Codon-anticodon interaction in tRNAPhe

J. Mol. Biol.

(1980) 142, 219-230

Codon-Anticodon

Interaction in tRNAPhe

II. A Nuclear Magnetic Resonance Study of the Binding of the Codon WC H.A.

MCEERDES,~

lDepartm.ent

J.H.

VAN BOOMS AND C.W.HILBERS~

of Biophysical Chemistry, University Toernooiveld, 6525 ED Nijlnegen The Netherlands

of Nijmegen

2Gorlaeus Laboratoria, State University of Leiden P.O. Box 9502, 2300 AA Leiden The Netherlands (Received

10 January

1980, and in revised form 28 May 1980)

The effect of binding of the codon WC to yeast tRNAPhe was investigated by means of n.m.r.t spectroscopy and analytical ultracentrifugation. Binding of UUC to the transfer RNA anticodon tends to promote the aggregation of tRNA molecules; this is manifest from a line broadening in the n.m.r. experiments as well as from an increase in szo,w in the ultracentrifuge experiments. Such an aggregation of tRNA molecules wae not observed upon addition of different oligonucleotides, as described in the accompanying paper. In addition to the specific resonances in the general broadening observed in the n.m.r. spectra, methyl proton spectrum as well as in the hydrogen-bonded proton spectrum are broadened or shifted upon binding of UUC. These results are explained on the basis of the premise that two different tRNA-WC complexes can exist in solution. It is suggested that the binding of UUC tends to promote a disruption of the m7G46.miG22 base-pair and its neighbouring base-pairs. In studying the binding of U-U-U-U to yeast tRNAPhe no resonances of protons hydrogen-bonded between the oligonucleotide and the tRNA could be detected at, low temperatures. This indicates, that at these temperatures the lifetime of the tRNA-U-U-U-U complex is substantially shorter than the lifetime of the other tRNA-oligonucleotide complexes studied in this and the accompanying paper under these conditions.

1. Introduction It has been demonstrated in the accompanying paper, that the anticodon loop rather easily adopts a conformation in which it can form a double helix together with complementary oligonucleotides. These oligonucleotides consisted of codons having additional base(s) at their 3’ side, which favor the formation of a base-pair with U,, in the anticodon loop. With the anticodon loop forced in such a conformation no indications were found of the disruption of the T-loop-D-loop interactions, a process t Abbreviations

used: n.m.r., nuclear magnetic resonance;

p.p.m., parts per million.

219 0022-283f3/80/260219-12

$02.00/O

0 1980 Academic

Press Inc. (London)

Ltd.

220

H.

A. M. GEERDES

ET

AL.

which may take place during protein synthesis. It has been argued (Kurland et al., 1975) that transfer RNA conformations favourable to ribosome binding are only generated when the anticodon bases are properly matched with the codon. A recent experiment by Moller et al. (1978) indicates that the tetranucleotide U-U-C-A may not fulfill this requirement. The authors demonstrated that the yeast tRNAPheU-U-C-A complex does hardly bind to the 30 S ribosomal subunit, while UUC and U-U-U-U stimulate tRNA binding to this ribosomal subunit. Therefore in this paper the association of UUC and U-U-U-U with yeast tRNA Phe is investigated by n.m.r.t spectroscopy. It will be shown that the trinucleotide has an interesting effect on the tRNA structure not observed for the other nucleotides. At concentrations used for n.m.r. UUC tends to promote the formation of tRNA dimers (aggregates), an effect that can be undone by binding the tetranucleotide U-U-C-A. The lifetime of the tRNAU-U-U-U complex is too short to allow the observation of the hydrogen-bonded protons in the codon-anticodon complex.

2. Materials and Methods Yeast tRNAPhe and U-U-U-U were purchased from Boehringer, Mannheim. The tRNA had an amino acid acceptance of 1400 pmol/o.n. 26,, unit as determined by the manufacturer. Mgs+ -free tRNA samples were prepared by dialysing the tRNA according to Thiebe (1975) ; subsequently the tRNA was dialysed 4 times against a solution of 1 mivr-Na,S&,. it contained 0.1 mol Mg2 + /mol tRNA as After dialysis the tRNA was lyophilized; determined by atomic absorption spectrophotometry. tRNA concentrations were unit/ml = 1.8 PM. calculated using 1 o.D.~~,, The oligonucleotides UUC and U-U-C-A were synthesized according to the phosphocontained no impurities triester method (Van Boom et al., 1977a,b). The preparations detectable by thin-layer chromatography, high-performance liquid chromatography and 31P n.m.r. A slight impurity (1%) was occasionally detected in lH n.m.r. giving rise to a resonance at 1.9 p.p.m. The oligonucleotides were used in the Na+ form; excess salt was removed by Sephadex GlO gel filtration. Concentrations of UUC, U-U-C-A and unit/ml equal to 40 PM, 25.6 PLMand 25 pM, U-U-U-U were calculated taking 1 O.D.260 respectively. The n.m.r. samples were prepared by dissolving the lyophilized tRNA in 0.2 ml buffer; solution conditions are indicated in the Figure legends. The H,O samples contained 5% 2H,0 to lock the field of the n.m.r. spectrometer to the resonance frequency of deuterium. lH n.m.r. spectra were recorded on a Bruker 360 MHz n.m.r. spectrometer, operating in the correlation spectroscopy mode; 500 to 2000 scans of 2 s each were accumulated in a Nicolet BNC 12 computer. Sample temperatures were held constant within 1 deg. C. Chemical shifts are indicated with respect to 4-4-dimethyl-4-silapentane-1-sulfonate. Centrifugation experiments were performed on a Spinco model E analytical ultracentrifuge using the Spinco four-place rotor An-F Ti. Sedimentation experiments were performed at 60,000 revs/min; equilibrium centrifugation was done at 30,000 revs/min. Rather high tRNA concentrations (0.2 to 0.9 IIIM) were used to facilitate a direct comparison with the n.m.r. results.

3. Results (a) Effect of UUC binding

on the methyl proton

spectra

of yeast tRNAPhe

n.m.r. spectra of the methyl protons (0 to 4 p.p,m.) of yeast tRNAPhe were recorded at various concentrations of UUC. Figure 1 shows an example of the effect of codont see footnote

to p. 219.

CODON-ANTICODON

INTERACTION.

II

221

p.p.m. FIG. 1. 360 MHz ‘H n.m.r. spectra of the methyl protons of yeast tRNAPhB at 15°C. (a) Without oligonucleotide added; (b) after addition of 4.2 mm-UUC; and (c) after addition of 4.2 mu-UUC and 3.9 m&r-U-U-C-A. Assignment of the resonances is according to the data of Davanloo et al. (1979); 1, Cs of D16.17; 2, mZGlo and C, of D16.17; 3, m;G,,; 4, Y,,(Cc~~,--CHd; 5, m%s; 6, m%o; 7, T,,. Asterisks indicate impurities present in the oligonucleotide. Solution conditions: 0.6 mM-tRNA, 0.1 M-NaCl, 2 mivr-NazSz03, 20 mM-MgCl,, 30 mM-sodium phosphate at pH 7.4 in ZHaO.

anticodon interaction. When the codon UUC binds to the tRNA, a significant line broadening of all resonances is observed and the resonance (number 4), located at 2.0 p.p.m. in the absence of oligonucleotide, shifts upfield (Fig. l(b)). This resonance has been assigned to the CI1 methyl group of the Y-base, which is located next to the anticodon. The shift has also been observed after complex formation of other oligonucleotides (Geerdes et al., 1978 and accompanying paper) and binding of Escherichia coli tRNAGrU (Davanloo et al., 1979). As discussed in the accompanying paper this demonstrates that the oligonucleotide binds to the anticodon loop of the tRNA.

222

H.

A. M. GEERDES

ET

1 12

, 13

AL.

, II

V,b IO

P.PJn

FIQ. 2. 360 MHz ‘H n.m.r. spectra of the hydrogen-bonded protons in yeast tRNAPhe before (a) and after (b) the additions of 3.2 mM-UUC, recorded at 2’C. (b) Represents a situation in which 81 y, of the anticodon sites is occupied. tRNA concentration wa8 1.4 mm in a solution containing 0.1 M-NaCI, 10 mM-MgCl,, 12 mM-Na&OB, 10 mM-sodium phosphate at pH 7.0.

The chemical shift of the Y-base Cl1 methyl resonance measured as a function of oligonucleotide concentration was used to determine the association constant for the equilibrium : (tRNAPhe-WC) . tRNAPhe + UUC i The binding constant obtained in this way amounted to 2OOOf1OOO M-~ at 2°C. This is in good agreement with data reported in the literature (Pongs & Reinwald, 1973; Eisinger et al., 1971). The upfield shift for the Y-base C,, methyl resonance was 0*15&0*02 p.p.m. upon full saturation of the anticodon sites. As mentioned above, binding of WC to yeast tRNAPhe also induces a substantial line broadening of all resonances of the methyl proton spectrum (Fig. l(b)). In addition resonances at 2.9 p.p.m. (number 1) and 2.4 p.p.m. (number 3) are specifically broadened or shifted. The tetranucleotide U-U-C-A competes very effectively with UUC for binding to the anticodon, because at the temperature of the experiment in Figure 1 the binding constant of U-U-C-A is 50 times that of WC (Yoon et al., 1975). Therefore, U-U-C-A was added to the tRNAPhe-UUC complex and the original spectral resolution was recovered (Fig. l(c)). This finding demonstrates that the broadening of resonances after complex formation of yeast tRNAPhe with WC is due to codon-anticodon interaction.

CODON-ANTICODON

I

I 14

INTERACTION.

I

I 12

I

223

II

,

h

IO

6w.m. FIG. 3. 360 MHz (a), (b) and after (c) tRNAPhe; (b) identical multiplication in the anticodon sites are legend to Fig. 2.

lH n.m.r. spectra of the hydrogen-bonded protons in yeast tRNAPhe before the add’t’1 ion of UUC to a concentration of 25 mu, recorded at 22°C. (a) Yeast with (a) except that all resonances were broadened by 12 Hz, by exponential time domain; (c) yeast tRNAPhe + 25 mM-UUC; in this situation 88% of the occupied. tRNA concentration was 1 mM; for other solution conditions see

(b) Effect of UUC binding

on the hydrogen-bonded

proton

spectra

Complex formation of yeast tRNAPhe with UUC also results in an overall line broadening of the resonances in the low-field n.m.r. spectrum. This is shown in Figure 2, where the yeast tRNAPhe spectrum (Fig. 2(a)) is compared with the spectrum of yeast tRNA Phe-UUC complex (Fig. 2(b)); both spectra were recorded at 2°C in the presence of 10 mm-Mg 2+. At a higher temperature the line broadening is less severe. Figure 3 shows the low-field n.m.r. spectra of yeast tRNAPhe before and after (Fig. 3(a) and (c), respectively) addition of UUC, recorded at 22°C. The spectrum in

224

H.

A. M. GEERDES

ET AL.

(c)-(b)

I

I

I

14

12

IO

txfm. FIQ. 4. 360 lH MHz n.m.r. spectra of the hydrogen-bonded protons in (a), (b) and after the addition of 2-7 mm-UUC (c), recorded at 2°C. (a) Yeast to (a), except that the resonances are broadened by 12 Hz, by exponential time domain; (c) yeast tRNArhs + 2.7 mm-UUC. The difference spectrum Figure represents spectrum (c) minus spectrum (b). Solution conditions: NaCl, 20 mm-Na,S,G,, 10 mhx-sodium phosphate at pH 7.0.

yeast tRNAPhe before tRNAPbe; (b) identical multiplication in the at the bottom of this 1.4 mM-tRNA, 0.1 M-

Figure 3(b) is a simulated spectrum obtained by increasing the line widths of the resonances of the spectrum in (a) by 12 Hz, by exponential filtering in the time domain. Comparison of the speotre of Figure 3(b) and (c) shows that both spectra are nearly identical, except for some extra decrease in the resolution around the main peak centred at 12.5 p.p.m. We cannot say whether this is due to an extra broadening or to slight shifts of particular resonances. In the absence of Mgs + the line broadening is not as pronounced as in Figure 2.

CODON-ANTICODON

INTERACTION.

II

225

FIG. 5. Plot of the extinction (E) as a function of the distance to the axis (r) after equilibrium contrifugation of yeast tRNA PM (lower trace) and yeast tRNAPhe + 13 mu-UUC (upper trace; 950/b of the anticodon sites occupied). At the detection wavelength (310 nm) only the tRNA contributes to the extinction. The centrifugation proceeded for more than 24 h at 30,000 revs/min and at 5°C; m, indicates the meniscus. Solution conditions: 0.9 mix-tRN.4, 100 miv-NaCl, 10 muXgCl,, 30 mM-sodium phosphate at pH 7.4.

This permitted the resolution of the resonances from the protons hydrogen-bonded between the tRNA and UUC. The low-field n.m.r. spectra of yeast tRNAPhe before and after addition of UUC, recorded at 2°C are shown in Figure 4(a) and (c), respectively. Again these two spectra cannot be subtracted immediately because of the difference in line widths. Subtraction was achieved by simulating the spectrum of Figure 4(b), which was obtained by broadening the resonances of Figure 4(a) by 12 Hz, by exponential multiplication in the time domain. The difference spectrum at the bottom of this Figure represents the difference, (c) minus (b). Clearly, also upon complex formation of yeast tRNAPhe with UUC, resonances hydrogen-bonded between the codon and the tRNA become visible. Extra resonance intensity is observed at about 13.3 p.p.m. in the difference spectrum, which is assigned to U(N3) protons hydrogen-bonded between WC and the tRNA in analogy with the findings for ot,her complementary oligonucleotides (accompanying paper). The G(N1) proton is not so well-resolved. At higher temperatures the lifetime of the codon-anticodon complex in tRNA Phe-UUC is too short ((5 ms) to allow detection of these proton resonances. (c) Ultracentrifuge

experiments

of complex

formation

of yeast tRNAPhe with UUC

Rather surprisingly the binding of UUC to yeast tRNAPhe induces a significant line broadening of all resonance of the lH n.m.r. spectra. Among other things such a line broadening may be the result of a decrease in the rotational correlation of the tRNA molecule after the binding process. Such a change in correlation time may be caused by a change in molecular conformation or by dimerization or aggregation of the tRNA molecules. To distinguish between these possibilities we measured the

226

H.

A. M. GEERDES

ET

AL

sedimentation rate of the tRNA and its complexes with UUC and U-U-C-A in an analytical ultracentrifuge experiment. It followed that the sedimentation rates of the tRNA and the tRNA-U-U-C-A complex are equal, i.e. szO,w = 3.4 S as determined at 8°C in a solution containing 0.2 m&r-tRNA, IO mw-sodium phosphate buffer (pH 7.0) and 10 mM-MgCl,. This sedimentation value is lower than values reported in the literature (~,a,~ = 4-O S; Lindahl et al., 1966). This is not surprising since the latter represents sedimentation rates extrapolated to infinite dilution, a correction that has not been made for our experiments. After complex formation of tRNA with UUC the szO,w value amounted to 3.9 S. Such an increase is not expected to be caused by a change (unfolding) of the tRNA conformation (Henley et al., 1966). Rather the results suggest that tRNA dimers or aggregates are formed. This is corroborated by the results of an equilibrium centrifugation experiment. As shown in Figure 5 the concentration gradient obtained for the tRNA-WC complex is much steeper than in the case of the tRNA alone. This provides strong evidence that tRNA dimers (or aggregates) are formed upon binding of UUC to yeast tRNA Phe. The equilibrium centrifugation experiments were performed at a rather high tRNA concentration (0.9 mM) to facilitate a direct comparison with the n.m.r. experiments. Since the apparent molecular weight is strongly dependent upon the concentration the derivation of the extent of aggregation quantitatively is not justified. (d) Effect of U-U-U-U

binding

on the proton spectra of yeast tRNAPhe

The n.m.r. experiments described for WC were repeated with U-U-U-U. After addition of this tetranucleotide to the tRNA solution the changes observed in the methyl spectra were similar to those obtained for the other oligonucleotides, i.e. the C,, methyl group of the Y-base shifts while the other methyl resonances remain at their positions. The hydrogen-bonded proton spectra of the yeast tRNAPhe-U-U-U-U complex recorded at 2°C do not show an increase in resonance intensity as a result of the formation of hydrogen bonds in the complex. From this we conclude that the lifetime of the complex is so short that the hydrogen-bonded proton resonances of the ring nitrogen protons of the tetranucleotide will not be observable.

4. Discussion (a) Kinetics

and thermodynamics

of codon-anticodon

complex formation

The trinucleotide WC is unique among the other nucleotides studied by us in that after complex formation with yeast tRNA Phe it induces a substantial line broadening in the lH n.m.r. spectra (Figs 1 and 3). This line broadening is due specifically to the binding of WC to the anticodon loop of the tRNA because addition of U-U-C-A to the tRNA-WC complex results in a spectrum in which the original line widths are recovered (Fig. 1). As described in the preceding section the line broadening is most likely caused by a decrease in the molecular tumbling time as a result of dimerization or aggregation of tRNA molecules. A rough estimate of the extent of aggregation can be obtained by considering the increase in line widt,h upon complex formation with WC. In a first approximation the line width will be proportional to the rotational correlation time, i.e. 6v = CT, where SV is the line width at half height, c is a constant

CODON-ANTICODON

INTERACTIOS.

II

227

a.nd 7 is the rotational correlation time. The rotational correlation time will increase wit,h the molecular weight and for spherical molecules is approximately proportional to the molecular weight. At 2°C the line width increases by about 3074, so that, roughly 30% of the material is dimerized if we assume that dimers are formed. It is interesting to see how tRNA molecules can possibly form dimers (aggregates) that can be reversibly dissociated. Before considering plausible possibilities, we shall first discuss some kinetic aspects of the binding of U-U-U-U, UUC and U-U-C-A. Resonances from hydrogen-bonded protons between oligonucleotides and tRNA are observable in n.m.r. experiments when the lifetime of the complex exceeds 5 ms. so that the line broadening is 50 Hz or less (Crothers et al., 1973; Geerdes et al., 1978). For all oligonucleotide-tRNA complexes studied, except for the tetranucleotide U-U-U-U, resonances could be observed at 1°C. U-U-U-U forms a wobble base-pair the lifetime of (Crick, 1966) with G,, in the anticodon of yeast tRNA Phe. Apparently the complex is substantially shorter than those of the complexes formed with oligonucleotides containing the “correct” UUC sequence. Similar observations have been anticodons. made by Grosjean et aZ. (1978) in a study of tRNAs with complementary Therefore the suggestion of Mizuno & Sundaralingam (1978) that a wobble G.U base-pair provides a stabilization of the codon-anticodon complex is at variance with t,hese two observations. Yoon et al. (1975) det,ermined the thermodynamic and kinet)ic parameters characterizing the binding of UUC and U-U-C-A to yeast tRNAPhe by means of temperaturejump experiments using the fluorescence of the Y-base to follow the kinet’ics. The stability of these tRNA-oligonucleotide complexes is much higher than that found for the corresponding tri- and tetranucleotide double helices. It turned out that the enhanced stability of the tRNA-oligonucleotide complexes does not, derive from the rea,ction enthalpy but is caused by the entropy terms, which are less negative t,han t,hose representative of the formation of the corresponding tri- and tetranucleotide double helices. The kinetic parameters characterizing the association of U-U-C-A to tRNA are similar to the values found for corresponding model systems. On the other hand the kinetics of the binding of UUC to the tRNA are anomalous. In particular, t,he association rate constant is lower by one order of magnitude, which was rationalized (Yoon et al., 1975) by assuming that the formation of the second base-pair is rate-limiting when UUC binds to the tRNA. In our n.m.r. experiments the binding of UUC leads to the rather unexpected effect of aggregation of tRNA molecules. This suggests a different explanation for the rather low association rate constant of yeast, t’RNAPhe-UUC complex formation. In order for the tRNA to form dimers (aggregates) t,he structure of the tRNA-UUC complex must somehow differ from the structure of a single tRNA molecule. We therefore consider the following reaction: k

T+C


k

TC -t2

k-2

TC* >

(1)

where T and C represent the tRNA and the trinucleotide UUC, respectively; TC and TC* denote different conformations of the tRNA-UUC complex. TC is the complex with normal association and dissociation rate constants; in the TC* complex the tRNA has undergone a conformational change so that it, is able to form dimers (or aggregates). Under the conditions of the n.m.r. experiments (i.e. high concentrations x

228

H.

A. M. GEERDES

ET

AL.

of tRNA) the complexes can be “trapped” in the dimeric state. A detailed analysis (Geerdes, 1979) shows that reaction (1) can account for the available data characterizing UUC binding (Yoon et al., 1975), for the changes in the anticodon loop formation (Urbanke & Maass, 1978) and for the n.m.r. results presented in this paper. It was necessary to assume that the first step in reaction (1) is a fast pre-equilibrium. This step is characterized by a reaction enthalpy, AH, expected for trinucleotide double helix formation (Borer et al., 1974). Very recently, using temperature-jump kinetics, Labuda & Porschke (personal communication) have observed a conformational change in yeast tRNAPhe upon binding of WC. This also leads to dimer formation of the tRNA molecules. Reaction schemes similar to that in equation (1) have been proposed by these investigators to explain their observations. (b) Effect of codon-anticodon

interaction

on the structure of the tRNA

Binding of UUC to yeast tRNA Phs has a profound influence on the n.m.r. spectra. At low temperature (Figs 1 and Z), where the exchange of the trinucleotide is considerably slowed down, a general line broadening is observed in the hydrogen-bonded as well as in the methyl proton spectra. Although the observation of particular changes in the spectra is somewhat hindered, the results do indicate that resonances of the anticodon stem are specifically influenced. This is seen, for instance, in Figure 3 where the resonances around the main peak are somewhat less well-resolved than in the simulated spectrum. On the basis of ring-current shift calculations (Geerdes & Hilbers, 1977) and also using the assignments of resonances from anticodon hairpins (Lightfoot et al., 1973 ; Rordorf, 1975), these resonances most likely come from hydrogen-bonded protons in the G * C, 7, G. C, s and G.C,, base-pairs of the anticodon stem. Similarly, the resonance at 2.4 p.p.m., specifically affected in the methyl proton spectrum, comes from the m:G,, residue, which is located on the top of the anticodon stem. Apparently binding of UUC can affect the resonance position of protons as far as 20 A away from the codon binding site. This conclusion is reinforced by the broadening of resonance 1 in Figure 1, which comes from C, of D, 6 and D, 7 (Davanloo et al., 1979). The present results do not give any insight into the precise structural adjustments of the anticodon stem upon oligonucleotide binding. An “unwinding” as well as an expanding of the anticodon stem can give rise to the effects experimentally observed in the n.m.r. spectra. Mizuno & Sundaralingam (1978) pointed out that the flipping of the anticodon loop from the 3’ stacked to the 5’ stacked conformation and vice versa can only take pla,ce together with a simultaneous flipping of the tilts of the anticodon stem base-pairs with respect to the helical axis of the anticodon stem. This may provide another explanation for the specific changes observed in the n.m.r. spectra. We may now ask whether these changes can furnish an explanation for the observed dimerization (aggregation) of the tRNA molecules and concomitant line broadening of the n.m.r. spectra. The changes observed for the methyl resonances may reflect a disrupture of the miGs,.A,, base-pair. If we postulate that such an effect causes a further disruption of base-pairs formed by residues of the extra arm this could make the G,5-m7G46-U,, sequence available for base-pairing to the A,,-C&,-C,, sequence of other tRNA molecules, thereby explaining the dimerization process. It has been

CODON-ANTICODON

INTERACTION.

II

229

shown that the latter sequence is available for intermolecular base-pairing with the trinucleotide GGU with an association constant of 3 x lo* M-I at -2°C (Pongs et al., 1973). In this respect it is interesting to note that the activation energy associated with a transition from the TC to the TC* conformation (see equation (1)) was estimated to be ~18 kcal.mol-l (Geerdes, 1979). This is just the amount of energS required to disrupt two or three base-pairs. Alt,hough necessarily speculative the model has some attractive features. First,. it forms a good working hypothesis for designing experiments to test the proposed conformational changes. Secondly, it forms a basis for the explanation of the behavior of the so-called Hirsch suppressor tRNB (&de infra). Thirdly, it is in agreement’ w&h recent experiments of Wagner & Garrett (1979), who studied the effect of codonanticodon interaction on the structure of E. coli tRNALY” by means of chemical modification experiments (vide in&a). The Hirsch suppressor tRNA (Hirsch, 1971) is able to read the UGA stop codon as the triplet coding for tryptophan. This suppressor tRNA turned out to be a mutant of tRNATrP having the normal tryptophan anticodon (CCA) complementary t,o the codon UGG. The structural alteration of this tRNA was found in its D-stem, where G,, is replaced by an A. Apparently, one way or another the D-stem is involved in the correct reading of the mRNA. It is most remarkable that in the three-dimensional structure of yeast tRNAPhe the m:G,,, which seems to be influenced by binding of the trinucleotide UUC to yeast tRNA Phe, is situated rather close to the G,, residue (Rich & RajBhandary, 1976), which is the site of mutation in the Hirsch suppressor tRNS. Moreover, the G,, residue is “sandwiched” between residues G,, and m7G4,, which may be involved in the intermolecular aggregation of tRNA molecules, as discussed above. Wagner & Garrett (1978,1979) studied the effect of codon-anticodon interaction on the structure of yeast tRNAPhe and E. coli tRNALYs by means of chemical modification experiments. They monitored the modification of G residues of these tRNAs with kethoxal in the presence and absence of “codons”. They conclude (Wagner & Garrett, LSTS), that the m7G,,.Gz2 base-pair is disrupted upon codon-anticodon interaction. This is in agreement with our suggestions made above. They also found that the G,, residue of the D-stem of E. coEi tRNALYs was slightly more exposed to the solut,ion upon codon binding. Interestingly, we also found an effect on proton resonances coming from the D-loop upon binding (Fig. 1) of UUC to yeast tRNAPhe. At present, it is hard to see why the binding of UUC to yeast tRNAPhe leads to more profound structural adjustment within the tRNA molecule than binding of the other oligonucleotides, e.g. U-U-C-A. Whatever the reason, the structure of the t,RNA-UUC complex is special. This follows not only from the n.m.r. experiment,s described in this paper, but also from the experiments of MGller et al. (1978), while it is also consistent with the kinetic data of Yoon et al. (1975). The correct triplettripleb interactions between tRNA and mRNA may thus favor special tRNA conformations, which may be important in the selection of tRNAs on the ribosome (Kurland et al., 1975). We are much indebted to Mm G. Wille Hazeleger and to ,J. van Kessel for experimental assistance.

for the synthesis of the oligonucleotides We thank Dr J. Walters for his help

230

H.

a.

M. GEERDES

ET

AL.

in the interpretation of the U.C results. We gratefully acknowledge support of the 360 MHz n.m.r. facility at Groningen and Dr R. Kaptein for keeping the instrument in excellent condition.

Z. W. 0. for the and K. Dijkstra

REFERENCES Borer, P. N., Dengler, B., Tinoco, I. & Uhlenbeck, 0. C’. (1974). J. Mol. Biol. 86, 843-853. Crick, F. H. C. (1966). J. Mol. Biol. 18, 548-55.5. Crothers, D. M., Hilbers, C. W. & Shulman, R. G. (1973). Proc. Nat. Acad. Sci., U.S.A. 70, 2899-2901. Davanloo, P., Sprinzl, M. &Cramer, F. (1979). Biochemistry, 18, 3189-3199. Eisinger, J., Feuer, B. & Yamane, T. (1971). Nuture (London), 231, 126-128. Geerdes, H. A. M. (1979). Thesis, University of Nijmegen. Geerdes, H. A. M. & Hilbers, C. W. ( 1977). Nucl. Acids Res. 4, 207-221. Geerdes, H. A. M., Van Boom, J. H. & Hilbers, C. W. (1978). FEBS Letters, 88, 27-32. Grosjean, H. J., de Henan, S. & Crothers, D. M. (1978). Proc. Nat. Acad. Sci., U.S.A., 75, 610-614. Henley, D. D., Lindahl, T. & Fresco, J. R. (1966). Proc. flat. Acad. Sci., U.S.A. 55, 191-198. Hirsch, D. (1971). J. Mol. Biol. 58, 459-468. Kurland, C. G., Rigler, R., Ehrenberg, M. & Blomberg, C. (1975). Proc. Nat. Acad. Sk., U.S.A. 72, 4248-4251. D. R., Reid, B. R. & Shulman, R. G. (1973). Lightfoot, D. R., Wong, K. L., Kearns, J. Mol. Biol. 78, 71-89. Lindahl, T., Henley, D. D. & Fresco, J. R. (1966). J. Amer. Chem. Sot. 87, 4961-4963. Mizuno, H. & Sundaralingham, M. (1978). Nucl. Acids Res. 5, 4451-4461. MBller, A., Schwarz, U., Lipecky, R. & Gassen, H. G. (1978). FEBS Letters, 89, 263-266. 50, 3~5.~363. Pongs, 0. & Reinwald, E. (1973). Biochem. Biophys. Res. Commun. Pongs, O., Bald, R. & Reinwald, E. (1973). Eur. J. Biochem. 32, 1177125. Rich, A. & RajBhandary, U. L. (1976). Annu. Rev. Biochem. 45, 805-860. Rordorf, F. (1975). Thesis, University of Riverside, California. Thiebe, R. (1975). FEBS Letters, 51, 259-261. Urbanke, C. & Maass, G. (1978). Sucl. Acids Res. 5, 1551-1560. Van Boom, J. H., Burgers, P. J. M., Crea, R., Van der Marel, G. & Wille Hazeleger, G. (1977a). Nucl. Acids Res. 4, 747-759. Van Boom, J. H., Burgers, P. J. M., Van der Mare& G., Verdegaal, C. H. M. & Wille Hazeleger, G. (1977b). Nucl. Acids Res. 4, 1047-1064. Wagner, R. & Garrett, R. A. (1978). PEBS Letters, 85, 291-295. Wagner, R. & Garrett, R. A. (1979). Eur. J. Biochem. 97, 615-621. Yoon, K., Turner, D. H. & Tinoco, I. (1975). J. Mol. Riol. 99, 5077518.