High-resolution Fourier-transform nuclear magnetic resonance at 15.18 MHz was used to observe the proton-decoupled natural-abundance (13)C spectra of aqueous unfractionated tRNA from baker's yeast in the presence of Mg(+2) (8 ions per tRNA molecule), as a function of temperature in the range of 27-82 degrees C. The spectrum of thermally denatured tRNA at 82 degrees C showed numerous sharp resonances, which were assigned to specific types of carbon atoms by comparison with the (13)C spectra of mononucleotides. Only the resonance of carbon 4' of the ribose rings was appreciably shifted (by about 1.5 ppm upfield) with its average position in the mononucleotides. This effect was also seen in the spectrum of poly(A). In the spectrum of folded tRNA (52 degrees C), carbon 4' was further shifted upfield by about 1.5 ppm, and carbons 2' and 3', which yielded a single resonance at 82 degrees C, now showed two partly resolved peaks. The variation of linewidths with temperature (200 mg/ml of tRNA) was gradual in the range 27-82 degrees C, and did not reflect the expected unfolding behavior of tRNA. Moreover, dilution to 80 mg/ml at 27 degrees C had the same effect as an increase in temperature to about 45 degrees C. The line-width changes below 60 degrees were ascribed to tRNA aggregation. In contrast to the behavior of the linewidths, the (13)C spin-lattice relaxation times (T(1)) of individual ribose carbon atoms, measured by means of partially relaxed Fourier-transform spectra were practically independent of temperature up to about 60 degrees C, and increased rapidly at higher temperatures. The T(1) values indicated that the backbone of thermally denatured tRNA is undergoing rapid segmental motion, with an effective correlation time of (2.6 +/- 0.5) x 10(-10) sec. The T(1) values of folded tRNA yielded no evidence of segmental motion. The correlation time for overall rotational reorientation is about (3 +/- 1) x 10(-8) sec in the range 35-54 degrees C. Within experimental error, the T(1) values of methine carbons of the bases were equal to those of the methine carbons on the backbone at all temperatures. Only an upper limit to the rate of internal rotation of the bases could be established.