28–32 and A38–C42 are indicative of the typical C3'–endo sugar pucker. None of the or torsional angles within the converged structures deviated from gauche–.
The most remarkable feature of the stem is the minimal structural perturbation caused by 32. Although the combination of A31 H8-32 H6 and A31 H2-32 H6 NOEs is unusual, the relative positions of the adjacent A31 and 32 bases that give rise to the interactions are accommodated within the A-form helix.
DISCUSSION
We are using the anticodon stem–loop of E.coli tRNAPhe as a model to probe the cumulative physical effects of sequential modifications in RNA molecules. Pseudouridine is the most common base modification found in RNA and the anticodon stem-loop of E.coli tRNAPhe has two of these nucleotides, 32 and 39. In this study, we have examined the thermodynamic and structural effects of 32 on the otherwise unmodified tRNAPhe anticodon stem–loop.
The structures of ACSLPhe and 32-ACSLPhe are similar
The unmodified anticodon stem–loop of ACSLPhe forms a highly ordered tri-loop conformation (42) and the 32-ACSLPhe adopts a similar structure. The stems of the two hairpins are continuous through base pair U33–A37, but this loop-closing base pair is weaker in the 32-modified molecule indicating a slightly less compact tri-loop. Also, fewer constraints involving G34 and A35 in 32-ACSLPhe lead to greater variability in the positions of these residues compared to the unmodified molecule, including excursions of A35 to the minor groove side of the loop. However, the lower number of experimental constraints in the loop is due to spectral overlap of G34 and A35 H1' resonances and probably does not reflect conformational heterogeneity inherent in the loop. Nonetheless, the small chemical shift differences between unmodified and 32-modified ACSLPhe molecules indicate that the loop structures are not identical. The energy difference between the converged conformations of 32-ACSLPhe is small (<15 kcal/mole), and the RMSD between the 32-ACSLPhe and ACSLPhe minimized average structures is 1.13 ?.
The pseudouridylation of U32 increases the overall stability of the RNA hairpin. This is consistent with studies that have examined the thermodynamic effects of pseudouridine incorporation (9,13,43,44). Pseudouridine within a helix can increase the melting temperature of the helix 3–5°C depending upon the neighboring base pairs and improves local base stacking (9). A contributing factor to the stabilization effect of is its ability to form a water-mediated hydrogen bond with the phosphate backbone (9). This hydrogen bond interaction has been inferred from difference maps of X-ray crystallographic studies of unmodified and fully modified tRNAGln molecules (19) and is present in molecular dynamics simulations of yeast tRNAAsp (45). The strong exchange-protected 32 N1H resonance (Figure 3) supports this interaction within the ACSLPhe and is modeled in the average structure (Figure 7). Pseudouridine also confers a smaller degree of helix stabilization when present in single strand regions adjacent to a helix (46) or is on the 3' end of the loop at loop-helix junctions such as 39 (46). 32 is adjacent to the helix-loop junction (U33–A37), but is at the 5' end of the loop. Thus, pseudouridine has a comparable stabilizing effect when positioned on either the 3' or 5' edge of a helix.
Pseudouridine tends to form a Watson–Crick base pair with adenine in helical contexts and the 32 is no exception (43). The HNN-COSY spectrum confirms the hydrogen bond configurations of the 32-A38 and U33–A37 base pairs (Figure 5). However, the H3 resonances of U33 and 32 are weaker than the corresponding H3 resonances of unmodified ACSLPhe (Figure 3). This weakening may reflect altered alignment of the hydrogen bonds of these base pairs as the base of 32 is stabilized toward the major groove by the water-mediated N1H-phosphate hydrogen bond (Figure 7). A 32 H6-A31 H2 NOE is consistent with the major groove displacement of the 32 base. This NOE is not observed in the unmodified molecule (42) and is not typical of 5'-AU-3' base stacking in helices with regular A-form geometry.
The 32–38 bp in tRNA
The nucleotide composition of residues 32 and 38 is not an equal distribution of the pyrimidine and purine bases. tRNA sequence data show that position 38 is often adenine (67%), but a significant fraction (30%) of position 38 nt are pyrimidines (14). Residue 32 is almost exclusively occupied by a pyrimidine base (>97%) and is modified in approximately a third of sequenced tRNA molecules on either the base or ribose group. The most common base modification at position 32 is pseudouridine and although this base does not alter the overall structure of the ACSLPhe loop compared to uridine, it partially destabilizes the tri-loop conformation and increases the stem melting temperature by about 3.5°C. Other modifications at position 32 are frequently those that favor the C3' endo conformation of the ribose, such as pyrimidine 2'-O-methyl and the pyrimidine C2-thiol (47), which may help to maintain the helical integrity of the stem of the anticodon arm (48). Whether these modifications can alter or stabilize anticodon loop conformations or function as pseudouridine primarily to stabilize the stem has not been investigated.
The geometry of the 32 base in solution is different from that observed for position 32 residues in crystal structures of tRNAs. The crystal structures of several tRNAs reveal a highly conserved structural motif, the bifurcated hydrogen bond, between residues 32 and 38 (49) (Figure 8). The bifurcated hydrogen bond frequently involves proton donation by an exocyclic NH2 group and acceptance by the O2 of C32 or U32. This hydrogen bond configuration imposes an underwinding of the helix geometry at the 32–38 base step. For E.coli tRNAPhe, the configuration would correspond to hydrogen bonding between A38 N6H2 and 32 O4 (adjacent to the C5-C1' glycosidic bond). This base arrangement of 32 and A38 is isosteric with other 32–38 pairs, such as C32–A38 and U32–C38, that adopt the bifurcated hydrogen bond (49) (Figure 8). Auffinger and Westhof (49) have proposed the geometry of nt 32 and 38 resulting from the bifurcated hydrogen bond creates a transition between the stem and loop in the anticodon arm and facilitates formation of the U-turn. Thus, the 32–38 interaction may allow an open conformation of the loop without distortion of the canonical A-form stem. The results of this study indicate that if the bifurcated hydrogen exists in solution and functions in U-turn stabilization, the 32 modification alone is not sufficient to organize this motif.
The crystal structure of fully modified Cys-tRNACys in complex with EF-Tu and GDPNP presents an alternative conformation for the 32-A38 base pair. In this structure, the 32 base adopts the syn configuration about the glycosidic bond and does not form any interactions with A38 (17). This configuration of the base can accommodate the 32 water-mediated hydrogen bond to the phosphate backbone and thus should retain the stabilizing effect conferred by pseudouridine in the anti configuration. Notably though, the anticodon loop in this complex adopts the characteristic U-turn motif with stacking of the anticodon bases G, C, and A. The EF-Tu protein-binds at the acceptor stem and has no contacts in the region of the anticodon arm, but the G and C nucleotides of the anticodon form intermolecular base pairs within the unit cell of the crystal. The structure of this molecule is of particular interest since the anticodon arm of tRNACys contains the same nucleotide base modifications found in fully modified E.coli tRNAPhe, specifically 32, 39, and ms2i6A37. These additional modifications could affect the conformation of the 32 base, however, neither 39 nor i6A37 alone are sufficient to induce the syn conformation of 32 in solution (J.Cabello and E.P.Nikonowicz, unpublished data).
Pseudouridylation within the anticodon arm of tRNA is important to the overall fitness of an organism and loss of this modification can lead to a spectrum of growth defects in bacteria and yeast (50–54). Disruption of the truA gene which catalyzes formation of 38, 39, and 40 decreases polypeptide chain elongation rates and reduces cell growth rate 30% (1,55). A similar effect is observed for the deg1 gene of yeast that catalyzes 38 and 39 formation (54,56). E.coli that contain a catalytically inactive truB gene which catalyzes formation of 55 exhibit no growth defects (52), even though 55 is universally conserved (14). Pseudouridylation at position 32 is intermediate in its physiological effects. Cells that lack RluA activity have near normal growth rates, but are strongly selected against when grown in competition with wild type cells (51). However, given that RluA modifies only six tRNA isoaccpetors, unlike TruA and TruB that modify their target sites in all tRNA molecules in E.coli, the physiological effects of its loss are remarkable and suggest a critical role of 32 in proper tRNA function. The specific translational defects associated with loss of 32 have not been determined, but studies of position 32 mutations in other tRNAs suggest translational fidelity and wobble base recognition are likely to be impaired (23–25,57). The studies presented here indicate that a primary role of 32 is the stabilization of the stem of the anticodon arm. Thus, 32 may serve as a buffer to preserve the stability and the architecture of the stem within the context of other loop-destabilizing nucleotide modifications (42) or after formation of the U-turn.
SUPPLEMENTARY DATA
Supplementary Data are available at NAR Online.
ACKNOWLEDGEMENTS
We thank M. Michnicka for preparation of the T7 RNA polymerase and isotopically labeled 5'-rNTPs. This work was supported by NSF grant MCB-0078501 and by a grant from the Robert A. Welch Foundation (C1277) to E.P.N. Funding to pay the Open Access publication charges for this article was provided by the Robert A. Welch Foundation.
REFERENCES
Bj?rk, G.R. (1995) Biosynthesis and function of modified nucleosides In S?ll, D. and RajBhandary, U.L. (Eds.). tRNA Structure, Biosynthesis, and Function, Washington, D.C. American Society for Microbiology Press pp. 165–206 .
Garcia, G.A. and Goodenough-Lashua, D.M. (1998) Mechanisms of modifying enzymes In Grosjean, H. and Benne, R. (Eds.). Modification and Editing of RNA: The Alteration of RNA Structure and Function, Washington, D.C. ASM Press pp. 135–168 .
Winkler, M.E. (1998) Genetics and regulation of tRNA and rRNA modification In Grosjean, H. and Benne, R. (Eds.). Modification and Editing of RNA: The Alteration of RNA Structure and Function, Washington, D.C. ASM Press pp. 441–469 .
Agris, P.F. (1996) The importance of being modified: roles of modified nucleosides and Mg2+ in RNA structure and function Prog. Nucleic Acid Res. Mol. Biol, . 53, 79–129 .
Charette, M. and Gray, M.W. (2000) Pseudouridine in RNA: what, where, how, and why IUBMB Life, 49, 341–351 .
Ofengand, J. (2002) Ribosomal RNA pseudouridines and pseudouridine synthases FEBS Lett, . 514, 17–25 .
Neumann, J.M., Bernassau, J.M., Gueron, M., Tran-Dinh, S. (1980) Comparative conformations of uridine and pseudouridine and their derivatives Eur. J. Biochem, . 108, 457–463 .
Davis, D.R. (1998) Biophysical and conformational properties of modified nucleosides in RNA (NMR Studies) In Grosjean, H. and Benne, R. (Eds.). Modification and Editing of RNA, Washington, D.C. ASM Press pp. 85–101 .
Davis, D.R. (1995) Stabilization of RNA stacking by pseudouridine Nucleic Acids Res, . 23, 5020–5026 .
Roy, S., Papastavros, M.Z., Sanchez, V., Redfield, A.G. (1984) Nitrogen-15-labeled yeast tRNAPhe: double and two-dimensional heteronuclear NMR of guanosine and uracil ring NH groups Biochemistry, 23, 4395–4400 .
Griffey, R.H., Davis, D., Yamaizumi, Z., Nishimura, S., Bax, A., Hawkins, B., Poulter, C.D. (1985) 15N-labeled Escherichia coli tRNAfMet, tRNAGlu, tRNATyr, and tRNAPhe. Double resonance and two-dimensional NMR of N1-labeled pseudouridine J. Biol. Chem, . 260, 9734–9741 .
Hruska, F.E., Grey, A.A., Smith, I.C. (1970) A nuclear magnetic resonance study of the molecular conformation of ?-pseudouridine in aqueous solution J. Am. Chem. Soc, . 92, 4088–4094 .
Yarian, C.S., Basti, M.M., Cain, R.J., Ansari, G., Guenther, R.H., Sochacka, E., Czerwinska, G., Malkiewicz, A., Agris, P.F. (1999) Structural and functional roles of the N1- and N3-protons of at tRNA's position 39 Nucleic Acids Res, . 27, 3543–3549 .
Sprinzl, M., Horn, C., Brown, M., Loudovitch, A., Steinberg, S. (1998) Compilation of tRNA sequences and sequences of tRNA genes Nucleic Acids Res, . 26, 148–153 .
Robertus, J.D., Ladner, J.E., Finch, J.R., Rhodes, D., Brown, R.S., Clark, B.F., Klug, A. (1974) Structure of yeast phenylalanine tRNA at 3 ? resolution Nature, 250, 546–551 .
Kim, S.-H., Suddath, F.L., Quigley, G.J., McPherson, A., Sussman, J.L., Wang, A.H.J., Seeman, N.C., Rich, A. (1974) Three-dimensional tertiary structure of yeast phenylalanine transfer RNA Science, 185, 435–440 .
Nissoen, P., Thirup, S., Kjeldgaard, M., Nyborg, J. (1999) The crystal structure of Cys-tRNA-EF-Tu-GDPNP reveals general and specific features in the ternary complex and in tRNA Structure, 7, 143–156 .
Moras, D., Comarmond, M.B., Fischer, J., Weiss, R., Thierry, J.C., Ebel, J.P., Giege, R. (1980) Crystal structure of yeast tRNAAsp Nature, 288, 669–674 .
Arnez, J.G. and Steitz, T.A. (1994) Crystal structure of unmodified tRNAGln complexed with glutaminyl-tRNA synthetase and ATP suggests a possible role for pseudo-uridines in stabilization of RNA structure Biochemistry, 33, 7560–7567 .
Shi, H. and Moore, P.B. (2000) The crystal structure of yeast phenylalanine tRNA at 1.93 ? resolution: a classic structure revisited RNA, 6, 1091–1105 .
Westhof, E., Dumas, P., Moras, D. (1985) Crystallographic refinement of yeast aspartic acid transfer RNA J. Mol. Biol, . 184, 119–145 .
Yarus, M. (1982) Translational efficiency of transfer RNAs: uses of an extended anticodon Science, 218, 646–652 .
Yarus, M., Cline, S., Raftery, L., Wier, P., Bradley, D. (1986) The translational efficiency of tRNA is a property of the anticodon arm J. Biol. Chem, . 261, 10496–10505 .
Lustig, F., Boren, T., Claesson, C., Simonsson, C., Barciszewska, M., Lagerkvist, U. (1993) The nucleotide in position 32 of the tRNA anticodon loop determines ability of anticodon UCC to discriminate among glycine codons Proc. Natl Acad. Sci. USA, 90, 3343–3347 .
Claesson, C., Lustig, F., Boren, T., Simonsson, C., Barciszewska, M., Lagerkvist, U. (1995) Glycine codon discrimination and the nucleotide in position 32 of the anticodon loop J. Mol. Biol, . 247, 191–196 .
O'Connor, M. (1998) tRNA imbalance promotes -1 frameshifting via near-cognate decoding J. Mol. Biol, . 279, 727–736 .
Davanloo, P., Rosenberg, A.H., Dunn, J.J., Sturdier, F.W. (1984) Cloning and expression of the gene for bacteriophage T7 RNA polymerase Proc. Natl Acad. Sci. USA, 81, 2035–2039 .
Wrzesinski, J., Nurse, K., Bakin, A., Lane, B.G., Ofengand, J. (1995) A dual-specificity pseudouridine synthase: an Escherichia coli synthase purified and cloned on the basis of its specificity for 746 in 23S RNA is also specific for 32 in tRNAPhe RNA, 4, 437–448 .
Nikonowicz, E.P., Sirr, A., Legault, P., Jucker, F.M., Baer, L.M., Pardi, A. (1992) Preparation of 13C and 15N Labelled RNAs for heteronuclear multi-dimensional NMR studies Nucleic Acids Res, . 20, 4507–4513 .
Markley, J.L., Bax, A., Arata, Y., Hilbers, C.W., Kaptein, R., Sykes, B.D., Wright, P.E., Wuthrich, K. (1998) Recommendations for the presentation of NMR structures of proteins and nucleic acids. IUPAC-IUBMB-IUPAB Inter-Union Task Group on the Standardization of Data Bases of Protein and Nucleic Acid Structures Determined by NMR Spectroscopy J. Biomol. NMR, 12, 1–23 .
Nikonowicz, E.P. and Pardi, A. (1993) An efficient procedure for assignment of the proton, carbon, and nitrogen resonances in 13C/15N labeled nucleic acids J. Mol. Biol, . 232, 1141–1156 .
Legault, P., Farmer, B.T., II, Mueller, L., Pardi, A. (1994) Through-bond correlation of adenine protons in a 13C-labeled ribozyme J. Am. Chem. Soc, . 116, 2203–2204 .
Hennig, M. and Williamson, J.R. (2000) Detection of N-H???N hydrogen bonding in RNA via scalar couplings in the absence of observable imino proton resonances Nucleic Acids Res, . 28, 1585–1593 .
Kellogg, G.W. and Schweitzer, B.I. (1993) Two- and three-dimensional 31P-driven NMR procedures for complete assignment of backbone resonances in oligodeoxyribonucleotides J. Biomol. NMR, 3, 577–595 .
Altona, C. (1982) Conformational analysis of nucleic acids. Determination of backbone geometry of single-helical RNA and DNA in aqueous solution Recl. Trav. Chim. Pays-Bas, 101, 413–433 .
Smith, J.S. and Nikonowicz, E.P. (1998) NMR structure and dynamics of an RNA motif common to the spliceosome branch-point helix and the RNA-binding site for phage GA coat protein Biochemistry, 37, 13486–13498 .
Varani, G., Aboul-ela, F., Allain, F.H.-T. (1996) NMR investigation of RNA structure Prog. NMR Spectroscopy, 29, 51–127 .
Gorenstein, D.G. (1981) Nucleotide conformational analysis by 31P nuclear magnetic resonance spectroscopy Annu Rev Biophys Bioeng, . 10, 355–386 .
Brünger, A.T. X-PLOR Version 3.1 Manual, (1992) NH Yale University .
Pardi, A. (1995) Multidimensional heteronuclear NMR experiments for structure determination of isotopically labeled RNA Methods Enzymol, . 261, 350–380 .
Sklenar, V., Peterson, R.D., Rejante, M.R., Feigon, J. (1994) Correlation of nucleotide base and sugar protons in a 15N-labeled HIV-1 RNA oligonucleotide by 1H-15N HSQC experiments J. Biomol. NMR, 4, 117–122 .
Cabello-Villegas, J., Winkler, M.E., Nikonowicz, E.P. (2002) Solution conformations of unmodified and A37N6-dimethylallyl modified anticodon stem–loops of Escherichia coli tRNAPhe J. Mol. Biol, . 319, 1015–1034 .
Hall, K.B. and McLaughlin, L.W. (1992) Properties of pseudouridine N1 imino protons located in the major groove of an A-form RNA duplex Nucleic Acids Res, . 20, 1883–1889 .
Meroueh, M., Grohar, P.J., Qiu, J., SantaLucia, J., Jr, Scaringe, S.A., Chow, C.S. (2000) Unique structural and stabilizing roles for the individual pseudouridine residues in the 1920 region of Escherichia coli 23S rRNA Nucleic Acids Res, . 28, 2075–2083 .
Auffinger, P. and Westhof, E. (1997) RNA hydration: three nanoseconds of multiple molecular dynamics simulations of the solvated tRNAAsp anticodon hairpin J. Mol. Biol, . 269, 326–341 .
Davis, D.R., Veltri, C.A., Nielsen, L. (1998) An RNA model system for investigation of pseudouridine stabilization of the codon-anticodon interaction in tRNALys, tRNAHis and tRNATyr J. Biomol. Struct. Dyn, . 15, 1121–1132 .
Yokoyama, S. and Nishimura, S. (1995) Modified nucleosides and codon recognition In S?ll, D. and RajBhandary, U.L. (Eds.). tRNA Structure, Biosynthesis, and Function, Washington, D.C. American Society for Microbiology Press pp. 207–224 .
Kawai, G., Yamamoto, Y., Kamimura, T., Masegi, T., Sekine, M., Hata, T., Iimori, T., Watanabe, T., Miyazawa, T., Yokoyama, S. (1992) Conformational rigidity of specific pyrimidine residues in tRNA arises from posttranscriptional modifications that enhance steric interaction between the base and the 2'-hydroxyl group Biochemistry, 31, 1040–1046 .
Auffinger, P. and Westhof, E. (1999) Singly and bifurcated hydrogen-bonded base-pairs in tRNA anticodon hairpins and ribozymes J. Mol. Biol, . 292, 467–483 .
Tsui, H.C., Arps, P.J., Connolly, D.M., Winkler, M.E. (1991) Absence of hisT-mediated tRNA pseudouridylation results in a uracil requirement that interferes with Escherichia coli K-12 cell division J Bacteriol, 173, 7395–7400 .
Raychaudhuri, S., Niu, L., Conrad, J., Lane, B.G., Ofengand, J. (1999) Functional effect of deletion and mutation of the Escherichia coli ribosomal RNA and tRNA pseudouridine synthase RluA J. Biol. Chem, . 274, 18880–18886 .
Gutgsell, N., Englund, N., Niu, L., Kaya, Y., Lane, B.G., Ofengand, J. (2000) Deletion of the Escherichia coli pseudouridine synthase gene truB blocks formation of pseudouridine 55 in tRNA in vivo, does not affect exponential growth, but confers a strong selective disadvantage in competition with wild-type cells RNA, 6, 1870–1881 .
Del Campo, M., Kaya, Y., Ofengand, J. (2001) Identification and site of action of the remaining four putative pseudouridine synthases in Escherichia coli RNA, 7, 1603–1615 .
Lecointe, F., Simos, G., Sauer, A., Hurt, E.C., Motorin, Y., Grosjean, H. (1998) Characterization of yeast protein Deg1 as pseudouridine synthase (Pus3) catalyzing the formation of 38 and 39 in tRNA anticodon loop J. Biol. Chem, . 273, 1316–1323 .
Singer, C.E., Smith, G.R., Cortese, R., Ames, B.N. (1972) Mutant tRNAHis ineffective in repression and lacking two pseudouridine modifications Nature New Biol, . 238, 72–74 .
Carbone, M.L., Solinas, M., Sora, S., Panzeri, L. (1991) A gene tightly linked to CEN6 is important for growth of Saccharomyces cerevisiae Curr. Genet, . 19, 1–8 .
Yarus, M., Cline, S.W., Wier, P., Breeden, L., Thompson, R.C. (1986) Actions of the anticodon arm in translation on the phenotypes of RNA mutants J. Mol. Biol, . 192, 235–255 .(Javier Cabello-Villegas and Edward P. Ni)