Effect of cleaving the dihydrouridine loop and the ribothymidine loop on the amino acid acceptor activity of yeast phenylalanine transfer ribonucleic acid.

Abstract Incubation of pure yeast phenylalanine transfer RNA with high levels of ribonuclease T1 at 37° in 20 mm MgCl2 leads initially to the production of a 3' three-quarter-molecule and a 5' quarter-molecule as a result of rapid cleavage in the dihydrouridine loop. Subsequently, scission in the ribothymidine loop occurs, generating 3' quarter-molecules and half-molecules; the latter comprise the central half of the molecule extending from the dihydrouridine loop to the ribothymidine loop. The identity of these fragments was established by column "fingerprinting" of the oligonucleotides produced by complete digestion with ribonuclease T1. The central half-molecule, the 5' quarter-molecule, and the 3' quarter-molecule can be reannealed to form an aggregate the size of intact tRNA as judged by Sephadex G-100 chromatography. This reconstituted tRNAphe can be charged to approximately 20% with phenylalanine by partially purified phenylalanyl-tRNA synthetase despite the cleavages in the dihydrouridine and ribothymidine loops, thus indicating that these loops need not be intact for enzyme recognition. Neither individual fragments nor aggregates with large sections of the tRNA missing could be aminoacylated.

A great deal of work in recent years has been concerned with the elucidation of the structural features of transfer RNA molecules, which are responsible for their specific aminoacylation by aminoacyl-tRNA synthetase enzymes. The assumed synthetase recognition site has been probed by various enzymatic and chemical methods (for a recent review, see Zachau (1)). One approach, involving cleavage of tRNA molecules and reassembly of the resultant fragments, has been used to investigate the possible participation of the single stranded regions in enzyme recognition.
It has been reported that scission in the anticodon loop (2-7) or in the dihydrouridine-containing loop (8), the ribothymidine-containing loop (9), or the extra loop (10) does not abolish the amino acid acceptor activity. * This work was supported in part by a grant from the University of California Cancer Coordinating Committee.
The relative ease of purification of yeast tRNAPhe and the elucidation of its primary structure by RajBhandary and Chang (11) prompted us to attempt fragmentation and reconstitution experiments with this tRNA.
The anticodon loop of this molecule is resistant to ribonuclease T1 since it contains only 1 guanosine that is 2'Gmethylated.
We have studied the degradation of yeast tRNA Phe by ribonuclease T1 and demonstrated cleavage into large fragments the size of half-and quarter-molecules; these fragments recombine upon mixing to produce an aggregate that accepts phenylalanine and is the size of intact tRNA (12). Further studies reported here have led to the identification of these fragments as being a 5' quarter extending into the dihydrouridine loop, a 3' quarter extending into the ribothymidine loop, and a half-molecule comprising the central region from the dihydrouridine loop to the ribothymidine loop. More detailed experiments on the reconstitution of this "horizontally bisected" tRNA are now reported, and their significance with regard to enzyme recognition is discussed.

MATERIALS AND METHODS
Purification of Yeast tRNAPh8-Brewers' yeast tRN~4 (Boehringer Mannheim) was fractionated on benzoylated DEAEcellulose columns according to the procedure of Wimmer, Maxwell,and Tener (13) to yield partially purified tRNAPhe. This material was preparatively aminoacylated with 14C-phenylalanine and rechromatographed on benzoylated DEAE-cellulose according to the method of Litt (14), to yield a symmetrical peak of pure tRNAPhe.
The stripped tRNAPhe which was used for fragmentation studies, was approximately 95 y0 homogeneous as judged by quantitative 14C-phenylalanine acceptance and complete ribonuclease T1 digestion patterns.
Pur$cation of Phe-tRNA Synthetase from Yeast-Aliquots, 450 g, of Fleischmann's pressed bakers' yeast (purchased frozen from Standard Brands, Inc., New York) were thawed in 250 ml of 50 mM Tris-Cl, pH 7.6, 10 nlM MgC12, and 10 mM mercaptoethanol buffer, and the suspension was homogenized with glass beads in a Bronwill MSK homogenizer.
The homogenate was centrifuged twice at 15,000 x g at 2" for 15 min, and the supernatant was centrifuged at 200,000 x g for 1 hour to remove ribosomes. The post-ribosomal supernatant proteins that precipitated between 40 and 70% saturation with ammonium sulfate at 0" were centrifuged, redissolved in 10 mM potassium phosphate, 1 mu 5744 Side Loop Cleavage of tRNAPhe Vol. 245,No. 21 EDTA, pH 7.2, to a concentration of 20 mg per ml, and dialyzed overnight into this buffer.
The dialyzed material was applied to a column of DEAE-Sephadex A25 (6 X 10 cm) equilibrated with the same buffer.
Elution with this buffer removed most of the protein, and the Phe-tRNA synthetase was subsequently eluted with the same buffer containing 0.2 M NaCl.
The eluted enzyme was concentrated by ammonium sulfate precipitation, redissolved in 10 Inn/r potassium phosphate, pH 7.2, and stored at -20" in 50% glycerol.
At this stage of purification, the enzyme was stable for several months, and the specific activity was routinely greater than 100 enzyme units per mg of protein (1 enzyme unit catalyzes the esterification of 1 nmole of i*C-phenylalanine to tRNA in 10 min at 37" under the standard assay conditions; see below under "Assay for l*C-Phenylalanine Acceptance").
Although specific activities of 2,000 to 3,000 units per mg have been obtained by further purification steps, the highly purified enzyme is unstable upon storage, and such preparations were not used in these experiments.
Partial Ribonuclease T1 Digestion of tRNAPhe--Pure tRNAPhe, 1,320 As60 units,l was incubated in 30 ml of 0.05 M Tris-HCl, pH 7.5, 0.02 M MgC& with 300,000 enzyme units (15) of ribonuclease T1 (Worthington) at 37" for 10 min. At the end of the incubation period, the enzyme was removed by five consecutive extractions with 30-ml aliquots of phenol saturated with 0.05 M Tris-Cl, pH 7.5. The pooled phenolic phases were back-extracted with 30 ml of 0.05 M Tris-Cl, pH 7.5. Finally, the combined aqueous phases were extracted 10 times with 30-ml aliquots of ether (to remove residual phenol); the ether-extracted solution was then concentrated by flash evaporation to approximately 3 ml and processed immediately by chromatography on Sephadex G-100 at 57". For analytical experiments involving the investigation of the time course of production of various fragment sizes, the above quantities were scaled down by a factor of 100.
Separation of Large Fragments Produced from tRNAPhe--The ribonuclease T1-treated tRNAPhe was applied to a column of Sephadex G-100 (2.2 x 105 cm) equilibrated with 10 mM potassium phosphate, pH 7.5, containing 0.1 M NaCl, following the procedure of Imura, Schwam, and Chambers (6). The tRNA fragments were chromatographed under denaturing conditions by maintaining the column at 57". The column was eluted with boiled buffer from a reservoir which was maintained at approximately 80" during the chromatography.
Fractions, 5 ml, of the eluate were collected.
Rechromatography of the individual fragments obtained by this procedure was performed under identical conditions.
The columns were calibrated with the use of intact tRNA, tRNAPhe halves (4), the tRNAPhe anticodon dodecanucleotide (17), and GMP. Identi$cation of Large Fragments by Complete Ribonuclease Tl Digestion-The large fragments obtained from tRNAPhe and purified by hot Sephadex G-100 chromatography were identified by means of their complete ribonuclease T1 fingerprints obtained by following the digestion procedure described by RajBhandary,Stuart,and Chang (16). Solid urea was added to the complete digest to give a final concentration of 7 M, and the sample was applied to a column of DEAE-cellulose (0.9 x 95 cm) equilibrated with the starting Tris-urea buffer.
Elution was carried out with a linear gradient between 500 ml of 0.02 M Tris-Cl, pH 7.5, 7 M urea, and 500 ml of the same buffer containing 0.3 M 1 One ARM) unit is the amount of material in 1 ml of solution that gives an absorbance of 1.0 at 260 nm in a l-cm light path.

NaCl.
A constant flow rate of 70 ml per hour was maintained; the effluent was passed through a flow cell assembly, and the A 260 was continuously recorded.
The reference fingerprint pattern of intact tRNBPhe was obtained in the same way.
Identi$cation of Complete Ribonuclease T1 Fingerprint Oligonucleotides-The identity of the various oligonucleotides obtained upon complete ribonuclease T1 digestion was based on the DEAE-cellulose elution profile of RajBhandary et al. (16) with the two exceptions mentioned under "Results." Base Composition of Oligonucleotides-The oligonucleotides were hydrolyzed at 37" in 0.25 N KOH for 16 hours, and the hydrolysate was adjusted to pH 10 by the addition of Dowex 50 (Hf form).
An aliquot was placed on a Dowex I-formate column (0.9 X 25 cm) equilibrated with water and eluted with the double exponential formic acid gradient system described by Carbon (17). The eluted mononucleotides were identified by elution positions and ultraviolet spectra. Characterization of 7-methyl-GMP was obtained by the ultraviolet spectrum of the alkaline breakdown product (2-amino-4-hydroxy-5-N-methylformamido-6-(N-&phosphoribofuranosylamino)pyrimidine), which eluted from the column in front of GMP.
Demonstration of Spec$c Reaggregation of Fragments--The various fragments used in reconstitution experiments were combined at room temperature in the concentrations indicated in the text in 10 m MgCIZ. The mixture was immediately applied to a column of Sephadex G-100 equilibrated with 0.05 M Tris-Cl, pH 7.5, 0.01 M MgCl*, which was eluted at room temperature with the same buffer.
Fractions, 5 ml, of the eluate were collected and monitored for absorbance at 260 nm. i*C-Phenylalanine acceptance was determined on aliquots of the eluted fractions with the use of partially purified phenylalanyl-tRNA synthetase under the standard tRNA-charging conditions.
Assay of 14C-Phenylalanine Acceptance-For assay, 0.01 to 2.0 Az6,, units of tRNA or recombined fragments were incubated with 0.08 mM l*C-phenylalanine (18 PCi per pmole, or 90 PCi per pmole in the case of small amounts of tRNA when greater sensitivity is required) and 10 enzyme units of partially purified phenylalanyl-tRNA synthetase (100 to 200 units per mg) in a medium containing 10 mM ATP, 50 mM MgC12, 100 mM Tris-Cl (pH 7.5), and 10 mM reduced glutathione.
The reaction mixture (usually 0.2 ml) was incubated at 37", and 40.~1 aliquots were removed at various time intervals onto filter paper discs (Schleicher and Schuell, 593-A) which were immediately dropped into ice-cold 10% trichloracetic acid. The discs were washed, dried, and counted in scintillation fluid according to previously published procedures (18).

RESULTS
RNase Tl cleavage of tRNAphe- Fig.  1 shows the results of a small scale experiment designed to find, at a given substrate concentration, magnesium ion concentration, temperature, and incubation time, the optimal RNase T1 concentration for the breakdown of tRNAPhe into fragments that can reanneal to form functional aggregates.
Following a suggestion by Oda et al. (3), tRNAPhe was incubated with increasing amounts of RNase T1 under the specified conditions; the reaction was stopped by 40-fold dilution of the incubation medium with water, and aliquots were assayed, immediately or after a heating-quick cooling step, for phenylalanine acceptor activity. As can be seen, the proportion of functional, albeit cleaved, molecules, i.e. molecules that retain chargeability when kept cold after the nuclease treatment but lose it upon disruption by heating-quick cooling, reaches a maximum at 10,000 enzyme units per ml. The usefulness of this result was tested by an independent procedure. At the high enzyme level indicated, tRNAPhe was incubated for various lengths of time, and the production of large fragments followed directly by chromatography of the partial digest on hot Sephadex G-100 columns.
As can be seen in Figs. 2 and 3, initially three-quarter-and quarter-sized fragments are formed in approximately equal molar quantities. Subsequently, halves are generated while the amount of three-quarter-molecules decreases and the percentage of quarters keeps rising.
Ultimately, all large fragments are degraded to the complete digest size, which ranges from GMP to the dodecanucleotide containing the anticodon loop. This experiment not only demonstrates that, at the previously determined enzyme concentration, the amount of large fragments reaches a maximum in a conveniently short time, but it also allows one to make statements as to the probable identity of the fragments produced.
Assuming that the base-paired regions of the molecule ("stems") are less susceptible to the enzyme than are the single stranded sections ("loops"), the most plausible sites for enzyme attack appear to be in the dihydrouridine-and ribothymidine-containing loops and perhaps in the minor loop since, as a result of the absence of RNase T1-susceptible residues in the anticodon loop, a vertical bisection of the molecule as reported for a number of other tRNAs (2,3,(5)(6)(7)19), is very unlikely.
Random attack, therefore, should lead to the formation of three-quarter-and quarter-sized molecules in the initial stages of digestion, with subsequent attack giving rise to half-molecules and more quarters. This is indeed observed.
The half-molecule would therefore be expected to extend from the dihydrouridine loop to the ribothymidine loop, while the quarter population should consist of 5' and 3' quarter fragments from the top half of the molecule. Cleavage of the extra loop apparently does not occur, as intermediate-size fragments are not observed.
For preparative scale fragment production, 70 mg of pure  Fig. 4 shows the elution pattern of the partial digest after chromatography on a hot Sephadex G-100 column. Sft.er double rechromatography of the individual peaks, the fragment populations appear homogeneous and free of cross-contamination, as can be seen from Fig. 5.
Identijification of Fragments-The half-and quarter molecules were subjected to complete ribonuclease T1 digestion to allow comparison of their oligonucleotide elution patterns on DEAEcellulose with that of a ribonuclease T1 fingerprint of intact tRNAPh".
The complete digest pattern of intact tRNAPh", shown at the top of Fig. 6 for 10 min. The sample was prepared for chromatography and fractionated at 57" on a column of Sephadex G-100 as described under "Materials and Methods." Fractions of 4.8 ml were collected and monitored for absorbance at 260 nm after appropriate dilution. places the CpApCpC found by them; Peaks 11 and 12 in our elution profile were found to be the individual components of their Peak 13, i.e. 7MeGpUpCp5MeCpUpGp and CpUpCp-ApGp, respectively (see Table I).
The T1 digestion patterns obtained under identical conditions from the half-molecule and quarter-molecule fragments are also shown in Fig. 6, together with the primary sequence of yeast tRNAPhe.
Although little interpretation is possible from the early eluting small oligonucleotides, the intermediate and large oligonucleotides are very diagnostic. The quarter fragment exhibits the following features.
The half-molecule fragment exhibits a complementary pattern, as follows.
These patterns establish that the quarter fraction is a mixture of 5' and 3' fragments comprising the top half of the cloverleaf structure, and that the half-molecule is the bottom half of the molecule extending from the dihydrouridine loop to the ribothymidine loop and containing an intact anticodon loop. This finding, which is in agreement with the hypothesis outlined earlier, indicates cleavage of the tRNA in both the dihydrouridine loop and the ribothymidine loop. Reconstitution of Fragments to Form Functional tRNAphe--One series of experiments deals with the restoration of function in mixtures in which one fragment is kept constant while the other is added in increasing amounts and vice versa. Acceptor activity, for a given amount of constant component, is plotted as a function of the ratio of A260 units of variable component to A260 units of constant component (Fig. 7). Instead of the expected break in the slope at a ratio of 1 (equal amounts of halves and quarters), one observes a linear increase in chargeability up to an AeGO ratio of quarters to halves of approximately 3 (i.e. a molar ratio of six quarters per half) when increasing  a The presence of GMP in this peak is due to the complete overlap of Peak 7 and Peak 6 (DpDpGp) in the preparative fractionation from which the oligonucleotide was obtained. b Based on the E,,, of the base-catalyzed, ring fission product of 7-methyl-GMP (20).
amounts of quarters are added. In a corresponding manner, the addition of increasing amounts of halves to a constant amount of quarters produces a linear response in chargeabiiity only up to an &,a ratio of halves to quarters of approximately 0.3. These results suggested that the quarter-molecule fraction might contain unequal amounts of the 5' quarter and 3' quarter.
A rough estimate of the relative proportion of the two quarters can be obtained from the observation that a g-fold molar excess of quarters is required to titrate the half-molecules, which would suggest a value of approximately 5 : 1 for the ratio of nonlimiting to limiting quarters. This interpretation should be verifiable by analyzing the complete ribonuclease T1 digestion pattern obtained from the mixed quarter population.
From the fingerprints shown in Fig. 6, one observes that the 2 moles of oligonucleotide in Peak 15 are both derived from the 3' quarter (14 nucleotides total), whereas the hexanucleotide comprising 6 columns as described under "Materials and Methods." On the right is shown the complete cloverleaf structure of yeast tRNAPhe, as elucidated by RajBhandary and Chang (12), indicating the positions in the molecule from which the individual oligonucleotide peaks are derived. FIG. 7. Restoration of function by adding increasing amounts of half-molecules to quarter fragments and vice versa. a, 0.11 A 260 unit of quarter fragments were mixed with increasing amounts of half-molecules in the ratios indicated and assayed for phenylalanine acceptor activity as described in the text. b, 0.037 A260 unit of half-molecules were mixed with increasing amounts of quarter fragments in the ratios indicated and assayed for phenylalanine acceptor activity as described in the text. Peak 14 is derived from the 5' quarter.
Hence, in an equal mixture of 3' and 5' quarters, the ratio (in At60 units) of Peak 15 to Peak 14 should be approximately 2.4:1, which is in fact observed in the fingerprint of intact tRNAPhe. In the fingerprint derived from the mixed quarter population, however, the ratio of Peak 15 to Peak 14 is only approximately 0.6:1, indicating a 4-fold lower amount of the Peak 15-containing (i.e. 3') quarter. Similarly, fingerprints of the three-quarter-molecule fraction were found to contain negligible amounts of Peaks 12 a.nd 14 indicating a marked predominance of the 3' three-quarter-molecule resulting from preferential cleavage of the dihydrouridinecontaining loop. As a corollary, an excess of 5' quarter frag-  Vol. 245,No. 21 ments is expected to be formed.
Thus, our interpretation involving the presence of a limited quantity of the 3' quarter fragment appears to be corroborated by direct chemical evidence. A further experiment designed to show the dependence on the half-molecule as well as the two types of quarters for phenylalanine acceptance was carried out by the continuous variation mixing method in which halves and quarters are reciprocally varied in the mixture.
Such an experiment is shown in Fig. 8. It should be noted in this experiment that the stock solution of quarters used had 4 times the As60 of the half-molecule solution. At the optimum of this curve, acceptor activity is obviously limited both by the half and by the limiting 3' quarter; i.e. their molar concentrations should be equal. At this point, the AZeO ratio of quarters to halves is 3 : 1, indicating a molar ratio of quarters to halves of 6: 1 under conditions in which the limiting 3' quarter and the half-molecule are present in equimolar amounts.
This again indicates a 5-fold molar excess of 5' quarters over the limiting 3' quarter.
Thus we conclude that the mixed quarter population comprises 16 to 20% 3' quarters and 80 to 84yo 5' quarters.
These experiments prove that isolated quarter and half fragments can be recombined to form a functional complex that can be recognized by phenylalanyl-tRNA synthetase despite the scission of both the dihydrouridine and ribothymidine-containing loops. It was interesting to note in these experiments that preliminary incubation of the mixed fragments for various periods at 37" in buffer (50 mM Tris-Cl, pH 7.5, 20 mM MgCl,), had no effect on the level of charging, indicating that the functional reconstitution between fragments is quite rapid.
In all of the foregoing experiments, the molar acceptance of 14C-phenylalanine was found to be 2O7,, based on the molar concentration of the limiting fragment. Since the synthetase recognition site may be located exclusively in the "upper half" of the molecule, an attempt was made to charge the mixed quarter fraction in the absence of the "lower half" of the molecule under conditions favoring the reannealing of the two quarters.
Results of these experiments are shown in Table II. As can be seen, the level of phenylalanine acceptance is barely detectable.