Determination of Secondary and Tertiary Structural Features of Transfer RNA Molecules in Solution by Nuclear Magnetic Resonance

Abstract

High-resolution 300-MHz proton nuclear magnetic resonance spectra of the hydrogen-bonded protons in three different purified tRNA molecules are presented. The resonances in the region between -11 and -15 ppm from 2,2-dimethyl-2-silapentane-5-sulfonate (DSS) are assigned to the ring NH protons of specific base pairs by two approaches. First, intrinsic positions of -14.8 ppm and -13.7 ppm are taken for the AU and GC ring NH protons, respectively, and the spectra are calculated by including ring current shifts from the nearest neighbors.The spectra calculated in this way on the basis of the cloverleaf are in good agreement with the observed. Second, fragments of yeast tRNAPhe were obtained, which helped in assignments of the spectrum of intact molecules. The close agreement strongly supports the cloverleaf model.Tertiary structural features were determined in a few cases where the ring currents at the terminal base pairs of helical regions depended upon stacking of the helices.In this way, we were able to show that in Escherichia colt tRNAGlu the CCA stem forms a continuous helix with the TVt-C stem, which is in accord with the preliminary x-ray structure of yeast tRNAPhe, suggesting that this stacking is observed in solution and may be a general property of different tRNA molecules. Similar reasoning suggests that in E. coli tRNAfMet G-27 is stacked upon the dihydrouridine helix.In order to understand in detail how tRNA functions in protein synthesis, the structure of tRNA in solution must be determined. The present note summarizes results of our recent high-resolution proton nuclear magnetic resonance (NMR) investigations (1-4) of tRNA in which we determined that the cloverleaf model is an accurate description of the hydrogen-bonded secondary structure of several purified tRNAs, including yeast tRNAPhe, Escherichia coli tRNAGlU, and E. coli tRNAfMet. In addition, we obtained information about certain important tertiary structural features of each molecule. In this paper we emphasize these structural aspects to show how the NMR measurements in solution complement and supplement the crystal structure presently emerging from the x-ray studies (5).The groundwork for the present study was developed in previous studies of the NMR spectra of tRNA molecules in solution where we showed that resonances observed in the region between -11 and -15 ppm from 2,2-dimethyl-2-Abbreviations: NMR, nuclear magnetic resonance; DSS, 2,2-dimethyl-2-silapentane-5-sulfonate.

Method for predicting RNA secondary structure.

“…First, all known tRNA primary sequences can be arranged in the form of a cloverleaf structure by using classical Watson-Crick base pairing rules with the occasional inclusion of G U pairs. The cloverleaf model is consistent with a large body of spectral data, including circular dichroism-optical rotatory dispersion and nuclear magnetic resonance studies (15,16). For at least one case, in which a specific tRNA was isolated and crystallized, yeast phenylalanine tRNA has been shown to be in the form of a cloverleaf by x-ray diffraction (17,18).…”

The “Speedy” Synthesis of Atom‐Specific 15N Imino/Amido‐Labeled RNA

“…NMR spectroscopy offers the unique possibility to study these interactions in solution with simple one-dimensional 1 H NMR experiments, whereby the amido/imino proton region between 9 and 15 ppm directly reflects nucleobase pairing in the RNA. [3–6] More sophisticated experiments like 1 H/ 15 N/ 15 N (HNN) COSY experiments, which require 15 N-labeled RNA, allow for observing the hydrogen bonds in Watson–Crick base pairs directly by internucleotide 2 J NN couplings. [7, 8] In general, NMR spectroscopy has become a powerful tool for studying nucleic acids in order to obtain information not only about their solution structure.…”

Quantitative determination of the number of secondary and tertiary structure base pairs in transfer RNA in solution

“…Each Watson-Crick base pair, and some tertiary structure base pairs, contribute one resonance (due to hydrogen-bonded ring nitrogen protons of G and U) to this spectral region, so, in principle, the total number of base pairs per molecule can be directly determined from integration of the low-field spectrum.While seemingly simple, integration of the low-field spectra of tRNA has been difficult due to lack of an accurate intensity standard. Initially, we integrated spectra by comparison with an external standard (Kearns et al, 1971a,b) or with a resolved resonance(s) in the low-field region that was assumed to have a known intensity (Shulman et al, 1973b;Lightfoot et al, 1973) . Neither method was completely satisfactory and the results indicated that several class I tRNA contained ~19 base pairs (not counting GU pairs), or about the number predicted by the cloverleaf model.…”

“…While seemingly simple, integration of the low-field spectra of tRNA has been difficult due to lack of an accurate intensity standard. Initially, we integrated spectra by comparison with an external standard (Kearns et al, 1971a,b) or with a resolved resonance(s) in the low-field region that was assumed to have a known intensity (Shulman et al, 1973b;Lightfoot et al, 1973) . Neither method was completely satisfactory and the results indicated that several class I tRNA contained ~19 base pairs (not counting GU pairs), or about the number predicted by the cloverleaf model.…”

“…# Lightfoot et al, 1973. * Shulman et al, 1973b. 1 Kearns et al, 1971b. > Robillard et al, 1976 table II: Comparison of the Calculated and Observed Ratio of the Intensities in the Aromatic (9-6.5 ppm) and Ribose (6.5-5.0 ppm) Regions of the Spectra of tRNA.…”