The secondary structure of transfer RNAs can be represented as a cloverleaf structure. The common features include:

1)  a 7-base pair (bp) acceptor arm or stem,
2) a 3- to 4-bp stem followed by the D-loop (D stands for the modified base called dihydrouridine),
3)  a 5-bp anticodon stem followed by the anticodon loop,
4)  a variable arm in which the number of nucleotides varies greatly among tRNAs,
5)  a 5-bp TYC arm followed by the TYC loop (the symbol Y denotes the modified base pseudouridine),
6)  the ubiquitous CCA sequence at the 3'-terminus in which the 3'-OH is the (final) site for attachment of a unique amino acid by a specific aminoacyl-tRNA synthetase.

Overall structure

     The tertiary structure of a transfer RNA resembles an inverted letter L. The vertical stem is formed by coaxial stacking of the acceptor and T-stems to form one contiguous helix, and the horizontal stem is likewise formed by stacking of the D-stem and the anticodon stem. The D-loop and the TYC loop meet at the corner of the L. In the crystal structure of yeast phenylalanine transfer RNA (tRNA^Phe), complex tertiary interactions can be seen. This summary is based on the structure with PDB code 6TNA. (Note: some minor differences have been found in a more recent crystallographic structure obtained in spring 2000.)

Figure 1A. Crystal structure of yeast tRNAPhe, wireframe representation.	Figure 1B. (Right image). Backbone representation. Acceptor arm:  yellow; TYC - loop:  blue; Variable loop:  orange; Anticodon arm:  red;D - loop:  green; Mg2+ ions:  cyan.

1. Double-stranded helices

     The stem regions of transfer RNA are formed by double-stranded helices. The double- stranded helices in the transfer RNA are in the A-form. Most of the base pairs in the helical regions are normal Watson-Crick base pairs, although the stem regions also contain non-Watson-Crick base pairs such as GU base pairs.

2. Base stacking interactions

     Overall tertiary structure is stabilized by base stacking interactions which can be clearly seen in one axis through the acceptor stem and the TYC loop and in another axis through the anticodon stem and the D-loop.

3. Tertiary base pairing interactions.

Most of the invariant and semi-invariant bases are located in the loop regions. The invariant bases are conserved

Figure 2.

among different tRNAs. The semi-invariant bases are either conserved purines or pyrimidines. These conserved bases participate in tertiary base pairing interactions that maintain the loop structures and hold the D-loop, the variable loop and the TYC loop together. Unlike the typical Watson-Crick base pairs in the stem regions, many of the tertiary base pairing interactions in the loop regions are those of non-Watson-Crick base pairs. The invariant U8 on the base of D stem and the invariant A14 on the D-loop participate in non-Watson-Crick base pairing called a reverse Hoogsteen pair to stabilize a sharp turn seen in the D-loop. In the Watson-Crick base pair, N1 of adenine serves as a hydrogen bonding acceptor in the AU pair. In the Hoogsteen base pair, N7 of adenine serves as a hydrogen acceptor. In another example, the tertiary hydrogen bonding between G18 in the D-loop and Y55 on the TYC loop stabilizes the interaction of these separate loops. The non-helical regions of the tRNA is very complex. The bases, the phosphate backbone, and 2'-OH of the ribose ring (which distinguishes RNA from DNA) participate in complex interactions to hold the tRNA compact.

4. Base intercalation

Figure 3.

Base intercalation may also stabilize the tertiary structure of tRNAs. In the crystal structure, two examples can be seen. G57 on TYC loop intercalates between G18 and G19 on the D-loop. A9 on the base of D-stem intercalates between G45 and 7mG46 on the variable loop.

5. The Anticodon.

Figure 4.

Most bases in the tRNA are utilized in either secondary hydrogen bonding or tertiary hydrogen bonding interactions to stabilize the overall structure. On the other hand, the three bases (Gm34, A35, and A36) which constitute the anticodon are accessible in solution. This allows the anticodon bases to pair with the corresponding codon in a messenger RNA. The invariant purine on the 3' side of the anticodon is typically modified. This modified purine base (Y37) is also accessible in solution and may assist the interaction between the anticodon and the codon on the mRNA.

6. Magnesium binding sites.

     Divalent metal ions such as magnesium ions (Mg²⁺) can associate with the phosphate backbone of RNA through electrostatic interaction. It is proposed that divalent metal ions may assist RNA folding and may stabilize the tertiary structure. In the crystal structure, four Mg²⁺ binding sites can be seen. One Mg²⁺ binding site is located in the pocket formed by a sharp bend of D-loop. Another site is located in the anticodon loop. Two Mg²⁺ binding sites are located in the corner of the L. These Mg²⁺ ions may stabilize the interaction between the D-loop and the TYC loop. In solution, Mg²⁺ ions are typically coordinated by six water molecules which occupy the apices of  an octahedron. In the crystal structure, the Mg²⁺ ion in the pocket of the D-loop are coordinated by six water molecules. These water molecules in turn participate in hydrogen bonding interactions with the phosphate oxygens. The Mg²⁺ ions in the other sites are directly coordinated by one or two phosphate oxygens and the remaining four or five sites are occupied by the water molecules. These water molecules participate in hydrogen bonding interactions with nitrogens or oxygens of the bases.

7. Hairpin loops or "U-Turns."

     Among the three hairpin loops in tRNA^Phe, two especially sharp hairpin loops, the TYC loop and the anticodon loop, are called an uridine turn or U-turn because of the key role of the uridine residues in these

Figure 5.

loops. In the U-turn, a sharp turn occurs following the uracil residue (U33 in the anticodon loop and Y55 in the TYC loop). In each case, the U-turn is stabilized through interactions between the uridine residue and the phosphate oxygen three residues downstream toward 3' terminus (U33 and Y55 interact P36 and P58 respectively). The phosphate oxygen two residues down stream is positioned directly beneath the base (see the position of P57 and the base Y55 in the figure). The base stacking interactions can be seen between the residue 53 to 55 and between the residues 56 to 58. The bases, 56, 57, and 58, point to the opposite direction with respect to the three bases preceding the U-turn.

The tertiary structure of transfer RNA^Phe

2. Quigley, G. J. and Rich, A. Structural domains of transfer RNA molecules. Science (1976) 194, 796-806.

3. Soll, D. Transfer RNA: an RNA for all seasons,  Ch. 7 in The RNA world, Gesteland, R. F. and Atkins, J. F., ed., Cold Spring Harbor Laboratory Press (1993)

Interaction between tRNA and aminoacyltransfer RNA synthetase

2. Kim, S. and Schimmel, P. Functional independence of microhelix aminoacylation from anticodon binding in a class I tRNA synthetase. J. Biol. Chem. (1992) 267, 15563-15567.

3. Perona, J. J., Swanson, R. N., Rould, M. A., Steitz, T. A., and Soll, D. Structural basis for  misaminoacylation by mutant E. coli glutaminyl-tRNA synthetase enzymes. Science (1989) 246, 1152-1154.

v4. Rould, M. A., Perona, J. J., Soll, D., and Steitz, T. A. Structure of E. coli glutaminyl-tRNA synthetase complexed with tRNA^Gln and ATP at 2.8 angstrom resolution. Science (1989) 246, 1135-1142.

5. Ruff, M., Krishnaswamy, S., Boeglin, A., Poterszman, A., Mitschler, A., Podjarny, A., Rees, B., Thierry, J. C., and Moras, D. Class II aminoacyl transfer RNA synthetases: Crystal structure of yeast aspartyl-tRNA synthetase complexed with tRNA^Asp. Science (1991) 252, 1682-1689.