Importance of the reverse Hoogsteen base pair 54-58 for tRNA function

Abstract

To elucidate the general constraints imposed on the structure of the D- and T-loops in functional tRNAs, active suppressor tRNAs were selected in vivo from a combinatorial tRNA gene library in which several nucleotide positions of these loops were randomized. Analysis of the nucleotide sequences of the selected clones demonstrates that among the randomized nucleotides, the most conservative are nucleotides 54 and 58 in the T-loop. In most cases, they make the combination U54-A58, which allows the formation of the normal reverse Hoogsteen base pair. Surprisingly, other clones have either the combination G54-A58 or G54-G58. However, molecular modeling shows that these purine-purine base pairs can very closely mimic the reverse Hoogsteen base pair U-A and thus can replace it in the T-loop of a functional tRNA. This places the reverse Hoogsteen base pair 54-58 as one of the most important structural aspects of tRNA functionality. We suggest that the major role of this base pair is to preserve the conformation of dinucleotide 59-60 and, through this, to maintain the general architecture of the tRNA L-form.

Figure 1

The standard tRNA L-form. Rectangles represent individual nucleotides. The DT region at the outer corner of the molecule is boxed. Cross-hatched and filled rectangles represent nucleotides of the D- and T-loop, respectively. Unpaired nucleotides as well as nucleotides at the beginning and the end of the helical regions are numbered in accordance with the standard tRNA nomenclature (1). Nucleotides of the anticodon loop, non-stacked nucleotides of the D-loop and nucleotide 47 are not shown. There are two base pairs, G18-Ψ55 and G19-C56, formed between the D- and T-loops. The reverse Hoogsteen base pair T54-A58, whose structure is seen in Figure 4, is formed within the T-loop. Dinucleotide 59–60 bulges from the double helical stem between base pairs G53-C61 and T54-A58. Nucleotide 59 stacks on the tertiary base pair 15-48, constituting the last layer of the D/anticodon helical domain. This interaction fixes the perpendicular arrangement of the two helical domains called the L-form.

Figure 2

Construction of the tRNA gene library. In the nucleotide sequence of E.coli tRNA^Ala_UGC, each of the two enclosed regions, 16–19 in the D-loop and 54–58 in the T-loop, was replaced by six fully randomized positions, while nucleotide G20 and the anticodon TGC (boxed) were replaced by T20 and CTA, respectively. Nucleotides 54 and 58, which form the reverse Hoogsteen base pair in the T-loop, are connected by a line. The EcoRI and PstI restriction sites that are seen flanking the 5′ and 3′ termini were used for cloning the library into the pGFIB-1 plasmid.

Figure 3

Northern blot showing the presence in the cytosol and the level of aminoacylation of some suppressor tRNAs. For each clone the – and + lanes correspond to the samples not treated and treated with Tris. In the – lanes the aminoacylated and deacylated forms of the suppressor tRNA move as individual bands, while in the + lanes the total tRNA is deacylated and the suppressor tRNA moves as one band. In all – lanes the bands corresponding to the aminoacylated form of the tRNA are much larger than those corresponding to the deacylated form and are comparable to the bands in the + lanes, representing the total amount of the suppressor tRNA. This indicates that in all clones most of the tRNA is present in the aminoacylated form. A smaller size of the bands of the suppressor tRNAs compared to tRNA^Ala_su+ indicates a notably lower presence of the selected tRNAs in the cytosol. 5S rRNA was visualized to monitor the amount of total RNA in each sample. Because the signal from 5S rRNA was much stronger than that from suppressor tRNAs, the upper and lower parts of the same membrane have been exposed, respectively, for 4 h and overnight. The nucleotide sequence of clone K8, due to its low β-galactosidase activity, is not included in Table 1, but is available upon request.

Figure 4

Juxtaposition of the bases in RH-GA, RH-GG and other alternative base pair candidates for replacement of RH-UA. (A) Positions of the glycosidic bonds in the alternative base pairs compared to that in RH-UA. In each base pair the position of the glycosidic bond corresponding to the base on the right is superimposed on that of U in RH-UA. The glycosidic bond of the other nucleotide will thus occupy a particular place depending on the structure of the base pair. The numbers indicating particular positions of the glycosidic bonds correspond to the base pairs in (B).

Figure 5

The model of the structure of the DT region for clone K31 (red) superimposed on the corresponding region in yeast tRNA^Phe (green). The figure also includes the T-stem and the tertiary base pair 15-48. For both tRNAs, the ribbon follows the sugar–phosphate backbone. Explicitly shown are base pairs 15-48 and 54-58 and nucleotide 59 in tRNA^Phe, as well as all nucleotides of the DT region and pair 15-48 in K31. Comparison of the modeled structure with tRNA^Phe demonstrates a good superposition of the T-stem and base pairs RH and 15-48, as well as nucleotide 59. The proper arrangements of the nucleotides in the RH base pair thus guarantees the proper position of nucleotide 59, whose stacking to base pair 15-48 would fix the juxtaposition of the two helical domains known as the L-form. Still, one can notice a difference in the conformation of the backbone in the two structures, which is highest between nucleotides 58 and 59. Such a difference makes a universal interaction of this region with a particular protein factor unlikely.